3-Dimensional Gradient and Spin-Echo Magnetic Resonance Cholangiopancreatography with Deep Learning Reconstruction at 3 T: Achieving Superior Image Quality with Reduced Acquisition Time

3-Dimensional Gradient and Spin-Echo Magnetic Resonance Cholangiopancreatography with Deep Learning Reconstruction at 3 T: Achieving Superior Image Quality with Reduced Acquisition Time | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article 3-Dimensional Gradient and Spin-Echo Magnetic Resonance Cholangiopancreatography with Deep Learning Reconstruction at 3 T: Achieving Superior Image Quality with Reduced Acquisition Time Kumi Ozaki, Takafumi Iyoda, Eri Sugioka, Jihun Kwon, Yasutomo Katsumata, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9286494/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose To evaluate image quality and clinical feasibility of breath-hold 3D magnetic resonance cholangiopancreatography (MRCP) using a gradient and spin-echo (GRASE) technique with deep learning reconstruction (GRASE-DLR) versus GRASE without DLR and turbo spin-echo with DLR (TSE-DLR) at 3 T. Methods Sixty-five consecutive patients who underwent 3D MRCP on a 3-T system were retrospectively enrolled. Three protocols were compared: GRASE-DLR, GRASE without DLR, and TSE-DLR. Acquisition time, quantitative metrics (SNR and CNR), and five-point qualitative scores for overall image quality, artefact reduction, background suppression, and ductal visualization were independently assessed by two radiologists. Interobserver agreement was evaluated using Cohen's weighted kappa. Diagnostic performance for biliary and pancreatic disease and anatomical variations was evaluated. Subgroup analysis was performed for patients with poor breath-hold capacity (n = 8). Results Mean acquisition time for GRASE-DLR was 8.9 s, representing reductions of 45.4% versus TSE-DLR and 49.1% versus GRASE (both p < 0.001). Despite comparable SNR across all three protocols, GRASE-DLR achieved significantly superior overall image quality, artefact reduction, and major duct visualization (all p < 0.001), with moderate-to-substantial interobserver agreement. Sensitivity for biliary disease was markedly higher with GRASE-DLR (93.8%) than TSE-DLR (56.2–62.5%) and GRASE (68.8–75.0%), with accuracy of 89.5–94.7%. For pancreatic disease, sensitivity was 87.5% with GRASE-DLR versus 55.0–75.0% for comparators, with accuracy of 84.8–87.0%. Indeterminate biliary anatomical variation findings were nearly eliminated with GRASE-DLR (0–1.5% vs. 21.5–26.2% for TSE-DLR). In patients with poor breath-hold capacity, GRASE-DLR demonstrated pronounced improvements in image quality and artefact suppression. Conclusions GRASE-DLR achieves superior image quality with substantially reduced acquisition time and improved diagnostic confidence, particularly in patients with limited breath-hold tolerance. Gradient and spin-echo Magnetic resonance cholangiopancreatography Deep learning reconstruction Turbo spin-echo Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Magnetic resonance cholangiopancreatography (MRCP) is an essential sequence for evaluating pancreaticobiliary disease because it facilitates noninvasive evaluation of the anatomy and abnormalities of the pancreaticobiliary tree with high spatial resolution [ 1 – 4 ]. Respiratory-triggered or navigator-gated three-dimensional (3D) turbo spin-echo (TSE)-based sequences have commonly been used in 3D MRCP because they provide a higher signal-to-noise ratio (SNR) and superior spatial resolution [ 5 , 6 ]. However, these techniques require lengthy acquisition times to obtain images during the expiratory phase, particularly in patients with shallow or irregular breathing patterns, which often compromises image quality. To solve this problem, a single-breath-hold protocol is highly desirable and can be performed using fast imaging techniques. Therefore, various strategies have been employed to achieve better image quality in breath-hold MRCP sequences. Parallel imaging, compressed sensing (CS), or other rapid sequences are commonly employed for rapid 3D breath-hold MRCP. With parallel imaging, 3D breath-hold MRCP has become feasible, yielding acceptable image quality [ 7 , 8 ]. However, parallel imaging techniques have limitations in visualizing small ductal structures, including peripheral intrahepatic bile ducts, pancreatic duct branches, and communications with cystic lesions. Additionally, parallel imaging alone has not fully achieved both short acquisition times and high image quality simultaneously [ 7 , 8 ]. CS, which substantially decreases the scan time through high k-space undersampling [ 8 – 10 ], is another approach for 3D breath-hold MRCP. Previous studies have shown that, compared with standard respiratory-triggered MRCP, breath-hold MRCP using CS can improve temporal resolution while maintaining image quality [ 11 , 12 ]. However, in clinical practice, high acceleration factors can result in diminished image quality due to insufficient noise removal by CS reconstruction [ 11 , 12 ]. The combination of gradient and spin–echo sequences (GRASE) enables faster image acquisition with a more homogeneous B0 field compared with TSE-based sequences in high-field MRI, thereby reducing motion artefacts and improving image uniformity [ 13 , 14 ]. However, despite these advantages, GRASE MRCP has been limited by inadequate visualization of small ductal structures, such as peripheral intrahepatic bile ducts and pancreatic duct branches, due to T2* decay introduced by the gradient echo component [ 13 – 15 ]. Recent technological advances in MR, both in reconstruction algorithms and hardware capabilities, have aimed to address these limitations. Deep learning reconstruction (DLR)-based approaches have been developed to improve image quality from undersampled MRI k-space data, with recent studies demonstrating that TSE-MRCP with DLR achieves image quality equivalent or superior to respiratory-triggered MRCP, while substantially reducing the acquisition time [ 16 – 18 ]. Concurrently, hardware advances, including enhanced SNR, improved gradient systems, and optimized field homogeneity, have enabled further improvements in GRASE MRCP image quality [ 13 , 14 ]. In this context, we employed a recently developed DLR algorithm to acquire GRASE sequences with a substantially reduced acquisition time while preserving image quality. To our knowledge, no previous study has evaluated the combined use of GRASE and DLR for MRCP. Therefore, this study aimed to evaluate whether GRASE-DLR can achieve superior image quality and diagnostic performance compared with GRASE without DLR and conventional TSE-DLR, while substantially reducing acquisition time. Materials and methods The Research Ethics Committee of our institution approved this study (Approval No. 24–159) and waived the requirement for written informed consent because of its retrospective design. Study Population We identified 70 consecutive patients who underwent abdominal 3-T MRI between July and September 2024. We excluded patients who underwent cholecystectomy (n = 3), whose examinations were interrupted (n = 1), and those with metal artefacts due to gastrectomy (n = 1) (Fig. 1 ). The indications for 3D MRCP were as follows: evaluation or follow-up of pancreatic cystic lesions (n = 33), biliary malignancy (n = 11), chronic pancreatitis (n = 5), gallstones or bile duct stone (n = 6), pancreatic neuroendocrine tumor (n = 5), pancreatic ductal adenocarcinoma (n = 3), primary sclerosing cholangitis (n = 1), and pancreaticobiliary maljunction (n = 1). MRCP Parameters and Model-Based DLR MRI was performed using a 3-T system (MR7700; Philips Healthcare, Best, the Netherlands) equipped with a 32-channel dS Torso coil. The 3D-MRCP protocol comprised the following sequences in coronal orientation: GRASE-DLR, GRASE without DLR, and TSE-DLR. CS reconstruction was applied only to the GRASE sequence without DLR. In contrast, GRASE-DLR and TSE-DLR used a model-based DLR algorithm that incorporates CS principles within a deep learning framework, rather than conventional CS reconstruction. For the GRASE protocols, the number of spin echoes was set to 8, and the number of gradient echoes per spin echo was set to 9, resulting in an echo planar imaging (EPI) factor of 72. For GRASE-DLR, the undersampling (acceleration) factor was set to be twice that of the standard GRASE protocol recommended by the vendor to halve the acquisition time; all other imaging parameters, including the EPI factor, were kept identical between the GRASE and GRASE-DLR sequences. The following model-based DLR technique was incorporated into GRASE-DLR and TSE-DLR. The prototype CS-DLR used an Adaptive-CS-Network scheme [ 19 , 20 , 21 ], which was based on CS theory, to reconstruct the images. The network architecture was identical across all sequences, integrating a multiscale convolutional neural network with CS principles. Multiscale sparsification based on wavelet transforms was replaced with a learned representation, while domain-specific constraints, such as data consistency, were preserved [ 22 ]. Moreover, neighboring slices were incorporated together with the center slice as inputs to the sparsifying transform, along with several soft priors encoding MRI domain knowledge. The Adaptive-CS-Network used in this study was trained and validated on approximately 740,000 MR images of various anatomies and contrast settings. The three sequences were acquired in random order prior to contrast administration. A detailed description of the DLR reconstruction algorithm is provided in a previous report [ 19 ]. Identical field-of-view and voxel size were used across all 3D MRCP sequences to ensure a fair comparison. The actual acquisition time for each protocol was recorded. The imaging parameters for all MRCP sequences are summarized in Table 1 . Table 1 Parameters for MRCP imaging protocols Sequence 3D_MRCP 2D_CSMRCP 2D_SRMRCP 3D-TSE 3D-GRASE 3D-GRASE Respiratory compensation Breath-hold Breath-hold Breath-hold Echo time (ms) 600* 73 73 Repetition time (ms) 2500 286 286 Acceleration technique DLR Compressed sensing DLR Acceleration factor 12 4 8 Flip angle (degree) 90 90 90 Slice thickness (mm) 2.0 2.4 2.4 Slice number 80 67 67 Field-of-view (mm) 320 320 320 Acquisition matrix 256×256 256×256 256×256 Acquisition voxel size (mm³) 1.25×1.25 1.25×1.25 1.25×1.25 Reconstruction matrix 512 512 512 Reconstruction voxel size (mm³) 0.62 0.62 0.62 Deep learning reconstruction Yes No Yes Echo train length 180 8 8 Echo planar imaging factor N/A 72 72 Note. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE = Gradient and Spin-Echo without DLR, GRASE-DLR = GRASE with DLR. *Effective echo time for TSE-DLR. Quantitative Image Analysis A board-certified radiologist with 23 years’ experience in MRCP performed the quantitative image analysis. We selected three representative slice levels [upper, middle, and lower common bile duct (CBD)] that depicted the center of the CBD in each patient. Signal intensity (SI) was measured by placing circular regions of interest (ROIs) on the CBD and periductal tissues (Fig. 2 ). The ROIs for the SI of the CBD, which were at least 5 mm 2 , were placed in homogeneous, artefact-free areas in the middle third of the course of the CBD. The ROIs for the SI of the periductal tissue, which were at least 20 mm 2 , were placed in homogeneous, artefact-free areas adjacent to the ROI of the CBD. Image noise was defined as the standard deviation (SD) of the signal within the CBD ROI rather than background noise, as deep learning reconstruction alters the spatial noise distribution and homogeneity, rendering background noise measurements unreliable and non-representative of local image noise. To ensure methodological consistency across all three protocols, the same ROI-based SD metric was applied uniformly. The SNR, contrast ratio, and contrast-to-noise ratio (CNR) of the three MRCP protocols were calculated. The SNR of the CBD was calculated using the formula [ 23 ]: SNR = mean SI CBD /mean SD CBD The contrast between the CBD and periductal tissues on 3D MRCP was evaluated quantitatively using the following formula: Contrast = (mean SI CBD – mean SI periductal tissue )/(mean SI CBD + mean SI periductal tissue ) The CNR between the CBD and periductal tissues was calculated using the following formula: CNR = (mean SI CBD – mean SI periductal tissue )/mean SD CBD Qualitative Image Evaluation Two independent radiologists (23 and 15 years of abdominal MRI experience), blinded to sequence type, evaluated images using five-point scales (5 = excellent, 4 = good, 3 = acceptable, 2 = suboptimal, 1 = unacceptable) for overall image quality, artefacts, background suppression, and visibility of biliary and pancreatic duct segments [ 11 , 18 ]. Diagnostic Performance Analysis The same two readers independently evaluated the presence of anatomical variations and diseases of the biliary and pancreatic ducts. Specific assessments included the presence of stenosis, ductal dilatation, pancreatic cystic lesions, and filling defects attributable to stones in the CBD or intrahepatic bile duct. All findings pertaining to anatomical variations and diseases were reported and specified. In this study, cases were classified as indeterminate when readers were unable to determine anatomical variation with sufficient diagnostic confidence, corresponding to a qualitative score of ≤ 2 (suboptimal or unacceptable). Standard of Reference for Anatomical Variation and Diseases of the Bile Duct and Pancreas The standard of reference for biliary and pancreatic ductal anatomical variations and disease was determined by consensus between two radiologists (23 and 27 years’ experience, respectively). Anatomical variations were diagnosed based on a comprehensive evaluation of all MRCP source images (n = 65), contrast-enhanced pancreatobiliary MRI sequences (n = 35), prior imaging studies (n = 37), and available clinical information including endoscopic retrograde cholangiopancreatography (ERCP) findings (n = 13). Neoplastic lesions were confirmed by surgery (n = 13), endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) (n = 6), or follow-up imaging and clinical assessment. Non-neoplastic conditions were diagnosed using a combination of contrast-enhanced MRI and/or follow-up imaging, ERCP, and clinical information. Patients with poor breath-hold capacity were defined in consensus by two independent radiologists who were not involved in the qualitative image analysis and were analyzed as a subgroup. Statistical Analysis Continuous variables were expressed as the mean ± SD, while discrete variables for qualitative assessment were expressed as the median and interquartile range (IQR) (25–75th percentile) in parentheses. The normality of the continuous variables was assessed using the Shapiro–Wilk test. Differences among groups were analyzed using the one-way analysis of variance, followed by Tukey's honest significant difference test for post-hoc pairwise comparisons when statistically significant differences were detected. Qualitative comparisons among the three MRCP protocols were performed using Friedman's test. When significant differences were detected, post-hoc pairwise comparisons were performed using the Wilcoxon signed-rank test with Bonferroni correction (corrected α = 0.017). For diagnostic performance analysis, the proportion of “indeterminate” cases and the diagnostic accuracy of ductal variation and ductal focal lesions were compared among the three MRCP protocols using Cochran's Q test in a per-patient analysis. Inter-reader agreement among the readers was assessed using Cohen's weighted Kappa analysis (0.01–0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1.00, almost perfect agreement) [ 24 ]. Confidence intervals exceeding the theoretical upper limit were truncated at 1.0. Differences in the area under the receiver operating characteristic curve (AUROC) among the three protocols were assessed using the DeLong method, with Bonferroni correction applied for pairwise comparisons. For all statistical analyses, two-sided p values < 0.05 denoted statistical significance. All p values for post-hoc pairwise comparisons are Bonferroni-corrected. All statistical analyses were performed using the SPSS software (version 27.0; IBM Corp., Armonk, NY, USA). Results Patient Characteristics A total of 65 patients (31 men and 34 women; mean age, 68.7 ± 12.4 years, range: 34–88 years) treated at our institution were enrolled. Biliary anatomical variations identified in this cohort included posterior right hepatic duct draining into the left hepatic duct (n = 1), posterior right hepatic duct draining directly into the common hepatic duct (n = 1), and trifurcation of the hepatic ducts (n = 1). One patient had pancreas divisum as a pancreatic anatomical variation. One patient with pancreaticobiliary maljunction, classified as both a biliary disease and an anatomical variation, was included as part of ongoing follow-up surveillance. The final diagnosis included biliary malignancy (n = 11; confirmed by surgery [n = 8] or EUS-FNA [n = 3]), pancreatic neuroendocrine tumor (n = 5; surgery [n = 3] or EUS-FNA [n = 2]), pancreatic ductal adenocarcinoma (n = 3; surgery [n = 2] or EUS-FNA [n = 1]), pancreatic cystic lesion (n = 33; diagnosed using contrast-enhanced MRI [n = 20] and/or follow-up imaging [n = 21]), gallstones or bile duct stones (n = 6; detected on MRCP in 2 of 6 cases), chronic pancreatitis (n = 5; all confirmed on follow-up imaging), and primary sclerosing cholangitis (n = 1; diagnosed based on multiple biliary strictures on MRCP, ERCP, and follow-up studies). Representative cases are shown in Figs. 3 – 5 . Quantitative Evaluation The mean acquisition times were 16.3 s for TSE-DLR, 17.4 s for GRASE, and 8.9 s for GRASE-DLR. GRASE-DLR reduced acquisition time by 45.4% vs. TSE-DLR (p < 0.001) and by 49.1% vs. GRASE (p < 0.001), while no significant difference was observed between TSE-DLR and GRASE (p = 0.264). SNR was comparable across all three protocols (TSE-DLR: 8.82 ± 3.79, GRASE: 9.50 ± 3.72, GRASE-DLR: 8.44 ± 3.79; p = 0.612). TSE-DLR demonstrated superior contrast ratio (14.63 ± 5.12) and CNR (25.13 ± 9.28) compared to both GRASE (5.76 ± 2.18 and 14.66 ± 4.56, respectively; both p < 0.001) and GRASE-DLR (6.43 ± 2.88 and 16.77 ± 6.28, respectively; both p < 0.001). Between GRASE and GRASE-DLR, no significant differences were observed in contrast ratio (p = 1.000) or CNR (p = 0.183). Duct Visualization Qualitative Assessment: Overall Study Population (n = 65) GRASE-DLR achieved superior overall image quality compared with conventional protocols. The median overall image quality scores were 3–4 (2–4) for TSE-DLR and 4 (3–4) for GRASE, with GRASE-DLR achieving superior scores of 5 (4–5 to 5–5) across both readers (p < 0.001). Artefact reduction improved significantly with GRASE-DLR (median 4–5 [ 4 – 5 ]), with significant benefits over both TSE-DLR (p < 0.001) and GRASE (p = 0.003–0.034) (Table 2 ). Background suppression did not differ significantly among the three techniques, with a consistent median score of 4 across all sequences. Table 2 Comparison of the average visual scores of three MRCP protocols of all cases n = 65 TSE-DLR GRASE GRASE-DLR p value TSE-DLR vs. GRASE GRASE vs. GRASE-DLR TSE-DLR vs. GRASE-DLR Overall image quality Reader1 3 (2, 3) 4 (3, 4) 5 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 4 (4, 4) 4 (4, 4) 5 (5, 5) < 0.001 < 0.001 < 0.001 κ value 0.473 (0.310–0.644) 0.562 (0.421–0.723) 0.544 (0.407–0.708) – Artefacts Reader1 3 (2, 3) 4 (3, 4) 5 (4, 5) < 0.001 0.003 < 0.001 Reader2 3 (3, 3) 4 (4, 4) 4 (4, 4) < 0.001 0.034 < 0.001 κ value 0.667 (0.527–0.815) 0.631 (0.470–0.788) 0.730 (0.594–0.860) – Background suppression Reader1 4 (4, 4) 4 (4, 4) 4 (4, 4) 0.389 Reader2 4 (4, 4) 4 (4, 4) 4 (4, 4) 0.275 κ value 0.629 (0.443–0.761) 0.592 (0.436–0.751) 0.613 (0.485–0.753) – CBD Reader1 4 (4, 4) 4 (4, 4) 5 (4, 5) < 0.001 0.041 < 0.001 Reader2 4 (4, 4) 4 (4, 4) 5 (5, 5) < 0.001 0.016 < 0.001 κ value 0.667 (0.516–0.807) 0.588 (0.433–0.740) 0.660 (0.519–0.803) – RHD Reader1 4 (3, 4) 4 (4, 4) 5 (4, 5) 0.076 < 0.001 < 0.001 Reader2 4 (4, 4) 4 (4, 4) 5 (4, 5) 0.068 < 0.001 < 0.001 κ value 0.688 (0.519–0.817) 0.714 (0.561–0.850) 0.714 (0.553–0.855) – LHD Reader1 4 (4,4) 4 (4, 4) 5 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 4 (4, 4) 4 (4, 4) 5 (5, 5) < 0.001 < 0.001 < 0.001 κ value 0.623 (0.461–0.770) 0.642 (0.482–0.778) 0.619 (0.443–0.771) – Anterior branch Reader1 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.724 Reader2 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.874 κ value 0.904 (0.798–0.977) 0.905 (0.803–0.977) 0.903 (0.792–0.977) – Posterior branch Reader1 3 (2, 4) 3 (2 ,4) 3 (2, 4) 0.362 Reader2 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.491 κ value 0.715 (0.563–0.853) 0.735 (0.587–0.859) 0.744 (0.602–0.864) – Segment 2 branch Reader1 3 (2, 3) 3 (2, 3) 3 (2, 3) 0.621 Reader2 3 (2, 3) 3 (2, 3) 3 (2, 3) 0.430 κ value 0.700 (0.555–0.841) 0.699 (0.549–0.845) 0.715 (0.569–0.853) – Segment 3 branch Reader1 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.195 Reader2 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.293 κ value 0.621 (0.462–0.758) 0.622 (0.462–0.764) 0.651 (0.497–0.791) – Segment 4 branch Reader1 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.492 Reader2 3 (2, 4) 3 (2, 4) 3 (2, 4) 0.539 κ value 0.619 (0.471–0.712) 0.683 (0.583–0.799) 0.704 (0.531–0.795) – Cystic duct Reader1 3 (2, 3) 4 (4, 5) 5 (4, 5) < 0.001 0.078 < 0.001 Reader2 3 (2, 3) 4 (4, 5) 5 (5, 5) < 0.001 0.058 < 0.001 κ value 0.654 (0.499–0.806) 0.635 (0.467–0.789) 0.594 (0.420–0.763) – Proximal MPD Reader1 4 (3, 5) 4 (3, 5) 4 (3, 4) 0.057 Reader2 4 (3, 5) 4 (3, 5) 4 (3, 4) 0.062 κ value 0.757 (0.612–0.876) 0.763 (0.628–0.880) 0.766 (0.628–0.881) – Middle MPD Reader1 4 (3, 4) 4 (3, 4) 4 (3, 4) 0.091 Reader2 4 (3, 4) 4 (3, 4) 4 (3, 4) 0.103 κ value 0.633 (0.479–0.767) 0.634 (0.471–0.767) 0.618 (0.452–0.756) – Distal MPD Reader1 3 (3 ,4) 3 (2, 4) 3 (3, 4) 0.015 0.001 0.064 Reader2 3 (3 ,4) 3 (2, 4) 3 (3, 4) 0.026 0.001 0.071 κ value 0.699 (0.528–0.804) 0.675 (0.528–0.810) 0.651 (0.494–0.796) – Note. Data for categorical variables are expressed as the median (interquartile range). Kappa values are presented with the 95% confidence intervals shown in parentheses. Significant p-values are shown in bold. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE = Gradient and Spin-Echo without DLR, GRASE-DLR = GRASE with DLR, CBD = Common Bile Duct, RHD = Right Hepatic Duct, LHD = Left Hepatic Duct, MPD = Main Pancreatic Duct. GRASE-DLR achieved median scores of 5 (4–5 to 5–5) for CBD and left hepatic duct visualization, compared with 4 (4, 4) for both TSE-DLR and GRASE (p < 0.001 for all pairwise comparisons). For the right hepatic duct, GRASE-DLR similarly achieved a median score of 5 (4, 5), significantly superior to both TSE-DLR and GRASE (both p < 0.001), although TSE-DLR and GRASE did not differ significantly from each other (p = 0.068–0.076). Cystic duct visualization was significantly better with both GRASE (median 4, 4–5) and GRASE-DLR (median 5, 4–5) than with TSE-DLR (median 3, 2–3) (p < 0.001), whereas no significant difference was observed between GRASE and GRASE-DLR (p = 0.058–0.078). Visualization of the intrahepatic bile duct branches and the proximal and middle pancreatic duct was adequate across all techniques, with median scores of 3–4. For distal pancreatic duct visualization, GRASE-DLR was superior to GRASE (p < 0.001 for both readers), while TSE-DLR and GRASE-DLR did not differ significantly (p = 0.064–0.071). Interobserver agreement ranged from moderate to substantial across all qualitative parameters (κ = 0.473–0.905) (Table 2 ). Qualitative Assessment: Subgroup with Poor Breath-Hold (n = 8) In patients with poor breath-hold, the overall image quality scores were substantially higher with GRASE-DLR (median 4, 4–5) compared with TSE-DLR (median 2, 1–3) and GRASE (median 3, 3–4) (p < 0.001). Artefact reduction was the most pronounced advantage of DLR in this challenging subgroup. The median artefact scores reached 4 (4–5) for GRASE-DLR versus 2 (2–3) for TSE-DLR and 3 (3–4) for GRASE (p < 0.001), indicating a dramatic improvement in noise suppression in patients with motion-related image degradation. GRASE-DLR yielded clinically significant improvements in major duct visualization in this population. CBD, right, and left hepatic duct visualization improved to median scores of 5 (4–5 to 5–5) with GRASE-DLR compared with 2 (2–3) with TSE-DLR (p < 0.001). GRASE-DLR showed a particularly pronounced improvement in cystic duct visualization (median 4, 4–5) compared with TSE-DLR (median 1, 1–2) (p < 0.001). GRASE-DLR demonstrated consistent improvement for visualizing the intrahepatic bile duct branches (anterior, posterior, and segmental branches) compared with TSE-DLR, yielding median scores of 3–4 and 1–2, respectively. Pancreatic duct visualization improved from a median of 1–2 with TSE-DLR to a median of 3–4 with GRASE-DLR (p < 0.001). Interobserver agreement ranged from substantial to almost perfect across all qualitative parameters (κ = 0.702–0.912) in patients with poor breath-hold capacity (Table 3 ). Table 3 Comparison of the average visual scores of three MRCP protocols of patients with inadequate breath-hold during acquisition n = 8 TSE-DLR GRASE GRASE-DLR p value TSE-DLR vs. GRASE GRASE vs. GRASE-DLR TSE-DLR vs. GRASE-DLR Overall image quality Reader1 2 (1, 2) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 2 (1, 3) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 κ value 0.761 (0.629–0.879) 0.789 (0.655–0.831) 0.801 (0.717–0.858) – Artefacts Reader1 2 (2, 3) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 2 (2, 3) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 κ value 0.857 (0.777–0.910) 0.884 (0.801–0.930) 0.903 (0.792–0.977) – Background suppression Reader1 3 (3, 4) 4 (3, 4) 4 (4, 4) < 0.001 0.766 < 0.001 Reader2 3 (3, 4) 4 (4, 4) 4 (4, 4) < 0.001 0.612 < 0.001 κ value 0.832 (0.763–0.890) 0.800 (0.739–0.878) 0.776 (0.710–0.842) CBD Reader1 2 (2, 3) 4 (4, 4) 5 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 2 (2, 3) 4 (4, 4) 5 (5, 5) < 0.001 < 0.001 < 0.001 κ value 0.777 (0.719–0.852) 0.806 (0.748–0.866) 0.863 (0.800–0.911) – RHD Reader1 2 (2, 3) 4 (4, 4) 5 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 2 (2, 3) 4 (4, 4) 5 (5, 5) < 0.001 < 0.001 < 0.001 κ value 0.827 (0.730–0.899) 0.808 (0.718–0.864) 0.722 (0.650–0.852) – LHD Reader1 2 (2, 3) 4 (4, 4) 5 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 2 (2, 3) 4 (4, 4) 5 (5, 5) < 0.001 < 0.001 < 0.001 κ value 0.765 (0.701–0.830) 0.702 (0.622–0.779) 0.811 (0.743–0.888) – Anterior branch Reader1 1 (1, 2) 3 (2, 4) 4 (3, 4) < 0.001 < 0.001 < 0.001 Reader2 1 (1, 2) 3 (2, 4) 4 (3, 4) < 0.001 < 0.001 < 0.001 κ value 0.904 (0.798–0.977) 0.905 (0.803–0.977) 0.891 (0.791–0.959) – Posterior branch Reader1 1 (1, 2) 3 (2, 4) 3 (2, 4) < 0.001 0.054 < 0.001 Reader2 1 (1, 2) 3 (2, 4) 3 (2, 4) < 0.001 0.077 < 0.001 κ value 0.803 (0.766–0.881) 0.782 (0.717–0.844) 0.814 (0.729–0.876) – Segment 2 branch Reader1 1 (1, 2) 3 (2, 3) 3 (2, 3) < 0.001 0.063 < 0.001 Reader2 1 (1, 2) 3 (2, 3) 3 (2, 3) < 0.001 0.059 < 0.001 κ value 0.795 (0.715–0.848) 0.800 (0.743–0.856) 0.782 (0.701–0.866) – Segment 3 branch Reader1 1 (1, 2) 2 (2, 3) 3 (2, 4) < 0.001 0.582 < 0.001 Reader2 1 (1, 2) 2 (2, 3) 3 (2, 4) < 0.001 0.568 < 0.001 κ value 0.790 (0.712–0.852) 0.767 (0.702–0.840) 0.792 (0.711–0.841) – Segment 4 branch Reader1 1 (1, 2) 2 (2, 3) 3 (2, 3) < 0.001 < 0.001 < 0.001 Reader2 1 (1, 2) 2 (2, 3) 3 (2, 3) < 0.001 < 0.001 < 0.001 κ value 0.790 (0.712–0.856) 0.722 (0.667–0.801) 0.802 (0.762–0.868) Cystic duct Reader1 1 (1, 2) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 Reader2 1 (1, 2) 3 (3, 4) 4 (4, 5) < 0.001 < 0.001 < 0.001 κ value 0.836 (0.766–0.891) 0.902 (0.835–0.944) 0.912 (0.848–0.960) – Proximal MPD Reader1 2 (1, 2) 3 (2, 3) 4 (3, 4) < 0.001 < 0.001 < 0.001 Reader2 2 (1, 2) 3 (2, 3) 4 (3, 4) < 0.001 < 0.001 < 0.001 κ value 0.769 (0.699–0.836) 0.766 (0.694–0.829) 0.823 (0.748–0.899) – Middle MPD Reader1 2 (1, 2) 3 (3, 4) 4 (3, 4) < 0.001 < 0.001 < 0.001 Reader2 2 (1, 2) 3 (3, 4) 4 (3, 4) < 0.001 < 0.001 < 0.001 κ value 0.795 (0.718–0.857) 0.740 (0.679–0.793) 0.764 (0.702–0.834) – Distal MPD Reader1 1 (1, 2) 2 (2, 3) 3 (3, 4) < 0.001 < 0.001 < 0.001 Reader2 2 (1, 2) 2 (2, 3) 3 (3, 4) 0.001 < 0.001 < 0.001 κ value 0.865 (0.808–0.924) 0.772 (0.688–0.843) 0.794 (0.714–0.862) – Note. Data for categorical variables are expressed as the median (interquartile range). Kappa values are presented with the 95% confidence intervals shown in parentheses. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE = Gradient and Spin-Echo without DLR, GRASE-DLR = GRASE with DLR, CBD = Common Bile Duct, RHD = Right Hepatic Duct, LHD = Left Hepatic Duct, MPD = Main Pancreatic Duct. Diagnostic Performance Analysis GRASE-DLR significantly reduced indeterminate findings for biliary anatomical variation assessment compared with both TSE-DLR (Reader 1: 26.2% to 0%, Reader 2: 21.5% to 1.5%; both p ≤ 0.001) and GRASE (both p < 0.01). For biliary disease, GRASE-DLR improved sensitivity and accuracy, reaching statistical significance for Reader 2 (Cochran's Q: p = 0.021 and 0.010, respectively), with consistently higher AUROC values across both readers (Table 4 ). For pancreatic disease, GRASE-DLR showed a trend toward improved sensitivity and AUROC, although differences did not reach statistical significance (Cochran's Q: p = 0.084–0.162). Interobserver agreement improved progressively across protocols for both biliary disease (moderate to substantial: TSE-DLR: κ = 0.469, GRASE: κ = 0.477, GRASE-DLR: κ = 0.650) and pancreatic disease (substantial to almost perfect: TSE-DLR: κ = 0.617, GRASE: κ = 0.687, GRASE-DLR: κ = 0.901) (Table 5 ). Table 4 Number of cases for accurate evaluation of anatomical variation in three MRCP protocols Indeterminate p value TSE-DLR GRASE GRASE-DLR TSE-DLR vs. GRASE GRASE vs. GRASE-DLR TSE-DLR vs. GRASE-DLR Reader 1 26.2% (17/65) 15.4% (10/65) 0% (0/65) 0.063 0.004 < 0.001 Reader 2 21.5% (14/65) 16.9% (11/65) 1.5% (1/65) 0.571 0.003 0.001 κ value 0.879 (0.745-1.000) 0.943 (0.833-1.000) 0.999 (0.902-1.000) – Note. Data are presented as percentages, with the numbers in parentheses indicating the number of cases. Kappa values are presented with the 95% confidence intervals shown in parentheses. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE = Gradient and Spin-Echo without DLR, GRASE-DLR = GRASE with DLR. P-values that are statistically significant are shown in bold. Table 5 Diagnostic performance for anatomic variation and disease: Number of cases with accurate diagnosis and determinable cases in three MRCP protocols TSE-DLR GRASE GRASE-DLR Cochran Q p value TSE−DLR vs. GRASE GRASE vs. GRASE−DLR TSE−DLR vs. GRASE−DLR Biliary disease (n = 19) Sensitivity Reader 1 62.5 (38.6–81.5) 68.8 (44.4–85.8) 93.8 (71.7–98.9) 0.050 ― Reader 2 56.2 (33.2–76.9) 75.0 (50.5–89.8) 93.8 (71.7–98.9) 0.021 0.371 0.248 0.041 Specificity Reader 1 66.7 (20.8–93.9) 66.7 (20.8–93.9) 66.7 (20.8–93.9) 1.000 ― Reader 2 66.7 (20.8–93.9) 66.7 (20.8–93.9) 100.0 (43.8–100.0) 0.368 ― PPV Reader 1 66.7 (20.8–93.9) 66.7 (20.8–93.9) 76.7 (43.8-96.89) 1.000 ― Reader 2 66.7 (20.8–93.9) 66.7 (20.8–93.9) 100.0 (43.8–100.0) 0.368 ― NPV Reader 1 25.0 (7.1–59.1) 28.6 (8.2–64.1) 66.7 (20.8–93.9) 1.000 ― Reader 2 22.2 (6.3–54.7) 33.3 (9.7–70.0) 75.0 (30.1–95.4) 0.368 ― Accuracy Reader 1 63.2 (41.0-80.9) 68.4 (46.0-84.6) 89.5 (68.6–97.1) 0.050 ― Reader 2 57.9 (36.3–76.9) 73.7 (51.2–88.2) 94.7 (75.4–99.1) 0.010 0.371 0.134 0.023 AUROC Reader 1 0.646 (0.297–0.995) 0.677 (0.330-1.000) 0.802 (0.470-1.000) ― 0.572 0.025 0.038 Reader 2 0.615 (0.265–0.965) 0.708 (0.364-1.000) 0.969 (0.908-1.000) ― 0.168 0.135 0.047 κ value 0.469 (0.073–0.866) 0.477 (0.003–0.979) 0.650 (0.292-1.000) ― Pancreatic disease (n = 46) Sensitivity Reader 1 55.0 (39.8–69.3) 62.5 (47.0-75.8) 87.5 (73.9–94.5) 0.162 ― Reader 2 72.5 (57.2–83.9) 75.0 (59.8–85.8) 87.5 (73.9–94.5) 0.162 ― Specificity Reader 1 50.0 (18.8–81.2) 50.0 (18.8–81.2) 66.7 (30.0-90.3) 0.472 ― Reader 2 50.0 (18.8–81.2) 66.7 (30.0-90.3) 83.3 (43.6–97.0) 0.472 ― PPV Reader 1 88.0 (70.0-95.8) 89.3 (72.8–96.3) 94.6 (82.3–98.5) 0.162 ― Reader 2 90.6 (75.8–96.8) 93.8 (79.9–98.3) 97.2 (85.8–99.5) 0.162 ― NPV Reader 1 14.3 (5.0-34.6) 16.7 (5.8–39.2) 44.4 (18.9–73.3) 0.472 ― Reader 2 21.4 (7.6–47.6) 28.6 (11.7–54.6) 50.0 (23.7–76.3) 0.472 ― Accuracy Reader 1 54.3 (40.2–67.8) 60.9 (46.5–73.6) 84.8 (71.8–92.4) 0.084 ― Reader 2 69.6 (55.2–80.9) 73.9 (59.7–84.4) 87.0 (74.3–93.9) 0.084 ― AUROC Reader 1 0.613 (0.382–0.843) 0.708 (0.491–0.926) 0.854 (0.683-1.000) ― 0.304 0.362 0.158 Reader 2 0.682 (0.513–0.808) 0.743 (0.558–0.960) 0.890 (0.708-1.000) ― 0.304 0.362 0.158 κ value 0.617 (0.338–0.895) 0.687 (0.325–0.953) 0.901 (0.768-1.000) ― Note. Values are presented as percentages with 95% confidence intervals. P-values that are statistically significant are shown in bold. Kappa values are presented with the 95% confidence intervals shown in parentheses. Pairwise comparisons were performed only when the overall Cochran's Q test indicated a statistically significant difference; '―' indicates that pairwise comparisons were not performed due to a non-significant overall test. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE = Gradient and Spin-Echo without DLR, GRASE-DLR = GRASE with DLR, PPV = positive predictive value, NPV = negative predictive value, AUROC = Area Under the Receiver Operating Characteristic Curve. Comparison of 3D MRCP images: (a) TSE-DLR, (b) GRASE, and (c) GRASE-DLR. GRASE sequences (b, c) provide better visualization of the cystic duct (arrows) than TSE-DLR. GRASE-DLR (c) yields superior image quality for a pancreatic cystic lesion of the uncinate process (arrowheads) compared with GRASE (b). Discussion This study demonstrated that GRASE-DLR achieved superior qualitative image quality and improved diagnostic performance despite a 45–49% reduction in acquisition time, although quantitative CNR remained lower than that of TSE-DLR. These findings indicate that rapid acquisition combined with deep learning reconstruction can provide clinically robust MRCP imaging within a single breath-hold. Importantly, GRASE-DLR employed DLR rather than CS for reconstruction of the undersampled k-space data, allowing the specific contribution of DLR—as opposed to CS-based reconstruction—to be evaluated in the context of a rapid GRASE acquisition. Therefore, the observed improvements primarily reflect the added value of DLR when applied to a rapid GRASE acquisition framework. GRASE sequences inherently enable faster imaging by acquiring multiple k-space lines per radiofrequency pulse through the integration of gradient and spin echoes [ 13 , 14 ]. However, this acceleration is typically associated with reduced contrast and increased susceptibility to artefacts [ 13 – 15 ]. In this context, DLR appears to compensate for these limitations by suppressing noise and preserving structural continuity, thereby enabling diagnostically reliable images even under aggressive acceleration. An apparent discrepancy was observed between quantitative and qualitative findings: GRASE-DLR achieved superior perceived image quality despite lower CNR than TSE-DLR. This can be explained by both technical and perceptual factors. From a technical perspective, the higher CNR of TSE-DLR is attributable to its long echo train length, which enhances T2 weighting and fluid signal. In contrast, GRASE incorporates gradient echoes that introduce T2* decay, inherently reducing contrast [ 13 , 14 ]. However, in clinical settings, image interpretability is strongly influenced by motion artefacts and field inhomogeneity, which are more pronounced in long-echo-train TSE acquisitions. The shorter acquisition time and reduced artefact burden with GRASE-DLR likely outweigh its lower intrinsic contrast. Furthermore, the more homogeneous B0 field characteristics inherent to GRASE sequences may additionally contribute to improved image uniformity compared with long-echo-train TSE acquisitions [ 13 , 14 ]. From a perceptual standpoint, conventional metrics such as SNR and CNR do not fully capture clinically relevant image quality, including edge definition, continuity of ductal structures, and artefact suppression. DLR may enhance these perceptual features by selectively preserving anatomical information while suppressing noise [ 22 ]. Consistent with prior studies, deep learning-based MRCP techniques have demonstrated the ability to maintain or improve diagnostic image quality despite accelerated acquisition [ 16 , 18 ]. These findings support the growing recognition that conventional pixel-based metrics are insufficient for evaluating DLR-reconstructed images and that task-based or perceptual assessments may be more appropriate [ 22 ]. A key clinical finding was the marked reduction in indeterminate cases for anatomical variation assessment. The near elimination of indeterminate findings with GRASE-DLR represents a meaningful improvement in diagnostic confidence, which is particularly relevant for preoperative planning and therapeutic decision-making. Improved interobserver agreement further supports the robustness and reproducibility of this approach. The benefits of GRASE-DLR were most pronounced in patients with poor breath-hold capacity. In this subgroup, GRASE-DLR substantially improved image quality, artefact suppression, and duct visualization. These results suggest that rapid acquisition combined with DLR can mitigate motion-related degradation, thereby expanding the applicability of MRCP to patients who are traditionally challenging to image. However, given the small number of patients in this subgroup (n = 8), these findings should be interpreted as preliminary, and inter-reader agreement estimates in particular may not be statistically robust. GRASE-DLR also demonstrated improved diagnostic performance for biliary disease and a trend toward improved performance for pancreatic disease, although the latter did not reach statistical significance. These findings should be interpreted with caution given the limited sample size and heterogeneity of the study population; in particular, specificity estimates for biliary disease are unreliable due to the small number of true-negative cases (n = 3). This study has several limitations. First, this study had a retrospective single-center design with a relatively small sample size, limiting generalizability. Notably, the biliary disease cohort (n = 19) included only three true-negative cases, yielding an extremely wide 95% confidence interval for specificity (20.8–93.9%); estimates of diagnostic performance should therefore be interpreted with caution. Second, the reference standard was heterogeneous, incorporating surgical, histopathological, endoscopic, and imaging follow-up data, which may introduce verification bias. Third, differences in acquisition and acceleration strategies among the sequences may confound the isolated evaluation of DLR effects. Fourth, the subgroup analysis was limited to eight patients, and kappa estimates in this subgroup may be statistically unstable; these findings should be regarded as exploratory. Fifth, the difference in slice thickness between TSE-DLR (2.0 mm) and the GRASE-based protocols (2.4 mm) may have systematically influenced quantitative measurements including SNR and CNR, as well as qualitative scores for small ductal structures. Finally, this study was conducted using a single-vendor system and DLR algorithm, and further validation across platforms is warranted. In conclusion, GRASE-DLR enables rapid, high-quality MRCP with improved diagnostic confidence, particularly in patients with limited breath-hold capacity. By combining fast acquisition with deep learning-based reconstruction, this approach has the potential to enhance clinical workflow and broaden the applicability of MRCP in routine practice. Declarations Conflict of interest statement Jihun Kwon and Yasutomo Katsumata are employee of the Philips. Other authors declare no conflicts of interest directly relevant to the content of this article. Ethics approval The institutional review board of our institution approved this retrospective study (Approval No. 24–159). The study was conducted in accordance with the ethical standards of the IRB and with the 1964 Helsinki Declaration and its later amendments. Informed consent The requirement to obtain patients' informed consent was waived owing to the retrospective nature of the study. Fundings The authors declare that no funding was received for this work. Author Contribution All authors contributed to the study conception and design. Data curation and formal analysis were performed by Kumi Ozaki, Takafumi Iyoda, Eri Sugioka, and Yukichi Tanahashi. MRI scan protocol optimization was performed by Jihun Kwon and Yasutomo Katsumata. The first draft of the manuscript was written by Kumi Ozaki and was reviewed by Satoshi Goshima. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Acknowledgement We would like to thank Mr. Naoki Ohishi and Mr. Yoji Yamada, radiological technologists, for their technical assistance with the MRI examinations. Data Availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. References Schindera ST, Merkle EM (2007) MR cholangiopancreatography: 1.5T versus 3T. Magn Reson Imaging Clin N Am 15:355–364. https://doi.org/10.1016/j.mric.2007.06.009 Andriulli A, Loperfido S, Napolitano G, Niro G, Valvano MR, Spirito F et al (2007) Incidence rates of post-ERCP complications: a systematic survey of prospective studies. Am J Gastroenterol 102:1781–1788. https://doi.org/10.1111/j.1572-0241.2007.01279.x Nandalur KR, Hussain HK, Weadock WJ, Wamsteker EJ, Johnson TD, Khan AS et al (2008) Possible biliary disease: diagnostic performance of high-spatial-resolution isotropic 3D T2-weighted MRCP. Radiology 249:883–890. https://doi.org/10.1148/radiol.2493080389 Anupindi SA, Victoria T (2008) Magnetic resonance cholangiopancreatography: techniques and applications. Magn Reson Imaging Clin N Am 16:453–466. https://doi.org/10.1016/j.mric.2008.04.005 Yoon JH, Nickel MD, Peeters JM, Lee JM (2019) Rapid imaging: recent advances in abdominal MRI for reducing acquisition time and its clinical applications. Korean J Radiol 20:1597–1615. https://doi.org/10.3348/kjr.2018.0931 Sun B, Chen Z, Duan Q, Xue Y, Zheng E, He Y et al (2020) Rapid 3D navigator-triggered MR cholangiopancreatography with SPACE sequence at 3T: only one-third acquisition time of conventional 3D SPACE navigator-triggered MRCP. Abdom Radiol 45:134–140. https://doi.org/10.1007/s00261-019-02342-3 Chen Z, Sun B, Duan Q, Xue Y, Zheng E, He Y et al (2019) Three-dimensional breath-hold MRCP using SPACE pulse sequence at 3 T: comparison with conventional navigator-triggered technique. AJR Am J Roentgenol 213:1247–1252. https://doi.org/10.2214/AJR.19.21399 Chandarana H, Doshi AM, Shanbhogue A, Babb JS, Bruno MT, Zhao T et al (2016) Three-dimensional MR cholangiopancreatography in a breath hold with sparsity-based reconstruction of highly undersampled data. Radiology 280:585–594. https://doi.org/10.1148/radiol.2016151935 Feng L, Benkert T, Block KT, Sodickson DK, Otazo R, Chandarana H (2017) Compressed sensing for body MRI. J Magn Reson Imaging 45:966–987. https://doi.org/10.1002/jmri.25547 Yoon JH, Lee SM, Kang HJ, Weiland E, Raithel E, Son Y et al (2017) Clinical feasibility of 3-dimensional magnetic resonance cholangiopancreatography using compressed sensing: comparison of image quality and diagnostic performance. Invest Radiol 52:612–619. https://doi.org/10.1097/RLI.0000000000000380 Blaise H, Remen T, Ambarki K, Weiland E, Kuehn B, Orry X et al (2021) Comparison of respiratory-triggered 3D MR cholangiopancreatography and breath-hold compressed-sensing 3D MR cholangiopancreatography at 1.5 T and 3 T and impact of individual factors on image quality. Eur J Radiol 142:109873. https://doi.org/10.1016/j.ejrad.2021.109873 Henninger B, Steurer M, Plaikner M, Weiland E, Jaschke W, Kremser C (2020) Magnetic resonance cholangiopancreatography with compressed sensing at 1.5 T: clinical application for the evaluation of branch duct IPMN of the pancreas. Eur Radiol 30:6014–6021. https://doi.org/10.1007/s00330-020-06996-2 Yoshida M, Nakaura T, Inoue T, Tanoue S, Takada S, Utsunomiya D et al (2018) Magnetic resonance cholangiopancreatography with GRASE sequence at 3.0T: does it improve image quality and acquisition time as compared with 3D TSE? Eur Radiol 28:2436–2443. https://doi.org/10.1007/s00330-017-5240-y Nam JG, Lee JM, Kang HJ, Lee SM, Kim E, Peeters JM et al (2018) GRASE revisited: breath-hold three-dimensional (3D) magnetic resonance cholangiopancreatography using a Gradient and Spin Echo (GRASE) technique at 3T. Eur Radiol 28:3721–3728. https://doi.org/10.1007/s00330-017-5275-0 Chien CP, Chiu FM, Shen YC, Chen YH, Chung HW (2020) Magnetic resonance cholangiopancreatography at 3T in a single breath-hold: comparative effectiveness between three-dimensional (3D) gradient- and spin-echo and two-dimensional (2D) thick-slab fast spin-echo acquisitions. Quant Imaging Med Surg 10:1265–1274. https://doi.org/10.21037/qims.2020.04.14 Tajima T, Akai H, Sugawara H, Yasaka K, Kunimatsu A, Yoshioka N et al (2021) Breath-hold 3D magnetic resonance cholangiopancreatography at 1.5 T using a deep learning-based noise-reduction approach: comparison with the conventional respiratory-triggered technique. Eur J Radiol 144:109994. https://doi.org/10.1016/j.ejrad.2021.109994 Chen Z, Sun B, Xue Y, Duan Q, Zheng E, He Y et al (2021) Comparing compressed sensing breath-hold 3D MR cholangiopancreatography with two parallel imaging MRCP strategies in main pancreatic duct and common bile duct. Eur J Radiol 142:109833. https://doi.org/10.1016/j.ejrad.2021.109833 Zhang Y, Peng W, Xiao Y, Ming Y, Ma K, Hu S et al (2023) Rapid 3D breath-hold MR cholangiopancreatography using deep learning-constrained compressed sensing reconstruction. Eur Radiol 33:2500–2509. https://doi.org/10.1007/s00330-022-09227-y Pezzotti N, Yousefi S, Elmahdy MS, van Gemert J, Schülke C, Doneva M et al (2020) An adaptive intelligence algorithm for undersampled knee MRI reconstruction: application to the 2019 fastMRI Challenge. IEEE Access 8:204825–204838. https://doi.org/10.1109/ACCESS.2020.3034287 Pezzotti N, de Weerdt E, Yousefi S, Elmahdy MS, van Gemert J, Schülke C et al (2019) Adaptive-CS-Net: FastMRI with Adaptive Intelligence. arXiv. https://arxiv.org/pdf/1912.12259.pdf Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Springer, Cham, pp 234–241. https://doi.org/10.1007/978-3-319-24574-4_28 Knoll F, Hammernik K, Zhang C, Moeller S, Pock T, Sodickson DK, Akçakaya M (2020) Deep-learning methods for parallel magnetic resonance imaging reconstruction: a survey of the current approaches, trends, and issues. IEEE Signal Process Mag 37:128–140. https://doi.org/10.1109/MSP.2019.2950640 Nakaura T, Kidoh M, Maruyama N, Kawahara T, Namimoto T, Sakai Y et al (2013) Usefulness of the SPACE pulse sequence at 1.5T MR cholangiography: comparison of image quality and image acquisition time with conventional 3D-TSE sequence. J Magn Reson Imaging 38:1014–1019. https://doi.org/10.1002/jmri.24033 Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159–174. https://doi.org/10.2307/2529310 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9286494","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":618670040,"identity":"796ae6ff-6a4e-4414-9ffb-1639eda66d84","order_by":0,"name":"Kumi Ozaki","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0ElEQVRIiWNgGAWjYHACZoaEAyA6GUgakKYlLYEELQxgLTlEqufnP/zY4MEZhnz+9pxvUjcK7jDwSx+/gFeLZMMx44SEGwyWM8683SadY/CMQbIvpwCvFoODDcYHEj78N2C4kQvScpjB4AxPAn4th9k/A7UwGMjfyHlGpJZjPGCHGRjcyGGDamE/gN8vPTzFBglnGAwMzzwztgZq4QGK4NUBDLHjmyV/HGMwkDue/PB2zp/Dcvw87A/w60EHQCt4iIwgJECqLaNgFIyCUTDcAQCrBUaFgGV7kQAAAABJRU5ErkJggg==","orcid":"","institution":"Hamamatsu University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Kumi","middleName":"","lastName":"Ozaki","suffix":""},{"id":618670041,"identity":"872668bf-937a-45f2-bce4-e0d40853a6a3","order_by":1,"name":"Takafumi Iyoda","email":"","orcid":"","institution":"Hamamatsu University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Takafumi","middleName":"","lastName":"Iyoda","suffix":""},{"id":618670042,"identity":"af68586c-6649-4a14-adef-4f2a649d2686","order_by":2,"name":"Eri Sugioka","email":"","orcid":"","institution":"Hamamatsu University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Eri","middleName":"","lastName":"Sugioka","suffix":""},{"id":618670043,"identity":"6734a548-ea9d-4c1c-ac63-89946011f920","order_by":3,"name":"Jihun Kwon","email":"","orcid":"","institution":"Philips","correspondingAuthor":false,"prefix":"","firstName":"Jihun","middleName":"","lastName":"Kwon","suffix":""},{"id":618670044,"identity":"c8197c05-2a61-4e22-97f0-1168131ee119","order_by":4,"name":"Yasutomo Katsumata","email":"","orcid":"","institution":"Philips","correspondingAuthor":false,"prefix":"","firstName":"Yasutomo","middleName":"","lastName":"Katsumata","suffix":""},{"id":618670045,"identity":"cd540bd0-59be-4b9a-8972-676205cd6257","order_by":5,"name":"Yukichi Tanahashi","email":"","orcid":"","institution":"Hamamatsu University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yukichi","middleName":"","lastName":"Tanahashi","suffix":""},{"id":618670046,"identity":"48561b87-ee3d-4cd6-a8ef-d950fe3ed31a","order_by":6,"name":"Satoshi Goshima","email":"","orcid":"","institution":"Hamamatsu University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Satoshi","middleName":"","lastName":"Goshima","suffix":""}],"badges":[],"createdAt":"2026-04-01 04:38:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9286494/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9286494/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106399820,"identity":"35335346-b5c7-40f1-9ca8-53036a650757","added_by":"auto","created_at":"2026-04-08 08:32:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":83332,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of patient enrollment in the study\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/3047b4cbb4e76cf703d63959.png"},{"id":106399822,"identity":"4c69a7ff-df8a-4da5-8c6c-2370301d864d","added_by":"auto","created_at":"2026-04-08 08:32:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":303020,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative slice used for quantitative analysis in each patient\u003c/p\u003e\n\u003cp\u003eIt shows the placement of the regions of interest (ROIs) in the common bile duct (black circles) and periductal tissue (white circles). The ROI of the common bile duct was as large as possible while avoiding partial volume effects.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/a6e754018a7754e64873b20a.png"},{"id":106399821,"identity":"566be59a-8040-43d1-bcbe-9771a0bc142a","added_by":"auto","created_at":"2026-04-08 08:32:19","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":273107,"visible":true,"origin":"","legend":"\u003cp\u003eA 70-year-old woman with pancreatic cystic lesions\u003c/p\u003e\n\u003cp\u003eComparison of 3D MRCP images: (a) TSE-DLR, (b) GRASE, and (c) GRASE-DLR. GRASE sequences (b, c) provide better visualization of the cystic duct (arrows) than TSE-DLR. GRASE-DLR (c) yields superior image quality for a pancreatic cystic lesion of the uncinate process (arrowheads) compared with GRASE (b).\u003c/p\u003e\n\u003cp\u003eNote. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = turbo spin-echo with deep learning reconstruction (DLR), GRASE = gradient and spin-echo without DLR, GRASE-DLR = GRASE with DLR.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/e4276e8b5aaa00da9a21ca68.png"},{"id":106399823,"identity":"ce3d55d5-8cf0-4f97-b282-f378dca4a8ed","added_by":"auto","created_at":"2026-04-08 08:32:20","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":311307,"visible":true,"origin":"","legend":"\u003cp\u003eA 39-year-old man with primary sclerosing cholangitis\u003c/p\u003e\n\u003cp\u003eComparison of 3D MRCP images: (a) TSE-DLR, (b) GRASE, and (c) GRASE-DLR. Due to motion artefacts, image quality with TSE-DLR (a) was extremely poor, and multiple stenosis of the biliary duct was not captured. Although GRASE (b) demonstrated improved image quality compared with image (a), GRASE-DLR (c) clearly depicts multiple stenosis of the intrahepatic biliary ducts (arrowheads) and common bile duct (arrows).\u003c/p\u003e\n\u003cp\u003eNote. MRCP = magnetic resonance cholangiopancreatography, TSE-DLR = turbo spin-echo with deep learning reconstruction (DLR), GRASE = gradient and spin-echo without DLR, GRASE-DLR = GRASE with DLR.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/18ebc15ec19499cbfa0ad545.png"},{"id":106399824,"identity":"317ef7f6-21f6-49c4-8752-931cf79920f9","added_by":"auto","created_at":"2026-04-08 08:32:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":308693,"visible":true,"origin":"","legend":"\u003cp\u003eA 70-year-old woman with pancreatic cystic lesions\u003c/p\u003e\n\u003cp\u003eComparison of 3D MRCP images: (a) TSE-DLR, (b) GRASE, and (c) GRASE-DLR. Inadequate breath-holding during acquisition resulted in significant motion artefacts on both TSE-DLR (a) and GRASE (b), which precluded assessment of the common bile duct, intrahepatic bile ducts, and main pancreatic duct. Conversely, GRASE-DLR (c), requiring only brief breath-holding (\u0026lt;10 seconds), provides adequate visualization of the overall ductal system including the main pancreatic duct (arrows) despite the same challenging breathing conditions. In GRASE-DLR (c), the cystic duct is clearly visualized (arrowhead). The absence of a gallbladder signal may reflect gallbladder contraction with concentrated intraluminal bile.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/caa731c744a9b24c4dc4c647.png"},{"id":106961885,"identity":"02293853-32a5-4e49-bc4c-e89172c7179e","added_by":"auto","created_at":"2026-04-15 09:27:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3137562,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9286494/v1/a300e460-a5ac-4d21-ac22-e078fa9e73be.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"3-Dimensional Gradient and Spin-Echo Magnetic Resonance Cholangiopancreatography with Deep Learning Reconstruction at 3 T: Achieving Superior Image Quality with Reduced Acquisition Time","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMagnetic resonance cholangiopancreatography (MRCP) is an essential sequence for evaluating pancreaticobiliary disease because it facilitates noninvasive evaluation of the anatomy and abnormalities of the pancreaticobiliary tree with high spatial resolution [\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRespiratory-triggered or navigator-gated three-dimensional (3D) turbo spin-echo (TSE)-based sequences have commonly been used in 3D MRCP because they provide a higher signal-to-noise ratio (SNR) and superior spatial resolution [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However, these techniques require lengthy acquisition times to obtain images during the expiratory phase, particularly in patients with shallow or irregular breathing patterns, which often compromises image quality.\u003c/p\u003e \u003cp\u003eTo solve this problem, a single-breath-hold protocol is highly desirable and can be performed using fast imaging techniques. Therefore, various strategies have been employed to achieve better image quality in breath-hold MRCP sequences. Parallel imaging, compressed sensing (CS), or other rapid sequences are commonly employed for rapid 3D breath-hold MRCP. With parallel imaging, 3D breath-hold MRCP has become feasible, yielding acceptable image quality [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. However, parallel imaging techniques have limitations in visualizing small ductal structures, including peripheral intrahepatic bile ducts, pancreatic duct branches, and communications with cystic lesions. Additionally, parallel imaging alone has not fully achieved both short acquisition times and high image quality simultaneously [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCS, which substantially decreases the scan time through high k-space undersampling [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], is another approach for 3D breath-hold MRCP. Previous studies have shown that, compared with standard respiratory-triggered MRCP, breath-hold MRCP using CS can improve temporal resolution while maintaining image quality [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. However, in clinical practice, high acceleration factors can result in diminished image quality due to insufficient noise removal by CS reconstruction [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. The combination of gradient and spin\u0026ndash;echo sequences (GRASE) enables faster image acquisition with a more homogeneous B0 field compared with TSE-based sequences in high-field MRI, thereby reducing motion artefacts and improving image uniformity [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, despite these advantages, GRASE MRCP has been limited by inadequate visualization of small ductal structures, such as peripheral intrahepatic bile ducts and pancreatic duct branches, due to T2* decay introduced by the gradient echo component [\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRecent technological advances in MR, both in reconstruction algorithms and hardware capabilities, have aimed to address these limitations. Deep learning reconstruction (DLR)-based approaches have been developed to improve image quality from undersampled MRI k-space data, with recent studies demonstrating that TSE-MRCP with DLR achieves image quality equivalent or superior to respiratory-triggered MRCP, while substantially reducing the acquisition time [\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Concurrently, hardware advances, including enhanced SNR, improved gradient systems, and optimized field homogeneity, have enabled further improvements in GRASE MRCP image quality [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. In this context, we employed a recently developed DLR algorithm to acquire GRASE sequences with a substantially reduced acquisition time while preserving image quality.\u003c/p\u003e \u003cp\u003eTo our knowledge, no previous study has evaluated the combined use of GRASE and DLR for MRCP. Therefore, this study aimed to evaluate whether GRASE-DLR can achieve superior image quality and diagnostic performance compared with GRASE without DLR and conventional TSE-DLR, while substantially reducing acquisition time.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003eThe Research Ethics Committee of our institution approved this study (Approval No. 24\u0026ndash;159) and waived the requirement for written informed consent because of its retrospective design.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Population\u003c/h2\u003e \u003cp\u003eWe identified 70 consecutive patients who underwent abdominal 3-T MRI between July and September 2024. We excluded patients who underwent cholecystectomy (n\u0026thinsp;=\u0026thinsp;3), whose examinations were interrupted (n\u0026thinsp;=\u0026thinsp;1), and those with metal artefacts due to gastrectomy (n\u0026thinsp;=\u0026thinsp;1) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The indications for 3D MRCP were as follows: evaluation or follow-up of pancreatic cystic lesions (n\u0026thinsp;=\u0026thinsp;33), biliary malignancy (n\u0026thinsp;=\u0026thinsp;11), chronic pancreatitis (n\u0026thinsp;=\u0026thinsp;5), gallstones or bile duct stone (n\u0026thinsp;=\u0026thinsp;6), pancreatic neuroendocrine tumor (n\u0026thinsp;=\u0026thinsp;5), pancreatic ductal adenocarcinoma (n\u0026thinsp;=\u0026thinsp;3), primary sclerosing cholangitis (n\u0026thinsp;=\u0026thinsp;1), and pancreaticobiliary maljunction (n\u0026thinsp;=\u0026thinsp;1).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMRCP Parameters and Model-Based DLR\u003c/h3\u003e\n\u003cp\u003eMRI was performed using a 3-T system (MR7700; Philips Healthcare, Best, the Netherlands) equipped with a 32-channel dS Torso coil. The 3D-MRCP protocol comprised the following sequences in coronal orientation: GRASE-DLR, GRASE without DLR, and TSE-DLR. CS reconstruction was applied only to the GRASE sequence without DLR. In contrast, GRASE-DLR and TSE-DLR used a model-based DLR algorithm that incorporates CS principles within a deep learning framework, rather than conventional CS reconstruction. For the GRASE protocols, the number of spin echoes was set to 8, and the number of gradient echoes per spin echo was set to 9, resulting in an echo planar imaging (EPI) factor of 72. For GRASE-DLR, the undersampling (acceleration) factor was set to be twice that of the standard GRASE protocol recommended by the vendor to halve the acquisition time; all other imaging parameters, including the EPI factor, were kept identical between the GRASE and GRASE-DLR sequences.\u003c/p\u003e \u003cp\u003eThe following model-based DLR technique was incorporated into GRASE-DLR and TSE-DLR. The prototype CS-DLR used an Adaptive-CS-Network scheme [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], which was based on CS theory, to reconstruct the images. The network architecture was identical across all sequences, integrating a multiscale convolutional neural network with CS principles. Multiscale sparsification based on wavelet transforms was replaced with a learned representation, while domain-specific constraints, such as data consistency, were preserved [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Moreover, neighboring slices were incorporated together with the center slice as inputs to the sparsifying transform, along with several soft priors encoding MRI domain knowledge. The Adaptive-CS-Network used in this study was trained and validated on approximately 740,000 MR images of various anatomies and contrast settings.\u003c/p\u003e \u003cp\u003eThe three sequences were acquired in random order prior to contrast administration. A detailed description of the DLR reconstruction algorithm is provided in a previous report [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Identical field-of-view and voxel size were used across all 3D MRCP sequences to ensure a fair comparison. The actual acquisition time for each protocol was recorded. The imaging parameters for all MRCP sequences are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eParameters for MRCP imaging protocols\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSequence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3D_MRCP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2D_CSMRCP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2D_SRMRCP\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3D-TSE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3D-GRASE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3D-GRASE\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRespiratory compensation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBreath-hold\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBreath-hold\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBreath-hold\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e600*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRepetition time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e286\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e286\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcceleration technique\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCompressed sensing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDLR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcceleration factor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFlip angle (degree)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice thickness (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice number\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eField-of-view (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e320\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e320\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e320\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcquisition matrix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e256\u0026times;256\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e256\u0026times;256\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e256\u0026times;256\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcquisition voxel size (mm\u0026sup3;)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.25\u0026times;1.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.25\u0026times;1.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.25\u0026times;1.25\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReconstruction matrix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e512\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e512\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e512\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReconstruction voxel size (mm\u0026sup3;)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep learning reconstruction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho train length\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho planar imaging factor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e72\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eNote. MRCP\u0026thinsp;=\u0026thinsp;magnetic resonance cholangiopancreatography, TSE-DLR\u0026thinsp;=\u0026thinsp;Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE\u0026thinsp;=\u0026thinsp;Gradient and Spin-Echo without DLR, GRASE-DLR\u0026thinsp;=\u0026thinsp;GRASE with DLR. *Effective echo time for TSE-DLR.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eQuantitative Image Analysis\u003c/h3\u003e\n\u003cp\u003eA board-certified radiologist with 23 years\u0026rsquo; experience in MRCP performed the quantitative image analysis. We selected three representative slice levels [upper, middle, and lower common bile duct (CBD)] that depicted the center of the CBD in each patient. Signal intensity (SI) was measured by placing circular regions of interest (ROIs) on the CBD and periductal tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The ROIs for the SI of the CBD, which were at least 5 mm\u003csup\u003e2\u003c/sup\u003e, were placed in homogeneous, artefact-free areas in the middle third of the course of the CBD. The ROIs for the SI of the periductal tissue, which were at least 20 mm\u003csup\u003e2\u003c/sup\u003e, were placed in homogeneous, artefact-free areas adjacent to the ROI of the CBD. Image noise was defined as the standard deviation (SD) of the signal within the CBD ROI rather than background noise, as deep learning reconstruction alters the spatial noise distribution and homogeneity, rendering background noise measurements unreliable and non-representative of local image noise. To ensure methodological consistency across all three protocols, the same ROI-based SD metric was applied uniformly. The SNR, contrast ratio, and contrast-to-noise ratio (CNR) of the three MRCP protocols were calculated. The SNR of the CBD was calculated using the formula [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]:\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSNR\u0026thinsp;=\u0026thinsp;mean SI\u003csub\u003eCBD\u003c/sub\u003e/mean SD\u003csub\u003eCBD\u003c/sub\u003e\u003c/p\u003e \u003cp\u003eThe contrast between the CBD and periductal tissues on 3D MRCP was evaluated quantitatively using the following formula:\u003c/p\u003e \u003cp\u003eContrast = (mean SI\u003csub\u003eCBD\u003c/sub\u003e\u0026ndash; mean SI\u003csub\u003eperiductal tissue\u003c/sub\u003e)/(mean SI\u003csub\u003eCBD\u003c/sub\u003e + mean SI\u003csub\u003eperiductal tissue\u003c/sub\u003e)\u003c/p\u003e \u003cp\u003eThe CNR between the CBD and periductal tissues was calculated using the following formula:\u003c/p\u003e \u003cp\u003eCNR = (mean SI\u003csub\u003eCBD\u003c/sub\u003e\u0026ndash; mean SI\u003csub\u003eperiductal tissue\u003c/sub\u003e)/mean SD\u003csub\u003eCBD\u003c/sub\u003e\u003c/p\u003e\n\u003ch3\u003eQualitative Image Evaluation\u003c/h3\u003e\n\u003cp\u003eTwo independent radiologists (23 and 15 years of abdominal MRI experience), blinded to sequence type, evaluated images using five-point scales (5\u0026thinsp;=\u0026thinsp;excellent, 4\u0026thinsp;=\u0026thinsp;good, 3\u0026thinsp;=\u0026thinsp;acceptable, 2\u0026thinsp;=\u0026thinsp;suboptimal, 1\u0026thinsp;=\u0026thinsp;unacceptable) for overall image quality, artefacts, background suppression, and visibility of biliary and pancreatic duct segments [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e\n\u003ch3\u003eDiagnostic Performance Analysis\u003c/h3\u003e\n\u003cp\u003eThe same two readers independently evaluated the presence of anatomical variations and diseases of the biliary and pancreatic ducts. Specific assessments included the presence of stenosis, ductal dilatation, pancreatic cystic lesions, and filling defects attributable to stones in the CBD or intrahepatic bile duct. All findings pertaining to anatomical variations and diseases were reported and specified. In this study, cases were classified as indeterminate when readers were unable to determine anatomical variation with sufficient diagnostic confidence, corresponding to a qualitative score of \u0026le;\u0026thinsp;2 (suboptimal or unacceptable).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStandard of Reference for Anatomical Variation and Diseases of the Bile Duct and Pancreas\u003c/h2\u003e \u003cp\u003eThe standard of reference for biliary and pancreatic ductal anatomical variations and disease was determined by consensus between two radiologists (23 and 27 years\u0026rsquo; experience, respectively). Anatomical variations were diagnosed based on a comprehensive evaluation of all MRCP source images (n\u0026thinsp;=\u0026thinsp;65), contrast-enhanced pancreatobiliary MRI sequences (n\u0026thinsp;=\u0026thinsp;35), prior imaging studies (n\u0026thinsp;=\u0026thinsp;37), and available clinical information including endoscopic retrograde cholangiopancreatography (ERCP) findings (n\u0026thinsp;=\u0026thinsp;13). Neoplastic lesions were confirmed by surgery (n\u0026thinsp;=\u0026thinsp;13), endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) (n\u0026thinsp;=\u0026thinsp;6), or follow-up imaging and clinical assessment. Non-neoplastic conditions were diagnosed using a combination of contrast-enhanced MRI and/or follow-up imaging, ERCP, and clinical information.\u003c/p\u003e \u003cp\u003ePatients with poor breath-hold capacity were defined in consensus by two independent radiologists who were not involved in the qualitative image analysis and were analyzed as a subgroup.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eContinuous variables were expressed as the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD, while discrete variables for qualitative assessment were expressed as the median and interquartile range (IQR) (25\u0026ndash;75th percentile) in parentheses. The normality of the continuous variables was assessed using the Shapiro\u0026ndash;Wilk test. Differences among groups were analyzed using the one-way analysis of variance, followed by Tukey's honest significant difference test for post-hoc pairwise comparisons when statistically significant differences were detected.\u003c/p\u003e \u003cp\u003eQualitative comparisons among the three MRCP protocols were performed using Friedman's test. When significant differences were detected, post-hoc pairwise comparisons were performed using the Wilcoxon signed-rank test with Bonferroni correction (corrected α\u0026thinsp;=\u0026thinsp;0.017). For diagnostic performance analysis, the proportion of \u0026ldquo;indeterminate\u0026rdquo; cases and the diagnostic accuracy of ductal variation and ductal focal lesions were compared among the three MRCP protocols using Cochran's Q test in a per-patient analysis. Inter-reader agreement among the readers was assessed using Cohen's weighted Kappa analysis (0.01\u0026ndash;0.20, poor; 0.21\u0026ndash;0.40, fair; 0.41\u0026ndash;0.60, moderate; 0.61\u0026ndash;0.80, substantial; and 0.81\u0026ndash;1.00, almost perfect agreement) [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Confidence intervals exceeding the theoretical upper limit were truncated at 1.0. Differences in the area under the receiver operating characteristic curve (AUROC) among the three protocols were assessed using the DeLong method, with Bonferroni correction applied for pairwise comparisons. For all statistical analyses, two-sided p values\u0026thinsp;\u0026lt;\u0026thinsp;0.05 denoted statistical significance. All p values for post-hoc pairwise comparisons are Bonferroni-corrected. All statistical analyses were performed using the SPSS software (version 27.0; IBM Corp., Armonk, NY, USA).\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003ePatient Characteristics\u003c/h2\u003e \u003cp\u003eA total of 65 patients (31 men and 34 women; mean age, 68.7\u0026thinsp;\u0026plusmn;\u0026thinsp;12.4 years, range: 34\u0026ndash;88 years) treated at our institution were enrolled. Biliary anatomical variations identified in this cohort included posterior right hepatic duct draining into the left hepatic duct (n\u0026thinsp;=\u0026thinsp;1), posterior right hepatic duct draining directly into the common hepatic duct (n\u0026thinsp;=\u0026thinsp;1), and trifurcation of the hepatic ducts (n\u0026thinsp;=\u0026thinsp;1). One patient had pancreas divisum as a pancreatic anatomical variation. One patient with pancreaticobiliary maljunction, classified as both a biliary disease and an anatomical variation, was included as part of ongoing follow-up surveillance. The final diagnosis included biliary malignancy (n\u0026thinsp;=\u0026thinsp;11; confirmed by surgery [n\u0026thinsp;=\u0026thinsp;8] or EUS-FNA [n\u0026thinsp;=\u0026thinsp;3]), pancreatic neuroendocrine tumor (n\u0026thinsp;=\u0026thinsp;5; surgery [n\u0026thinsp;=\u0026thinsp;3] or EUS-FNA [n\u0026thinsp;=\u0026thinsp;2]), pancreatic ductal adenocarcinoma (n\u0026thinsp;=\u0026thinsp;3; surgery [n\u0026thinsp;=\u0026thinsp;2] or EUS-FNA [n\u0026thinsp;=\u0026thinsp;1]), pancreatic cystic lesion (n\u0026thinsp;=\u0026thinsp;33; diagnosed using contrast-enhanced MRI [n\u0026thinsp;=\u0026thinsp;20] and/or follow-up imaging [n\u0026thinsp;=\u0026thinsp;21]), gallstones or bile duct stones (n\u0026thinsp;=\u0026thinsp;6; detected on MRCP in 2 of 6 cases), chronic pancreatitis (n\u0026thinsp;=\u0026thinsp;5; all confirmed on follow-up imaging), and primary sclerosing cholangitis (n\u0026thinsp;=\u0026thinsp;1; diagnosed based on multiple biliary strictures on MRCP, ERCP, and follow-up studies). Representative cases are shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eQuantitative Evaluation\u003c/h2\u003e \u003cp\u003eThe mean acquisition times were 16.3 s for TSE-DLR, 17.4 s for GRASE, and 8.9 s for GRASE-DLR. GRASE-DLR reduced acquisition time by 45.4% vs. TSE-DLR (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and by 49.1% vs. GRASE (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while no significant difference was observed between TSE-DLR and GRASE (p\u0026thinsp;=\u0026thinsp;0.264). SNR was comparable across all three protocols (TSE-DLR: 8.82\u0026thinsp;\u0026plusmn;\u0026thinsp;3.79, GRASE: 9.50\u0026thinsp;\u0026plusmn;\u0026thinsp;3.72, GRASE-DLR: 8.44\u0026thinsp;\u0026plusmn;\u0026thinsp;3.79; p\u0026thinsp;=\u0026thinsp;0.612). TSE-DLR demonstrated superior contrast ratio (14.63\u0026thinsp;\u0026plusmn;\u0026thinsp;5.12) and CNR (25.13\u0026thinsp;\u0026plusmn;\u0026thinsp;9.28) compared to both GRASE (5.76\u0026thinsp;\u0026plusmn;\u0026thinsp;2.18 and 14.66\u0026thinsp;\u0026plusmn;\u0026thinsp;4.56, respectively; both p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and GRASE-DLR (6.43\u0026thinsp;\u0026plusmn;\u0026thinsp;2.88 and 16.77\u0026thinsp;\u0026plusmn;\u0026thinsp;6.28, respectively; both p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Between GRASE and GRASE-DLR, no significant differences were observed in contrast ratio (p\u0026thinsp;=\u0026thinsp;1.000) or CNR (p\u0026thinsp;=\u0026thinsp;0.183).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eDuct Visualization\u003c/h2\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003eQualitative Assessment: Overall Study Population (n\u0026thinsp;=\u0026thinsp;65)\u003c/h2\u003e \u003cp\u003eGRASE-DLR achieved superior overall image quality compared with conventional protocols. The median overall image quality scores were 3\u0026ndash;4 (2\u0026ndash;4) for TSE-DLR and 4 (3\u0026ndash;4) for GRASE, with GRASE-DLR achieving superior scores of 5 (4\u0026ndash;5 to 5\u0026ndash;5) across both readers (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Artefact reduction improved significantly with GRASE-DLR (median 4\u0026ndash;5 [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]), with significant benefits over both TSE-DLR (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and GRASE (p\u0026thinsp;=\u0026thinsp;0.003\u0026ndash;0.034) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Background suppression did not differ significantly among the three techniques, with a consistent median score of 4 across all sequences.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of the average visual scores of three MRCP protocols of all cases\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003en\u0026thinsp;=\u0026thinsp;65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTSE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTSE-DLR vs. GRASE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003eGRASE vs. GRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTSE-DLR vs. GRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eOverall image quality\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.473\u003c/p\u003e \u003cp\u003e(0.310\u0026ndash;0.644)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.562\u003c/p\u003e \u003cp\u003e(0.421\u0026ndash;0.723)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.544\u003c/p\u003e \u003cp\u003e(0.407\u0026ndash;0.708)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eArtefacts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.003\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (3, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.034\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003cp\u003e(0.527\u0026ndash;0.815)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.631\u003c/p\u003e \u003cp\u003e(0.470\u0026ndash;0.788)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.730\u003c/p\u003e \u003cp\u003e(0.594\u0026ndash;0.860)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eBackground suppression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.389\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.275\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.629\u003c/p\u003e \u003cp\u003e(0.443\u0026ndash;0.761)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.592\u003c/p\u003e \u003cp\u003e(0.436\u0026ndash;0.751)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.613\u003c/p\u003e \u003cp\u003e(0.485\u0026ndash;0.753)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eCBD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.041\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.016\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003cp\u003e(0.516\u0026ndash;0.807)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.588\u003c/p\u003e \u003cp\u003e(0.433\u0026ndash;0.740)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.660\u003c/p\u003e \u003cp\u003e(0.519\u0026ndash;0.803)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eRHD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.076\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.068\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.688\u003c/p\u003e \u003cp\u003e(0.519\u0026ndash;0.817)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.714\u003c/p\u003e \u003cp\u003e(0.561\u0026ndash;0.850)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.714\u003c/p\u003e \u003cp\u003e(0.553\u0026ndash;0.855)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eLHD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003cp\u003e(0.461\u0026ndash;0.770)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.642\u003c/p\u003e \u003cp\u003e(0.482\u0026ndash;0.778)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.619\u003c/p\u003e \u003cp\u003e(0.443\u0026ndash;0.771)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eAnterior branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.724\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.874\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.904\u003c/p\u003e \u003cp\u003e(0.798\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.905\u003c/p\u003e \u003cp\u003e(0.803\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003cp\u003e(0.792\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003ePosterior branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2 ,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.362\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.491\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.715\u003c/p\u003e \u003cp\u003e(0.563\u0026ndash;0.853)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.735\u003c/p\u003e \u003cp\u003e(0.587\u0026ndash;0.859)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003cp\u003e(0.602\u0026ndash;0.864)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 2 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.621\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.430\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.700\u003c/p\u003e \u003cp\u003e(0.555\u0026ndash;0.841)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.699\u003c/p\u003e \u003cp\u003e(0.549\u0026ndash;0.845)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.715\u003c/p\u003e \u003cp\u003e(0.569\u0026ndash;0.853)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 3 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.195\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.293\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.621\u003c/p\u003e \u003cp\u003e(0.462\u0026ndash;0.758)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.622\u003c/p\u003e \u003cp\u003e(0.462\u0026ndash;0.764)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.651\u003c/p\u003e \u003cp\u003e(0.497\u0026ndash;0.791)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 4 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.492\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.539\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.619\u003c/p\u003e \u003cp\u003e(0.471\u0026ndash;0.712)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.683\u003c/p\u003e \u003cp\u003e(0.583\u0026ndash;0.799)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.704\u003c/p\u003e \u003cp\u003e(0.531\u0026ndash;0.795)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eCystic duct\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e0.078\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003e0.058\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.654\u003c/p\u003e \u003cp\u003e(0.499\u0026ndash;0.806)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.635\u003c/p\u003e \u003cp\u003e(0.467\u0026ndash;0.789)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.594\u003c/p\u003e \u003cp\u003e(0.420\u0026ndash;0.763)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eProximal MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (3, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.057\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (3, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.062\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003cp\u003e(0.612\u0026ndash;0.876)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.763\u003c/p\u003e \u003cp\u003e(0.628\u0026ndash;0.880)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.766\u003c/p\u003e \u003cp\u003e(0.628\u0026ndash;0.881)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eMiddle MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.091\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e0.103\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003cp\u003e(0.479\u0026ndash;0.767)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003cp\u003e(0.471\u0026ndash;0.767)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.618\u003c/p\u003e \u003cp\u003e(0.452\u0026ndash;0.756)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eDistal MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (3 ,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.015\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.064\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (3 ,4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.026\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.071\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.699\u003c/p\u003e \u003cp\u003e(0.528\u0026ndash;0.804)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.675\u003c/p\u003e \u003cp\u003e(0.528\u0026ndash;0.810)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.651\u003c/p\u003e \u003cp\u003e(0.494\u0026ndash;0.796)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eNote. Data for categorical variables are expressed as the median (interquartile range). Kappa values are presented with the 95% confidence intervals shown in parentheses. Significant p-values are shown in bold. MRCP\u0026thinsp;=\u0026thinsp;magnetic resonance cholangiopancreatography, TSE-DLR\u0026thinsp;=\u0026thinsp;Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE\u0026thinsp;=\u0026thinsp;Gradient and Spin-Echo without DLR, GRASE-DLR\u0026thinsp;=\u0026thinsp;GRASE with DLR, CBD\u0026thinsp;=\u0026thinsp;Common Bile Duct, RHD\u0026thinsp;=\u0026thinsp;Right Hepatic Duct, LHD\u0026thinsp;=\u0026thinsp;Left Hepatic Duct, MPD\u0026thinsp;=\u0026thinsp;Main Pancreatic Duct.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eGRASE-DLR achieved median scores of 5 (4\u0026ndash;5 to 5\u0026ndash;5) for CBD and left hepatic duct visualization, compared with 4 (4, 4) for both TSE-DLR and GRASE (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 for all pairwise comparisons). For the right hepatic duct, GRASE-DLR similarly achieved a median score of 5 (4, 5), significantly superior to both TSE-DLR and GRASE (both p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), although TSE-DLR and GRASE did not differ significantly from each other (p\u0026thinsp;=\u0026thinsp;0.068\u0026ndash;0.076). Cystic duct visualization was significantly better with both GRASE (median 4, 4\u0026ndash;5) and GRASE-DLR (median 5, 4\u0026ndash;5) than with TSE-DLR (median 3, 2\u0026ndash;3) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), whereas no significant difference was observed between GRASE and GRASE-DLR (p\u0026thinsp;=\u0026thinsp;0.058\u0026ndash;0.078). Visualization of the intrahepatic bile duct branches and the proximal and middle pancreatic duct was adequate across all techniques, with median scores of 3\u0026ndash;4. For distal pancreatic duct visualization, GRASE-DLR was superior to GRASE (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 for both readers), while TSE-DLR and GRASE-DLR did not differ significantly (p\u0026thinsp;=\u0026thinsp;0.064\u0026ndash;0.071). Interobserver agreement ranged from moderate to substantial across all qualitative parameters (κ\u0026thinsp;=\u0026thinsp;0.473\u0026ndash;0.905) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eQualitative Assessment: Subgroup with Poor Breath-Hold (n\u0026thinsp;=\u0026thinsp;8)\u003c/h2\u003e \u003cp\u003eIn patients with poor breath-hold, the overall image quality scores were substantially higher with GRASE-DLR (median 4, 4\u0026ndash;5) compared with TSE-DLR (median 2, 1\u0026ndash;3) and GRASE (median 3, 3\u0026ndash;4) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Artefact reduction was the most pronounced advantage of DLR in this challenging subgroup. The median artefact scores reached 4 (4\u0026ndash;5) for GRASE-DLR versus 2 (2\u0026ndash;3) for TSE-DLR and 3 (3\u0026ndash;4) for GRASE (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating a dramatic improvement in noise suppression in patients with motion-related image degradation. GRASE-DLR yielded clinically significant improvements in major duct visualization in this population. CBD, right, and left hepatic duct visualization improved to median scores of 5 (4\u0026ndash;5 to 5\u0026ndash;5) with GRASE-DLR compared with 2 (2\u0026ndash;3) with TSE-DLR (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). GRASE-DLR showed a particularly pronounced improvement in cystic duct visualization (median 4, 4\u0026ndash;5) compared with TSE-DLR (median 1, 1\u0026ndash;2) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). GRASE-DLR demonstrated consistent improvement for visualizing the intrahepatic bile duct branches (anterior, posterior, and segmental branches) compared with TSE-DLR, yielding median scores of 3\u0026ndash;4 and 1\u0026ndash;2, respectively. Pancreatic duct visualization improved from a median of 1\u0026ndash;2 with TSE-DLR to a median of 3\u0026ndash;4 with GRASE-DLR (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Interobserver agreement ranged from substantial to almost perfect across all qualitative parameters (κ\u0026thinsp;=\u0026thinsp;0.702\u0026ndash;0.912) in patients with poor breath-hold capacity (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of the average visual scores of three MRCP protocols of patients with inadequate breath-hold during acquisition\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003en\u0026thinsp;=\u0026thinsp;8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTSE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTSE-DLR vs. GRASE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eGRASE vs. GRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTSE-DLR vs. GRASE-DLR\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eOverall image quality\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.761\u003c/p\u003e \u003cp\u003e(0.629\u0026ndash;0.879)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.789\u003c/p\u003e \u003cp\u003e(0.655\u0026ndash;0.831)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.801\u003c/p\u003e \u003cp\u003e(0.717\u0026ndash;0.858)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eArtefacts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.857\u003c/p\u003e \u003cp\u003e(0.777\u0026ndash;0.910)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.884\u003c/p\u003e \u003cp\u003e(0.801\u0026ndash;0.930)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003cp\u003e(0.792\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eBackground suppression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.766\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.832\u003c/p\u003e \u003cp\u003e(0.763\u0026ndash;0.890)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003cp\u003e(0.739\u0026ndash;0.878)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.776\u003c/p\u003e \u003cp\u003e(0.710\u0026ndash;0.842)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eCBD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.777\u003c/p\u003e \u003cp\u003e(0.719\u0026ndash;0.852)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.806\u003c/p\u003e \u003cp\u003e(0.748\u0026ndash;0.866)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.863\u003c/p\u003e \u003cp\u003e(0.800\u0026ndash;0.911)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eRHD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.827\u003c/p\u003e \u003cp\u003e(0.730\u0026ndash;0.899)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003cp\u003e(0.718\u0026ndash;0.864)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003cp\u003e(0.650\u0026ndash;0.852)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eLHD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (4, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (5, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.765\u003c/p\u003e \u003cp\u003e(0.701\u0026ndash;0.830)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.702\u003c/p\u003e \u003cp\u003e(0.622\u0026ndash;0.779)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.811\u003c/p\u003e \u003cp\u003e(0.743\u0026ndash;0.888)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eAnterior branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.904\u003c/p\u003e \u003cp\u003e(0.798\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.905\u003c/p\u003e \u003cp\u003e(0.803\u0026ndash;0.977)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.891\u003c/p\u003e \u003cp\u003e(0.791\u0026ndash;0.959)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003ePosterior branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.803\u003c/p\u003e \u003cp\u003e(0.766\u0026ndash;0.881)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003cp\u003e(0.717\u0026ndash;0.844)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.814\u003c/p\u003e \u003cp\u003e(0.729\u0026ndash;0.876)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 2 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.063\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.059\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.795\u003c/p\u003e \u003cp\u003e(0.715\u0026ndash;0.848)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.800\u003c/p\u003e \u003cp\u003e(0.743\u0026ndash;0.856)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003cp\u003e(0.701\u0026ndash;0.866)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 3 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.582\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.568\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.790\u003c/p\u003e \u003cp\u003e(0.712\u0026ndash;0.852)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.767\u003c/p\u003e \u003cp\u003e(0.702\u0026ndash;0.840)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.792\u003c/p\u003e \u003cp\u003e(0.711\u0026ndash;0.841)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSegment 4 branch\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.790\u003c/p\u003e \u003cp\u003e(0.712\u0026ndash;0.856)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003cp\u003e(0.667\u0026ndash;0.801)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.802\u003c/p\u003e \u003cp\u003e(0.762\u0026ndash;0.868)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eCystic duct\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (4, 5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.836\u003c/p\u003e \u003cp\u003e(0.766\u0026ndash;0.891)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.902\u003c/p\u003e \u003cp\u003e(0.835\u0026ndash;0.944)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.912\u003c/p\u003e \u003cp\u003e(0.848\u0026ndash;0.960)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eProximal MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.769\u003c/p\u003e \u003cp\u003e(0.699\u0026ndash;0.836)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.766\u003c/p\u003e \u003cp\u003e(0.694\u0026ndash;0.829)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.823\u003c/p\u003e \u003cp\u003e(0.748\u0026ndash;0.899)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eMiddle MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.795\u003c/p\u003e \u003cp\u003e(0.718\u0026ndash;0.857)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.740\u003c/p\u003e \u003cp\u003e(0.679\u0026ndash;0.793)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.764\u003c/p\u003e \u003cp\u003e(0.702\u0026ndash;0.834)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eDistal MPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (1, 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (3, 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.865\u003c/p\u003e \u003cp\u003e(0.808\u0026ndash;0.924)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.772\u003c/p\u003e \u003cp\u003e(0.688\u0026ndash;0.843)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.794\u003c/p\u003e \u003cp\u003e(0.714\u0026ndash;0.862)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"8\"\u003eNote. Data for categorical variables are expressed as the median (interquartile range). Kappa values are presented with the 95% confidence intervals shown in parentheses.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"8\"\u003eMRCP\u0026thinsp;=\u0026thinsp;magnetic resonance cholangiopancreatography, TSE-DLR\u0026thinsp;=\u0026thinsp;Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE\u0026thinsp;=\u0026thinsp;Gradient and Spin-Echo without DLR, GRASE-DLR\u0026thinsp;=\u0026thinsp;GRASE with DLR, CBD\u0026thinsp;=\u0026thinsp;Common Bile Duct, RHD\u0026thinsp;=\u0026thinsp;Right Hepatic Duct, LHD\u0026thinsp;=\u0026thinsp;Left Hepatic Duct, MPD\u0026thinsp;=\u0026thinsp;Main Pancreatic Duct.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eDiagnostic Performance Analysis\u003c/h2\u003e \u003cp\u003eGRASE-DLR significantly reduced indeterminate findings for biliary anatomical variation assessment compared with both TSE-DLR (Reader 1: 26.2% to 0%, Reader 2: 21.5% to 1.5%; both p\u0026thinsp;\u0026le;\u0026thinsp;0.001) and GRASE (both p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). For biliary disease, GRASE-DLR improved sensitivity and accuracy, reaching statistical significance for Reader 2 (Cochran's Q: p\u0026thinsp;=\u0026thinsp;0.021 and 0.010, respectively), with consistently higher AUROC values across both readers (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). For pancreatic disease, GRASE-DLR showed a trend toward improved sensitivity and AUROC, although differences did not reach statistical significance (Cochran's Q: p\u0026thinsp;=\u0026thinsp;0.084\u0026ndash;0.162). Interobserver agreement improved progressively across protocols for both biliary disease (moderate to substantial: TSE-DLR: κ\u0026thinsp;=\u0026thinsp;0.469, GRASE: κ\u0026thinsp;=\u0026thinsp;0.477, GRASE-DLR: κ\u0026thinsp;=\u0026thinsp;0.650) and pancreatic disease (substantial to almost perfect: TSE-DLR: κ\u0026thinsp;=\u0026thinsp;0.617, GRASE: κ\u0026thinsp;=\u0026thinsp;0.687, GRASE-DLR: κ\u0026thinsp;=\u0026thinsp;0.901) (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eNumber of cases for accurate evaluation of anatomical variation in three MRCP protocols\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eIndeterminate\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTSE-DLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGRASE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGRASE-DLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTSE-DLR vs. GRASE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGRASE vs. GRASE-DLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTSE-DLR vs. GRASE-DLR\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.2% (17/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.4% (10/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0% (0/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.063\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.004\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21.5% (14/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.9% (11/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5% (1/65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.571\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.003\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.879\u003c/p\u003e \u003cp\u003e(0.745-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.943\u003c/p\u003e \u003cp\u003e(0.833-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.999\u003c/p\u003e \u003cp\u003e(0.902-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote. Data are presented as percentages, with the numbers in parentheses indicating the number of cases. Kappa values are presented with the 95% confidence intervals shown in parentheses.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eMRCP\u0026thinsp;=\u0026thinsp;magnetic resonance cholangiopancreatography, TSE-DLR\u0026thinsp;=\u0026thinsp;Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE\u0026thinsp;=\u0026thinsp;Gradient and Spin-Echo without DLR, GRASE-DLR\u0026thinsp;=\u0026thinsp;GRASE with DLR. P-values that are statistically significant are shown in bold.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDiagnostic performance for anatomic variation and disease: Number of cases with accurate diagnosis and determinable cases in three MRCP protocols\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTSE-DLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGRASE-DLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCochran Q\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003csup\u003eTSE\u0026minus;DLR vs. GRASE\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003csup\u003eGRASE vs. GRASE\u0026minus;DLR\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003csup\u003eTSE\u0026minus;DLR vs. GRASE\u0026minus;DLR\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"9\" nameend=\"c9\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBiliary disease (n\u0026thinsp;=\u0026thinsp;19)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e62.5\u003c/p\u003e \u003cp\u003e(38.6\u0026ndash;81.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68.8\u003c/p\u003e \u003cp\u003e(44.4\u0026ndash;85.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e93.8\u003c/p\u003e \u003cp\u003e(71.7\u0026ndash;98.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.050\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e56.2\u003c/p\u003e \u003cp\u003e(33.2\u0026ndash;76.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.0 (50.5\u0026ndash;89.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e93.8\u003c/p\u003e \u003cp\u003e(71.7\u0026ndash;98.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.021\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.371\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.248\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0.041\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100.0\u003c/p\u003e \u003cp\u003e(43.8\u0026ndash;100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e76.7\u003c/p\u003e \u003cp\u003e(43.8-96.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100.0\u003c/p\u003e \u003cp\u003e(43.8\u0026ndash;100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25.0\u003c/p\u003e \u003cp\u003e(7.1\u0026ndash;59.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e28.6\u003c/p\u003e \u003cp\u003e(8.2\u0026ndash;64.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(20.8\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.2\u003c/p\u003e \u003cp\u003e(6.3\u0026ndash;54.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e33.3\u003c/p\u003e \u003cp\u003e(9.7\u0026ndash;70.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e75.0\u003c/p\u003e \u003cp\u003e(30.1\u0026ndash;95.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63.2\u003c/p\u003e \u003cp\u003e(41.0-80.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68.4\u003c/p\u003e \u003cp\u003e(46.0-84.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e89.5\u003c/p\u003e \u003cp\u003e(68.6\u0026ndash;97.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.050\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e57.9\u003c/p\u003e \u003cp\u003e(36.3\u0026ndash;76.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73.7\u003c/p\u003e \u003cp\u003e(51.2\u0026ndash;88.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e94.7\u003c/p\u003e \u003cp\u003e(75.4\u0026ndash;99.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.010\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.371\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0.023\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAUROC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.646\u003c/p\u003e \u003cp\u003e(0.297\u0026ndash;0.995)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.677\u003c/p\u003e \u003cp\u003e(0.330-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.802\u003c/p\u003e \u003cp\u003e(0.470-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.572\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0.025\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0.038\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.615\u003c/p\u003e \u003cp\u003e(0.265\u0026ndash;0.965)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.708\u003c/p\u003e \u003cp\u003e(0.364-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.969\u003c/p\u003e \u003cp\u003e(0.908-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.135\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0.047\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.469\u003c/p\u003e \u003cp\u003e(0.073\u0026ndash;0.866)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.477\u003c/p\u003e \u003cp\u003e(0.003\u0026ndash;0.979)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.650\u003c/p\u003e \u003cp\u003e(0.292-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"9\" nameend=\"c9\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePancreatic disease (n\u0026thinsp;=\u0026thinsp;46)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e55.0\u003c/p\u003e \u003cp\u003e(39.8\u0026ndash;69.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e62.5\u003c/p\u003e \u003cp\u003e(47.0-75.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e87.5\u003c/p\u003e \u003cp\u003e(73.9\u0026ndash;94.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.162\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e72.5\u003c/p\u003e \u003cp\u003e(57.2\u0026ndash;83.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.0\u003c/p\u003e \u003cp\u003e(59.8\u0026ndash;85.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e87.5\u003c/p\u003e \u003cp\u003e(73.9\u0026ndash;94.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.162\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.0\u003c/p\u003e \u003cp\u003e(18.8\u0026ndash;81.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50.0\u003c/p\u003e \u003cp\u003e(18.8\u0026ndash;81.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(30.0-90.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.0\u003c/p\u003e \u003cp\u003e(18.8\u0026ndash;81.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66.7\u003c/p\u003e \u003cp\u003e(30.0-90.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e83.3\u003c/p\u003e \u003cp\u003e(43.6\u0026ndash;97.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e88.0\u003c/p\u003e \u003cp\u003e(70.0-95.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e89.3\u003c/p\u003e \u003cp\u003e(72.8\u0026ndash;96.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e94.6\u003c/p\u003e \u003cp\u003e(82.3\u0026ndash;98.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.162\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e90.6\u003c/p\u003e \u003cp\u003e(75.8\u0026ndash;96.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e93.8\u003c/p\u003e \u003cp\u003e(79.9\u0026ndash;98.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e97.2\u003c/p\u003e \u003cp\u003e(85.8\u0026ndash;99.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.162\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.3\u003c/p\u003e \u003cp\u003e(5.0-34.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.7\u003c/p\u003e \u003cp\u003e(5.8\u0026ndash;39.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e44.4\u003c/p\u003e \u003cp\u003e(18.9\u0026ndash;73.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21.4\u003c/p\u003e \u003cp\u003e(7.6\u0026ndash;47.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e28.6\u003c/p\u003e \u003cp\u003e(11.7\u0026ndash;54.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e50.0\u003c/p\u003e \u003cp\u003e(23.7\u0026ndash;76.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.472\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.3\u003c/p\u003e \u003cp\u003e(40.2\u0026ndash;67.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60.9\u003c/p\u003e \u003cp\u003e(46.5\u0026ndash;73.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e84.8\u003c/p\u003e \u003cp\u003e(71.8\u0026ndash;92.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.084\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e69.6\u003c/p\u003e \u003cp\u003e(55.2\u0026ndash;80.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73.9\u003c/p\u003e \u003cp\u003e(59.7\u0026ndash;84.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e87.0\u003c/p\u003e \u003cp\u003e(74.3\u0026ndash;93.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.084\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c9\" namest=\"c7\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAUROC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.613\u003c/p\u003e \u003cp\u003e(0.382\u0026ndash;0.843)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.708\u003c/p\u003e \u003cp\u003e(0.491\u0026ndash;0.926)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.854\u003c/p\u003e \u003cp\u003e(0.683-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.304\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.362\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.158\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.682\u003c/p\u003e \u003cp\u003e(0.513\u0026ndash;0.808)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.743\u003c/p\u003e \u003cp\u003e(0.558\u0026ndash;0.960)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.890\u003c/p\u003e \u003cp\u003e(0.708-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.304\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.362\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.158\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eκ value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003cp\u003e(0.338\u0026ndash;0.895)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.687\u003c/p\u003e \u003cp\u003e(0.325\u0026ndash;0.953)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.901\u003c/p\u003e \u003cp\u003e(0.768-1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003e―\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eNote. Values are presented as percentages with 95% confidence intervals. P-values that are statistically significant are shown in bold. Kappa values are presented with the 95% confidence intervals shown in parentheses. Pairwise comparisons were performed only when the overall Cochran's Q test indicated a statistically significant difference; '―' indicates that pairwise comparisons were not performed due to a non-significant overall test.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eMRCP\u0026thinsp;=\u0026thinsp;magnetic resonance cholangiopancreatography, TSE-DLR\u0026thinsp;=\u0026thinsp;Turbo Spin-Echo with Deep Learning Reconstruction (DLR), GRASE\u0026thinsp;=\u0026thinsp;Gradient and Spin-Echo without DLR, GRASE-DLR\u0026thinsp;=\u0026thinsp;GRASE with DLR, PPV\u0026thinsp;=\u0026thinsp;positive predictive value, NPV\u0026thinsp;=\u0026thinsp;negative predictive value, AUROC\u0026thinsp;=\u0026thinsp;Area Under the Receiver Operating Characteristic Curve.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eComparison of 3D MRCP images: (a) TSE-DLR, (b) GRASE, and (c) GRASE-DLR. GRASE sequences (b, c) provide better visualization of the cystic duct (arrows) than TSE-DLR. GRASE-DLR (c) yields superior image quality for a pancreatic cystic lesion of the uncinate process (arrowheads) compared with GRASE (b).\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study demonstrated that GRASE-DLR achieved superior qualitative image quality and improved diagnostic performance despite a 45\u0026ndash;49% reduction in acquisition time, although quantitative CNR remained lower than that of TSE-DLR. These findings indicate that rapid acquisition combined with deep learning reconstruction can provide clinically robust MRCP imaging within a single breath-hold.\u003c/p\u003e \u003cp\u003eImportantly, GRASE-DLR employed DLR rather than CS for reconstruction of the undersampled k-space data, allowing the specific contribution of DLR\u0026mdash;as opposed to CS-based reconstruction\u0026mdash;to be evaluated in the context of a rapid GRASE acquisition. Therefore, the observed improvements primarily reflect the added value of DLR when applied to a rapid GRASE acquisition framework.\u003c/p\u003e \u003cp\u003eGRASE sequences inherently enable faster imaging by acquiring multiple k-space lines per radiofrequency pulse through the integration of gradient and spin echoes [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, this acceleration is typically associated with reduced contrast and increased susceptibility to artefacts [\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. In this context, DLR appears to compensate for these limitations by suppressing noise and preserving structural continuity, thereby enabling diagnostically reliable images even under aggressive acceleration.\u003c/p\u003e \u003cp\u003eAn apparent discrepancy was observed between quantitative and qualitative findings: GRASE-DLR achieved superior perceived image quality despite lower CNR than TSE-DLR. This can be explained by both technical and perceptual factors. From a technical perspective, the higher CNR of TSE-DLR is attributable to its long echo train length, which enhances T2 weighting and fluid signal. In contrast, GRASE incorporates gradient echoes that introduce T2* decay, inherently reducing contrast [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, in clinical settings, image interpretability is strongly influenced by motion artefacts and field inhomogeneity, which are more pronounced in long-echo-train TSE acquisitions. The shorter acquisition time and reduced artefact burden with GRASE-DLR likely outweigh its lower intrinsic contrast. Furthermore, the more homogeneous B0 field characteristics inherent to GRASE sequences may additionally contribute to improved image uniformity compared with long-echo-train TSE acquisitions [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eFrom a perceptual standpoint, conventional metrics such as SNR and CNR do not fully capture clinically relevant image quality, including edge definition, continuity of ductal structures, and artefact suppression. DLR may enhance these perceptual features by selectively preserving anatomical information while suppressing noise [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Consistent with prior studies, deep learning-based MRCP techniques have demonstrated the ability to maintain or improve diagnostic image quality despite accelerated acquisition [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. These findings support the growing recognition that conventional pixel-based metrics are insufficient for evaluating DLR-reconstructed images and that task-based or perceptual assessments may be more appropriate [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eA key clinical finding was the marked reduction in indeterminate cases for anatomical variation assessment. The near elimination of indeterminate findings with GRASE-DLR represents a meaningful improvement in diagnostic confidence, which is particularly relevant for preoperative planning and therapeutic decision-making. Improved interobserver agreement further supports the robustness and reproducibility of this approach.\u003c/p\u003e \u003cp\u003eThe benefits of GRASE-DLR were most pronounced in patients with poor breath-hold capacity. In this subgroup, GRASE-DLR substantially improved image quality, artefact suppression, and duct visualization. These results suggest that rapid acquisition combined with DLR can mitigate motion-related degradation, thereby expanding the applicability of MRCP to patients who are traditionally challenging to image. However, given the small number of patients in this subgroup (n\u0026thinsp;=\u0026thinsp;8), these findings should be interpreted as preliminary, and inter-reader agreement estimates in particular may not be statistically robust.\u003c/p\u003e \u003cp\u003eGRASE-DLR also demonstrated improved diagnostic performance for biliary disease and a trend toward improved performance for pancreatic disease, although the latter did not reach statistical significance. These findings should be interpreted with caution given the limited sample size and heterogeneity of the study population; in particular, specificity estimates for biliary disease are unreliable due to the small number of true-negative cases (n\u0026thinsp;=\u0026thinsp;3).\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, this study had a retrospective single-center design with a relatively small sample size, limiting generalizability. Notably, the biliary disease cohort (n\u0026thinsp;=\u0026thinsp;19) included only three true-negative cases, yielding an extremely wide 95% confidence interval for specificity (20.8\u0026ndash;93.9%); estimates of diagnostic performance should therefore be interpreted with caution. Second, the reference standard was heterogeneous, incorporating surgical, histopathological, endoscopic, and imaging follow-up data, which may introduce verification bias. Third, differences in acquisition and acceleration strategies among the sequences may confound the isolated evaluation of DLR effects. Fourth, the subgroup analysis was limited to eight patients, and kappa estimates in this subgroup may be statistically unstable; these findings should be regarded as exploratory. Fifth, the difference in slice thickness between TSE-DLR (2.0 mm) and the GRASE-based protocols (2.4 mm) may have systematically influenced quantitative measurements including SNR and CNR, as well as qualitative scores for small ductal structures. Finally, this study was conducted using a single-vendor system and DLR algorithm, and further validation across platforms is warranted.\u003c/p\u003e \u003cp\u003eIn conclusion, GRASE-DLR enables rapid, high-quality MRCP with improved diagnostic confidence, particularly in patients with limited breath-hold capacity. By combining fast acquisition with deep learning-based reconstruction, this approach has the potential to enhance clinical workflow and broaden the applicability of MRCP in routine practice.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eConflict of interest statement\u003c/h2\u003e\n\u003cp\u003eJihun Kwon and Yasutomo Katsumata are employee of the Philips. Other authors declare no conflicts of interest directly relevant to the content of this article.\u003c/p\u003e\n\u003ch2\u003eEthics approval\u003c/h2\u003e\n\u003cp\u003eThe institutional review board of our institution approved this retrospective study (Approval No. 24\u0026ndash;159). The study was conducted in accordance with the ethical standards of the IRB and with the 1964 Helsinki Declaration and its later amendments.\u003c/p\u003e\n\u003ch2\u003eInformed consent\u003c/h2\u003e\n\u003cp\u003eThe requirement to obtain patients\u0026apos; informed consent was waived owing to the retrospective nature of the study.\u003c/p\u003e\n\u003ch2\u003eFundings\u003c/h2\u003e\n\u003cp\u003eThe authors declare that no funding was received for this work.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eAll authors contributed to the study conception and design. Data curation and formal analysis were performed by Kumi Ozaki, Takafumi Iyoda, Eri Sugioka, and Yukichi Tanahashi. MRI scan protocol optimization was performed by Jihun Kwon and Yasutomo Katsumata. The first draft of the manuscript was written by Kumi Ozaki and was reviewed by Satoshi Goshima. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003ch2\u003eAcknowledgement\u003c/h2\u003e\n\u003cp\u003eWe would like to thank Mr. Naoki Ohishi and Mr. Yoji Yamada, radiological technologists, for their technical assistance with the MRI examinations.\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSchindera ST, Merkle EM (2007) MR cholangiopancreatography: 1.5T versus 3T. Magn Reson Imaging Clin N Am 15:355\u0026ndash;364. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.mric.2007.06.009\u003c/span\u003e\u003cspan address=\"10.1016/j.mric.2007.06.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndriulli A, Loperfido S, Napolitano G, Niro G, Valvano MR, Spirito F et al (2007) Incidence rates of post-ERCP complications: a systematic survey of prospective studies. Am J Gastroenterol 102:1781\u0026ndash;1788. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1572-0241.2007.01279.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1572-0241.2007.01279.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNandalur KR, Hussain HK, Weadock WJ, Wamsteker EJ, Johnson TD, Khan AS et al (2008) Possible biliary disease: diagnostic performance of high-spatial-resolution isotropic 3D T2-weighted MRCP. Radiology 249:883\u0026ndash;890. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/radiol.2493080389\u003c/span\u003e\u003cspan address=\"10.1148/radiol.2493080389\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAnupindi SA, Victoria T (2008) Magnetic resonance cholangiopancreatography: techniques and applications. Magn Reson Imaging Clin N Am 16:453\u0026ndash;466. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.mric.2008.04.005\u003c/span\u003e\u003cspan address=\"10.1016/j.mric.2008.04.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoon JH, Nickel MD, Peeters JM, Lee JM (2019) Rapid imaging: recent advances in abdominal MRI for reducing acquisition time and its clinical applications. Korean J Radiol 20:1597\u0026ndash;1615. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3348/kjr.2018.0931\u003c/span\u003e\u003cspan address=\"10.3348/kjr.2018.0931\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun B, Chen Z, Duan Q, Xue Y, Zheng E, He Y et al (2020) Rapid 3D navigator-triggered MR cholangiopancreatography with SPACE sequence at 3T: only one-third acquisition time of conventional 3D SPACE navigator-triggered MRCP. Abdom Radiol 45:134\u0026ndash;140. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00261-019-02342-3\u003c/span\u003e\u003cspan address=\"10.1007/s00261-019-02342-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Z, Sun B, Duan Q, Xue Y, Zheng E, He Y et al (2019) Three-dimensional breath-hold MRCP using SPACE pulse sequence at 3 T: comparison with conventional navigator-triggered technique. AJR Am J Roentgenol 213:1247\u0026ndash;1252. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2214/AJR.19.21399\u003c/span\u003e\u003cspan address=\"10.2214/AJR.19.21399\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChandarana H, Doshi AM, Shanbhogue A, Babb JS, Bruno MT, Zhao T et al (2016) Three-dimensional MR cholangiopancreatography in a breath hold with sparsity-based reconstruction of highly undersampled data. Radiology 280:585\u0026ndash;594. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/radiol.2016151935\u003c/span\u003e\u003cspan address=\"10.1148/radiol.2016151935\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeng L, Benkert T, Block KT, Sodickson DK, Otazo R, Chandarana H (2017) Compressed sensing for body MRI. J Magn Reson Imaging 45:966\u0026ndash;987. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jmri.25547\u003c/span\u003e\u003cspan address=\"10.1002/jmri.25547\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoon JH, Lee SM, Kang HJ, Weiland E, Raithel E, Son Y et al (2017) Clinical feasibility of 3-dimensional magnetic resonance cholangiopancreatography using compressed sensing: comparison of image quality and diagnostic performance. Invest Radiol 52:612\u0026ndash;619. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/RLI.0000000000000380\u003c/span\u003e\u003cspan address=\"10.1097/RLI.0000000000000380\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlaise H, Remen T, Ambarki K, Weiland E, Kuehn B, Orry X et al (2021) Comparison of respiratory-triggered 3D MR cholangiopancreatography and breath-hold compressed-sensing 3D MR cholangiopancreatography at 1.5 T and 3 T and impact of individual factors on image quality. Eur J Radiol 142:109873. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejrad.2021.109873\u003c/span\u003e\u003cspan address=\"10.1016/j.ejrad.2021.109873\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenninger B, Steurer M, Plaikner M, Weiland E, Jaschke W, Kremser C (2020) Magnetic resonance cholangiopancreatography with compressed sensing at 1.5 T: clinical application for the evaluation of branch duct IPMN of the pancreas. Eur Radiol 30:6014\u0026ndash;6021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-020-06996-2\u003c/span\u003e\u003cspan address=\"10.1007/s00330-020-06996-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoshida M, Nakaura T, Inoue T, Tanoue S, Takada S, Utsunomiya D et al (2018) Magnetic resonance cholangiopancreatography with GRASE sequence at 3.0T: does it improve image quality and acquisition time as compared with 3D TSE? Eur Radiol 28:2436\u0026ndash;2443. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-017-5240-y\u003c/span\u003e\u003cspan address=\"10.1007/s00330-017-5240-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNam JG, Lee JM, Kang HJ, Lee SM, Kim E, Peeters JM et al (2018) GRASE revisited: breath-hold three-dimensional (3D) magnetic resonance cholangiopancreatography using a Gradient and Spin Echo (GRASE) technique at 3T. Eur Radiol 28:3721\u0026ndash;3728. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-017-5275-0\u003c/span\u003e\u003cspan address=\"10.1007/s00330-017-5275-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChien CP, Chiu FM, Shen YC, Chen YH, Chung HW (2020) Magnetic resonance cholangiopancreatography at 3T in a single breath-hold: comparative effectiveness between three-dimensional (3D) gradient- and spin-echo and two-dimensional (2D) thick-slab fast spin-echo acquisitions. Quant Imaging Med Surg 10:1265\u0026ndash;1274. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21037/qims.2020.04.14\u003c/span\u003e\u003cspan address=\"10.21037/qims.2020.04.14\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTajima T, Akai H, Sugawara H, Yasaka K, Kunimatsu A, Yoshioka N et al (2021) Breath-hold 3D magnetic resonance cholangiopancreatography at 1.5 T using a deep learning-based noise-reduction approach: comparison with the conventional respiratory-triggered technique. Eur J Radiol 144:109994. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejrad.2021.109994\u003c/span\u003e\u003cspan address=\"10.1016/j.ejrad.2021.109994\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Z, Sun B, Xue Y, Duan Q, Zheng E, He Y et al (2021) Comparing compressed sensing breath-hold 3D MR cholangiopancreatography with two parallel imaging MRCP strategies in main pancreatic duct and common bile duct. Eur J Radiol 142:109833. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejrad.2021.109833\u003c/span\u003e\u003cspan address=\"10.1016/j.ejrad.2021.109833\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Peng W, Xiao Y, Ming Y, Ma K, Hu S et al (2023) Rapid 3D breath-hold MR cholangiopancreatography using deep learning-constrained compressed sensing reconstruction. Eur Radiol 33:2500\u0026ndash;2509. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-022-09227-y\u003c/span\u003e\u003cspan address=\"10.1007/s00330-022-09227-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePezzotti N, Yousefi S, Elmahdy MS, van Gemert J, Sch\u0026uuml;lke C, Doneva M et al (2020) An adaptive intelligence algorithm for undersampled knee MRI reconstruction: application to the 2019 fastMRI Challenge. IEEE Access 8:204825\u0026ndash;204838. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ACCESS.2020.3034287\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2020.3034287\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePezzotti N, de Weerdt E, Yousefi S, Elmahdy MS, van Gemert J, Sch\u0026uuml;lke C et al (2019) Adaptive-CS-Net: FastMRI with Adaptive Intelligence. arXiv. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://arxiv.org/pdf/1912.12259.pdf\u003c/span\u003e\u003cspan address=\"https://arxiv.org/pdf/1912.12259.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRonneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention \u0026ndash; MICCAI 2015. Springer, Cham, pp 234\u0026ndash;241. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-319-24574-4_28\u003c/span\u003e\u003cspan address=\"10.1007/978-3-319-24574-4_28\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKnoll F, Hammernik K, Zhang C, Moeller S, Pock T, Sodickson DK, Ak\u0026ccedil;akaya M (2020) Deep-learning methods for parallel magnetic resonance imaging reconstruction: a survey of the current approaches, trends, and issues. IEEE Signal Process Mag 37:128\u0026ndash;140. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/MSP.2019.2950640\u003c/span\u003e\u003cspan address=\"10.1109/MSP.2019.2950640\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNakaura T, Kidoh M, Maruyama N, Kawahara T, Namimoto T, Sakai Y et al (2013) Usefulness of the SPACE pulse sequence at 1.5T MR cholangiography: comparison of image quality and image acquisition time with conventional 3D-TSE sequence. J Magn Reson Imaging 38:1014\u0026ndash;1019. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jmri.24033\u003c/span\u003e\u003cspan address=\"10.1002/jmri.24033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159\u0026ndash;174. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2529310\u003c/span\u003e\u003cspan address=\"10.2307/2529310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Gradient and spin-echo, Magnetic resonance cholangiopancreatography, Deep learning reconstruction, Turbo spin-echo","lastPublishedDoi":"10.21203/rs.3.rs-9286494/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9286494/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose\u003c/h2\u003e \u003cp\u003eTo evaluate image quality and clinical feasibility of breath-hold 3D magnetic resonance cholangiopancreatography (MRCP) using a gradient and spin-echo (GRASE) technique with deep learning reconstruction (GRASE-DLR) versus GRASE without DLR and turbo spin-echo with DLR (TSE-DLR) at 3 T.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eSixty-five consecutive patients who underwent 3D MRCP on a 3-T system were retrospectively enrolled. Three protocols were compared: GRASE-DLR, GRASE without DLR, and TSE-DLR. Acquisition time, quantitative metrics (SNR and CNR), and five-point qualitative scores for overall image quality, artefact reduction, background suppression, and ductal visualization were independently assessed by two radiologists. Interobserver agreement was evaluated using Cohen's weighted kappa. Diagnostic performance for biliary and pancreatic disease and anatomical variations was evaluated. Subgroup analysis was performed for patients with poor breath-hold capacity (n\u0026thinsp;=\u0026thinsp;8).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eMean acquisition time for GRASE-DLR was 8.9 s, representing reductions of 45.4% versus TSE-DLR and 49.1% versus GRASE (both p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Despite comparable SNR across all three protocols, GRASE-DLR achieved significantly superior overall image quality, artefact reduction, and major duct visualization (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), with moderate-to-substantial interobserver agreement. Sensitivity for biliary disease was markedly higher with GRASE-DLR (93.8%) than TSE-DLR (56.2\u0026ndash;62.5%) and GRASE (68.8\u0026ndash;75.0%), with accuracy of 89.5\u0026ndash;94.7%. For pancreatic disease, sensitivity was 87.5% with GRASE-DLR versus 55.0\u0026ndash;75.0% for comparators, with accuracy of 84.8\u0026ndash;87.0%. Indeterminate biliary anatomical variation findings were nearly eliminated with GRASE-DLR (0\u0026ndash;1.5% vs. 21.5\u0026ndash;26.2% for TSE-DLR). In patients with poor breath-hold capacity, GRASE-DLR demonstrated pronounced improvements in image quality and artefact suppression.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eGRASE-DLR achieves superior image quality with substantially reduced acquisition time and improved diagnostic confidence, particularly in patients with limited breath-hold tolerance.\u003c/p\u003e","manuscriptTitle":"3-Dimensional Gradient and Spin-Echo Magnetic Resonance Cholangiopancreatography with Deep Learning Reconstruction at 3 T: Achieving Superior Image Quality with Reduced Acquisition Time","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-08 08:32:05","doi":"10.21203/rs.3.rs-9286494/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ec5a96d8-f072-4407-ab8a-e5f4789bb715","owner":[],"postedDate":"April 8th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-15T03:09:44+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-08 08:32:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9286494","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9286494","identity":"rs-9286494","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}