ADC-Based MRI Achieves Superior Accuracy in Preoperative LVSI Prediction for Endometrial Cancer

ADC-Based MRI Achieves Superior Accuracy in Preoperative LVSI Prediction for Endometrial Cancer | 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 ADC-Based MRI Achieves Superior Accuracy in Preoperative LVSI Prediction for Endometrial Cancer Alisa Mohebbi, Mehrad Zare, Kimia Darmiani, Ahmadreza Shekarchian, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9151794/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 Background: Preoperative identification of lymphovascular space invasion (LVSI) in endometrial cancer remains challenging yet critical for surgical planning and adjuvant therapy. This study compared qualitative diffusion-weighted imaging (DWI) and quantitative apparent diffusion coefficient (ADC) analysis for preoperative LVSI detection. Methods: A total of 95 patients with histologically confirmed endometrial cancer underwent preoperative DWI. Two blinded radiologists performed a visual assessment of high b-value images, and manual regions of interest were placed on the ADC maps to derive the mean ADC values. Optimal ADC thresholds were determined and compared with published cut-offs. Diagnostic performance metrics were calculated for both methods in the overall cohort and stratified by tumor size, FIGO stage, and grade. Results: Qualitative DWI yielded 85.4% sensitivity, 21.3% specificity, and 53.7% accuracy. ADC at a threshold of ≤0.690×10⁻³ mm²/s achieved 75.0% sensitivity, 76.6% specificity, and 75.8% accuracy, with an AUC of 0.770. Published ADC cut-offs (0.690-0.820×10⁻³ mm²/s) consistently outperformed visual assessment (accuracy 62.1-71.6%). In subgroup analyses, ADC maintained superior accuracy across all tumor sizes (4 cm: 80.3%), FIGO stages (I: 71.4%; II: 82.1%; III: 92.9%), and grades (1: 72.4%; 2: 78.4%; 3: 84.0%). Combining visual and quantitative assessments did not improve the AUC beyond that of ADC alone. Conclusions: Quantitative ADC analysis significantly outperforms qualitative DWI for preoperative detection of LVSI in endometrial cancer, providing an objective biomarker that enhances risk stratification and informs surgical and adjuvant treatment strategies. Nuclear Medicine & Medical Imaging Endometrial cancer Lymphovascular space invasion (LVSI) Diffusion-weighted imaging (DWI) Apparent diffusion coefficient (ADC) Preoperative MRI Figures Figure 1 Figure 2 Figure 3 Introduction Endometrial cancer is the most common gynecologic cancer in developed countries, with a rising incidence over recent decades. Despite generally favorable outcomes for early-stage disease, the identification of high-risk features remains crucial for optimal treatment planning and prognosis assessment [ 1 , 2 ]. Lymphovascular space invasion (LVSI), defined as the presence of tumor cells within blood or lymphatic vessels beyond the core tumor, represents one of the most significant prognostic indicators in endometrial cancer [ 3 ]. According to the International Federation of Gynecology and Obstetrics (FIGO) staging system (2023), substantial LVSI (defined as involvement of ≥ 5 vessels) is a key determinant of endometrial cancer staging and prognosis [ 2 , 4 ]. Currently, LVSI assessment relies predominantly on postoperative histopathological examination, limiting its utility in preoperative surgical planning and treatment stratification. The inability to reliably determine LVSI status preoperatively presents significant clinical challenges, as this information could guide decisions regarding the extent of lymphadenectomy, adjuvant therapy recommendations, and overall treatment approach [ 5 ]. Therefore, developing reliable preoperative imaging biomarkers for LVSI detection represents a critical need in endometrial cancer management [ 6 ]. Magnetic resonance imaging (MRI) has emerged as the mainstay of preoperative staging for endometrial cancer, providing excellent soft tissue contrast and accurate assessment of myometrial invasion, cervical involvement, and lymph node status. Diffusion-weighted imaging (DWI), an advanced functional MRI technique, exploits the principle that malignant tissues demonstrate restricted water diffusion due to increased cellularity and disrupted cellular architecture, resulting in characteristic signal patterns that differ from normal myometrial tissue [ 7 – 9 ]. DWI can be evaluated through two distinct approaches: qualitative visual assessment and quantitative analysis using the apparent diffusion coefficient (ADC). Qualitative DWI assessment relies on subjective visual interpretation of signal intensity changes across different b-values [ 10 ]. This approach identifies areas of restricted diffusion by recognizing persistent high signal intensity on high b-value images, which in endometrial cancer may correlate with tumor aggressiveness and LVSI presence [ 11 – 13 ]. The primary advantages of qualitative assessment include rapid interpretation, widespread accessibility without specialized software, and the ability to incorporate overall tumor morphology into the diagnostic evaluation [ 10 ]. However, qualitative DWI suffers from inherent limitations, including significant interobserver variability, subjective interpretation bias, and susceptibility to the T2 shine-through effect, where tissues with prolonged T2 relaxation times can mimic restricted diffusion which lead to false-positive interpretations, particularly in cystic or necrotic tumor components [ 14 – 16 ]. On the other hand, quantitative DWI analysis using ADC measurements provides an objective, numerical assessment of water diffusion. ADC values offer a standardized metric that is less susceptible to subjective interpretation and demonstrates superior reproducibility across different observers and imaging centers [ 17 , 18 ]. Lower ADC values typically correlate with higher tumor cellularity, increased aggressiveness, and the presence of adverse pathological features, including LVSI [ 19 ]. The quantitative nature of ADC allows for the establishment of specific threshold values for diagnostic purposes and enables standardized comparison across studies and institutions. Despite these advantages, quantitative ADC assessment faces several limitations. ADC values can be influenced by technical factors, including scanner manufacturer, magnetic field strength, acquisition parameters, and sequence protocols, potentially affecting the transferability of threshold values across different imaging systems [ 20 , 21 ]. The measurement process requires careful region-of-interest placement and may be affected by tumor heterogeneity, partial volume effects, and motion artifacts [ 22 – 24 ]. Furthermore, while ADC provides excellent objectivity, it may not capture subtle morphological features that experienced radiologists can appreciate through visual assessment. Previous studies investigating DWI in endometrial cancer have primarily focused on individual assessment approaches, with limited direct comparison between qualitative and quantitative methods specifically for LVSI detection. While ADC has shown promise in predicting various endometrial cancer characteristics, including tumor grade and myometrial invasion depth, its comparative diagnostic performance against qualitative assessment for LVSI remains incompletely characterized. Understanding the relative strengths and limitations of each approach is essential for determining the optimal DWI evaluation strategy in clinical practice. Therefore, the primary aim of this study was to conduct a comprehensive head-to-head comparison between qualitative visual assessment and quantitative ADC analysis for preoperative LVSI detection in endometrial cancer. Secondary aims included evaluating this comparison across different tumor sizes, FIGO stages, and pathological grades to determine which approach maintains diagnostic value in challenging clinical scenarios. Additionally, a comprehensive literature review was performed regarding ADC cut-off points for LVSI determination, with subsequent validation of their diagnostic efficacy in our dataset. Materials & Methods This retrospective, single-center study was conducted with the approval of the institutional review board, in accordance with the ethical standards outlined in the Declaration of Helsinki. Given the use of anonymized medical records, the ethics committee waived the requirement for informed consent. This investigation aimed to compare visually based DWI assessment and quantitative ADC analysis for preoperative detection of LVSI in patients with confirmed endometrial cancer. Patient Selection & Study Cohort Medical records were reviewed to identify women who underwent surgical treatment for histologically confirmed endometrial carcinoma between January 2019 and November. Inclusion criteria required a definitive histopathological diagnosis of endometrial carcinoma, availability of preoperative pelvic MRI scans including diffusion-weighted sequences, and complete pathological documentation of LVSI status. Exclusion criteria encompassed cases with suspected non-endometrial malignancy or mixed histology, the presence of distant metastases at initial staging (e.g., FIGO stage IV), incomplete or poor-quality MRI examinations that precluded reliable DWI evaluation, the absence of recorded ADC or visual DWI assessments, and missing postoperative LVSI data. The initial dataset comprised 184 consecutive patients. Application of these criteria resulted in the exclusion of 89 patients: 42 due to the absence of LVSI pathology reports, 27 for inadequate MRI quality, and 20 lacking ADC values or visual DWI assessments. Thus, a final cohort of 95 patients was included for analysis. MRI Acquisition Protocol All MRI examinations were performed on a three-Tesla system (GE Discovery™ 750 GEM; GE Healthcare, Chicago, IL, USA) equipped with an 8-channel phased-array pelvic coil. Patients were instructed to fast for at least four hours prior to scanning to minimize bowel peristalsis, and an intravenous injection of 20 mg hyoscine N-butyl bromide was administered immediately before the examination unless contraindicated. Vaginal gel was instilled to distend the vaginal canal unless patient intolerance precluded its use. The imaging protocol conformed to institutional standards for endometrial cancer evaluation and included axial and sagittal T2-weighted turbo spin-echo sequences (repetition time (TR) = 4,500 ms; echo time (TE) = 100 ms; slice thickness = 4 mm; interslice gap = 1 mm; field of view (FOV) = 24 cm; matrix = 320 × 256), axial T1-weighted spin-echo images (TR = 650 ms; TE = 12 ms; slice thickness = 4 mm; interslice gap = 1 mm; FOV = 24 cm; matrix = 256 × 224), and diffusion-weighted imaging acquired in the axial plane using single-shot echo-planar imaging with b-values of 0, 500, and 1,000 s/mm² (TR = 4,000 ms; TE = 70 ms; slice thickness = 4 mm; interslice gap = 1 mm; FOV = 24 cm; matrix = 128 × 128). Diffusion gradients were applied in three orthogonal directions, and ADC maps were automatically generated by the vendor-supplied software. In all cases, contrast-enhanced sequences were performed using a standard gadolinium-based contrast agent (0.1 mmol/kg). The agent was injected at a dose of 0.2 ml/kg body weight, at a rate of 3 ml/sec using an injector, followed by 10 cc of normal saline. Post-contrast images were obtained dynamically at intervals of 30, 60, 120, 180, and 300 seconds after injection. Image Retrieval & Blinding All MRI studies were retrieved from the institution’s Picture Archiving and Communication System (PACS). Two centers with independent board-certified radiologists with ten and fifteen years of experience in gynecological imaging, contributed to the study. There was no requirement for inter-observer agreement analysis or result combining because each case was assessed by a single radiologist who was related to their center. Every case from Center 1 was assessed by Reader 1, and each one from Center 2 was assessed by Reader 2. Radiologists were blinded to clinical and pathological information, including patient age, tumor grade, FIGO stage, and especially LVSI status. Each radiologist conducted visual DWI assessment and quantitative ADC measurement in separate sessions at least two weeks apart. Qualitative evaluation of diffusion-weighted images was based on visual assessment of signal intensity on high b-value (1,000 s/mm²) images. Qualitative evaluation focused on detecting restricted diffusion within the endometrial tumor on images, serving as a surrogate marker for LVSI potential. Tumors were categorized as positive if they exhibited heterogeneous signal patterns with focal areas of significant diffusion restriction. This method assesses tissue microstructural alterations linked to aggressive tumor characteristics rather than directly evaluating anatomical invasion. The visual assessment specifically aimed to identify restricted diffusion patterns correlating with tumor aggressiveness and the presence of LVSI, rather than directly visualizing vascular invasion. Quantitative analysis was performed using the same DWI source data. On ADC maps, regions of interest (ROIs) were manually placed to encompass the solid component of the tumor. Three circular ROIs (minimum 20mm²) were manually placed on ADC maps in areas showing the lowest ADC values within the tumor, avoiding necrotic areas, blood vessels, and artifacts. The median ADC value of the three measurements was used for analysis, consistent with established protocols for endometrial cancer ADC measurement [25, 26]. The representative ADC for that tumor expressed in units of ×10⁻³ mm²/s. ADC thresholds for LVSI detection were determined in two ways: first, by calculating the optimal cut-off value from the study cohort using Youden’s index applied to receiver operating characteristic (ROC) analysis; and second, by applying previously published ADC thresholds (e.g., 690, 745, and 820 ×10⁻ 6 mm²/s) to the dataset to assess external validity [27-29]. Pathological Reference Standard Postoperative specimens were processed according to standard histopathological protocols. Tumor type, FIGO grade, depth of myometrial invasion, cervical stromal involvement, and LVSI status were reported by experienced gynecologic pathologists blinded to MRI findings. LVSI was defined as the presence of tumor cells within endothelial-lined vascular channels (blood vessels or lymphatics) beyond the main tumor mass. Only “substantial” LVSI, entailing invasion of five or more vessels per slide, was recorded as positive, in alignment with the 2023 FIGO staging guidelines. Cases with ambiguous vascular invasion on initial review were subjected to immunohistochemical staining for endothelial markers (CD31 and D2-40) to confirm LVSI. Tumor Size, Stage, and Grade Stratification Radiological tumor size was measured on T2-weighted images as the maximum straight-line diameter in axial, sagittal, or coronal planes. Tumors were stratified into three size categories; 4 cm to evaluate potential size-dependent variation in diagnostic performance [30, 31]. Clinical staging was assigned preoperatively based on MRI findings and recorded according to the 2023 FIGO classification: stage I (tumor confined to corpus uteri), stage II (cervical stromal invasion without extrauterine spread), and stage III (extension to serosa, adnexa, vagina, parametria, or pelvic/para-aortic lymph nodes). Pathological grade was classified as grade 1 (well differentiated, ≤5% solid growth), grade 2 (moderately differentiated, 6-50% solid growth), or grade 3 (poorly differentiated, >50% solid growth), with upgrading by one grade for severe nuclear atypia when appropriate. Statistical Analysis All statistical analyses were conducted using MedCalc version 23.2 (MedCalc Software, Ostend, Belgium) and Stata version 18.0 (StataCorp, College Station, TX, USA). Continuous variables were reported as mean ± standard deviation or median and interquartile range, depending on the normality of the distribution, assessed by the Shapiro–Wilk test. Categorical data were expressed as counts and percentages. The diagnostic performance of qualitative DWI and quantitative ADC analysis for LVSI detection was evaluated by calculating sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and the area under the ROC curve (AUC) with 95% confidence intervals. Sensitivity, specificity, and accuracy comparisons were performed using dependent proportion difference tests. This approach provides both point estimates and confidence intervals for direct comparison of diagnostic performance. For all analyses, a true positive was defined as LVSI identified by both MRI and pathology, a true negative as LVSI absent on both MRI and pathology, a false positive as MRI-positive but pathology-negative, and a false negative as MRI-negative but pathology-positive. Optimal ADC cut-off values were determined by maximizing Youden’s index. Pairwise comparisons of AUCs between qualitative and quantitative methods, as well as between different ADC thresholds, were performed using the DeLong test. Subgroup analyses were performed using AUC comparisons rather than separate cutoff values to maintain standardization and validity; these analyses assessed diagnostic metrics within tumor size, FIGO stage, and pathological grade categories. A p-value of <0.05 was considered statistically significant for all comparisons. Results Demographic and clinicopathological characteristics The final study cohort comprised 95 patients with histologically confirmed endometrial cancer who met all inclusion criteria (Table 1). The mean age was 55.85 ± 10.34 years (range: 30-77 years). The mean tumor size on MRI was 5.42 ± 3.27 cm. Histopathological examination revealed that 82 (86.3%) of tumors were endometrioid carcinoma, while 13 (13.7%) were of non-endometrioid subtypes. Pathological grading revealed that 29 (31.8%) were grade 1, 37 (40.6%) were grade 2, and 25 (27.6%) were grade 3. The LVSI was present in 48 (51%) of patients and absent in 47 (49%). Overall Performance of Qualitative vs. Quantitative DWI Qualitative visual assessment of diffusion-weighted images demonstrated a limited overall accuracy of 53.7% for preoperative prediction of LVSI with a high sensitivity (85.4%) but low specificity (21.3%). In contrast, quantitative ADC analysis yielded markedly higher performance. Using an optimal ADC threshold of ≤0.690 × 10⁻³ mm²/s, quantitative assessment achieved an overall accuracy of 75.8%, with a sensitivity of 75.0% and a specificity of 76.6%. Moreover, the AUC for ADC-based detection was 0.770 (95% CI, 0.673 to 0.850; p < 0.001), indicating moderate discriminative ability and a notable improvement over visual evaluation (Figure 1A) (Table 2)[28, 29, 32]. When applying previously published ADC cut-off values of 0.690, 0.745, and 0.820 ×10⁻³ mm²/s to this dataset, accuracies ranged between 62.1% and 71.6%, consistently surpassing the performance of qualitative assessment (Table 2). The pairwise comparisons among literature cut-offs revealed no significant differences between all these cut-point performances (all p > 0.1). However, when directly compared with qualitative DWI, quantitative analysis using thresholds of ≤0.690 and ≤0.745 demonstrated significantly better diagnostic performance (p = 0.017 and p = 0.011, respectively), whereas a threshold of < 0.820 did not (p = 0.241). These findings, based on external validation, reinforce the observed trend that quantitative ADC assessment provides greater diagnostic reliability for LVSI detection than visual interpretation. In addition, when both quantitative and qualitative DWI assessments were combined using binary logistic regression, the diagnostic performance of quantitative ADC analysis alone was identical to that of the combined ADC and qualitative visual assessment model (AUC = 0.770, CI: 0.672 to 0.850; p < 0.001) (Figure 1B) (p = 0.956) indicating that further visual assessment to ADC does not provide additional discriminatory and informativeness value for detecting LVSI. Figures 2 and 3 illustrate two clinical examples. Performance Stratified by Radiological Tumor Size In the subgroup with the smallest tumors (<2 cm), qualitative DWI demonstrated high sensitivity but poor specificity, resulting in an overall accuracy of 40.0%. Quantitative ADC analysis, with an optimal threshold of ≤1.013 × 10⁻³ mm²/s in this subset, achieved a significant increase in accuracy to 90.0% and an AUC of 0.688 (CI: 0.336 to 0.927) (all p > 0.05). For tumors of intermediate size (2-4 cm), qualitative assessment again suffered from low specificity, corresponding to a modest accuracy of 48.3%. Quantitative ADC measurement at an optimized cut-off of ≤0.609 ×10⁻³ mm²/s improved accuracy to 79.3% and produced an AUC of 0.803 (95% CI, 0.614 to 0.926), reinforcing the superiority of cohort-specific ADC calibration compared to visual interpretation alone (all p > 0.05). In larger tumors (>4 cm), qualitative DWI achieved an accuracy of 58.9%. ADC analysis at the optimal threshold of ≤0.690 ×10⁻³ mm²/s increased overall accuracy to 80.3% and reached an AUC of 0.831 (95% CI, 0.707 to 0.918), the highest among size-based subgroups (all p > 0.05). The detailed diagnostic metrics are presented in Table 3. Diagnostic Performance by FIGO Stage In stage I (n = 49), visual DWI exhibited an overall accuracy of 47.9%. Quantitative ADC evaluation, applying a threshold of ≤0.710 ×10⁻³ mm²/s, improved accuracy to 71.4% and achieved an AUC of 0.679 (CI, 0.530 to 0.805) (all p > 0.05). Among patients with stage II disease (n = 28), qualitative DWI accuracy reached only 57.1%. ADC analysis at the threshold of ≤0.656 ×10⁻³ mm²/s yielded an accuracy of 82.1% and an AUC of 0.847 (CI, 0.661 to 0.954) (all p > 0.05). In stage III (n = 14), qualitative visual assessment achieved an accuracy of 69.2%. ADC measurement using an optimal threshold of ≤0.813 ×10⁻³ mm²/s achieved the highest subgroup accuracy of 92.9%, although the small sample size limits the precision of this estimate. The corresponding AUC was 0.625 (CI, 0.335 to 0.861) (all p > 0.05). The detailed diagnostic metrics are presented in Table 3. Diagnostic performance by pathological grading In grade 1 tumors, qualitative DWI accuracy was 44.8%. ADC analysis using a threshold of ≤0.710 ×10⁻³ mm²/s improved accuracy to 72.4% and produced an AUC of 0.616 (CI, 0.418 to 0.789) (all p > 0.05). For grade 2 lesions, qualitative assessment achieved an accuracy of 46.0%, whereas ADC measurement at ≤0.690 ×10⁻³ mm²/s delivered an accuracy of 78.4% and an AUC of 0.810 (CI: 0.648 to 0.920) (all p > 0.05). In grade 3 tumors, qualitative DWI reached an accuracy of 72.0%. Quantitative ADC analysis, with an optimal threshold of ≤0.711 ×10⁻³ mm²/s, attained an accuracy of 84.0% and an AUC of 0.740 (CI, 0.527 to 0.893) (all p > 0.05). The detailed diagnostic metrics are presented in Table 3. Discussion In this study, the head-to-head comparison in 95 patients demonstrated that ADC-based evaluation not only achieved superior overall accuracy (75.8%) compared with qualitative assessment (53.7%), but also maintained consistent performance across tumor size, FIGO stage, and pathological grade subgroups. These findings underscore the critical importance of integrating quantitative diffusion metrics into routine preoperative MRI protocols to inform surgical planning and the selection of adjuvant therapy. Our study’s principal observation was the substantially higher specificity of ADC analysis relative to visual assessment, which suffered from very low specificity (21.3%) despite high sensitivity (85.4%). This discrepancy reflects the inherent subjectivity and susceptibility of qualitative DWI to T2 shine-through and interobserver variability [ 7 , 28 , 33 ]. By contrast, ADC offers standardized, numerical thresholds (optimized in our cohort at ≤ 0.690 × 10⁻³ mm²/s) that balance sensitivity (75.0%) and specificity (76.6%), translating into a more reliable preoperative indicator of LVSI. Notably, when applying previously published thresholds (≤ 0.690, ≤ 0.745, ≤ 0.820 × 10⁻³ mm²/s), ADC performance consistently surpassed that of qualitative DWI, reinforcing its robustness across diverse settings [ 34 , 35 ]. Crucially, combining visual assessment with ADC did not improve diagnostic performance beyond quantitative analysis alone (AUC = 0.770), indicating that ADC encapsulates the salient diffusion characteristics necessary for LVSI detection. This aligns with prior reports that highlight the limited incremental value of visual ADC map evaluation once quantitative metrics are established [ 7 , 36 ]. In clinical practice, this suggests that resource-intensive, dual-reader visual assessments may be supplanted by streamlined quantitative protocols without a loss of diagnostic fidelity. Our subgroup analyses revealed that ADC maintained its diagnostic advantage irrespective of tumor size, stage, or grade. In tumors < 2 cm, ADC analysis with a threshold of ≤ 1.013×10⁻³ mm²/s achieved 90.0% accuracy (AUC = 0.688), indicating its resilience in small lesions often challenging for visual interpretation. Similarly, in FIGO stage II and III tumors, ADC thresholds of ≤ 0.656 and ≤ 0.813 × 10⁻³ mm²/s yielded accuracies of 82.1% and 92.9%, respectively, underscoring the applicability of ADC across disease extents. These findings echo those of Ma et al., who reported ADC’s stable performance across heterogeneous pathological groups [ 37 ]. These findings align with previous research. For instance, Petrila et al. observed that the mean ADC inversely correlated with histological grade, deep myometrial invasion, and LVSI in endometrial cancer, underlining its association with tumor aggressiveness [ 7 ]. Similarly, Wang et al. reported links between ADC values and high-risk features, including LVSI, reinforcing the diagnostic relevance of ADC across varying tumor behaviors [ 34 ]. Additionally, a multicenter radiomics study by Liu et al. found that radiomic models incorporating ADC and texture features predicted LVSI more effectively than traditional imaging, highlighting the added value of quantitative data [ 6 ]. Similar advantages of ADC over qualitative interpretation have also been reported by Satta et al. [ 36 ] and Ma et al. [ 38 ], who found that quantitative diffusion metrics are strong predictors of adverse histopathological features in endometrial cancer. Collectively, these data support ADC’s role as a generalizable imaging biomarker that transcends conventional tumor stratifications. The clinical implications of our findings are considerable. Preoperative identification of LVSI can guide the extent of lymphadenectomy, inform neoadjuvant treatment decisions, and refine risk stratification within multidisciplinary tumor boards [ 39 ]. Existing nomograms incorporating ADC and radiomic features demonstrate AUCs up to 0.959 for LVSI prediction [ 40 ], yet our data suggest that even standalone ADC metrics without complex radiomics pipelines offer high diagnostic yield, facilitating broader adoption in settings lacking advanced software infrastructures. Moreover, our external validation of literature-derived ADC cut-offs mitigates concerns of cohort-specific overfitting. All published thresholds tested outperformed qualitative DWI, confirming ADC’s reproducibility across scanners and institutions despite known influences of magnetic field strength and acquisition parameters. Nonetheless, site-specific calibration of ADC thresholds remains advisable to account for technical variabilities and ensure optimal diagnostic accuracy. Our results complement the growing body of multiparametric MRI research. Studies integrating diffusion, perfusion (IVIM, DCE), and volumetric metrics report enhanced prognostic stratification, linking quantitative parameters to tumor grade, depth of invasion, and inflammatory infiltrates [ 13 , 36 ]. While such multiparametric approaches hold promise for comprehensive tumor characterization, our findings emphasize that ADC measurement alone provides a potent, objective marker for LVSI, warranting its prioritization in preoperative imaging protocols. This study has several limitations that may affect the generalizability and interpretation of our findings. First, its retrospective design may limit generalizability, and validation in larger, multicenter cohorts is warranted. Second, the relatively modest sample size of 95 patients, particularly in subgroup analyses, may reduce statistical power and precision of estimates, potentially limiting the robustness of conclusions in specific clinical scenarios. Third, the manual placement of regions of interest for ADC quantification introduces potential operator-dependent variability and may be affected by tumor heterogeneity, partial volume effects, and motion artifacts, despite our attempts to standardize the measurement protocol. Conclusion Quantitative ADC analysis outperformed qualitative DWI for the preoperative detection of LVSI in endometrial cancer, demonstrating higher accuracy and greater diagnostic reliability across the overall cohort and all clinical subgroups. Its consistent performance, even when using literature-derived cut-offs, and the lack of added value from combining it with visual assessment, support its integration into routine MRI protocols. Incorporating quantitative ADC measurement into clinical practice may improve preoperative risk stratification and guide individualized treatment planning in patients with endometrial cancer. Declarations All authors contributed significantly to this article. Funding: None Conflict of interest: None Informed consent: Not applicable. Animal study: N/A AI-based tools were used only for language editing; all scientific content, data analysis, and interpretations are the authors’ own. Acknowledgment: We would like to acknowledge all those who contributed to gathering the dataset. Data Availability: The datasets generated and analyzed during the current study are not publicly available due to patient privacy and institutional restrictions, but are available from the corresponding author on reasonable request. References E.J. Crosbie, S.J. Kitson, J.N. McAlpine, A. Mukhopadhyay, M.E. Powell, N. Singh, Endometrial cancer, Lancet (London, England) 399(10333) (2022) 1412-1428. Z. Li, J. Peng, B. Zhang, C. Zhao, Z. Chen, H. Xiao, Z. Zhang, X. Ma, F. Gao, R. Xin, W. Wang, S. Wang, L. Tang, Y. Zhang, B. Kong, L. Li, A. Yin, The prognostic and clinical significance of substantial lymphovascular space invasion in early-stage endometrial carcinoma, European journal of cancer (Oxford, England : 1990) 218 (2025) 115258. F. Yarandi, E. Shirali, S. Akhavan, F. Nili, S. Ramhormozian, The impact of lymphovascular space invasion on survival in early stage low-grade endometrioid endometrial cancer, European journal of medical research 28(1) (2023) 118. M. Menendez-Santos, C. Gonzalez-Baerga, D. Taher, R. Waters, M. Virarkar, P. Bhosale, Endometrial Cancer: 2023 Revised FIGO Staging System and the Role of Imaging, Cancers 16(10) (2024). M.R. Oliver-Perez, P. Padilla-Iserte, O. Arencibia-Sanchez, C. Martin-Arriscado, J.C. Muruzabal, B. Diaz-Feijóo, S. Cabrera, P. Coronado, M.B. Martín-Salamanca, M. Pantoja-Garrido, J. Marcos-Sanmartin, E. Cabezas-López, C. Lorenzo, D. Beric, J.R. Rodriguez-Hernandez, F. Roldan-Rivas, J. Gilabert-Estelles, L. Sanchez, M. Laseca-Modrego, C. Tauste-Rubio, B. Gil-Ibañez, A. Tejerizo-Garcia, G. On Behalf Of The Spain-Gog, Lymphovascular Space Invasion in Early-Stage Endometrial Cancer (LySEC): Patterns of Recurrence and Predictors. A Multicentre Retrospective Cohort Study of the Spain Gynecologic Oncology Group, Cancers 15(9) (2023). X.F. Liu, B.C. Yan, Y. Li, F.H. Ma, J.W. Qiang, Radiomics feature as a preoperative predictive of lymphovascular invasion in early-stage endometrial cancer: A multicenter study, Frontiers in oncology 12 (2022) 966529. O. Petrila, I. Nistor, N.S. Romedea, D. Negru, V. Scripcariu, Can the ADC Value Be Used as an Imaging "Biopsy" in Endometrial Cancer?, Diagnostics (Basel, Switzerland) 14(3) (2024). T. Ota, T. Tsuboyama, H. Onishi, A. Nakamoto, H. Fukui, K. Yano, T. Honda, K. Kiso, M. Tatsumi, N. Tomiyama, Diagnostic accuracy of MRI for evaluating myometrial invasion in endometrial cancer: a comparison of MUSE-DWI, rFOV-DWI, and DCE-MRI, La Radiologia medica 128(6) (2023) 629-643. A.S.L. Moreira, V. Ribeiro, G. Aringhieri, S.C. Fanni, L. Tumminello, L. Faggioni, D. Cioni, E. Neri, Endometrial Cancer Staging: Is There Value in ADC?, Journal of personalized medicine 13(5) (2023). V. Baliyan, C.J. Das, R. Sharma, A.K. Gupta, Diffusion weighted imaging: Technique and applications, World journal of radiology 8(9) (2016) 785-798. M. Xie, Z. Ren, D. Bian, D. Li, L. Yu, F. Zhu, R. Huang, Z. Zhang, S. Suye, C. Fu, High resolution diffusion-weighted imaging with readout segmentation of long variable echo-trains for determining myometrial invasion in endometrial carcinoma, Cancer imaging : the official publication of the International Cancer Imaging Society 20(1) (2020) 66. R. A, J.T. M, C. N, S. M, M.G. H, A. Gh, Signal Intensity of High B-value Diffusion-weighted Imaging for the Detection of Prostate Cancer, Journal of biomedical physics & engineering 9(4) (2019) 453-458. J.J. Wang, X.H. Zhang, X.H. Guo, Y. Ying, X. Wang, Z.H. Luan, W.Q. Lv, P.F. Wang, Prediction of Lymphovascular Space Invision in Endometrial Cancer based on Multi-parameter MRI Radiomics Model, Current medical imaging 20 (2024) e15734056266366. L. Quinn, K. Tryposkiadis, J. Deeks, H.C.W. De Vet, S. Mallett, L.B. Mokkink, Y. Takwoingi, S. Taylor-Phillips, A. Sitch, Interobserver variability studies in diagnostic imaging: a methodological systematic review, The British journal of radiology 96(1148) (2023) 20220972. Y. Kim, S.K. Lee, J.Y. Kim, J.H. Kim, Pitfalls of Diffusion-Weighted Imaging: Clinical Utility of T2 Shine-through and T2 Black-out for Musculoskeletal Diseases, Diagnostics (Basel, Switzerland) 13(9) (2023). K. Fukushima, K. Sano, H. Machida, T. Kariyasu, T. Yoshioka, S. Takahashi, A. Nakanishi, H. Kusahara, K. Yokoyama, Effect of an Ultrahigh b Value of 3000 s/mm(2) and the Minimal Echo-time on Image Quality and the T2 Shine-through Effect in Diffusion-weighted Imaging of the Liver at 3T: Phantom and Clinical Pilot Studies, Magnetic resonance in medical sciences : MRMS : an official journal of Japan Society of Magnetic Resonance in Medicine 22(2) (2023) 232-240. F. Zeinalkhani, P. Kamali Hakim, M. Aghasi, F. Mahdavi Sabet, R. Mortazavi Ardestani, Y. Baghban, M. Malek, H. Zeinalkhani, S. Delazar, F. Shakki Katouli, F. Azizinik, S. Moghaddamimonaghi, S. Mohammadzadeh, Direct comparison of full protocol MRI and modified non-contrast MRI in staging of cervical cancer: a retrospective study, Abdominal radiology (New York) (2025). D.C. Newitt, Z. Zhang, J.E. Gibbs, S.C. Partridge, T.L. Chenevert, M.A. Rosen, P.J. Bolan, H.S. Marques, S. Aliu, W. Li, L. Cimino, B.N. Joe, H. Umphrey, H. Ojeda-Fournier, B. Dogan, K. Oh, H. Abe, J. Drukteinis, L.J. Esserman, N.M. Hylton, Test-retest repeatability and reproducibility of ADC measures by breast DWI: Results from the ACRIN 6698 trial, Journal of magnetic resonance imaging : JMRI 49(6) (2019) 1617-1628. A. Keven, E.E. Yetim, A. Elmalı, A.G. Arslan, S.M. Çubuk, The role of diffusion magnetic resonance imaging in determining tumor aggressiveness during preoperative surgical planning in early-stage endometrial cancer, Radiologie (Heidelberg, Germany) 63(Suppl 2) (2023) 41-48. C. Moore, C. Bull, A. Darekar, D. Wilson, A. Goodall, P. Manoharan, P. Hoskin, M. van Herk, D.L. Buckley, D.J. McHugh, A. Datta, M.J. Dubec, Multi-institution longitudinal apparent diffusion coefficient measurements in a diffusion weighted imaging phantom at room temperature, Physics and imaging in radiation oncology 35 (2025) 100814. B. Moreau, A. Iannessi, C. Hoog, H. Beaumont, How reliable are ADC measurements? A phantom and clinical study of cervical lymph nodes, European radiology 28(8) (2018) 3362-3371. H. Takahashi, K. Yoshida, A. Kawashima, N.J. Lee, A.T. Froemming, D.A. Adamo, A. Khandelwal, C.W. Bolan, M.T. Heller, R.P. Hartman, B. Kim, K.A. Philbrick, R.E. Carter, L.A. Mynderse, M.R. Humphreys, J.C. Cai, N. Takahashi, Impact of measurement method on interobserver variability of apparent diffusion coefficient of lesions in prostate MRI, PloS one 17(5) (2022) e0268829. M. Gity, B. Moradi, R. Arami, A. Arabkheradmand, M.A. Kazemi, Two Different Methods of Region-of-Interest Placement for Differentiation of Benign and Malignant Breast Lesions by Apparent Diffusion Coefficient Value, Asian Pacific journal of cancer prevention : APJCP 19(10) (2018) 2765-2770. J.S. Son, H.S. Park, S. Park, Y.J. Kim, M.H. Yu, S.I. Jung, M. Paek, M.D. Nickel, Motion-Corrected versus Conventional Diffusion-Weighted Magnetic Resonance Imaging of the Liver Using Non-Rigid Registration, Diagnostics (Basel, Switzerland) 13(6) (2023). K. Tamai, T. Koyama, T. Saga, S. Umeoka, Y. Mikami, S. Fujii, K. Togashi, Diffusion-weighted MR imaging of uterine endometrial cancer, Journal of magnetic resonance imaging : JMRI 26(3) (2007) 682-7. G. Rechichi, S. Galimberti, M. Signorelli, C.T. Franzesi, P. Perego, M.G. Valsecchi, S. Sironi, Endometrial cancer: correlation of apparent diffusion coefficient with tumor grade, depth of myometrial invasion, and presence of lymph node metastases, AJR. American journal of roentgenology 197(1) (2011) 256-62. R. Yan, X. Zhang, Q. Cao, J. Xu, Y. Chen, S. Qin, S. Zhang, W. Zhao, X. Xing, W. Yang, N. Lang, Multitask Deep Learning for Automated Segmentation and Prognostic Stratification of Endometrial Cancer via Biparametric MRI, Journal of magnetic resonance imaging : JMRI (2025). G.A. Saleh, R. Abdelrazek, A. Hassan, O. Hamdy, M.S.I. Tantawy, Diagnostic utility of apparent diffusion coefficient in preoperative assessment of endometrial cancer: are we ready for the 2023 FIGO staging?, BMC medical imaging 24(1) (2024) 226. A. Akçay, M.A. Gültekin, F. Altıntaş, A.A. Peker, S. Balsak, B. Atasoy, Ö. Toluk, H. Toprak, Updated endometrial cancer FIGO staging: the role of MRI in determining newly included histopathological criteria, Abdominal radiology (New York) 49(10) (2024) 3711-3721. X. Jin, C. Shen, X. Yang, Y. Yu, J. Wang, X. Che, Association of Tumor Size With Myometrial Invasion, Lymphovascular Space Invasion, Lymph Node Metastasis, and Recurrence in Endometrial Cancer: A Meta-Analysis of 40 Studies With 53,276 Patients, Frontiers in oncology 12 (2022) 881850. R. Berretta, T.S. Patrelli, C. Migliavacca, M. Rolla, L. Franchi, M. Monica, A.B. Modena, S. Gizzo, Assessment of tumor size as a useful marker for the surgical staging of endometrial cancer, Oncol Rep 31(5) (2014) 2407-2412. C. Sun, J. Mao, Y. Xia, M. Li, Z. Zhao, Analysis of the Correlation between MRI Imaging Signs and Lymphovascular Space Invasion in Endometrial Cancer, Current medical imaging (2025). E. Maheshwari, S. Nougaret, E.B. Stein, G.M. Rauch, K.P. Hwang, R.J. Stafford, A.H. Klopp, P.T. Soliman, K.E. Maturen, A.G. Rockall, S.I. Lee, E.A. Sadowski, A.M. Venkatesan, Update on MRI in Evaluation and Treatment of Endometrial Cancer, Radiographics : a review publication of the Radiological Society of North America, Inc 42(7) (2022) 2112-2130. Y. Wang, M. He, P. Cao, P.P.C. Ip, C.Y. Lin, W. Liu, C.W. Lee, E.Y.P. Lee, Tissue Characteristics of Endometrial Carcinoma Analyzed by Quantitative Synthetic MRI and Diffusion-Weighted Imaging, Diagnostics (Basel, Switzerland) 12(12) (2022). K. Zhang, Y. Zhang, X. Fang, M. Fang, B. Shi, J. Dong, L. Qian, Nomograms of Combining Apparent Diffusion Coefficient Value and Radiomics for Preoperative Risk Evaluation in Endometrial Carcinoma, Frontiers in oncology 11 (2021) 705456. S. Satta, M. Dolciami, V. Celli, F. Di Stadio, G. Perniola, I. Palaia, A. Pernazza, C. Della Rocca, S. Rizzo, C. Catalano, S. Capuani, L. Manganaro, Quantitative diffusion and perfusion MRI in the evaluation of endometrial cancer: validation with histopathological parameters, The British journal of radiology 94(1125) (2021) 20210054. X. Ma, M. Shen, Y. He, F. Ma, J. Liu, G. Zhang, J. Qiang, The role of volumetric ADC histogram analysis in preoperatively evaluating the tumour subtype and grade of endometrial cancer, European journal of radiology 140 (2021) 109745. X. Ma, X. Ren, M. Shen, F. Ma, X. Chen, G. Zhang, J. Qiang, Volumetric ADC histogram analysis for preoperative evaluation of LVSI status in stage I endometrioid adenocarcinoma, European radiology 32(1) (2022) 460-469. B. Sun, X. Zhang, Y. Dong, X. Li, X. Yang, L. Zhao, J. Wang, Y. Cheng, Prognostic significance of lymphovascular space invasion in early-stage low-grade endometrioid endometrial cancer: a fifteen-year retrospective Chinese cohort study, World journal of surgical oncology 22(1) (2024) 203. C. Ma, Y. Zhao, Q. Song, X. Meng, Q. Xu, S. Tian, L. Chen, N. Wang, Q. Song, L. Lin, J. Wang, A. Liu, Multi-parametric MRI-based radiomics for preoperative prediction of multiple biological characteristics in endometrial cancer, Frontiers in oncology 13 (2023) 1280022. Tables Table 1: patient characteristics Characteristic Category/Value N % Age (years) Mean 55.85 - SD 10.34 - Radiological Tumor Size (cm) Mean SD 5.42 3.27 - - ADC Value (mm 2 /s) Mean 719.57 - SD 158.97 - Pathological Subtype Endometroid 82 86.3 Non-Endometroid 13 13.7 Pathological Grade Grade 1 29 30.5 Grade 2 37 38.8 Grade 3 25 26.3 Missing Data 4 4.2 FIGO Stage Stage 1 49 51.6 Stage 2 28 29.5 Stage 3 14 14.7 Missing Data 4 4.2 LVSI (visual assessment) Positive 78 82.1 Negative 17 17.9 LVSI (Pathological assessment) Positive 48 51 Negative 47 49 SD = Standard Deviation; ADC = Apparent diffusion coefficient; LVSI: Lymphovascular Space Invasion Table 2: Diagnostic performance of visual and quantitative assessment for LVSI Method Cutoff value Sensitivity, % (n/N; 95% CI) Specificity, % (n/N; 95% CI) Accuracy, % (n/N; 95% CI) Visual - 85.4 (41/48; 72.2-93.9) 21.3 (10/47; 10.7-35.7) 53.68 (53/95; 43.1-63.9) Quantitative 0.690 (Optimal cutoff) 75 (36/48; 60.4-86.4) 76.6 (36/47; 62-87.7) 75.78 (72/95; 65.9-84) 0.767 85.4 (41/48; 72.2-93.9) 55.3 (26/47; 40.1-69.8) 70.5 (67/95; 60.2-79.4) 0.820 89.6 (43/48; 77.3-96.5) 34 (16/47; 20.9-49.3) 62.1 (59/95; 51.5-71.8) 0.745 81.2 (39/48; 67.4-91.1) 61.7 (29/47; 46.4-75.5) 71.6 (68/95; 61.4-80.3) Cutoff 690 was determined as the optimal threshold by ROC analysis. CI = Confidence interval; ADC = Apparent diffusion coefficient. Table 3: Diagnostic performance by tumor size, FIGO stage, and pathological grade Category Group Method Sensitivity, % (n/N; 95% CI) Specificity, % (n/N; 95% CI) Accuracy, % (n/N; 95% CI) Tumor Size <2 cm Visual 100 (2/2; 15.8-100) 25 (2/8; 3.2-65.1) 40 (4/10; 12.1-73.7) Quantitative 50 (1/2; 1.3-98.7) 100 (8/8; 63.1-100) 90 (9/10; 55.5-99.7) 2-4 cm Visual 80 (8/10; 44.4-97.5) 31.6 (6/19; 12.6-56.6) 48.2 (14/29; 29.4-67.5) Quantitative 70 (7/10; 34.8-93.3) 84.2 (16/19; 60.4-96.6) 79.3 (23/29; 60.3-92.0) >4 cm Visual 86.1 (31/36; 70.5-95.3) 10 (2/20; 1.2-31.7) 58.92 (33/56; 44.97-71.90) Quantitative 77.8 (28/36; 60.8-89.9) 85.0 (17/20; 62.1-96.8) 80.3 (45/56; 67.5-89.7) FIGO Stage Stage I Visual 100 (14/14; 76.8-100) 26.5 (9/34; 12.9-44.4) 47.9 (23/48; 33.3-62.8) Quantitative 71.4 (10/14; 41.9-91.6) 71.4 (24/34; 53.7-85.4) 71.4 (34/48; 56.7-83.4) Stage II Visual 80 (16/20; 56.3-94.3) 0.0 (0/8; 0-36.9) 57.1 (16/28; 37.2-75.53) Quantitative 80 (16/20; 56.3-94.3) 87.5 (7/8; 47.3-99.7) 82.1 (23/28; 63.1-93.9) Stage III Visual 72.7 (8/11; 39-94) 50 (1/2; 1.3-98.7) 69.2 (9/13; 38.5-90.9) Quantitative 100 (11/11; 73.5-100) 50 (1/2; 1.3-98.7) 92.8 (12/13; 66.1-99.8) Pathological Grade Grade 1 Visual 100 (8/8; 63.1-100) 23.8 (5/21; 8.2-47.2) 44.8 (13/29; 26.4-64.3) Quantitative 50 (4/8; 15.7-84.3) 80.9 (17/21; 58.1-94.6) 72.4 (21/29; 52.7-87.2) Grade 2 Visual 76.5 (13/17; 50.1-93.2) 20 (4/20; 5.7-43.7) 45.9 (17/37; 29.4-63.0) Quantitative 88.2 (15/17; 63.6-98.5) 70 (14/20; 45.7-88.1) 78.4 (30/37; 61.7-90.1) Grade 3 Visual 85 (17/20; 62.1-96.8) 20 (1/5; 0.5-71.6) 72.0 (18/25; 50.6-87.9) Quantitative 90 (18/20; 68.3-98.8) 60 (3/5; 14.7-94.7) 84.0 (21/25; 63.9-95.4) Values are presented as percentages with numerator/denominator and 95% confidence intervals. Accuracy is defined as (TP + TN)/Total. CI = Confidence interval; LVSI = Lymphovascular space invasion Additional Declarations The authors declare no competing interests. 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-9151794","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":607815829,"identity":"3cc46d7c-f9e6-482b-b857-d58e462d3ce0","order_by":0,"name":"Alisa Mohebbi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIiWNgGAWjYDACdsYGEJXAIMHDcOCDgQ2Qzdh4AK8WZiQtB2dUpIG0NBDQAqHAWph5zhwG8/Bq4W9mbv7wM8cmj1+69+DBmW3n7da2HwbaUmMTjUuLxGHGNsnebWnFknPOJRz42HY7eduZRKCWY2m5Dbj0ALUw8G47nLjhRo4B0JbbyWYHgFoYGw7j1CJ/mLH541+glv1ALYd5284lm51/iF+LwWHGBmmwLRJALTxnDtiZ3SBgiyHQYdKy29ISZ9w5lwAM5OQEsxtAWxLw+EXuePvjj2+32ST2z+49/OGDgZ292fn0hw8+1Njg9j46SASrTCBWOQjYk6J4FIyCUTAKRgYAABcbb2V/JZvzAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0001-3393-3451","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":true,"prefix":"","firstName":"Alisa","middleName":"","lastName":"Mohebbi","suffix":""},{"id":607815830,"identity":"fba1c571-73a0-4bfe-8851-ea54fb745d4b","order_by":1,"name":"Mehrad Zare","email":"","orcid":"","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":false,"prefix":"","firstName":"Mehrad","middleName":"","lastName":"Zare","suffix":""},{"id":607815831,"identity":"160c7c83-576f-42fa-8ce2-1cb5f93a7ce3","order_by":2,"name":"Kimia Darmiani","email":"","orcid":"","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":false,"prefix":"","firstName":"Kimia","middleName":"","lastName":"Darmiani","suffix":""},{"id":607815832,"identity":"2d0d19ff-5c86-4508-b0d4-4caf3e239268","order_by":3,"name":"Ahmadreza Shekarchian","email":"","orcid":"","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":false,"prefix":"","firstName":"Ahmadreza","middleName":"","lastName":"Shekarchian","suffix":""},{"id":607815833,"identity":"8d606a01-2b18-4f4d-949c-92433e4faa9f","order_by":4,"name":"Fatemeh Shakki Katouli","email":"","orcid":"","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":false,"prefix":"","firstName":"Fatemeh","middleName":"Shakki","lastName":"Katouli","suffix":""},{"id":607815834,"identity":"e7253e82-8c3c-4694-ab21-8171fa968b57","order_by":5,"name":"Fahimeh Zeinalkhani","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIiWNgGAWjYBAC9gYgkcDAICMhAWJUMDAYENLCcwCihQei5QyxWhhgWhjbiNHCwP7wwcMddjySs5ufPXg477C8OXvzAYYfFdvwaGFINkg8k8wjLXPM3CBx22HDnT3HEhh7ztzGqcWegeGYRGIbM4+cRIKZBFAL44YbOQbMjG24tfAwMLb/SGyrB2pJ/yaROOewPRFamNkYEtsO80hL5ABtaTicSIQWNmagw47zSM7IKZNIOJaevOHMsYSD+PwCCrGPP9uq5SRupG+T/FFjbbvhePPBBz8qcGthkH+Awm0Gkwdwq8cEdaQoHgWjYBSMghECAHhpVGPuwPuVAAAAAElFTkSuQmCC","orcid":"","institution":"Tehran University of Medical Sciences, Tehran, Iran","correspondingAuthor":true,"prefix":"","firstName":"Fahimeh","middleName":"","lastName":"Zeinalkhani","suffix":""}],"badges":[],"createdAt":"2026-03-17 18:27:49","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9151794/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9151794/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105149544,"identity":"e2f26b26-9209-4777-b620-475393c33e6e","added_by":"auto","created_at":"2026-03-22 14:56:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":283794,"visible":true,"origin":"","legend":"\u003cp\u003e(A) ROC curve illustrating diagnostic performance of ADC values for the detection of LVSI. (B) ROC curve illustrating diagnostic performance of combined model incorporating ADC values and visual assessment for LVSI detection.\u003c/p\u003e\n\u003cp\u003e(B) ROC curve illustrating diagnostic performance of combined model incorporating ADC values and visual assessment for LVSI detection.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9151794/v1/1f204af996994a96696b35d9.png"},{"id":105149541,"identity":"6e732e0d-21b5-4884-965d-3d981d7c52b5","added_by":"auto","created_at":"2026-03-22 14:56:08","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1155284,"visible":true,"origin":"","legend":"\u003cp\u003eMultiparametric MRI of confirmed endometrioid carcinoma, grade 1, with lymphovascular invasion and \u0026lt;50% myometrial invasion (FIGO stage IIB). \u0026nbsp;(A) Axial T2-weighted image shows a heterogeneous endometrial mass.\u003c/p\u003e\n\u003cp\u003e(B) Contrast-enhanced axial T1-weighted image demonstrates heterogeneous enhancement.\u003c/p\u003e\n\u003cp\u003e(C) Diffusion-weighted image reveals marked diffusion restriction.\u003c/p\u003e\n\u003cp\u003e(D) Corresponding ADC map demonstrates low signal intensity (mean ADC = 0.653 × 10⁻3 mm²/s).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9151794/v1/e90bbae617d77e62d7b55827.png"},{"id":105149542,"identity":"b3ae963a-29ae-4871-bbf8-1db567952d20","added_by":"auto","created_at":"2026-03-22 14:56:08","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":665742,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Multiparametric MRI of confirmed endometrioid carcinoma, grade 1, without lymphovascular invasion and with \u0026lt;50% myometrial invasion (FIGO stage IA). (A) Axial T2-weighted image demonstrates a heterogeneous endometrial mass;\u003c/p\u003e\n\u003cp\u003e(B) Contrast-enhanced axial T1-weighted image shows heterogeneous enhancement;\u003c/p\u003e\n\u003cp\u003e(C) Diffusion-weighted image (echo-planar sequence) reveals diffusion restriction;\u003c/p\u003e\n\u003cp\u003e(D) ADC map demonstrates low signal intensity (mean ADC = 1.004 × 10⁻3mm²/s).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9151794/v1/e55a8876b8241d95e2ee79e6.png"},{"id":105563863,"identity":"cce1e20b-4b96-49ab-9f5d-75ec4ce81e83","added_by":"auto","created_at":"2026-03-27 12:48:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3337697,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9151794/v1/9e65842a-9f4c-4706-a456-ee475d8adc2d.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eADC-Based MRI Achieves Superior Accuracy in Preoperative LVSI Prediction for Endometrial Cancer\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEndometrial cancer is the most common gynecologic cancer in developed countries, with a rising incidence over recent decades. Despite generally favorable outcomes for early-stage disease, the identification of high-risk features remains crucial for optimal treatment planning and prognosis assessment [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Lymphovascular space invasion (LVSI), defined as the presence of tumor cells within blood or lymphatic vessels beyond the core tumor, represents one of the most significant prognostic indicators in endometrial cancer [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. According to the International Federation of Gynecology and Obstetrics (FIGO) staging system (2023), substantial LVSI (defined as involvement of \u0026ge;\u0026thinsp;5 vessels) is a key determinant of endometrial cancer staging and prognosis [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Currently, LVSI assessment relies predominantly on postoperative histopathological examination, limiting its utility in preoperative surgical planning and treatment stratification. The inability to reliably determine LVSI status preoperatively presents significant clinical challenges, as this information could guide decisions regarding the extent of lymphadenectomy, adjuvant therapy recommendations, and overall treatment approach [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Therefore, developing reliable preoperative imaging biomarkers for LVSI detection represents a critical need in endometrial cancer management [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eMagnetic resonance imaging (MRI) has emerged as the mainstay of preoperative staging for endometrial cancer, providing excellent soft tissue contrast and accurate assessment of myometrial invasion, cervical involvement, and lymph node status. Diffusion-weighted imaging (DWI), an advanced functional MRI technique, exploits the principle that malignant tissues demonstrate restricted water diffusion due to increased cellularity and disrupted cellular architecture, resulting in characteristic signal patterns that differ from normal myometrial tissue [\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. DWI can be evaluated through two distinct approaches: qualitative visual assessment and quantitative analysis using the apparent diffusion coefficient (ADC).\u003c/p\u003e \u003cp\u003eQualitative DWI assessment relies on subjective visual interpretation of signal intensity changes across different b-values [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. This approach identifies areas of restricted diffusion by recognizing persistent high signal intensity on high b-value images, which in endometrial cancer may correlate with tumor aggressiveness and LVSI presence [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. The primary advantages of qualitative assessment include rapid interpretation, widespread accessibility without specialized software, and the ability to incorporate overall tumor morphology into the diagnostic evaluation [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. However, qualitative DWI suffers from inherent limitations, including significant interobserver variability, subjective interpretation bias, and susceptibility to the T2 shine-through effect, where tissues with prolonged T2 relaxation times can mimic restricted diffusion which lead to false-positive interpretations, particularly in cystic or necrotic tumor components [\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOn the other hand, quantitative DWI analysis using ADC measurements provides an objective, numerical assessment of water diffusion. ADC values offer a standardized metric that is less susceptible to subjective interpretation and demonstrates superior reproducibility across different observers and imaging centers [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Lower ADC values typically correlate with higher tumor cellularity, increased aggressiveness, and the presence of adverse pathological features, including LVSI [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. The quantitative nature of ADC allows for the establishment of specific threshold values for diagnostic purposes and enables standardized comparison across studies and institutions. Despite these advantages, quantitative ADC assessment faces several limitations. ADC values can be influenced by technical factors, including scanner manufacturer, magnetic field strength, acquisition parameters, and sequence protocols, potentially affecting the transferability of threshold values across different imaging systems [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The measurement process requires careful region-of-interest placement and may be affected by tumor heterogeneity, partial volume effects, and motion artifacts [\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Furthermore, while ADC provides excellent objectivity, it may not capture subtle morphological features that experienced radiologists can appreciate through visual assessment.\u003c/p\u003e \u003cp\u003ePrevious studies investigating DWI in endometrial cancer have primarily focused on individual assessment approaches, with limited direct comparison between qualitative and quantitative methods specifically for LVSI detection. While ADC has shown promise in predicting various endometrial cancer characteristics, including tumor grade and myometrial invasion depth, its comparative diagnostic performance against qualitative assessment for LVSI remains incompletely characterized. Understanding the relative strengths and limitations of each approach is essential for determining the optimal DWI evaluation strategy in clinical practice.\u003c/p\u003e \u003cp\u003eTherefore, the primary aim of this study was to conduct a comprehensive head-to-head comparison between qualitative visual assessment and quantitative ADC analysis for preoperative LVSI detection in endometrial cancer. Secondary aims included evaluating this comparison across different tumor sizes, FIGO stages, and pathological grades to determine which approach maintains diagnostic value in challenging clinical scenarios. Additionally, a comprehensive literature review was performed regarding ADC cut-off points for LVSI determination, with subsequent validation of their diagnostic efficacy in our dataset.\u003c/p\u003e"},{"header":"Materials \u0026 Methods","content":"\u003cp\u003eThis retrospective, single-center study was conducted with the approval of the institutional review board, in accordance with the ethical standards outlined in the Declaration of Helsinki. Given the use of anonymized medical records, the ethics committee waived the requirement for informed consent. This investigation aimed to compare visually based DWI assessment and quantitative ADC analysis for preoperative detection of LVSI in patients with confirmed endometrial cancer.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient Selection \u0026amp; Study Cohort\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMedical records were reviewed to identify women who underwent surgical treatment for histologically confirmed endometrial carcinoma between January 2019 and November. Inclusion criteria required a definitive histopathological diagnosis of endometrial carcinoma, availability of preoperative pelvic MRI scans including diffusion-weighted sequences, and complete pathological documentation of LVSI status. Exclusion criteria encompassed cases with suspected non-endometrial malignancy or mixed histology, the presence of distant metastases at initial staging (e.g., FIGO stage IV), incomplete or poor-quality MRI examinations that precluded reliable DWI evaluation, the absence of recorded ADC or visual DWI assessments, and missing postoperative LVSI data. The initial dataset comprised 184 consecutive patients. Application of these criteria resulted in the exclusion of 89 patients: 42 due to the absence of LVSI pathology reports, 27 for inadequate MRI quality, and 20 lacking ADC values or visual DWI assessments. Thus, a final cohort of 95 patients was included for analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMRI Acquisition Protocol\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll MRI examinations were performed on a three-Tesla system (GE Discovery\u0026trade; 750 GEM; GE Healthcare, Chicago, IL, USA) equipped with an 8-channel phased-array pelvic coil. Patients were instructed to fast for at least four hours prior to scanning to minimize bowel peristalsis, and an intravenous injection of 20 mg hyoscine N-butyl bromide was administered immediately before the examination unless contraindicated. Vaginal gel was instilled to distend the vaginal canal unless patient intolerance precluded its use. The imaging protocol conformed to institutional standards for endometrial cancer evaluation and included axial and sagittal T2-weighted turbo spin-echo sequences (repetition time (TR) = 4,500 ms; echo time (TE) = 100 ms; slice thickness = 4 mm; interslice gap = 1 mm; field of view (FOV) = 24 cm; matrix = 320 \u0026times; 256), axial T1-weighted spin-echo images (TR = 650 ms; TE = 12 ms; slice thickness = 4 mm; interslice gap = 1 mm; FOV = 24 cm; matrix = 256 \u0026times; 224), and diffusion-weighted imaging acquired in the axial plane using single-shot echo-planar imaging with b-values of 0, 500, and 1,000 s/mm\u0026sup2; (TR = 4,000 ms; TE = 70 ms; slice thickness = 4 mm; interslice gap = 1 mm; FOV = 24 cm; matrix = 128 \u0026times; 128). Diffusion gradients were applied in three orthogonal directions, and ADC maps were automatically generated by the vendor-supplied software. In all cases, contrast-enhanced sequences were performed using a standard gadolinium-based contrast agent (0.1 mmol/kg). The agent was injected at a dose of 0.2 ml/kg body weight, at a rate of 3 ml/sec using an injector, followed by 10 cc of normal saline. Post-contrast images were obtained dynamically at intervals of 30, 60, 120, 180, and 300 seconds after injection.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImage Retrieval \u0026amp; Blinding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll MRI studies were retrieved from the institution\u0026rsquo;s Picture Archiving and Communication System (PACS). Two centers with independent board-certified radiologists with ten and fifteen years of experience in gynecological imaging, contributed to the study. There was no requirement for inter-observer agreement analysis or result combining because each case was assessed by a single radiologist who was related to their center. Every case from Center 1 was assessed by Reader 1, and each one from Center 2 was assessed by Reader 2. Radiologists were blinded to clinical and pathological information, including patient age, tumor grade, FIGO stage, and especially LVSI status. Each radiologist conducted visual DWI assessment and quantitative ADC measurement in separate sessions at least two weeks apart. Qualitative evaluation of diffusion-weighted images was based on visual assessment of signal intensity on high b-value (1,000 s/mm\u0026sup2;) images. Qualitative evaluation focused on detecting restricted diffusion within the endometrial tumor on images, serving as a surrogate marker for LVSI potential. Tumors were categorized as positive if they exhibited heterogeneous signal patterns with focal areas of significant diffusion restriction. This method assesses tissue microstructural alterations linked to aggressive tumor characteristics rather than directly evaluating anatomical invasion. The visual assessment specifically aimed to identify restricted diffusion patterns correlating with tumor aggressiveness and the presence of LVSI, rather than directly visualizing vascular invasion.\u003c/p\u003e\n\u003cp\u003eQuantitative analysis was performed using the same DWI source data. On ADC maps, regions of interest (ROIs) were manually placed to encompass the solid component of the tumor. Three circular ROIs (minimum 20mm\u0026sup2;) were manually placed on ADC maps in areas showing the lowest ADC values within the tumor, avoiding necrotic areas, blood vessels, and artifacts. The median ADC value of the three measurements was used for analysis, consistent with established protocols for endometrial cancer ADC measurement [25, 26]. The representative ADC for that tumor expressed in units of \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s. ADC thresholds for LVSI detection were determined in two ways: first, by calculating the optimal cut-off value from the study cohort using Youden\u0026rsquo;s index applied to receiver operating characteristic (ROC) analysis; and second, by applying previously published ADC thresholds (e.g., 690, 745, and 820 \u0026times;10⁻\u003csup\u003e6\u003c/sup\u003e mm\u0026sup2;/s) to the dataset to assess external validity [27-29].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathological Reference Standard\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePostoperative specimens were processed according to standard histopathological protocols. Tumor type, FIGO grade, depth of myometrial invasion, cervical stromal involvement, and LVSI status were reported by experienced gynecologic pathologists blinded to MRI findings. LVSI was defined as the presence of tumor cells within endothelial-lined vascular channels (blood vessels or lymphatics) beyond the main tumor mass. Only \u0026ldquo;substantial\u0026rdquo; LVSI, entailing invasion of five or more vessels per slide, was recorded as positive, in alignment with the 2023 FIGO staging guidelines. Cases with ambiguous vascular invasion on initial review were subjected to immunohistochemical staining for endothelial markers (CD31 and D2-40) to confirm LVSI.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTumor Size, Stage, and Grade Stratification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRadiological tumor size was measured on T2-weighted images as the maximum straight-line diameter in axial, sagittal, or coronal planes. Tumors were stratified into three size categories; \u0026lt;2 cm, 2-4 cm, and \u0026gt;4 cm to evaluate potential size-dependent variation in diagnostic performance [30, 31]. Clinical staging was assigned preoperatively based on MRI findings and recorded according to the 2023 FIGO classification: stage I (tumor confined to corpus uteri), stage II (cervical stromal invasion without extrauterine spread), and stage III (extension to serosa, adnexa, vagina, parametria, or pelvic/para-aortic lymph nodes). Pathological grade was classified as grade 1 (well differentiated, \u0026le;5% solid growth), grade 2 (moderately differentiated, 6-50% solid growth), or grade 3 (poorly differentiated, \u0026gt;50% solid growth), with upgrading by one grade for severe nuclear atypia when appropriate.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were conducted using MedCalc version 23.2 (MedCalc Software, Ostend, Belgium) and Stata version 18.0 (StataCorp, College Station, TX, USA). Continuous variables were reported as mean \u0026plusmn; standard deviation or median and interquartile range, depending on the normality of the distribution, assessed by the Shapiro\u0026ndash;Wilk test. Categorical data were expressed as counts and percentages. The diagnostic performance of qualitative DWI and quantitative ADC analysis for LVSI detection was evaluated by calculating sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and the area under the ROC curve (AUC) with 95% confidence intervals. Sensitivity, specificity, and accuracy comparisons were performed using dependent proportion difference tests. This approach provides both point estimates and confidence intervals for direct comparison of diagnostic performance. For all analyses, a true positive was defined as LVSI identified by both MRI and pathology, a true negative as LVSI absent on both MRI and pathology, a false positive as MRI-positive but pathology-negative, and a false negative as MRI-negative but pathology-positive. Optimal ADC cut-off values were determined by maximizing Youden\u0026rsquo;s index. Pairwise comparisons of AUCs between qualitative and quantitative methods, as well as between different ADC thresholds, were performed using the DeLong test. Subgroup analyses were performed using AUC comparisons rather than separate cutoff values to maintain standardization and validity; these analyses assessed diagnostic metrics within tumor size, FIGO stage, and pathological grade categories. A p-value of \u0026lt;0.05 was considered statistically significant for all comparisons.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eDemographic and clinicopathological characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe final study cohort comprised 95 patients with histologically confirmed endometrial cancer who met all inclusion criteria (Table 1). The mean age was 55.85 \u0026plusmn; 10.34 years (range: 30-77 years). The mean tumor size on MRI was 5.42 \u0026plusmn; 3.27 cm. Histopathological examination revealed that 82 (86.3%) of tumors were endometrioid carcinoma, while 13 (13.7%) were of non-endometrioid subtypes. Pathological grading revealed that 29 (31.8%) were grade 1, 37 (40.6%) were grade 2, and 25 (27.6%) were grade 3. The LVSI was present in 48 (51%) of patients and absent in 47 (49%).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverall Performance of Qualitative vs. Quantitative DWI\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQualitative visual assessment of diffusion-weighted images demonstrated a limited overall accuracy of 53.7% for preoperative prediction of LVSI with a high sensitivity (85.4%) but low specificity (21.3%). In contrast, quantitative ADC analysis yielded markedly higher performance. Using an optimal ADC threshold of \u0026le;0.690 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s, quantitative assessment achieved an overall accuracy of 75.8%, with a sensitivity of 75.0% and a specificity of 76.6%. Moreover, the AUC for ADC-based detection was 0.770 (95% CI, 0.673 to 0.850; p \u0026lt; 0.001), indicating moderate discriminative ability and a notable improvement over visual evaluation (Figure 1A) (Table 2)[28, 29, 32].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhen applying previously published ADC cut-off values of 0.690, 0.745, and 0.820 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s to this dataset, accuracies ranged between 62.1% and 71.6%, consistently surpassing the performance of qualitative assessment (Table 2). The pairwise comparisons among literature cut-offs revealed no significant differences between all these cut-point performances (all p \u0026gt; 0.1). However, when directly compared with qualitative DWI, quantitative analysis using thresholds of \u0026le;0.690 and \u0026le;0.745 demonstrated significantly better diagnostic performance (p = 0.017 and p = 0.011, respectively), whereas a threshold of \u0026lt; 0.820 did not (p = 0.241). These findings, based on external validation, reinforce the observed trend that quantitative ADC assessment provides greater diagnostic reliability for LVSI detection than visual interpretation.\u003c/p\u003e\n\u003cp\u003eIn addition, when both quantitative and qualitative DWI assessments were combined using binary logistic regression, the diagnostic performance of quantitative ADC analysis alone was identical to that of the combined ADC and qualitative visual assessment model (AUC = 0.770, CI: 0.672 to 0.850; p \u0026lt; 0.001) (Figure 1B) (p = 0.956) indicating that further visual assessment to ADC does not provide additional discriminatory and informativeness value for detecting LVSI. Figures 2 and 3 illustrate two clinical examples.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePerformance Stratified by Radiological Tumor Size\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the subgroup with the smallest tumors (\u0026lt;2 cm), qualitative DWI demonstrated high sensitivity but poor specificity, resulting in an overall accuracy of 40.0%. Quantitative ADC analysis, with an optimal threshold of \u0026le;1.013 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s in this subset, achieved a significant increase in accuracy to 90.0% and an AUC of 0.688 (CI: 0.336 to 0.927) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eFor tumors of intermediate size (2-4 cm), qualitative assessment again suffered from low specificity, corresponding to a modest accuracy of 48.3%. Quantitative ADC measurement at an optimized cut-off of \u0026le;0.609 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s improved accuracy to 79.3% and produced an AUC of 0.803 (95% CI, 0.614 to 0.926), reinforcing the superiority of cohort-specific ADC calibration compared to visual interpretation alone (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eIn larger tumors (\u0026gt;4 cm), qualitative DWI achieved an accuracy of 58.9%. ADC analysis at the optimal threshold of \u0026le;0.690 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s increased overall accuracy to 80.3% and reached an AUC of 0.831 (95% CI, 0.707 to 0.918), the highest among size-based subgroups (all p \u0026gt; 0.05). The detailed diagnostic metrics are presented in Table 3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiagnostic Performance by FIGO Stage\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn stage I (n = 49), visual DWI exhibited an overall accuracy of 47.9%. Quantitative ADC evaluation, applying a threshold of \u0026le;0.710 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s, improved accuracy to 71.4% and achieved an AUC of 0.679 (CI, 0.530 to 0.805) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eAmong patients with stage II disease (n = 28), qualitative DWI accuracy reached only 57.1%. ADC analysis at the threshold of \u0026le;0.656 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s yielded an accuracy of 82.1% and an AUC of 0.847 (CI, 0.661 to 0.954) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eIn stage III (n = 14), qualitative visual assessment achieved an accuracy of 69.2%. ADC measurement using an optimal threshold of \u0026le;0.813 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s achieved the highest subgroup accuracy of 92.9%, although the small sample size limits the precision of this estimate. The corresponding AUC was 0.625 (CI, 0.335 to 0.861) (all p \u0026gt; 0.05). The detailed diagnostic metrics are presented in Table 3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiagnostic performance by pathological grading\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn grade 1 tumors, qualitative DWI accuracy was 44.8%. ADC analysis using a threshold of \u0026le;0.710 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s improved accuracy to 72.4% and produced an AUC of 0.616 (CI, 0.418 to 0.789) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eFor grade 2 lesions, qualitative assessment achieved an accuracy of 46.0%, whereas ADC measurement at \u0026le;0.690 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s delivered an accuracy of 78.4% and an AUC of 0.810 (CI: 0.648 to 0.920) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eIn grade 3 tumors, qualitative DWI reached an accuracy of 72.0%. Quantitative ADC analysis, with an optimal threshold of \u0026le;0.711 \u0026times;10⁻\u0026sup3; mm\u0026sup2;/s, attained an accuracy of 84.0% and an AUC of 0.740 (CI, 0.527 to 0.893) (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003eThe detailed diagnostic metrics are presented in Table 3.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, the head-to-head comparison in 95 patients demonstrated that ADC-based evaluation not only achieved superior overall accuracy (75.8%) compared with qualitative assessment (53.7%), but also maintained consistent performance across tumor size, FIGO stage, and pathological grade subgroups. These findings underscore the critical importance of integrating quantitative diffusion metrics into routine preoperative MRI protocols to inform surgical planning and the selection of adjuvant therapy.\u003c/p\u003e \u003cp\u003eOur study\u0026rsquo;s principal observation was the substantially higher specificity of ADC analysis relative to visual assessment, which suffered from very low specificity (21.3%) despite high sensitivity (85.4%). This discrepancy reflects the inherent subjectivity and susceptibility of qualitative DWI to T2 shine-through and interobserver variability [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. By contrast, ADC offers standardized, numerical thresholds (optimized in our cohort at \u0026le;\u0026thinsp;0.690 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s) that balance sensitivity (75.0%) and specificity (76.6%), translating into a more reliable preoperative indicator of LVSI. Notably, when applying previously published thresholds (\u0026le;\u0026thinsp;0.690, \u0026le;\u0026thinsp;0.745, \u0026le; 0.820 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s), ADC performance consistently surpassed that of qualitative DWI, reinforcing its robustness across diverse settings [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCrucially, combining visual assessment with ADC did not improve diagnostic performance beyond quantitative analysis alone (AUC\u0026thinsp;=\u0026thinsp;0.770), indicating that ADC encapsulates the salient diffusion characteristics necessary for LVSI detection. This aligns with prior reports that highlight the limited incremental value of visual ADC map evaluation once quantitative metrics are established [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. In clinical practice, this suggests that resource-intensive, dual-reader visual assessments may be supplanted by streamlined quantitative protocols without a loss of diagnostic fidelity.\u003c/p\u003e \u003cp\u003eOur subgroup analyses revealed that ADC maintained its diagnostic advantage irrespective of tumor size, stage, or grade. In tumors\u0026thinsp;\u0026lt;\u0026thinsp;2 cm, ADC analysis with a threshold of \u0026le;\u0026thinsp;1.013\u0026times;10⁻\u0026sup3; mm\u0026sup2;/s achieved 90.0% accuracy (AUC\u0026thinsp;=\u0026thinsp;0.688), indicating its resilience in small lesions often challenging for visual interpretation. Similarly, in FIGO stage II and III tumors, ADC thresholds of \u0026le;\u0026thinsp;0.656 and \u0026le;\u0026thinsp;0.813 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s yielded accuracies of 82.1% and 92.9%, respectively, underscoring the applicability of ADC across disease extents. These findings echo those of Ma et al., who reported ADC\u0026rsquo;s stable performance across heterogeneous pathological groups [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. These findings align with previous research. For instance, Petrila et al. observed that the mean ADC inversely correlated with histological grade, deep myometrial invasion, and LVSI in endometrial cancer, underlining its association with tumor aggressiveness [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Similarly, Wang et al. reported links between ADC values and high-risk features, including LVSI, reinforcing the diagnostic relevance of ADC across varying tumor behaviors [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Additionally, a multicenter radiomics study by Liu et al. found that radiomic models incorporating ADC and texture features predicted LVSI more effectively than traditional imaging, highlighting the added value of quantitative data [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Similar advantages of ADC over qualitative interpretation have also been reported by Satta et al. [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] and Ma et al. [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e], who found that quantitative diffusion metrics are strong predictors of adverse histopathological features in endometrial cancer. Collectively, these data support ADC\u0026rsquo;s role as a generalizable imaging biomarker that transcends conventional tumor stratifications.\u003c/p\u003e \u003cp\u003eThe clinical implications of our findings are considerable. Preoperative identification of LVSI can guide the extent of lymphadenectomy, inform neoadjuvant treatment decisions, and refine risk stratification within multidisciplinary tumor boards [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Existing nomograms incorporating ADC and radiomic features demonstrate AUCs up to 0.959 for LVSI prediction [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e], yet our data suggest that even standalone ADC metrics without complex radiomics pipelines offer high diagnostic yield, facilitating broader adoption in settings lacking advanced software infrastructures. Moreover, our external validation of literature-derived ADC cut-offs mitigates concerns of cohort-specific overfitting. All published thresholds tested outperformed qualitative DWI, confirming ADC\u0026rsquo;s reproducibility across scanners and institutions despite known influences of magnetic field strength and acquisition parameters. Nonetheless, site-specific calibration of ADC thresholds remains advisable to account for technical variabilities and ensure optimal diagnostic accuracy.\u003c/p\u003e \u003cp\u003eOur results complement the growing body of multiparametric MRI research. Studies integrating diffusion, perfusion (IVIM, DCE), and volumetric metrics report enhanced prognostic stratification, linking quantitative parameters to tumor grade, depth of invasion, and inflammatory infiltrates [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. While such multiparametric approaches hold promise for comprehensive tumor characterization, our findings emphasize that ADC measurement alone provides a potent, objective marker for LVSI, warranting its prioritization in preoperative imaging protocols.\u003c/p\u003e \u003cp\u003eThis study has several limitations that may affect the generalizability and interpretation of our findings. First, its retrospective design may limit generalizability, and validation in larger, multicenter cohorts is warranted. Second, the relatively modest sample size of 95 patients, particularly in subgroup analyses, may reduce statistical power and precision of estimates, potentially limiting the robustness of conclusions in specific clinical scenarios. Third, the manual placement of regions of interest for ADC quantification introduces potential operator-dependent variability and may be affected by tumor heterogeneity, partial volume effects, and motion artifacts, despite our attempts to standardize the measurement protocol.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eQuantitative ADC analysis outperformed qualitative DWI for the preoperative detection of LVSI in endometrial cancer, demonstrating higher accuracy and greater diagnostic reliability across the overall cohort and all clinical subgroups. Its consistent performance, even when using literature-derived cut-offs, and the lack of added value from combining it with visual assessment, support its integration into routine MRI protocols. Incorporating quantitative ADC measurement into clinical practice may improve preoperative risk stratification and guide individualized treatment planning in patients with endometrial cancer.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAll authors contributed significantly to this article.\u003c/p\u003e\n\u003cp\u003eFunding: None\u003c/p\u003e\n\u003cp\u003eConflict of interest: None\u003c/p\u003e\n\u003cp\u003eInformed consent: Not applicable.\u003c/p\u003e\n\u003cp\u003eAnimal study: N/A\u003c/p\u003e\n\u003cp\u003eAI-based tools were used only for language editing; all scientific content, data analysis, and interpretations are the authors\u0026rsquo; own.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgment:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to acknowledge all those who contributed to gathering the dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;The datasets generated and analyzed during the current study are not publicly available due to patient privacy and institutional restrictions, but are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eE.J. Crosbie, S.J. Kitson, J.N. McAlpine, A. Mukhopadhyay, M.E. Powell, N. Singh, Endometrial cancer, Lancet (London, England) 399(10333) (2022) 1412-1428.\u003c/li\u003e\n\u003cli\u003eZ. Li, J. Peng, B. Zhang, C. Zhao, Z. Chen, H. Xiao, Z. Zhang, X. Ma, F. Gao, R. Xin, W. Wang, S. Wang, L. Tang, Y. Zhang, B. Kong, L. Li, A. Yin, The prognostic and clinical significance of substantial lymphovascular space invasion in early-stage endometrial carcinoma, European journal of cancer (Oxford, England : 1990) 218 (2025) 115258.\u003c/li\u003e\n\u003cli\u003eF. Yarandi, E. Shirali, S. Akhavan, F. Nili, S. Ramhormozian, The impact of lymphovascular space invasion on survival in early stage low-grade endometrioid endometrial cancer, European journal of medical research 28(1) (2023) 118.\u003c/li\u003e\n\u003cli\u003eM. Menendez-Santos, C. Gonzalez-Baerga, D. Taher, R. Waters, M. Virarkar, P. Bhosale, Endometrial Cancer: 2023 Revised FIGO Staging System and the Role of Imaging, Cancers 16(10) (2024).\u003c/li\u003e\n\u003cli\u003eM.R. Oliver-Perez, P. Padilla-Iserte, O. Arencibia-Sanchez, C. Martin-Arriscado, J.C. Muruzabal, B. Diaz-Feij\u0026oacute;o, S. Cabrera, P. Coronado, M.B. Mart\u0026iacute;n-Salamanca, M. Pantoja-Garrido, J. Marcos-Sanmartin, E. Cabezas-L\u0026oacute;pez, C. Lorenzo, D. Beric, J.R. Rodriguez-Hernandez, F. Roldan-Rivas, J. Gilabert-Estelles, L. Sanchez, M. Laseca-Modrego, C. Tauste-Rubio, B. Gil-Iba\u0026ntilde;ez, A. Tejerizo-Garcia, G. On Behalf Of The Spain-Gog, Lymphovascular Space Invasion in Early-Stage Endometrial Cancer (LySEC): Patterns of Recurrence and Predictors. A Multicentre Retrospective Cohort Study of the Spain Gynecologic Oncology Group, Cancers 15(9) (2023).\u003c/li\u003e\n\u003cli\u003eX.F. Liu, B.C. Yan, Y. Li, F.H. Ma, J.W. Qiang, Radiomics feature as a preoperative predictive of lymphovascular invasion in early-stage endometrial cancer: A multicenter study, Frontiers in oncology 12 (2022) 966529.\u003c/li\u003e\n\u003cli\u003eO. Petrila, I. Nistor, N.S. Romedea, D. Negru, V. Scripcariu, Can the ADC Value Be Used as an Imaging \u0026quot;Biopsy\u0026quot; in Endometrial Cancer?, Diagnostics (Basel, Switzerland) 14(3) (2024).\u003c/li\u003e\n\u003cli\u003eT. Ota, T. Tsuboyama, H. Onishi, A. Nakamoto, H. Fukui, K. Yano, T. Honda, K. Kiso, M. Tatsumi, N. Tomiyama, Diagnostic accuracy of MRI for evaluating myometrial invasion in endometrial cancer: a comparison of MUSE-DWI, rFOV-DWI, and DCE-MRI, La Radiologia medica 128(6) (2023) 629-643.\u003c/li\u003e\n\u003cli\u003eA.S.L. Moreira, V. Ribeiro, G. Aringhieri, S.C. Fanni, L. Tumminello, L. Faggioni, D. Cioni, E. Neri, Endometrial Cancer Staging: Is There Value in ADC?, Journal of personalized medicine 13(5) (2023).\u003c/li\u003e\n\u003cli\u003eV. Baliyan, C.J. Das, R. Sharma, A.K. Gupta, Diffusion weighted imaging: Technique and applications, World journal of radiology 8(9) (2016) 785-798.\u003c/li\u003e\n\u003cli\u003eM. Xie, Z. Ren, D. Bian, D. Li, L. Yu, F. Zhu, R. Huang, Z. Zhang, S. Suye, C. Fu, High resolution diffusion-weighted imaging with readout segmentation of long variable echo-trains for determining myometrial invasion in endometrial carcinoma, Cancer imaging : the official publication of the International Cancer Imaging Society 20(1) (2020) 66.\u003c/li\u003e\n\u003cli\u003eR. A, J.T. M, C. N, S. M, M.G. H, A. Gh, Signal Intensity of High B-value Diffusion-weighted Imaging for the Detection of Prostate Cancer, Journal of biomedical physics \u0026amp; engineering 9(4) (2019) 453-458.\u003c/li\u003e\n\u003cli\u003eJ.J. Wang, X.H. Zhang, X.H. Guo, Y. Ying, X. Wang, Z.H. Luan, W.Q. Lv, P.F. Wang, Prediction of Lymphovascular Space Invision in Endometrial Cancer based on Multi-parameter MRI Radiomics Model, Current medical imaging 20 (2024) e15734056266366.\u003c/li\u003e\n\u003cli\u003eL. Quinn, K. Tryposkiadis, J. Deeks, H.C.W. De Vet, S. Mallett, L.B. Mokkink, Y. Takwoingi, S. Taylor-Phillips, A. Sitch, Interobserver variability studies in diagnostic imaging: a methodological systematic review, The British journal of radiology 96(1148) (2023) 20220972.\u003c/li\u003e\n\u003cli\u003eY. Kim, S.K. Lee, J.Y. Kim, J.H. Kim, Pitfalls of Diffusion-Weighted Imaging: Clinical Utility of T2 Shine-through and T2 Black-out for Musculoskeletal Diseases, Diagnostics (Basel, Switzerland) 13(9) (2023).\u003c/li\u003e\n\u003cli\u003eK. Fukushima, K. Sano, H. Machida, T. Kariyasu, T. Yoshioka, S. Takahashi, A. Nakanishi, H. Kusahara, K. Yokoyama, Effect of an Ultrahigh b Value of 3000 s/mm(2) and the Minimal Echo-time on Image Quality and the T2 Shine-through Effect in Diffusion-weighted Imaging of the Liver at 3T: Phantom and Clinical Pilot Studies, Magnetic resonance in medical sciences : MRMS : an official journal of Japan Society of Magnetic Resonance in Medicine 22(2) (2023) 232-240.\u003c/li\u003e\n\u003cli\u003eF. Zeinalkhani, P. Kamali Hakim, M. Aghasi, F. Mahdavi Sabet, R. Mortazavi Ardestani, Y. Baghban, M. Malek, H. Zeinalkhani, S. Delazar, F. Shakki Katouli, F. Azizinik, S. Moghaddamimonaghi, S. Mohammadzadeh, Direct comparison of full protocol MRI and modified non-contrast MRI in staging of cervical cancer: a retrospective study, Abdominal radiology (New York) (2025).\u003c/li\u003e\n\u003cli\u003eD.C. Newitt, Z. Zhang, J.E. Gibbs, S.C. Partridge, T.L. Chenevert, M.A. Rosen, P.J. Bolan, H.S. Marques, S. Aliu, W. Li, L. Cimino, B.N. Joe, H. Umphrey, H. Ojeda-Fournier, B. Dogan, K. Oh, H. Abe, J. Drukteinis, L.J. Esserman, N.M. Hylton, Test-retest repeatability and reproducibility of ADC measures by breast DWI: Results from the ACRIN 6698 trial, Journal of magnetic resonance imaging : JMRI 49(6) (2019) 1617-1628.\u003c/li\u003e\n\u003cli\u003eA. Keven, E.E. Yetim, A. Elmalı, A.G. Arslan, S.M. \u0026Ccedil;ubuk, The role of diffusion magnetic resonance imaging in determining tumor aggressiveness during preoperative surgical planning in early-stage endometrial cancer, Radiologie (Heidelberg, Germany) 63(Suppl 2) (2023) 41-48.\u003c/li\u003e\n\u003cli\u003eC. Moore, C. Bull, A. Darekar, D. Wilson, A. Goodall, P. Manoharan, P. Hoskin, M. van Herk, D.L. Buckley, D.J. McHugh, A. Datta, M.J. Dubec, Multi-institution longitudinal apparent diffusion coefficient measurements in a diffusion weighted imaging phantom at room temperature, Physics and imaging in radiation oncology 35 (2025) 100814.\u003c/li\u003e\n\u003cli\u003eB. Moreau, A. Iannessi, C. Hoog, H. Beaumont, How reliable are ADC measurements? A phantom and clinical study of cervical lymph nodes, European radiology 28(8) (2018) 3362-3371.\u003c/li\u003e\n\u003cli\u003eH. Takahashi, K. Yoshida, A. Kawashima, N.J. Lee, A.T. Froemming, D.A. Adamo, A. Khandelwal, C.W. Bolan, M.T. Heller, R.P. Hartman, B. Kim, K.A. Philbrick, R.E. Carter, L.A. Mynderse, M.R. Humphreys, J.C. Cai, N. Takahashi, Impact of measurement method on interobserver variability of apparent diffusion coefficient of lesions in prostate MRI, PloS one 17(5) (2022) e0268829.\u003c/li\u003e\n\u003cli\u003eM. Gity, B. Moradi, R. Arami, A. Arabkheradmand, M.A. Kazemi, Two Different Methods of Region-of-Interest Placement for Differentiation of Benign and Malignant Breast Lesions by Apparent Diffusion Coefficient Value, Asian Pacific journal of cancer prevention : APJCP 19(10) (2018) 2765-2770.\u003c/li\u003e\n\u003cli\u003eJ.S. Son, H.S. Park, S. Park, Y.J. Kim, M.H. Yu, S.I. Jung, M. Paek, M.D. Nickel, Motion-Corrected versus Conventional Diffusion-Weighted Magnetic Resonance Imaging of the Liver Using Non-Rigid Registration, Diagnostics (Basel, Switzerland) 13(6) (2023).\u003c/li\u003e\n\u003cli\u003eK. Tamai, T. Koyama, T. Saga, S. Umeoka, Y. Mikami, S. Fujii, K. Togashi, Diffusion-weighted MR imaging of uterine endometrial cancer, Journal of magnetic resonance imaging : JMRI 26(3) (2007) 682-7.\u003c/li\u003e\n\u003cli\u003eG. Rechichi, S. Galimberti, M. Signorelli, C.T. Franzesi, P. Perego, M.G. Valsecchi, S. Sironi, Endometrial cancer: correlation of apparent diffusion coefficient with tumor grade, depth of myometrial invasion, and presence of lymph node metastases, AJR. American journal of roentgenology 197(1) (2011) 256-62.\u003c/li\u003e\n\u003cli\u003eR. Yan, X. Zhang, Q. Cao, J. Xu, Y. Chen, S. Qin, S. Zhang, W. Zhao, X. Xing, W. Yang, N. Lang, Multitask Deep Learning for Automated Segmentation and Prognostic Stratification of Endometrial Cancer via Biparametric MRI, Journal of magnetic resonance imaging : JMRI (2025).\u003c/li\u003e\n\u003cli\u003eG.A. Saleh, R. Abdelrazek, A. Hassan, O. Hamdy, M.S.I. Tantawy, Diagnostic utility of apparent diffusion coefficient in preoperative assessment of endometrial cancer: are we ready for the 2023 FIGO staging?, BMC medical imaging 24(1) (2024) 226.\u003c/li\u003e\n\u003cli\u003eA. Ak\u0026ccedil;ay, M.A. G\u0026uuml;ltekin, F. Altıntaş, A.A. Peker, S. Balsak, B. Atasoy, \u0026Ouml;. Toluk, H. Toprak, Updated endometrial cancer FIGO staging: the role of MRI in determining newly included histopathological criteria, Abdominal radiology (New York) 49(10) (2024) 3711-3721.\u003c/li\u003e\n\u003cli\u003eX. Jin, C. Shen, X. Yang, Y. Yu, J. Wang, X. Che, Association of Tumor Size With Myometrial Invasion, Lymphovascular Space Invasion, Lymph Node Metastasis, and Recurrence in Endometrial Cancer: A Meta-Analysis of 40 Studies With 53,276 Patients, Frontiers in oncology 12 (2022) 881850.\u003c/li\u003e\n\u003cli\u003eR. Berretta, T.S. Patrelli, C. Migliavacca, M. Rolla, L. Franchi, M. Monica, A.B. Modena, S. Gizzo, Assessment of tumor size as a useful marker for the surgical staging of endometrial cancer, Oncol Rep 31(5) (2014) 2407-2412.\u003c/li\u003e\n\u003cli\u003eC. Sun, J. Mao, Y. Xia, M. Li, Z. Zhao, Analysis of the Correlation between MRI Imaging Signs and Lymphovascular Space Invasion in Endometrial Cancer, Current medical imaging (2025).\u003c/li\u003e\n\u003cli\u003eE. Maheshwari, S. Nougaret, E.B. Stein, G.M. Rauch, K.P. Hwang, R.J. Stafford, A.H. Klopp, P.T. Soliman, K.E. Maturen, A.G. Rockall, S.I. Lee, E.A. Sadowski, A.M. Venkatesan, Update on MRI in Evaluation and Treatment of Endometrial Cancer, Radiographics : a review publication of the Radiological Society of North America, Inc 42(7) (2022) 2112-2130.\u003c/li\u003e\n\u003cli\u003eY. Wang, M. He, P. Cao, P.P.C. Ip, C.Y. Lin, W. Liu, C.W. Lee, E.Y.P. Lee, Tissue Characteristics of Endometrial Carcinoma Analyzed by Quantitative Synthetic MRI and Diffusion-Weighted Imaging, Diagnostics (Basel, Switzerland) 12(12) (2022).\u003c/li\u003e\n\u003cli\u003eK. Zhang, Y. Zhang, X. Fang, M. Fang, B. Shi, J. Dong, L. Qian, Nomograms of Combining Apparent Diffusion Coefficient Value and Radiomics for Preoperative Risk Evaluation in Endometrial Carcinoma, Frontiers in oncology 11 (2021) 705456.\u003c/li\u003e\n\u003cli\u003eS. Satta, M. Dolciami, V. Celli, F. Di Stadio, G. Perniola, I. Palaia, A. Pernazza, C. Della Rocca, S. Rizzo, C. Catalano, S. Capuani, L. Manganaro, Quantitative diffusion and perfusion MRI in the evaluation of endometrial cancer: validation with histopathological parameters, The British journal of radiology 94(1125) (2021) 20210054.\u003c/li\u003e\n\u003cli\u003eX. Ma, M. Shen, Y. He, F. Ma, J. Liu, G. Zhang, J. Qiang, The role of volumetric ADC histogram analysis in preoperatively evaluating the tumour subtype and grade of endometrial cancer, European journal of radiology 140 (2021) 109745.\u003c/li\u003e\n\u003cli\u003eX. Ma, X. Ren, M. Shen, F. Ma, X. Chen, G. Zhang, J. Qiang, Volumetric ADC histogram analysis for preoperative evaluation of LVSI status in stage I endometrioid adenocarcinoma, European radiology 32(1) (2022) 460-469.\u003c/li\u003e\n\u003cli\u003eB. Sun, X. Zhang, Y. Dong, X. Li, X. Yang, L. Zhao, J. Wang, Y. Cheng, Prognostic significance of lymphovascular space invasion in early-stage low-grade endometrioid endometrial cancer: a fifteen-year retrospective Chinese cohort study, World journal of surgical oncology 22(1) (2024) 203.\u003c/li\u003e\n\u003cli\u003eC. Ma, Y. Zhao, Q. Song, X. Meng, Q. Xu, S. Tian, L. Chen, N. Wang, Q. Song, L. Lin, J. Wang, A. Liu, Multi-parametric MRI-based radiomics for preoperative prediction of multiple biological characteristics in endometrial cancer, Frontiers in oncology 13 (2023) 1280022.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u0026nbsp; \u0026nbsp;Table 1: patient characteristics\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"526\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCategory/Value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eMean\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e55.85\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eSD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e10.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 243px;\"\u003e\n \u003cp\u003eRadiological Tumor Size (cm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003cp\u003eSD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e5.42\u003c/p\u003e\n \u003cp\u003e3.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 42px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003eADC Value (mm\u003csup\u003e2\u003c/sup\u003e/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eMean\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e719.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eSD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e158.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003ePathological Subtype\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eEndometroid\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e86.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eNon-Endometroid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e13.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003ePathological Grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eGrade 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e30.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eGrade 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e38.8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eGrade 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e26.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eMissing Data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e4.2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003eFIGO Stage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eStage 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e51.6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eStage 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e29.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eStage 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e14.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eMissing Data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e4.2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003eLVSI (visual assessment)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003ePositive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e82.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eNegative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e17.9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003eLVSI (Pathological assessment)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003ePositive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 243px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003eNegative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 42px;\"\u003e\n \u003cp\u003e49\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eSD = Standard Deviation; ADC = Apparent diffusion coefficient; LVSI: Lymphovascular Space Invasion\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 2: Diagnostic performance of visual and quantitative assessment for LVSI\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"746\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMethod\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCutoff value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003eVisual\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e85.4 (41/48; 72.2-93.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e21.3 (10/47; 10.7-35.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e53.68 (53/95; 43.1-63.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.690 (Optimal cutoff)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e75 (36/48; 60.4-86.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e76.6 (36/47; 62-87.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e75.78 (72/95; 65.9-84)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.767\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e85.4 (41/48; 72.2-93.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e55.3 (26/47; 40.1-69.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e70.5 (67/95; 60.2-79.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.820\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e89.6 (43/48; 77.3-96.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e34 (16/47; 20.9-49.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e62.1 (59/95; 51.5-71.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 204px;\"\u003e\n \u003cp\u003e81.2 (39/48; 67.4-91.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 168px;\"\u003e\n \u003cp\u003e61.7 (29/47; 46.4-75.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e71.6 (68/95; 61.4-80.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eCutoff 690 was determined as the optimal threshold by ROC analysis. CI = Confidence interval; ADC = Apparent diffusion coefficient. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 3: Diagnostic performance by tumor size, FIGO stage, and pathological grade\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"768\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCategory\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGroup\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMethod\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy, % (n/N; 95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eTumor Size\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026lt;2 cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e100 (2/2; 15.8-100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e25 (2/8; 3.2-65.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e40 (4/10; 12.1-73.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e50 (1/2; 1.3-98.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e100 (8/8; 63.1-100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e90 (9/10; 55.5-99.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e2-4 cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e80 (8/10; 44.4-97.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e31.6 (6/19; 12.6-56.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e48.2 (14/29; 29.4-67.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e70 (7/10; 34.8-93.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e84.2 (16/19; 60.4-96.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e79.3 (23/29; 60.3-92.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026gt;4 cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e86.1 (31/36; 70.5-95.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e10 (2/20; 1.2-31.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e58.92 (33/56; 44.97-71.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e77.8 (28/36; 60.8-89.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e85.0 (17/20; 62.1-96.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e80.3 (45/56; 67.5-89.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eFIGO Stage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eStage I\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e100 (14/14; 76.8-100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e26.5 (9/34; 12.9-44.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e47.9 (23/48; 33.3-62.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e71.4 (10/14; 41.9-91.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e71.4 (24/34; 53.7-85.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e71.4 (34/48; 56.7-83.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eStage II\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e80 (16/20; 56.3-94.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e0.0 (0/8; 0-36.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e57.1 (16/28; 37.2-75.53)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e80 (16/20; 56.3-94.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e87.5 (7/8; 47.3-99.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e82.1 (23/28; 63.1-93.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eStage III\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e72.7 (8/11; 39-94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e50 (1/2; 1.3-98.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e69.2 (9/13; 38.5-90.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e100 (11/11; 73.5-100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e50 (1/2; 1.3-98.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e92.8 (12/13; 66.1-99.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003ePathological Grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eGrade 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e100 (8/8; 63.1-100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e23.8 (5/21; 8.2-47.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e44.8 (13/29; 26.4-64.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e50 (4/8; 15.7-84.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e80.9 (17/21; 58.1-94.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e72.4 (21/29; 52.7-87.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eGrade 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e76.5 (13/17; 50.1-93.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e20 (4/20; 5.7-43.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e45.9 (17/37; 29.4-63.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e88.2 (15/17; 63.6-98.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e70 (14/20; 45.7-88.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e78.4 (30/37; 61.7-90.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eGrade 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eVisual\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e85 (17/20; 62.1-96.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e20 (1/5; 0.5-71.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e72.0 (18/25; 50.6-87.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003eQuantitative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e90 (18/20; 68.3-98.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003e60 (3/5; 14.7-94.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 150px;\"\u003e\n \u003cp\u003e84.0 (21/25; 63.9-95.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eValues are presented as percentages with numerator/denominator and 95% confidence intervals. Accuracy is defined as (TP + TN)/Total. CI = Confidence interval; LVSI = Lymphovascular space invasion\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Tehran University of Medical Sciences","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":"Endometrial cancer, Lymphovascular space invasion (LVSI), Diffusion-weighted imaging (DWI), Apparent diffusion coefficient (ADC), Preoperative MRI","lastPublishedDoi":"10.21203/rs.3.rs-9151794/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9151794/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003ePreoperative identification of lymphovascular space invasion (LVSI) in endometrial cancer remains challenging yet critical for surgical planning and adjuvant therapy. This study compared qualitative diffusion-weighted imaging (DWI) and quantitative apparent diffusion coefficient (ADC) analysis for preoperative LVSI detection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e A total of 95 patients with histologically confirmed endometrial cancer underwent preoperative DWI. Two blinded radiologists performed a visual assessment of high b-value images, and manual regions of interest were placed on the ADC maps to derive the mean ADC values. Optimal ADC thresholds were determined and compared with published cut-offs. Diagnostic performance metrics were calculated for both methods in the overall cohort and stratified by tumor size, FIGO stage, and grade.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Qualitative DWI yielded 85.4% sensitivity, \u0026nbsp;21.3% specificity, and 53.7% accuracy. ADC at a threshold of ≤0.690×10⁻³ mm²/s \u0026nbsp;achieved 75.0% sensitivity, 76.6% specificity, and 75.8% accuracy, with an AUC \u0026nbsp;of 0.770. Published ADC cut-offs (0.690-0.820×10⁻³ mm²/s) consistently \u0026nbsp;outperformed visual assessment (accuracy 62.1-71.6%). In subgroup analyses, \u0026nbsp;ADC maintained superior accuracy across all tumor sizes (\u0026lt;2 cm: 90.0%; 2-4 \u0026nbsp;cm: 79.3%; \u0026gt;4 cm: 80.3%), FIGO stages (I: 71.4%; II: 82.1%; III: 92.9%), \u0026nbsp;and grades (1: 72.4%; 2: 78.4%; 3: 84.0%). Combining visual and quantitative \u0026nbsp;assessments did not improve the AUC beyond that of ADC alone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eQuantitative ADC analysis significantly outperforms qualitative DWI for preoperative detection of LVSI in endometrial cancer, providing an objective biomarker that enhances risk stratification and informs surgical and adjuvant treatment strategies.\u003c/p\u003e","manuscriptTitle":"ADC-Based MRI Achieves Superior Accuracy in Preoperative LVSI Prediction for Endometrial Cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-22 14:56:03","doi":"10.21203/rs.3.rs-9151794/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":"1be2411e-2347-4cab-acb7-73889ac065c6","owner":[],"postedDate":"March 22nd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64676406,"name":"Nuclear Medicine \u0026 Medical Imaging"}],"tags":[],"updatedAt":"2026-03-22T14:56:03+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-22 14:56:03","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9151794","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9151794","identity":"rs-9151794","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}