Diagnostic Performance of Node-RADS for CT-based Assessment of Regional Lymph Nodes in Pancreatic Ductal Adenocarcinoma

Diagnostic Performance of Node-RADS for CT-based Assessment of Regional Lymph Nodes in Pancreatic Ductal Adenocarcinoma | 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 Diagnostic Performance of Node-RADS for CT-based Assessment of Regional Lymph Nodes in Pancreatic Ductal Adenocarcinoma Rolf Reiter, Sophie Roigas, Artemis Knittel, Marcus Bahra, Christian Schineis, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9115727/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 Objectives To investigate the diagnostic performance of Node Reporting and Data System (Node-RADS) for regional lymph node staging in pancreatic ductal adenocarcinoma (PDAC), using histopathology as reference standard, and to compare Node-RADS with individual size- and morphology-based criteria. Methods This retrospective multicenter study included 87 consecutive patients (median age, 72 years, range 43–91 years; 45 women) with histologically confirmed PDAC and preoperative CT. Two blinded readers independently assessed regional lymph nodes and assigned Node-RADS scores. Diagnostic performance (sensitivity, specificity, Youden’s index, area under the receiver operating characteristic curve (AUROC) for numeric size and Node-RADS) and interreader agreement (Cohen’s κ) were calculated. Results Eighty-seven patients were included (median age 72 years); 54/87 (62.1%) were node-positive. Overall, 257 regional lymph nodes were assessed on CT (median, 3 per patient), compared with a median of 16 lymph nodes per patient evaluated by histopathology. Node-RADS showed poor discrimination and did not outperform short-axis size (AUROC 0.58, CI 0.45–0.70 vs. 0.56, CI 0.43–0.68; p = 0.70). The PDAC-recommended threshold (Node-RADS ≥ 3) yielded 24.1% sensitivity and 81.8% specificity (Youden’s index 0.06). The best-performing size cutoff was short-axis ≥ 10 mm (sensitivity 25.9%, specificity 84.8%, Youden’s index 0.11). Individual morphologic features were specific but infrequent, resulting in low sensitivity. Interreader agreement was slight to substantial for size thresholds (κ = 0.32–0.65) but slight for Node-RADS (κ = 0.07 for ≥ 3) and most configuration features. Conclusion Node-RADS provides limited accuracy and low reproducibility for regional lymph node staging in PDAC and does not improve upon conventional size- or morphology-based criteria. Nuclear Medicine & Medical Imaging Node-RADS lymph node pancreatic ductal adenocarcinoma computed tomography Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Key Points Question Preoperative detection of lymph node metastases in pancreatic ductal adenocarcinoma remains challenging, as conventional size criteria show limited diagnostic performance. Findings Node Reporting and Data System showed similar diagnostic performance to lymph node short-axis size on computed tomography for detecting lymph node metastases. Clinical Relevance Statement Node Reporting and Data System does not improve detection of lymph node metastases compared with simple lymph node size measurements on CT, highlighting the limited value of current morphological imaging features and the need for improved imaging biomarkers. Introduction Pancreatic ductal adenocarcinoma (PDAC) is still one of the most lethal malignancies worldwide with a five-year survival rate that remains dismal despite tremendous advances in diagnostics and multimodal oncologic therapy [1; 2]. The incidence of about 500,000 cases per year is not much higher than its mortality with about 470,000 deaths yearly [ 3 ]. Only 15–20% of patients have resectable, potentially curable PDAC at diagnosis [ 4 ]. Therefore, accurate staging at the time of diagnosis is critical, as it directly affects treatment selection, prognostication, and patient outcome. Among staging parameters such as resectability of the primary tumor and metastasis, regional lymph node involvement is a key prognostic factor of survival, especially in potentially resectable disease [ 5 ]. Lymph node involvement in PDAC patients is an indication for neoadjuvant therapy and thus has important implications for the optimal oncologic treatment strategy [ 6 ]. Yet accurate lymph node staging remains one of the most challenging aspects of preoperative diagnostic workup [7; 8]. However, CT-based assessment of regional lymph node involvement is limited by the reliance on size, typically using a threshold of 10 mm short-axis diameter to identify metastatic lymph nodes. This criterion has suboptimal and variable diagnostic performance [ 9 – 12 ]. As a result, preoperative nodal staging often lacks consistency and accuracy, leading to variability in clinical decision-making. To overcome these shortcomings of oncologic imaging, the Node Reporting and Data System (Node-RADS) version 1.0 has been recently introduced for the standardized assessment of lymph node status in cancer patients [ 13 ]. By incorporating size and morphology into a unified scoring system ranging from very low (score 1) to very high (score 5) probability of nodal metastasis, Node-RADS aims to reduce interobserver variability and improve diagnostic performance across different cancer entities [ 14 – 21 ]. A recent meta-analysis found varying results for interobserver reliability and diagnostic performance [ 22 ]. While Node-RADS has shown promise in several malignancies, its diagnostic performance in patients with PDAC has not been fully established. Given the critical role of nodal status in staging and treatment planning for PDAC, validation of Node-RADS for this entity is warranted. Therefore, the purpose of this study is to investigate the diagnostic performance of Node-RADS compared to the use of various individual criteria alone in identifying regional lymph node involvement on CT scans of PDAC patients using histopathology as reference. Material and Methods Patients Consecutive patients with histologically confirmed PDAC who underwent oncologic surgical resection between 2021 and 2023 were included. A flowchart of patient recruitment with inclusion and exclusion criteria is presented in Fig. 1 . This retrospective multicenter study was approved by the institutional review boards of both recruiting centers, Charité – Universitätsmedizin Berlin, Germany and Waldfriede Hospital, Berlin, Germany, with waiver of informed consent (No. EA4/140/23). CT Imaging and Node-RADS Assessment All CT scans were independently reviewed by two blinded readers (R.R., reader 1: board-certified radiologist with > 10 years of experience; S.R., reader 2: radiology resident with 2 years of experience). Primary diagnostic performance results are reported for reader 1; the results for reader 2 are compiled in Supplementary Table S1; agreement metrics are based on both readers. Minimum image quality was defined as contrast-enhanced CT with slice thickness ≤ 5 mm. Regional lymph nodes were defined according to the American Joint Committee on Cancer staging system for PDAC (8th edition) and included peripancreatic, hepatoduodenal ligament, common hepatic artery, celiac axis, splenic hilum, splenic artery, and superior mesenteric artery lymph nodes [ 23 ]. Lymph node assessment was performed using Node-RADS version 1.0, assigning scores of 1 (very low), 2 (low), 3 (equivocal), 4 (high), and 5 (very high) for the probability of metastatic involvement. Node-RADS evaluation is based on lymph node size and configuration. Nodes were categorized as normal, enlarged (short axis ≥ 10 mm), or bulk (any axis ≥ 30 mm). Configuration was assessed using texture, border, and shape, with a summed configuration score ranging from 0 to 5 points. The final Node-RADS score was derived by weighting the configuration score according to lymph node size. All bulk lymph nodes were directly classified as Node-RADS 5. Detailed information on the scoring criteria is available in [ 13 ]. All Node-RADS criteria were additionally analyzed individually. Moreover, irregular or ill-defined borders were subclassified as lobulated, spiculated, or indistinct. Interreader agreement was calculated for all evaluations performed by the two readers. Surgery and Histopathology Patients underwent curative-intent surgical resection with standard regional lymphadenectomy according to current oncologic guidelines [ 6 ]. Treatment decisions were made by an interdisciplinary tumor board. Surgery was performed either upfront or following neoadjuvant chemotherapy. Patients undergoing preoperative radiotherapy were excluded. Histopathology reports of formalin-fixed surgical specimens were reviewed. Patients were classified as node-positive (pN+) if at least one regional lymph node metastasis was identified and node-negative (pN−) otherwise. A median of 16 lymph nodes per patient were histopathologically evaluated. Tumors were staged according to the 8th edition of the TNM classification [ 24 ]. Statistical Analysis Diagnostic performance of individual CT criteria and Node-RADS was assessed using sensitivity, specificity, and Youden’s index (sensitivity+specificity-1). Additionally, for short-axis size (continuous) and Node‑RADS (ordinal 1–5), area under the receiver operating characteristic curve (AUROC) was calculated using the maximum value per patient. DeLong’s test was used to compare AUROCs [ 25 ]. 95% confidence intervals (CIs) were calculated. CT examinations were considered node-positive (cN+) if at least one lymph node met the respective criterion. Histopathology served as the reference standard. Associations between the number of visible lymph nodes on CT and nodal status were evaluated using the Mann–Whitney U-test. Interreader agreement was calculated using Cohen’s kappa as follows: κ < 0, no agreement; κ = 0.00-0.20, slight agreement; κ = 0.21–0.40, fair agreement; κ = 0.41–0.60, moderate agreement; κ = 0.61–0.80, substantial agreement; κ = 0.81-1.00, near-perfect agreement. A p value ≤ 0.05 was considered statistically significant. Statistical analyses were conducted using SPSS Statistics (Version 25.0). ChatGPT (OpenAI, San Francisco, USA) was used as a non-decision-making AI support tool. Results A total of 87 patients with histologically confirmed PDAC were included in the final analysis. In all CT examinations, regional lymph nodes could be identified (patients: n = 87, 100%; lymph nodes: n = 257, median 3, interquartile range 2). Histopathology identified regional lymph node metastases (pN+) in 54 patients (62.1%), while 33 patients (37.9%) were node-negative (pN−). There was no significant association between the number of identified regional lymph nodes in preoperative CT examinations and histopathologic lymph node involvement (pN+, p = 0.28). Seventy patients (80.5%) underwent upfront surgery, and 17 patients had surgery following neoadjuvant chemotherapy (19.5%). The median interval between CT and surgery was 13 days (interquartile range 19). Demographic, clinical, and oncologic characteristics of the study population are presented in Table 1 . Table 1 Demographic, clinical, and oncologic characteristics of the study population. Patients n = 87 Age Median age (years) 72 Age range (years) 43–91 Sex Female 45 (51.7%) Male 42 (48.3%) Upfront surgery vs. neoadjuvant therapy first Primary resection 70 (80.5%) Preoperative chemotherapy 17 (19.5%) Interval between CT and surgery Median 13 days Interquartile range 19 days Localization within the pancreas Head 64 (73.6%) Body 12 (13.8%) Tail 10 (11.5%) Multifocal 1 (1.1%) Histopathologic staging pN- 33 (37.9%) pN+ 54 (62.1%) (y)pT1 12 (13.8%) (y)pT2 48 (55.2%) (y)pT3 24 (27.6%) (y)pT4 1 (1.1%) ypT0 2 (2.3%) Size Criterion Sensitivities, specificities, and Youden’s indices for each short-axis diameter cutoff value (3–16 mm) are compiled in Table 2 . Figure 2 displays the percentages of patients in whom a lymph node of the respective short-axis diameter was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. Lower size thresholds (≥ 3–7 mm) resulted in high sensitivity (up to 100%) but low specificity (< 40.0%). Increasing the size threshold improved specificity at the expense of sensitivity. A cutoff of ≥ 10 mm showed the highest Youden’s index (0.11) for the size criterion with a high specificity of 84.8% and a limited sensitivity of 25.9%. Thus, size-based criteria alone demonstrated limited diagnostic performance, with Youden’s indices remaining ≤ 0.11 across all thresholds. The overall AUROC value for the size criterion was 0.56 (CI: 0.43–0.68, p = 0.37). Table 2 Size criterion. Sensitivitiy, specificity, and Youden’s index by cutoff value of short-axis diameter (3–16 mm). Cutoff value with the highest Youden’s index is shown in bold. Criterion “Size“ Cutoff value Sensitivity Specificity Youden’s Index 3 mm 100% CI: 93.4-1.00 0% CI: 0.0-10.4 0.00 CI: 0.00–0.00 4 mm 100% CI: 93.4-1.00 0% CI: 0.0-10.4 0.00 CI: 0.00–0.00 5 mm 92.6% CI: 82.4–97.1 2.9% CI: 0.5–15.3 -0.04 CI: -0.14-0.05 6 mm 85.2% CI: 73.4–92.3 21.2% CI: 10.7–37.8 0.06 CI: -0.10-0.24 7 mm 70.4% CI: 57.2–80.9 36.4% CI: 22.2–53.4 0.07 CI: -0.14-0.26 8 mm 51.9% CI: 38.9–64.6 54.5% CI: 38.0-70.2 0.06 CI: -0.15-0.29 9 mm 37.0% CI: 25.4–50.4 72.7% CI: 55.8–84.9 0.10 CI: -0.11-0.29 10 mm 25.9% CI: 16.1–38.9 84.8% CI: 69.1–93.3 0.11 CI: 0.06–0.28 11 mm 13.0% CI: 6.5–24.4 90.9% CI: 76.4–96.9 0.04 CI: -0.10-0.17 12 mm 13.0% CI: 6.4–24.4 90.9% CI: 76.4–96.9 0.04 CI: -0.09-0.16 13 mm 5.6% CI: 1.9–15.1 97.0% CI: 84.7–99.5 0.03 CI: -0.06-0.11 14 mm 3.7% CI: 1.0-12.5 97.0% CI: 84.7–99.5 0.01 CI: -0.08-0.08 15 mm 3.7% CI: 1.0-12.5 97.0% CI: 84.7–99.5 0.01 CI: -0.08-0.08 16 mm 0% CI: 0.0-6.6 97.0% CI: 84.7–99.5 -0.03 CI: -0.10-0.00 Configuration Criterion Sensitivities, specificities, and Youden’s indices for all morphologic criteria are presented in Table 3 . Figure 3 displays the percentages of patients in whom a lymph node with the respective morphologic feature was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. For all individual morphologic features of lymph node texture and contour, specificity was high (> 80.0%). However, lymph nodes with such morphologic changes occur too rarely, resulting in sensitivities below 25.0% for all morphologic features individually. Therefore, the resulting diagnostic performance of these features was comparable to that of the short-axis diameter at higher cutoff values. The standalone overall diagnostic performance of morphologic features remains slightly below that of the size criterion with Youden’s indices remaining ≤ 0.06 across all criteria. No single morphologic feature showed a markedly higher Youden’s index than any of the other morphologic features. Formally, the sum criterion of any texture abnormality demonstrated the highest Youden’s index of 0.06 with a sensitivity of 24.1% and a specificity of 81.8%. No lymph node with focal necrosis was identified in our study population. Table 3 Configuration criterion. Sensitivity, specificity and Youden’s index by morphologic criterion. Morphologic criterion with best diagnostic performance is shown in bold. Criterion “Configuration“ Sensitivity Specificity Youden’s Index Texture – any change 24.1% CI: 14.6–36.9 81.8% CI: 65.6–91.4 0.06 CI: -0.11-0.23 Heterogeneous 20.4% CI: 11.8–32.9 84.9% CI: 69.1–93.3 0.05 CI: -0.12-0.21 Focal necrosis - - - Gross necrosis 3.7% CI: 1.0-12.5 97.0% CI: 84.7–99.5 0.01 CI: -0.07-0.08 Border contour – any change 33.3% CI: 22.2–46.6 69.7% CI: 52.7–82.6 0.03 CI: -0.17-0.23 Lobulated border 20.4% CI: 11.8–32.9 84.8% CI: 69.1–93.3 0.05 -0.11-0.21 Spiculated border 16.7% CI: 9.0-28.7 87.9% CI: 72.7–95.2 0.05 CI: -0.10-0.19 Indistinct border 5.6% CI: 1.9–15.1 90.9% CI: 76.4–96.9 -0.04 CI: -0.16-0.07 Shape Spheric shape 3.7% CI: 1.0-12.5 100% CI: 89.6–100 0.04 CI: 0.00-0.09 Node-RADS According to Elsholtz et al., for PDAC, Node-RADS scores of 3, 4 and 5 should be interpreted as cN + as opposed to scores of 4 and 5 for other cancers [ 13 ]. Sensitivities, specificities, and Youden’s indices for Node-RADS are presented in Table 4 . Figure 4 displays the percentage of patients in whom a lymph node of the respective Node- RADS score was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. While Node-RADS scores ≥ 3 are specific, they are rarely assigned, leading to a sensitivity of only 24.1%. Formally, the recommended threshold of ≥ 3 demonstrates the highest Youden’s index of 0.06. The overall AUROC value for Node-RADS was 0.58 (CI: 0.45–0.70, p = 0.22). In summary, Node-RADS scoring also has limited diagnostic performance due to low sensitivity and does not improve diagnostic performance compared with short-axis diameter or lymph node configuration in PDAC. Figure 5 shows representative examples of Node-RADS scores 1–5. Table 4 Node-RADS. Sensitivity, specificity and Youden’s index for Node-RADS scores ≥ 3 to 5. Score with best diagnostic performance is shown in bold. Node-RADS Sensitivity Specificity Youden’s Index Node-RADS score ≥ 3 24.1% CI: 14.6–37.0 81.8% CI: 65.6–91.4 0.06 CI: -0.12-0.23 Node-RADS score ≥ 4 11.1% CI: 5.2–22.2 87.9% CI: 72.7–95.2 -0.01 CI: -0.15-0.13 Node-RADS score = 5 7.4% CI: 2.9–17.6 93.9% CI: 80.4–98.3 0.01 CI: -0.10-0.12 There was no statistically significant difference between the AUROC values of Node-RADS and short-axis size (0.58, CI 0.45–0.70 vs. 0.56, CI 0.43–0.68), as assessed by the DeLong test (p = 0.70; Fig. 6 ). Interreader Agreement Table 5 summarizes interreader agreement results using Cohen’s kappa (κ) with corresponding p-values and provides the Youden’s indices for all short-axis cutoff values investigated, the morphologic criteria, and Node-RADS scores separately for the two readers. Across all CT scans, agreement for the short-axis diameter ranged from fair to substantial (κ = 0.32–0.65). In contrast, agreement for all other criteria was poor, ranging from no agreement to slight agreement (κ = −0.02–0.17). The lowest concordance was observed for morphologic features, particularly contour irregularity and internal texture characteristics. Supplementary Table S1 provides the complete diagnostic performance metrics (sensitivity, specificity, and Youden’s index) and κ statistics (with p-values) for all criteria for both readers. Table 5 Interreader agreement (Cohen’s kappa statistic, κ) with p-values and Youden’s indices by criterion, separately for each of the two readers. Criterion Cohen’s Kappaκ p-value Youden’s Index Reader 1 Youden’s Index Reader 2 >= 5 mm 0.46 = 6 mm 0.53 = 7 mm 0.57 = 8 mm 0.52 = 9 mm 0.53 = 10 mm 0.54 = 11 mm 0.38 = 12 mm 0.32 0.01 0.04 CI: -0.09-0.16 0,04 CI: -0.10-0.17 >= 13 mm 0.59 < 0.01 0.03 CI: -0.06-0.11 0,07 CI: -0.06-0.18 Texture – any change 0.02 1.00 0.06 CI: -0.11-0.23 0.16 CI: -0.06-36 Heterogeneous 0.00 1.00 0.05 CI: -0.12-0.21 0.12 CI: -0.08-0.34 Focal necrosis - - - 0.02 CI: 0.00-0.06 Gross necrosis -0.02 1.00 0.01 CI: -0.07-0.08 0,02 CI: 0.00-0.06 Border contour – any change 0.12 0.14 0.03 CI: -0.17-0.23 0.07 CI: -0.11-0.27 Lobulated border 0.06 0.38 0.05 -0.12-0.21 0,07 CI: -0.13-0.28 Spiculated border -0.02 1,00 0.05 CI: -0.10-0.19 0,02 CI: 0.00-0.06 Indistinct border -0.02 1,00 -0.04 CI: -0.16-0.07 0,02 CI: 0.00-0.06 Spheric shape 0.17 0.18 0.04 CI: 0.00-0.09 0,10 CI: 0.01–0.21 Node-RADS score ≥ 3 0.07 0.60 0.06 CI: -0.12-0.23 0.26 CI: 0.05–0.45 Node-RADS score ≥ 4 0.15 0.17 -0.01 CI: -0.15-0.13 0.09 CI: -0.06-0.23 Node-RADS score = 5 0.15 0.25 0.01 CI: -0.10-0.12 0.03 CI: -0.07-0.11 Subgroup Analysis - Primary Surgery vs. Neoadjuvant Chemotherapy Subgroup analysis was performed comparing preoperative lymph node status of patients with upfront surgery compared to those undergoing neoadjuvant therapy first. The Youden’s index with the highest value per category of criteria (size, configuration, and Node-RADS score) was consistently higher in the subgroup of patients who underwent neoadjuvant therapy (0.26 vs. 0.07 for size, 0.44 vs. 0.02 for configuration, and 0.33 vs. -0.01 for Node-RADS score). Sensitivities, specificities, and Youden’s indices for all criteria for the subgroups of patients with primary surgery versus neoadjuvant chemotherapy are presented in Supplementary Table S2. The results of this subgroup analysis must be interpreted cautiously as the subgroup of patients who underwent neoadjuvant therapy before surgery was small (n = 17; 19.5%). Discussion In this multicenter population of patients who underwent surgery for PDAC, we investigated whether standardized interpretation using Node-RADS improves CT-based regional lymph node staging compared with conventional size- and morphology-based assessment using histopathology as the reference standard. Both, Node-RADS scores and short‑axis diameters demonstrated poor discriminatory ability, and Node-RADS scoring did not outperform identification of metastatic involvement based on short-axis diameter. At the positivity threshold recommended for PDAC (Node-RADS score ≥ 3), specificity was acceptable (81.8%), but sensitivity remained low (24.1%), resulting in a small overall Youden’s index (0.06). The best-performing short-axis diameter cutoff (≥ 10 mm) similarly yielded high specificity (84.8%) but poor sensitivity (25.9%; Youden’s index 0.11). AUROC point estimates of Node‑RADS scoring and short‑axis diameter revealed no significant difference (AUROC 0.58 vs. 0.56, p = 0.70). Individual configuration features were generally specific (> 80%) but occurred infrequently and therefore provided low sensitivity (≤ 33.3% across criteria), and focal necrosis was not observed in any lymph node. Together, our findings indicate that a structured morphologic lymph node scoring system applied to routine CT scans cannot overcome the intrinsic limitations of morphologic nodal staging in PDAC. Specific disease-related factors likely contribute to this poor performance. Biologically, PDAC lymph node metastases are often microscopic and may occur in normal-sized lymph nodes without conspicuous architectural distortion, whereas benign reactive enlargement may occur in biliary obstruction, pancreatitis, or peritumoral inflammation –diminishing the discriminatory value of nodal size and subtle border or texture changes. When placed in the context of the literature, our results in PDAC patients lie at the lower end of reported diagnostic performance and underscore substantial heterogeneity across tumor entities. In a recent systematic review and meta-analysis covering nine cancers, Node-RADS showed high pooled discriminatory performance, with hierarchical summary AUCs of 0.92 for Node-RADS ≥ 3 and 0.91 for ≥ 4 as positivity thresholds; category-wise malignancy rates increased monotonically from 4% (Node-RADS 1) to 100% (Node-RADS 5), while inter-observer reliability ranged from fair to substantial across the studies analyzed [ 22 ]. In gastric cancer, Loch et al. reported good diagnostic performance for CT-based Node-RADS: Node-RADS ≥ 3 achieved 56.8% sensitivity and 90.7% specificity (Youden’s index 0.48), and Node-RADS ≥ 4 achieved 48.6% sensitivity and 98.1% specificity (Youden’s index 0.47), with an AUROC of 0.78 [ 15 ]. In bladder cancer, Leonardo et al. found accuracy to be moderate to high, with AUC 0.87 at the patient level and 0.91 at the lymph node level, and a cutoff of Node-RADS > 2 provided a balanced sensitivity/specificity of 78.6%/77.1% [ 14 ]. For mediastinal lymph node assessment in lung cancer, Meyer et al. reported an AUC of 0.94 for Node-RADS scoring, with a threshold of 2 yielding 0.74 sensitivity and 0.93 specificity [ 17 ]. Conversely, not all entities show robust performance: in colon cancer, Leonhardi et al. found a high interreader variability and limited diagnostic accuracy for CT-based Node-RADS (AUC 0.68), with a threshold of 2 yielding 0.62 sensitivity and 0.71 specificity [ 21 ]. In an earlier study of pancreatic head PDAC conducted before the introduction of Node-RADS, Loch et al. showed that combining nodal size with morphologic criteria improved staging performance (sensitivity 61%, specificity 82%, PPV 90%, Youden’s index 0.43 [95% CI 0.15–0.60]) compared with size alone (sensitivity 44.2%, specificity 82.4%, Youden’s index 0.27), yet overall accuracy remained only moderate, highlighting the intrinsic limitations of morphology-based lymph node assessment in PDAC. The fact that our PDAC findings are in agreement with these lower-performing settings supports the concept that nodal morphology, node conspicuity, and the incremental benefit of Node-RADS staging might vary with tumor biology and anatomic context. Reproducibility is a second key barrier to clinical implementation in patients with PDAC. In our study population, short-axis diameter thresholds showed fair interreader agreement, whereas agreement for configuration features and for Node-RADS scores was only slight to none (κ approximately − 0.02 to 0.17 for most configuration items; κ = 0.07 for Node-RADS ≥ 3). This is notably lower than the substantial agreement reported in gastric cancer for Node-RADS thresholds (κ 0.73 for ≥ 3 and κ = 0.67 for ≥ 4) and lower than the moderate agreement reported for overall Node-RADS scoring in the lung cancer study (κ = 0.48) [15; 17]. Even where overall performance was modest, interreader agreement for the aggregated configuration score was only slight (κ = 0.18), indicating poor overall reproducibility of morphology-based lymph node assessment [ 21 ]. Clinically, our findings support the assumption that CT-based cN staging in PDAC, whether performed using size thresholds or Node-RADS, should be interpreted cautiously. A “negative” nodal assessment on CT cannot rule out nodal metastasis and should not be used in isolation to downstage risk or to de-escalate therapy. Conversely, clearly suspicious lymph nodes (marked enlargement or overt abnormal configuration translating into higher Node-RADS categories) may still be useful because of relatively high specificity, supporting structured communication in multidisciplinary tumor boards and potentially guiding targeted confirmatory procedures in selected cases. However, given the low sensitivity observed even at the PDAC-specific Node-RADS threshold, meaningful improvement in preoperative nodal assessment will likely require approaches that extend beyond morphologic CT evaluation, such as multimodal imaging (e.g., PET-based or diffusion-based techniques, or MR elastography) [ 26 – 31 ]. Our study has several limitations. First, its retrospective design entails an inherent risk of selection bias. Second, we only included surgically treated patients, which may restrict the generalizability of our findings to individuals with unresectable primary or metastatic disease. Third, histopathology served as a robust reference standard, but correlation was performed on a per-patient basis rather than node-to-node matching; consequently, metastatic deposits within non-visualized or non-selected nodes could not be directly linked to specific imaging findings. Finally, CT acquisition parameters were not fully standardized across centers, and slice thickness of up to 5 mm was allowed, which may have degraded depiction of subtle morphologic signs. Conversely, key strengths include the multicenter design, the use of histopathology after standardized regional lymphadenectomy, and the blinded assessment by readers with different experience levels, providing a pragmatic estimate of real-world performance and reproducibility. In conclusion, Node-RADS demonstrates limited accuracy and interreader agreement for regional lymph node staging in PDAC. These results underscore the persistent need for improved imaging strategies to more reliably assess nodal status in this highly aggressive malignancy. Declarations Acknowledgements Support from the German Research Foundation (DFG; Ingolf Sack and Rolf Reiter: SFB1340 Matrix-in-Vision project number 372486779, FOR5628 project number 513752256, M5 project number 540759292) is gratefully acknowledged. Rolf Reiter is a participant of the BIH-Charité Digital Clinician Scientist Program funded by Charité – Universitätsmedizin Berlin, Berlin Institute of Health, and the DFG. This manuscript contains results of the thesis “Kategorisierung von Lymphknoten mittels moderner Schnittbildgebungs- und Auswertungsverfahren im Rahmen des prätherapeutischen Stagings des Pankreaskarzinoms”, which will be submitted to Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health by A.J.T. Knittel. References Conroy T, Castan F, Lopez A et al (2022) Five-Year Outcomes of FOLFIRINOX vs Gemcitabine as Adjuvant Therapy for Pancreatic Cancer: A Randomized Clinical Trial. JAMA Oncol 8:1571–1578 Siegel RL, Giaquinto AN, Jemal A (2024) Cancer statistics, 2024. Cancer J Clin 74:12–49 Sung H, Ferlay J, Siegel RL et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. Cancer J Clin 71:209–249 Vincent A, Herman J, Schulick R, Hruban RH, Goggins M (2011) Pancreatic cancer. Lancet 378:607–620 Allen PJ, Kuk D, Castillo CF et al (2017) Multi-institutional Validation Study of the American Joint Commission on Cancer (8th Edition) Changes for T and N Staging in Patients With Pancreatic Adenocarcinoma. Ann Surg 265:185–191 (2025) NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®): Pancreatic adenocarcinoma. National Comprehensive Cancer Network Version 2 Morales-Oyarvide V, Rubinson DA, Dunne RF et al (2017) Lymph node metastases in resected pancreatic ductal adenocarcinoma: predictors of disease recurrence and survival. Br J Cancer 117:1874–1882 Loch FN, Asbach P, Haas M et al (2020) Accuracy of various criteria for lymph node staging in ductal adenocarcinoma of the pancreatic head by computed tomography and magnetic resonance imaging. World J Surg Oncol 18:213 Du T, Bill KA, Ford J et al (2018) The diagnosis and staging of pancreatic cancer: A comparison of endoscopic ultrasound and computed tomography with pancreas protocol. Am J Surg 215:472–475 Tran Cao HS, Zhang Q, Sada YH et al (2017) Value of lymph node positivity in treatment planning for early stage pancreatic cancer. Surgery 162:557–567 Prenzel KL, Holscher AH, Vallbohmer D et al (2010) Lymph node size and metastatic infiltration in adenocarcinoma of the pancreatic head. Eur J Surg Oncol 36:993–996 Swords DS, Firpo MA, Johnson KM, Boucher KM, Scaife CL, Mulvihill SJ (2017) Implications of inaccurate clinical nodal staging in pancreatic adenocarcinoma. Surgery 162:104–111 Elsholtz FHJ, Asbach P, Haas M et al (2021) Introducing the Node Reporting and Data System 1.0 (Node-RADS): a concept for standardized assessment of lymph nodes in cancer. Eur Radiol 31:6116–6124 Leonardo C, Flammia RS, Lucciola S et al (2023) Performance of Node-RADS Scoring System for a Standardized Assessment of Regional Lymph Nodes in Bladder Cancer Patients. Cancers (Basel) 15 Loch FN, Beyer K, Kreis ME et al (2024) Diagnostic performance of Node Reporting and Data System (Node-RADS) for regional lymph node staging of gastric cancer by CT. Eur Radiol 34:3183–3193 Lucciola S, Pisciotti ML, Frisenda M et al (2022) Predictive role of node-rads score in patients with prostate cancer candidates for radical prostatectomy with extended lymph node dissection: comparative analysis with validated nomograms. Prostate Cancer Prostatic Dis. 10.1038/s41391-022-00564-z Meyer HJ, Schnarkowski B, Pappisch J et al (2022) CT texture analysis and node-RADS CT score of mediastinal lymph nodes - diagnostic performance in lung cancer patients. Cancer Imaging 22:75 Maggialetti N, Greco CN, Lucarelli NM et al (2024) Correction: Applications of new radiological scores: the Node–rads in colon cancer staging. Radiologia Med 129:524 Kim HJ, Chae EY, Eom HJ et al (2025) Node Reporting and Data System Evaluation of Axillary Nodes in Invasive Ductal and Lobular Carcinoma. Radiology 316:e243823 Pediconi F, Maroncelli R, Pasculli M et al (2024) Performance of MRI for standardized lymph nodes assessment in breast cancer: are we ready for Node-RADS? Eur Radiol 34:7734–7745 Leonhardi J, Mehdorn M, Stelzner S et al (2025) Diagnostic accuracy and reliability of CT-based Node-RADS for colon cancer. Abdom Radiol (NY) 50:1–7 Zhong J, Mao S, Chen H et al (2025) Node-RADS: a systematic review and meta-analysis of diagnostic performance, category-wise malignancy rates, and inter-observer reliability. Eur Radiol 35:2723–2735 Shin DW, Kim J (2020) The American Joint Committee on Cancer 8th edition staging system for the pancreatic ductal adenocarcinoma: is it better than the 7th edition? Hepatobiliary Surg Nutr 9:98–100 Brierley JDGM, Wittekind C (2017) UICC: TNM Classification of Malignant Tumours. Wiley-Blackwell, Oxford Elizabeth R, DeLong DMDaDLC, -P (1988) Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach. Int Biometric Soc 44:837–845 Yoon JK, Park MS, Kim SS et al (2022) Regional lymph node metastasis detected on preoperative CT and/or FDG-PET may predict early recurrence of pancreatic adenocarcinoma after curative resection. Sci Rep 12:17296 Chen K, Mei Y, Xu M et al (2025) Preoperative prediction of lymph node metastasis in pancreatic ductal adenocarcinoma using MRI-derived whole-tumor ADC histogram analysis. Front Med (Lausanne) 12:1736306 Bayerl C, Safraou Y, Reiter R et al (2024) Investigation of hepatic inflammation via viscoelasticity at low and high mechanical frequencies - A magnetic resonance elastography study. J Mech Behav Biomed Mater 160:106711 Bustin H, Meyer T, Reiter R et al (2025) ElastoNet: Neural network-based multicomponent MR elastography wave inversion with uncertainty quantification. Med Image Anal 105:103642 Neelsen C, Elgeti T, Meyer T et al (2024) Multifrequency Magnetic Resonance Elastography Detects Small Abdominal Lymph Node Metastasis by High Stiffness. Invest Radiol 59:787–793 Schattenfroh J, Almutawakel S, Bieling J et al (2025) Technical Recommendation on Multi-Driver Multifrequency MR Elastography for Tomographic Mapping of Abdominal Stiffness With a Focus on the Pancreas and Pancreatic Ductal Adenocarcinoma. Invest Radiol. 10.1097/RLI.0000000000001231 Additional Declarations The authors declare no competing interests. Supplementary Files SupplementaryMaterials.docx 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-9115727","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":605730637,"identity":"af8c32a3-20d8-441e-87ad-accf08c121a2","order_by":0,"name":"Rolf Reiter","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIie3PsWrDMBCA4TsE5+XcrC4E/AqaGgoheRWXgNcOXQItVMYgT94z5EEydJAxJItfoHRxF08ZvGZJKodOBdsdM+gHC2Tu4xCAy3WTCdXak+2HNSoAgt9rf5hsuhk7JOSV0D9JN0RBR2CMzLIkUfgxny69vFmfdovnuzA1cFz3k2lVJAqamJkPD595tXrURBFuq34SBE9ZDaZkDmL6Qi0kEUvh60Fit5gLc9jQC+p3Syat8M+jxNgtRAJ12W0B4asBwvYtkVkxV7G4z/XBklgW2/0A8dJCtWax9LI9tif9KsO0/K6Pb/3kWvT3hxkBLpfL5RrpB3+sR7+x6MibAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-9741-6736","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":true,"prefix":"","firstName":"Rolf","middleName":"","lastName":"Reiter","suffix":""},{"id":605730638,"identity":"3096783e-7378-4c5e-b7c6-8fe82e21eebb","order_by":1,"name":"Sophie Roigas","email":"","orcid":"","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Sophie","middleName":"","lastName":"Roigas","suffix":""},{"id":605730639,"identity":"d640f85c-cead-43d5-bc89-81cee1578118","order_by":2,"name":"Artemis Knittel","email":"","orcid":"","institution":"Waldfriede Hospital","correspondingAuthor":false,"prefix":"","firstName":"Artemis","middleName":"","lastName":"Knittel","suffix":""},{"id":605730640,"identity":"e2c28472-6bc5-4bde-8521-f7ff399ee145","order_by":3,"name":"Marcus Bahra","email":"","orcid":"","institution":"Waldfriede Hospital","correspondingAuthor":false,"prefix":"","firstName":"Marcus","middleName":"","lastName":"Bahra","suffix":""},{"id":605730641,"identity":"27386c35-ac3b-48bd-9904-4f473a6e4468","order_by":4,"name":"Christian Schineis","email":"","orcid":"","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Christian","middleName":"","lastName":"Schineis","suffix":""},{"id":605730642,"identity":"844803e4-d6e1-478b-9ed9-703f0eb68be7","order_by":5,"name":"Wael Rayya","email":"","orcid":"","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Wael","middleName":"","lastName":"Rayya","suffix":""},{"id":605730643,"identity":"601b6c6f-2cca-472f-9b9a-5f0077170671","order_by":6,"name":"Ingolf Sack","email":"","orcid":"https://orcid.org/0000-0003-2460-1444","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Ingolf","middleName":"","lastName":"Sack","suffix":""},{"id":605730644,"identity":"094bae08-bdc5-4cc4-9844-d8e307880472","order_by":7,"name":"Jens Vogel-Claussen","email":"","orcid":"https://orcid.org/0000-0001-5595-6948","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Jens","middleName":"","lastName":"Vogel-Claussen","suffix":""},{"id":605730645,"identity":"ae4e9293-b271-4175-b261-0ec8de7fc17d","order_by":8,"name":"Martin E. Kreis","email":"","orcid":"","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Martin","middleName":"E.","lastName":"Kreis","suffix":""},{"id":605730646,"identity":"7d3df751-41bc-445b-90f3-bc1eef38c4a4","order_by":9,"name":"Tim Vilz","email":"","orcid":"https://orcid.org/0000-0002-3763-7158","institution":"Charité – Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Tim","middleName":"","lastName":"Vilz","suffix":""},{"id":605730647,"identity":"16917d65-fc35-4345-a15b-0fdb3adfee0f","order_by":10,"name":"Carsten Kamphues","email":"","orcid":"https://orcid.org/0000-0002-5406-8540","institution":"Parkklinik Weißensee","correspondingAuthor":false,"prefix":"","firstName":"Carsten","middleName":"","lastName":"Kamphues","suffix":""},{"id":605730648,"identity":"a16c45b9-b792-42fe-8f28-c183f1ab8ae3","order_by":11,"name":"Florian N. Loch","email":"","orcid":"https://orcid.org/0000-0003-1513-8339","institution":"Martin-Luther-Krankenhaus","correspondingAuthor":false,"prefix":"","firstName":"Florian","middleName":"N.","lastName":"Loch","suffix":""}],"badges":[],"createdAt":"2026-03-13 14:24:42","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-9115727/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9115727/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104876552,"identity":"275970fc-b01e-4009-9678-85d1d2a0c513","added_by":"auto","created_at":"2026-03-18 08:42:52","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":251330,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of patient recruitment\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/5fa9f356a22132cff0fac5a4.png"},{"id":104876630,"identity":"d7ecdc7d-7326-4fc5-9b69-6be2af6bc77b","added_by":"auto","created_at":"2026-03-18 08:43:12","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":124744,"visible":true,"origin":"","legend":"\u003cp\u003eSize criterion. Percentage of patients with (pN+, red bars) and without (pN-, blue bars) histopathologic lymph node involvement in whom a lymph node with the respective short-axis diameter cutoff was visible on preoperative CT (3–16 mm). n = 257 lymph nodes in a total of 87 patients.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/f81b92eefb84467c140b87ed.png"},{"id":104876682,"identity":"fa67e022-23fe-4da5-98ca-b14c5cb90d2a","added_by":"auto","created_at":"2026-03-18 08:43:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":142153,"visible":true,"origin":"","legend":"\u003cp\u003eConfiguration criterion. Percentage of patients with (pN+, red bars) and without (pN-, blue bars) histopathologic lymph node involvement in whom a lymph node with the respective morphologic features was visible on preoperative CT. n = 257 lymph nodes in a total of 87 patients.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/6786d4cc93b453728eb99961.png"},{"id":104876550,"identity":"11f054c5-0508-47e4-bcba-dcd64adb11a6","added_by":"auto","created_at":"2026-03-18 08:42:52","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":140691,"visible":true,"origin":"","legend":"\u003cp\u003eNode-RADS. Percentage of patients with (pN+, red bars) and without (pN-, blue bars) histopathologic lymph node involvement in whom a lymph node with the respective Node-RADS score was visible on preoperative CT. n = 257 lymph nodes in a total of 87 patients.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/bd8b6927cd25eb6f31e0a900.png"},{"id":104876676,"identity":"498cb5bc-e8c9-4df5-8446-3e790b69505d","added_by":"auto","created_at":"2026-03-18 08:43:22","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":850564,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative examples of Node-RADS scores 1 - 5. White box indicates the zoomed area and arrow indicates the selected lymph node. (a, b) A 52-year-old woman with node-positive PDAC. The lymph node measures 18 × 5 mm and shows a homogeneous texture, smooth border, and oval shape. (c, d) A 73-year-old woman with node-negative PDAC. The lymph node measures 23 × 10 mm and demonstrates mild enlargement with otherwise homogeneous texture, smooth border, and oval configuration. (e, f) A 64-year-old man with node-positive PDAC. The lymph node measures 10 × 5 mm, showing heterogeneous texture and irregular border. (g, h) A 72-year-old man with node-positive PDAC. The lymph node measures 15 × 10 mm and is enlarged, showing heterogeneous texture and irregular border. Asterisk indicates a bowel loop. (i, j) A 69-year-old patient with node-positive PDAC. The lymph node measures 25 × 15 mm, is markedly enlarged, and shows pronounced heterogeneity with gross necrosis and irregular border.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/e4d614249fb491cfa7cd5917.png"},{"id":104876706,"identity":"ae3b8b65-7dfe-43f4-a49d-ca7032b8a7d6","added_by":"auto","created_at":"2026-03-18 08:43:28","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":483613,"visible":true,"origin":"","legend":"\u003cp\u003eArea under the receiver operating characteristic curve analysis of size (blue line) and Node-RADS (orange line).\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/d047f850ba2232a2e21eb17e.png"},{"id":104876835,"identity":"d2b93f34-86ae-4465-9e57-f70b5ab2171c","added_by":"auto","created_at":"2026-03-18 08:43:50","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3540910,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/7528d61b-365e-4d9d-a26a-feebec44c700.pdf"},{"id":104876775,"identity":"2f28d38e-31dd-4ffe-938b-31eeddcd8a6d","added_by":"auto","created_at":"2026-03-18 08:43:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":29902,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-9115727/v1/73cac0381f57ade114119e1b.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eDiagnostic Performance of Node-RADS for CT-based Assessment of Regional Lymph Nodes in Pancreatic Ductal Adenocarcinoma\u003c/p\u003e","fulltext":[{"header":"Key Points","content":"\u003cp\u003e\u003cstrong\u003eQuestion\u003c/strong\u003e Preoperative detection of lymph node metastases in pancreatic ductal adenocarcinoma remains challenging, as conventional size criteria show limited diagnostic performance.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFindings\u003c/strong\u003e Node Reporting and Data System showed similar diagnostic performance to lymph node short-axis size on computed tomography for detecting lymph node metastases.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Relevance Statement\u0026nbsp;\u003c/strong\u003eNode Reporting and Data System does not improve detection of lymph node metastases compared with simple lymph node size measurements on CT, highlighting the limited value of current morphological imaging features and the need for improved imaging biomarkers.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003ePancreatic ductal adenocarcinoma (PDAC) is still one of the most lethal malignancies worldwide with a five-year survival rate that remains dismal despite tremendous advances in diagnostics and multimodal oncologic therapy [1; 2]. The incidence of about 500,000 cases per year is not much higher than its mortality with about 470,000 deaths yearly [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Only 15\u0026ndash;20% of patients have resectable, potentially curable PDAC at diagnosis [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Therefore, accurate staging at the time of diagnosis is critical, as it directly affects treatment selection, prognostication, and patient outcome. Among staging parameters such as resectability of the primary tumor and metastasis, regional lymph node involvement is a key prognostic factor of survival, especially in potentially resectable disease [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Lymph node involvement in PDAC patients is an indication for neoadjuvant therapy and thus has important implications for the optimal oncologic treatment strategy [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Yet accurate lymph node staging remains one of the most challenging aspects of preoperative diagnostic workup [7; 8]. However, CT-based assessment of regional lymph node involvement is limited by the reliance on size, typically using a threshold of 10 mm short-axis diameter to identify metastatic lymph nodes. This criterion has suboptimal and variable diagnostic performance [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. As a result, preoperative nodal staging often lacks consistency and accuracy, leading to variability in clinical decision-making.\u003c/p\u003e \u003cp\u003eTo overcome these shortcomings of oncologic imaging, the Node Reporting and Data System (Node-RADS) version 1.0 has been recently introduced for the standardized assessment of lymph node status in cancer patients [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. By incorporating size and morphology into a unified scoring system ranging from very low (score 1) to very high (score 5) probability of nodal metastasis, Node-RADS aims to reduce interobserver variability and improve diagnostic performance across different cancer entities [\u003cspan additionalcitationids=\"CR15 CR16 CR17 CR18 CR19 CR20\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. A recent meta-analysis found varying results for interobserver reliability and diagnostic performance [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eWhile Node-RADS has shown promise in several malignancies, its diagnostic performance in patients with PDAC has not been fully established. Given the critical role of nodal status in staging and treatment planning for PDAC, validation of Node-RADS for this entity is warranted. Therefore, the purpose of this study is to investigate the diagnostic performance of Node-RADS compared to the use of various individual criteria alone in identifying regional lymph node involvement on CT scans of PDAC patients using histopathology as reference.\u003c/p\u003e"},{"header":"Material and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eConsecutive patients with histologically confirmed PDAC who underwent oncologic surgical resection between 2021 and 2023 were included. A flowchart of patient recruitment with inclusion and exclusion criteria is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. This retrospective multicenter study was approved by the institutional review boards of both recruiting centers, Charit\u0026eacute; \u0026ndash; Universit\u0026auml;tsmedizin Berlin, Germany and Waldfriede Hospital, Berlin, Germany, with waiver of informed consent (No. EA4/140/23).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCT Imaging and Node-RADS Assessment\u003c/h3\u003e\n\u003cp\u003eAll CT scans were independently reviewed by two blinded readers (R.R., reader 1: board-certified radiologist with \u0026gt;\u0026thinsp;10 years of experience; S.R., reader 2: radiology resident with 2 years of experience). Primary diagnostic performance results are reported for reader 1; the results for reader 2 are compiled in Supplementary Table S1; agreement metrics are based on both readers. Minimum image quality was defined as contrast-enhanced CT with slice thickness\u0026thinsp;\u0026le;\u0026thinsp;5 mm. Regional lymph nodes were defined according to the American Joint Committee on Cancer staging system for PDAC (8th edition) and included peripancreatic, hepatoduodenal ligament, common hepatic artery, celiac axis, splenic hilum, splenic artery, and superior mesenteric artery lymph nodes [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Lymph node assessment was performed using Node-RADS version 1.0, assigning scores of 1 (very low), 2 (low), 3 (equivocal), 4 (high), and 5 (very high) for the probability of metastatic involvement. Node-RADS evaluation is based on lymph node size and configuration. Nodes were categorized as normal, enlarged (short axis\u0026thinsp;\u0026ge;\u0026thinsp;10 mm), or bulk (any axis\u0026thinsp;\u0026ge;\u0026thinsp;30 mm). Configuration was assessed using texture, border, and shape, with a summed configuration score ranging from 0 to 5 points. The final Node-RADS score was derived by weighting the configuration score according to lymph node size. All bulk lymph nodes were directly classified as Node-RADS 5. Detailed information on the scoring criteria is available in [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. All Node-RADS criteria were additionally analyzed individually. Moreover, irregular or ill-defined borders were subclassified as lobulated, spiculated, or indistinct. Interreader agreement was calculated for all evaluations performed by the two readers.\u003c/p\u003e\n\u003ch3\u003eSurgery and Histopathology\u003c/h3\u003e\n\u003cp\u003ePatients underwent curative-intent surgical resection with standard regional lymphadenectomy according to current oncologic guidelines [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Treatment decisions were made by an interdisciplinary tumor board. Surgery was performed either upfront or following neoadjuvant chemotherapy. Patients undergoing preoperative radiotherapy were excluded. Histopathology reports of formalin-fixed surgical specimens were reviewed. Patients were classified as node-positive (pN+) if at least one regional lymph node metastasis was identified and node-negative (pN\u0026minus;) otherwise. A median of 16 lymph nodes per patient were histopathologically evaluated. Tumors were staged according to the 8th edition of the TNM classification [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eDiagnostic performance of individual CT criteria and Node-RADS was assessed using sensitivity, specificity, and Youden\u0026rsquo;s index (sensitivity+specificity-1). Additionally, for short-axis size (continuous) and Node‑RADS (ordinal 1\u0026ndash;5), area under the receiver operating characteristic curve (AUROC) was calculated using the maximum value per patient. DeLong\u0026rsquo;s test was used to compare AUROCs [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. 95% confidence intervals (CIs) were calculated. CT examinations were considered node-positive (cN+) if at least one lymph node met the respective criterion. Histopathology served as the reference standard. Associations between the number of visible lymph nodes on CT and nodal status were evaluated using the Mann\u0026ndash;Whitney U-test. Interreader agreement was calculated using Cohen\u0026rsquo;s kappa as follows: \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0, no agreement; \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.00-0.20, slight agreement; \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.21\u0026ndash;0.40, fair agreement; \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.41\u0026ndash;0.60, moderate agreement; \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.61\u0026ndash;0.80, substantial agreement; \u003cem\u003eκ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.81-1.00, near-perfect agreement. A \u003cem\u003ep\u003c/em\u003e value\u0026thinsp;\u0026le;\u0026thinsp;0.05 was considered statistically significant. Statistical analyses were conducted using SPSS Statistics (Version 25.0). ChatGPT (OpenAI, San Francisco, USA) was used as a non-decision-making AI support tool.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 87 patients with histologically confirmed PDAC were included in the final analysis. In all CT examinations, regional lymph nodes could be identified (patients: n\u0026thinsp;=\u0026thinsp;87, 100%; lymph nodes: n\u0026thinsp;=\u0026thinsp;257, median 3, interquartile range 2). Histopathology identified regional lymph node metastases (pN+) in 54 patients (62.1%), while 33 patients (37.9%) were node-negative (pN\u0026minus;). There was no significant association between the number of identified regional lymph nodes in preoperative CT examinations and histopathologic lymph node involvement (pN+, p\u0026thinsp;=\u0026thinsp;0.28). Seventy patients (80.5%) underwent upfront surgery, and 17 patients had surgery following neoadjuvant chemotherapy (19.5%). The median interval between CT and surgery was 13 days (interquartile range 19). Demographic, clinical, and oncologic characteristics of the study population are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic, clinical, and oncologic characteristics of the study population.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePatients\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en\u0026thinsp;=\u0026thinsp;87\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian age (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e72\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge range (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e43\u0026ndash;91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSex\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45 (51.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e42 (48.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eUpfront surgery vs. neoadjuvant therapy first\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrimary resection\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e70 (80.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePreoperative chemotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17 (19.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInterval between CT and surgery\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13 days\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInterquartile range\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19 days\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLocalization within the pancreas\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHead\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64 (73.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTail\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10 (11.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMultifocal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (1.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHistopathologic staging\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epN-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e33 (37.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epN+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e54 (62.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(y)pT1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(y)pT2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e48 (55.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(y)pT3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24 (27.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(y)pT4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (1.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eypT0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (2.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSize Criterion\u003c/h2\u003e \u003cp\u003eSensitivities, specificities, and Youden\u0026rsquo;s indices for each short-axis diameter cutoff value (3\u0026ndash;16 mm) are compiled in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e displays the percentages of patients in whom a lymph node of the respective short-axis diameter was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. Lower size thresholds (\u0026ge;\u0026thinsp;3\u0026ndash;7 mm) resulted in high sensitivity (up to 100%) but low specificity (\u0026lt;\u0026thinsp;40.0%). Increasing the size threshold improved specificity at the expense of sensitivity. A cutoff of \u0026ge;\u0026thinsp;10 mm showed the highest Youden\u0026rsquo;s index (0.11) for the size criterion with a high specificity of 84.8% and a limited sensitivity of 25.9%. Thus, size-based criteria alone demonstrated limited diagnostic performance, with Youden\u0026rsquo;s indices remaining\u0026thinsp;\u0026le;\u0026thinsp;0.11 across all thresholds. The overall AUROC value for the size criterion was 0.56 (CI: 0.43\u0026ndash;0.68, p\u0026thinsp;=\u0026thinsp;0.37).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSize criterion. Sensitivitiy, specificity, and Youden\u0026rsquo;s index by cutoff value of short-axis diameter (3\u0026ndash;16 mm). Cutoff value with the highest Youden\u0026rsquo;s index is shown in bold.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCriterion \u0026ldquo;Size\u0026ldquo;\u003c/p\u003e \u003cp\u003eCutoff value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYouden\u0026rsquo;s Index\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003cp\u003eCI: 93.4-1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003cp\u003eCI: 0.0-10.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003cp\u003eCI: 0.00\u0026ndash;0.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003cp\u003eCI: 93.4-1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003cp\u003eCI: 0.0-10.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003cp\u003eCI: 0.00\u0026ndash;0.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e92.6%\u003c/p\u003e \u003cp\u003eCI: 82.4\u0026ndash;97.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.9%\u003c/p\u003e \u003cp\u003eCI: 0.5\u0026ndash;15.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.04\u003c/p\u003e \u003cp\u003eCI: -0.14-0.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e85.2%\u003c/p\u003e \u003cp\u003eCI: 73.4\u0026ndash;92.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21.2%\u003c/p\u003e \u003cp\u003eCI: 10.7\u0026ndash;37.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.10-0.24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e70.4%\u003c/p\u003e \u003cp\u003eCI: 57.2\u0026ndash;80.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36.4%\u003c/p\u003e \u003cp\u003eCI: 22.2\u0026ndash;53.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003cp\u003eCI: -0.14-0.26\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51.9%\u003c/p\u003e \u003cp\u003eCI: 38.9\u0026ndash;64.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.5%\u003c/p\u003e \u003cp\u003eCI: 38.0-70.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.15-0.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37.0%\u003c/p\u003e \u003cp\u003eCI: 25.4\u0026ndash;50.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e72.7%\u003c/p\u003e \u003cp\u003eCI: 55.8\u0026ndash;84.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.10\u003c/p\u003e \u003cp\u003eCI: -0.11-0.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e10 mm\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e25.9%\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003eCI: 16.1\u0026ndash;38.9\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e84.8%\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003eCI: 69.1\u0026ndash;93.3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.11\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003eCI: 0.06\u0026ndash;0.28\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.0%\u003c/p\u003e \u003cp\u003eCI: 6.5\u0026ndash;24.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e90.9%\u003c/p\u003e \u003cp\u003eCI: 76.4\u0026ndash;96.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: -0.10-0.17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.0%\u003c/p\u003e \u003cp\u003eCI: 6.4\u0026ndash;24.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e90.9%\u003c/p\u003e \u003cp\u003eCI: 76.4\u0026ndash;96.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: -0.09-0.16\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.6%\u003c/p\u003e \u003cp\u003eCI: 1.9\u0026ndash;15.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.0%\u003c/p\u003e \u003cp\u003eCI: 84.7\u0026ndash;99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003cp\u003eCI: -0.06-0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e14 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.7%\u003c/p\u003e \u003cp\u003eCI: 1.0-12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.0%\u003c/p\u003e \u003cp\u003eCI: 84.7\u0026ndash;99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.08-0.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e15 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.7%\u003c/p\u003e \u003cp\u003eCI: 1.0-12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.0%\u003c/p\u003e \u003cp\u003eCI: 84.7\u0026ndash;99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.08-0.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e16 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003cp\u003eCI: 0.0-6.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.0%\u003c/p\u003e \u003cp\u003eCI: 84.7\u0026ndash;99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.03\u003c/p\u003e \u003cp\u003eCI: -0.10-0.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eConfiguration Criterion\u003c/h3\u003e\n\u003cp\u003eSensitivities, specificities, and Youden\u0026rsquo;s indices for all morphologic criteria are presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e displays the percentages of patients in whom a lymph node with the respective morphologic feature was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. For all individual morphologic features of lymph node texture and contour, specificity was high (\u0026gt;\u0026thinsp;80.0%). However, lymph nodes with such morphologic changes occur too rarely, resulting in sensitivities below 25.0% for all morphologic features individually. Therefore, the resulting diagnostic performance of these features was comparable to that of the short-axis diameter at higher cutoff values. The standalone overall diagnostic performance of morphologic features remains slightly below that of the size criterion with Youden\u0026rsquo;s indices remaining\u0026thinsp;\u0026le;\u0026thinsp;0.06 across all criteria. No single morphologic feature showed a markedly higher Youden\u0026rsquo;s index than any of the other morphologic features. Formally, the sum criterion of any texture abnormality demonstrated the highest Youden\u0026rsquo;s index of 0.06 with a sensitivity of 24.1% and a specificity of 81.8%. No lymph node with focal necrosis was identified in our study population.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eConfiguration criterion. Sensitivity, specificity and Youden\u0026rsquo;s index by morphologic criterion. Morphologic criterion with best diagnostic performance is shown in bold.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCriterion \u0026ldquo;Configuration\u0026ldquo;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYouden\u0026rsquo;s Index\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTexture \u0026ndash; any change\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.1%\u003c/p\u003e \u003cp\u003eCI: 14.6\u0026ndash;36.9\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81.8%\u003c/p\u003e \u003cp\u003eCI: 65.6\u0026ndash;91.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.11-0.23\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeterogeneous\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.4%\u003c/p\u003e \u003cp\u003eCI: 11.8\u0026ndash;32.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e84.9%\u003c/p\u003e \u003cp\u003eCI: 69.1\u0026ndash;93.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003eCI: -0.12-0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFocal necrosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGross necrosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.7%\u003c/p\u003e \u003cp\u003eCI: 1.0-12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e97.0%\u003c/p\u003e \u003cp\u003eCI: 84.7\u0026ndash;99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.07-0.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBorder contour \u0026ndash; any change\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e33.3%\u003c/p\u003e \u003cp\u003eCI: 22.2\u0026ndash;46.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e69.7%\u003c/p\u003e \u003cp\u003eCI: 52.7\u0026ndash;82.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003cp\u003eCI: -0.17-0.23\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLobulated border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.4%\u003c/p\u003e \u003cp\u003eCI: 11.8\u0026ndash;32.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e84.8%\u003c/p\u003e \u003cp\u003eCI: 69.1\u0026ndash;93.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003e-0.11-0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpiculated border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.7%\u003c/p\u003e \u003cp\u003eCI: 9.0-28.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e87.9%\u003c/p\u003e \u003cp\u003eCI: 72.7\u0026ndash;95.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003eCI: -0.10-0.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIndistinct border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.6%\u003c/p\u003e \u003cp\u003eCI: 1.9\u0026ndash;15.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90.9%\u003c/p\u003e \u003cp\u003eCI: 76.4\u0026ndash;96.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.04\u003c/p\u003e \u003cp\u003eCI: -0.16-0.07\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eShape\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpheric shape\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.7%\u003c/p\u003e \u003cp\u003eCI: 1.0-12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003cp\u003eCI: 89.6\u0026ndash;100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: 0.00-0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eNode-RADS\u003c/h3\u003e\n\u003cp\u003eAccording to Elsholtz et al., for PDAC, Node-RADS scores of 3, 4 and 5 should be interpreted as cN\u0026thinsp;+\u0026thinsp;as opposed to scores of 4 and 5 for other cancers [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Sensitivities, specificities, and Youden\u0026rsquo;s indices for Node-RADS are presented in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e displays the percentage of patients in whom a lymph node of the respective Node- RADS score was visible on preoperative CT without (pN-, blue bars) and with (pN+, red bars) histopathologic lymph node involvement. While Node-RADS scores\u0026thinsp;\u0026ge;\u0026thinsp;3 are specific, they are rarely assigned, leading to a sensitivity of only 24.1%. Formally, the recommended threshold of \u0026ge;\u0026thinsp;3 demonstrates the highest Youden\u0026rsquo;s index of 0.06. The overall AUROC value for Node-RADS was 0.58 (CI: 0.45\u0026ndash;0.70, p\u0026thinsp;=\u0026thinsp;0.22). In summary, Node-RADS scoring also has limited diagnostic performance due to low sensitivity and does not improve diagnostic performance compared with short-axis diameter or lymph node configuration in PDAC. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e shows representative examples of Node-RADS scores 1\u0026ndash;5.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eNode-RADS. Sensitivity, specificity and Youden\u0026rsquo;s index for Node-RADS scores\u0026thinsp;\u0026ge;\u0026thinsp;3 to 5. Score with best diagnostic performance is shown in bold.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eYouden\u0026rsquo;s Index\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;\u0026ge;\u0026thinsp;3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.1%\u003c/p\u003e \u003cp\u003eCI: 14.6\u0026ndash;37.0\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e81.8%\u003c/p\u003e \u003cp\u003eCI: 65.6\u0026ndash;91.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.12-0.23\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;\u0026ge;\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11.1%\u003c/p\u003e \u003cp\u003eCI: 5.2\u0026ndash;22.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e87.9%\u003c/p\u003e \u003cp\u003eCI: 72.7\u0026ndash;95.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.01\u003c/p\u003e \u003cp\u003eCI: -0.15-0.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;=\u0026thinsp;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.4%\u003c/p\u003e \u003cp\u003eCI: 2.9\u0026ndash;17.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e93.9%\u003c/p\u003e \u003cp\u003eCI: 80.4\u0026ndash;98.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.10-0.12\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThere was no statistically significant difference between the AUROC values of Node-RADS and short-axis size (0.58, CI 0.45\u0026ndash;0.70 vs. 0.56, CI 0.43\u0026ndash;0.68), as assessed by the DeLong test (p\u0026thinsp;=\u0026thinsp;0.70; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eInterreader Agreement\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e summarizes interreader agreement results using Cohen\u0026rsquo;s kappa (κ) with corresponding p-values and provides the Youden\u0026rsquo;s indices for all short-axis cutoff values investigated, the morphologic criteria, and Node-RADS scores separately for the two readers. Across all CT scans, agreement for the short-axis diameter ranged from fair to substantial (κ\u0026thinsp;=\u0026thinsp;0.32\u0026ndash;0.65). In contrast, agreement for all other criteria was poor, ranging from no agreement to slight agreement (κ = \u0026minus;0.02\u0026ndash;0.17). The lowest concordance was observed for morphologic features, particularly contour irregularity and internal texture characteristics. Supplementary Table S1 provides the complete diagnostic performance metrics (sensitivity, specificity, and Youden\u0026rsquo;s index) and κ statistics (with p-values) for all criteria for both readers.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eInterreader agreement (Cohen\u0026rsquo;s kappa statistic, κ) with p-values and Youden\u0026rsquo;s indices by criterion, separately for each of the two readers.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCriterion\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCohen\u0026rsquo;s Kappaκ\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYouden\u0026rsquo;s Index\u003c/p\u003e \u003cp\u003eReader 1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYouden\u0026rsquo;s Index\u003c/p\u003e \u003cp\u003eReader 2\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 5 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.04\u003c/p\u003e \u003cp\u003eCI: -0.14-0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: -0.10-0.20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 6 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.10-0.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,01\u003c/p\u003e \u003cp\u003eCI:-0.18-0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 7 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003cp\u003eCI: -0.14-0.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,04\u003c/p\u003e \u003cp\u003eCI: -0.18-0.25\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 8 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.15-0.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-0,01\u003c/p\u003e \u003cp\u003eCI: -0.22-0.23\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 9 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.10\u003c/p\u003e \u003cp\u003eCI: -0.11-0.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,02\u003c/p\u003e \u003cp\u003eCI: -0.17-0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 10 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003cp\u003eCI: 0.06\u0026ndash;0.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003eCI: -0.12-0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 11 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: -0.10-0.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,03\u003c/p\u003e \u003cp\u003eCI: -0.12-0.17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 12 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: -0.09-0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,04\u003c/p\u003e \u003cp\u003eCI: -0.10-0.17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;= 13 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003cp\u003eCI: -0.06-0.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,07\u003c/p\u003e \u003cp\u003eCI: -0.06-0.18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTexture \u0026ndash;\u003c/p\u003e \u003cp\u003eany change\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.11-0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.16\u003c/p\u003e \u003cp\u003eCI: -0.06-36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeterogeneous\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003eCI: -0.12-0.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.12\u003c/p\u003e \u003cp\u003eCI: -0.08-0.34\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFocal necrosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003cp\u003eCI: 0.00-0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGross necrosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.07-0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,02\u003c/p\u003e \u003cp\u003eCI: 0.00-0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBorder contour \u0026ndash; any change\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003cp\u003eCI: -0.17-0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003cp\u003eCI: -0.11-0.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLobulated border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003e-0.12-0.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,07\u003c/p\u003e \u003cp\u003eCI: -0.13-0.28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpiculated border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003cp\u003eCI: -0.10-0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,02\u003c/p\u003e \u003cp\u003eCI: 0.00-0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIndistinct border\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.04\u003c/p\u003e \u003cp\u003eCI: -0.16-0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,02\u003c/p\u003e \u003cp\u003eCI: 0.00-0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpheric shape\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003cp\u003eCI: 0.00-0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0,10\u003c/p\u003e \u003cp\u003eCI: 0.01\u0026ndash;0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;\u0026ge;\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003cp\u003eCI: -0.12-0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.26\u003c/p\u003e \u003cp\u003eCI: 0.05\u0026ndash;0.45\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;\u0026ge;\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.01\u003c/p\u003e \u003cp\u003eCI: -0.15-0.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003cp\u003eCI: -0.06-0.23\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNode-RADS score\u0026thinsp;=\u0026thinsp;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003cp\u003eCI: -0.10-0.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003cp\u003eCI: -0.07-0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSubgroup Analysis - Primary Surgery vs. Neoadjuvant Chemotherapy\u003c/h2\u003e \u003cp\u003eSubgroup analysis was performed comparing preoperative lymph node status of patients with upfront surgery compared to those undergoing neoadjuvant therapy first. The Youden\u0026rsquo;s index with the highest value per category of criteria (size, configuration, and Node-RADS score) was consistently higher in the subgroup of patients who underwent neoadjuvant therapy (0.26 vs. 0.07 for size, 0.44 vs. 0.02 for configuration, and 0.33 vs. -0.01 for Node-RADS score). Sensitivities, specificities, and Youden\u0026rsquo;s indices for all criteria for the subgroups of patients with primary surgery versus neoadjuvant chemotherapy are presented in Supplementary Table S2. The results of this subgroup analysis must be interpreted cautiously as the subgroup of patients who underwent neoadjuvant therapy before surgery was small (n\u0026thinsp;=\u0026thinsp;17; 19.5%).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this multicenter population of patients who underwent surgery for PDAC, we investigated whether standardized interpretation using Node-RADS improves CT-based regional lymph node staging compared with conventional size- and morphology-based assessment using histopathology as the reference standard. Both, Node-RADS scores and short‑axis diameters demonstrated poor discriminatory ability, and Node-RADS scoring did not outperform identification of metastatic involvement based on short-axis diameter. At the positivity threshold recommended for PDAC (Node-RADS score\u0026thinsp;\u0026ge;\u0026thinsp;3), specificity was acceptable (81.8%), but sensitivity remained low (24.1%), resulting in a small overall Youden\u0026rsquo;s index (0.06). The best-performing short-axis diameter cutoff (\u0026ge;\u0026thinsp;10 mm) similarly yielded high specificity (84.8%) but poor sensitivity (25.9%; Youden\u0026rsquo;s index 0.11).\u003c/p\u003e \u003cp\u003eAUROC point estimates of Node‑RADS scoring and short‑axis diameter revealed no significant difference (AUROC 0.58 vs. 0.56, p\u0026thinsp;=\u0026thinsp;0.70). Individual configuration features were generally specific (\u0026gt;\u0026thinsp;80%) but occurred infrequently and therefore provided low sensitivity (\u0026le;\u0026thinsp;33.3% across criteria), and focal necrosis was not observed in any lymph node. Together, our findings indicate that a structured morphologic lymph node scoring system applied to routine CT scans cannot overcome the intrinsic limitations of morphologic nodal staging in PDAC. Specific disease-related factors likely contribute to this poor performance. Biologically, PDAC lymph node metastases are often microscopic and may occur in normal-sized lymph nodes without conspicuous architectural distortion, whereas benign reactive enlargement may occur in biliary obstruction, pancreatitis, or peritumoral inflammation \u0026ndash;diminishing the discriminatory value of nodal size and subtle border or texture changes.\u003c/p\u003e \u003cp\u003eWhen placed in the context of the literature, our results in PDAC patients lie at the lower end of reported diagnostic performance and underscore substantial heterogeneity across tumor entities. In a recent systematic review and meta-analysis covering nine cancers, Node-RADS showed high pooled discriminatory performance, with hierarchical summary AUCs of 0.92 for Node-RADS\u0026thinsp;\u0026ge;\u0026thinsp;3 and 0.91 for \u0026ge;\u0026thinsp;4 as positivity thresholds; category-wise malignancy rates increased monotonically from 4% (Node-RADS 1) to 100% (Node-RADS 5), while inter-observer reliability ranged from fair to substantial across the studies analyzed [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. In gastric cancer, Loch et al. reported good diagnostic performance for CT-based Node-RADS: Node-RADS\u0026thinsp;\u0026ge;\u0026thinsp;3 achieved 56.8% sensitivity and 90.7% specificity (Youden\u0026rsquo;s index 0.48), and Node-RADS\u0026thinsp;\u0026ge;\u0026thinsp;4 achieved 48.6% sensitivity and 98.1% specificity (Youden\u0026rsquo;s index 0.47), with an AUROC of 0.78 [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. In bladder cancer, Leonardo et al. found accuracy to be moderate to high, with AUC 0.87 at the patient level and 0.91 at the lymph node level, and a cutoff of Node-RADS\u0026thinsp;\u0026gt;\u0026thinsp;2 provided a balanced sensitivity/specificity of 78.6%/77.1% [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. For mediastinal lymph node assessment in lung cancer, Meyer et al. reported an AUC of 0.94 for Node-RADS scoring, with a threshold of 2 yielding 0.74 sensitivity and 0.93 specificity [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Conversely, not all entities show robust performance: in colon cancer, Leonhardi et al. found a high interreader variability and limited diagnostic accuracy for CT-based Node-RADS (AUC 0.68), with a threshold of 2 yielding 0.62 sensitivity and 0.71 specificity [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. In an earlier study of pancreatic head PDAC conducted before the introduction of Node-RADS, Loch et al. showed that combining nodal size with morphologic criteria improved staging performance (sensitivity 61%, specificity 82%, PPV 90%, Youden\u0026rsquo;s index 0.43 [95% CI 0.15\u0026ndash;0.60]) compared with size alone (sensitivity 44.2%, specificity 82.4%, Youden\u0026rsquo;s index 0.27), yet overall accuracy remained only moderate, highlighting the intrinsic limitations of morphology-based lymph node assessment in PDAC. The fact that our PDAC findings are in agreement with these lower-performing settings supports the concept that nodal morphology, node conspicuity, and the incremental benefit of Node-RADS staging might vary with tumor biology and anatomic context.\u003c/p\u003e \u003cp\u003eReproducibility is a second key barrier to clinical implementation in patients with PDAC. In our study population, short-axis diameter thresholds showed fair interreader agreement, whereas agreement for configuration features and for Node-RADS scores was only slight to none (κ approximately\u0026thinsp;\u0026minus;\u0026thinsp;0.02 to 0.17 for most configuration items; κ\u0026thinsp;=\u0026thinsp;0.07 for Node-RADS\u0026thinsp;\u0026ge;\u0026thinsp;3). This is notably lower than the substantial agreement reported in gastric cancer for Node-RADS thresholds (κ 0.73 for \u0026ge;\u0026thinsp;3 and κ\u0026thinsp;=\u0026thinsp;0.67 for \u0026ge;\u0026thinsp;4) and lower than the moderate agreement reported for overall Node-RADS scoring in the lung cancer study (κ\u0026thinsp;=\u0026thinsp;0.48) [15; 17]. Even where overall performance was modest, interreader agreement for the aggregated configuration score was only slight (κ\u0026thinsp;=\u0026thinsp;0.18), indicating poor overall reproducibility of morphology-based lymph node assessment [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Clinically, our findings support the assumption that CT-based cN staging in PDAC, whether performed using size thresholds or Node-RADS, should be interpreted cautiously. A \u0026ldquo;negative\u0026rdquo; nodal assessment on CT cannot rule out nodal metastasis and should not be used in isolation to downstage risk or to de-escalate therapy. Conversely, clearly suspicious lymph nodes (marked enlargement or overt abnormal configuration translating into higher Node-RADS categories) may still be useful because of relatively high specificity, supporting structured communication in multidisciplinary tumor boards and potentially guiding targeted confirmatory procedures in selected cases. However, given the low sensitivity observed even at the PDAC-specific Node-RADS threshold, meaningful improvement in preoperative nodal assessment will likely require approaches that extend beyond morphologic CT evaluation, such as multimodal imaging (e.g., PET-based or diffusion-based techniques, or MR elastography) [\u003cspan additionalcitationids=\"CR27 CR28 CR29 CR30\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOur study has several limitations. First, its retrospective design entails an inherent risk of selection bias. Second, we only included surgically treated patients, which may restrict the generalizability of our findings to individuals with unresectable primary or metastatic disease. Third, histopathology served as a robust reference standard, but correlation was performed on a per-patient basis rather than node-to-node matching; consequently, metastatic deposits within non-visualized or non-selected nodes could not be directly linked to specific imaging findings. Finally, CT acquisition parameters were not fully standardized across centers, and slice thickness of up to 5 mm was allowed, which may have degraded depiction of subtle morphologic signs. Conversely, key strengths include the multicenter design, the use of histopathology after standardized regional lymphadenectomy, and the blinded assessment by readers with different experience levels, providing a pragmatic estimate of real-world performance and reproducibility.\u003c/p\u003e \u003cp\u003eIn conclusion, Node-RADS demonstrates limited accuracy and interreader agreement for regional lymph node staging in PDAC. These results underscore the persistent need for improved imaging strategies to more reliably assess nodal status in this highly aggressive malignancy.\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eSupport from the German Research Foundation (DFG; Ingolf Sack and Rolf Reiter: SFB1340 Matrix-in-Vision project number 372486779, FOR5628 project number 513752256, M5 project number 540759292) is gratefully acknowledged. Rolf Reiter is a participant of the BIH-Charit\u0026eacute; Digital Clinician Scientist Program funded by Charit\u0026eacute; \u0026ndash; Universit\u0026auml;tsmedizin Berlin, Berlin Institute of Health, and the DFG.\u003c/p\u003e\n\u003cp\u003eThis manuscript contains results of the thesis \u0026ldquo;Kategorisierung von Lymphknoten mittels moderner Schnittbildgebungs- und Auswertungsverfahren im Rahmen des pr\u0026auml;therapeutischen Stagings des Pankreaskarzinoms\u0026rdquo;, which will be submitted to \u003cem\u003eCharit\u0026eacute; - Universit\u0026auml;tsmedizin Berlin, corporate member of Freie Universit\u0026auml;t Berlin, Humboldt-Universit\u0026auml;t zu Berlin, and Berlin Institute of Health\u003c/em\u003e by A.J.T. Knittel.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eConroy T, Castan F, Lopez A et al (2022) Five-Year Outcomes of FOLFIRINOX vs Gemcitabine as Adjuvant Therapy for Pancreatic Cancer: A Randomized Clinical Trial. JAMA Oncol 8:1571\u0026ndash;1578\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSiegel RL, Giaquinto AN, Jemal A (2024) Cancer statistics, 2024. Cancer J Clin 74:12\u0026ndash;49\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. Cancer J Clin 71:209\u0026ndash;249\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVincent A, Herman J, Schulick R, Hruban RH, Goggins M (2011) Pancreatic cancer. Lancet 378:607\u0026ndash;620\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllen PJ, Kuk D, Castillo CF et al (2017) Multi-institutional Validation Study of the American Joint Commission on Cancer (8th Edition) Changes for T and N Staging in Patients With Pancreatic Adenocarcinoma. Ann Surg 265:185\u0026ndash;191\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e(2025) NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines\u0026reg;): Pancreatic adenocarcinoma. National Comprehensive Cancer Network Version 2\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorales-Oyarvide V, Rubinson DA, Dunne RF et al (2017) Lymph node metastases in resected pancreatic ductal adenocarcinoma: predictors of disease recurrence and survival. Br J Cancer 117:1874\u0026ndash;1882\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoch FN, Asbach P, Haas M et al (2020) Accuracy of various criteria for lymph node staging in ductal adenocarcinoma of the pancreatic head by computed tomography and magnetic resonance imaging. World J Surg Oncol 18:213\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDu T, Bill KA, Ford J et al (2018) The diagnosis and staging of pancreatic cancer: A comparison of endoscopic ultrasound and computed tomography with pancreas protocol. Am J Surg 215:472\u0026ndash;475\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTran Cao HS, Zhang Q, Sada YH et al (2017) Value of lymph node positivity in treatment planning for early stage pancreatic cancer. Surgery 162:557\u0026ndash;567\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrenzel KL, Holscher AH, Vallbohmer D et al (2010) Lymph node size and metastatic infiltration in adenocarcinoma of the pancreatic head. Eur J Surg Oncol 36:993\u0026ndash;996\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwords DS, Firpo MA, Johnson KM, Boucher KM, Scaife CL, Mulvihill SJ (2017) Implications of inaccurate clinical nodal staging in pancreatic adenocarcinoma. Surgery 162:104\u0026ndash;111\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElsholtz FHJ, Asbach P, Haas M et al (2021) Introducing the Node Reporting and Data System 1.0 (Node-RADS): a concept for standardized assessment of lymph nodes in cancer. Eur Radiol 31:6116\u0026ndash;6124\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeonardo C, Flammia RS, Lucciola S et al (2023) Performance of Node-RADS Scoring System for a Standardized Assessment of Regional Lymph Nodes in Bladder Cancer Patients. Cancers (Basel) 15\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoch FN, Beyer K, Kreis ME et al (2024) Diagnostic performance of Node Reporting and Data System (Node-RADS) for regional lymph node staging of gastric cancer by CT. Eur Radiol 34:3183\u0026ndash;3193\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLucciola S, Pisciotti ML, Frisenda M et al (2022) Predictive role of node-rads score in patients with prostate cancer candidates for radical prostatectomy with extended lymph node dissection: comparative analysis with validated nomograms. Prostate Cancer Prostatic Dis. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41391-022-00564-z\u003c/span\u003e\u003cspan address=\"10.1038/s41391-022-00564-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer HJ, Schnarkowski B, Pappisch J et al (2022) CT texture analysis and node-RADS CT score of mediastinal lymph nodes - diagnostic performance in lung cancer patients. Cancer Imaging 22:75\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaggialetti N, Greco CN, Lucarelli NM et al (2024) Correction: Applications of new radiological scores: the Node\u0026ndash;rads in colon cancer staging. Radiologia Med 129:524\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim HJ, Chae EY, Eom HJ et al (2025) Node Reporting and Data System Evaluation of Axillary Nodes in Invasive Ductal and Lobular Carcinoma. Radiology 316:e243823\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePediconi F, Maroncelli R, Pasculli M et al (2024) Performance of MRI for standardized lymph nodes assessment in breast cancer: are we ready for Node-RADS? Eur Radiol 34:7734\u0026ndash;7745\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeonhardi J, Mehdorn M, Stelzner S et al (2025) Diagnostic accuracy and reliability of CT-based Node-RADS for colon cancer. Abdom Radiol (NY) 50:1\u0026ndash;7\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhong J, Mao S, Chen H et al (2025) Node-RADS: a systematic review and meta-analysis of diagnostic performance, category-wise malignancy rates, and inter-observer reliability. Eur Radiol 35:2723\u0026ndash;2735\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShin DW, Kim J (2020) The American Joint Committee on Cancer 8th edition staging system for the pancreatic ductal adenocarcinoma: is it better than the 7th edition? Hepatobiliary Surg Nutr 9:98\u0026ndash;100\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrierley JDGM, Wittekind C (2017) UICC: TNM Classification of Malignant Tumours. Wiley-Blackwell, Oxford\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElizabeth R, DeLong DMDaDLC, -P (1988) Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach. Int Biometric Soc 44:837\u0026ndash;845\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoon JK, Park MS, Kim SS et al (2022) Regional lymph node metastasis detected on preoperative CT and/or FDG-PET may predict early recurrence of pancreatic adenocarcinoma after curative resection. Sci Rep 12:17296\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen K, Mei Y, Xu M et al (2025) Preoperative prediction of lymph node metastasis in pancreatic ductal adenocarcinoma using MRI-derived whole-tumor ADC histogram analysis. Front Med (Lausanne) 12:1736306\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBayerl C, Safraou Y, Reiter R et al (2024) Investigation of hepatic inflammation via viscoelasticity at low and high mechanical frequencies - A magnetic resonance elastography study. J Mech Behav Biomed Mater 160:106711\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBustin H, Meyer T, Reiter R et al (2025) ElastoNet: Neural network-based multicomponent MR elastography wave inversion with uncertainty quantification. Med Image Anal 105:103642\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNeelsen C, Elgeti T, Meyer T et al (2024) Multifrequency Magnetic Resonance Elastography Detects Small Abdominal Lymph Node Metastasis by High Stiffness. Invest Radiol 59:787\u0026ndash;793\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchattenfroh J, Almutawakel S, Bieling J et al (2025) Technical Recommendation on Multi-Driver Multifrequency MR Elastography for Tomographic Mapping of Abdominal Stiffness With a Focus on the Pancreas and Pancreatic Ductal Adenocarcinoma. Invest Radiol. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/RLI.0000000000001231\u003c/span\u003e\u003cspan address=\"10.1097/RLI.0000000000001231\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Charité – Universitätsmedizin Berlin","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":"Node-RADS, lymph node, pancreatic ductal adenocarcinoma, computed tomography","lastPublishedDoi":"10.21203/rs.3.rs-9115727/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9115727/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjectives\u003c/h2\u003e \u003cp\u003eTo investigate the diagnostic performance of Node Reporting and Data System (Node-RADS) for regional lymph node staging in pancreatic ductal adenocarcinoma (PDAC), using histopathology as reference standard, and to compare Node-RADS with individual size- and morphology-based criteria.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003e This retrospective multicenter study included 87 consecutive patients (median age, 72 years, range 43\u0026ndash;91 years; 45 women) with histologically confirmed PDAC and preoperative CT. Two blinded readers independently assessed regional lymph nodes and assigned Node-RADS scores. Diagnostic performance (sensitivity, specificity, Youden\u0026rsquo;s index, area under the receiver operating characteristic curve (AUROC) for numeric size and Node-RADS) and interreader agreement (Cohen\u0026rsquo;s κ) were calculated.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eEighty-seven patients were included (median age 72 years); 54/87 (62.1%) were node-positive. Overall, 257 regional lymph nodes were assessed on CT (median, 3 per patient), compared with a median of 16 lymph nodes per patient evaluated by histopathology. Node-RADS showed poor discrimination and did not outperform short-axis size (AUROC 0.58, CI 0.45\u0026ndash;0.70 vs. 0.56, CI 0.43\u0026ndash;0.68; p\u0026thinsp;=\u0026thinsp;0.70). The PDAC-recommended threshold (Node-RADS\u0026thinsp;\u0026ge;\u0026thinsp;3) yielded 24.1% sensitivity and 81.8% specificity (Youden\u0026rsquo;s index 0.06). The best-performing size cutoff was short-axis\u0026thinsp;\u0026ge;\u0026thinsp;10 mm (sensitivity 25.9%, specificity 84.8%, Youden\u0026rsquo;s index 0.11). Individual morphologic features were specific but infrequent, resulting in low sensitivity. Interreader agreement was slight to substantial for size thresholds (κ\u0026thinsp;=\u0026thinsp;0.32\u0026ndash;0.65) but slight for Node-RADS (κ\u0026thinsp;=\u0026thinsp;0.07 for \u0026ge;\u0026thinsp;3) and most configuration features.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eNode-RADS provides limited accuracy and low reproducibility for regional lymph node staging in PDAC and does not improve upon conventional size- or morphology-based criteria.\u003c/p\u003e","manuscriptTitle":"Diagnostic Performance of Node-RADS for CT-based Assessment of Regional Lymph Nodes in Pancreatic Ductal Adenocarcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-18 08:38:24","doi":"10.21203/rs.3.rs-9115727/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":"c8a906eb-88d9-46ab-aca3-74d2f42f6b55","owner":[],"postedDate":"March 18th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64465791,"name":"Nuclear Medicine \u0026 Medical Imaging"}],"tags":[],"updatedAt":"2026-03-18T08:38:24+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-18 08:38:24","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9115727","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9115727","identity":"rs-9115727","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}