A CT-based radiomics nomogram for the preoperative prediction of perineural invasion in pancreatic ductal adenocarcinoma

A CT-based radiomics nomogram for the preoperative prediction of perineural invasion 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 A CT-based radiomics nomogram for the preoperative prediction of perineural invasion in pancreatic ductal adenocarcinoma Yan Deng, Haopeng Yu, Xiuping Duan, Li Liu, Zixing Huang, Bin Song This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4161245/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Preoperative evaluation perineural invasion (PNI) affects the treatment and prognosis of patients with pancreatic ductal adenocarcinoma (PDAC). This study aims to develop a nomogram based on a CT radiomics nomogram for the preoperative prediction of PNI in PDAC patients. Methods A total of 217 patients with histologically confirmed PDAC were enrolled in this retrospective study. Radiomics features were extracted from the whole tumor. Univariate analysis and least absolute shrinkage and selection operator logistic regression were applied for feature selection and radiomics model construction. Finally, a nomogram combining the radiomics score (Rad-score) and clinical characteristics was established. Receiver operating characteristic curve analysis, calibration curve analysis and decision curve analysis (DCA) were used to evaluate the predictive performance of the radiomics nomogram. Results According to multivariate analysis, CT features, including the evaluation of radiologists regarding PNI status based on CECT (CTPNI) (OR = 0.315 [95% CI: 0.131, 0.761], P = 0.01), the lymph node status determined on CECT (CTLN) (OR = 0.169 [95% CI: 0.059, 0.479], P = 0.001) and the Rad-score (OR = 3.666 [95% CI: 2.069, 6.494], P < 0.001), were significantly associated with PNI. The area under the receiver operating characteristic curve (AUC) for the nomogram combined with the Rad-score, CTLN and CTPNI achieved favorable discrimination of PNI status, with AUCs of 0.846 and 0.778 in the training and testing cohorts, respectively, which were superior to those of the Rad-score (AUC of 0.720 in the training cohort and 0.640 in the testing cohort) and CTPNI (AUC of 0.610 in the training cohort and 0.675 in the testing cohort). The calibration plot and decision curve showed good results. Conclusion The CT-based radiomics nomogram has the potential to accurately predict PNI in patients with PDAC. pancreatic ductal adenocarcinoma perineural invasion computed tomography radiomics nomogram Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Key points 1. Preoperative evaluation perineural invasion (PNI) affects the treatment and prognosis of patients with pancreatic ductal adenocarcinoma (PDAC). 2. This retrospective study established a nomogram based on morphological features, and the quantitative radiomics score from contrast-enhanced CT has the potential to preoperatively predict PNI status in PDAC patients. 3. The performance of the nomogram was better than that of radiomics score or radiologists in evaluating PNI status. Introduction Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related death, and its prevalence has slowly increased among men, from 12.1 per 100000 in 2000 to 12.7 per 100000 in 2020, while its prevalence has remained stable in women at 9.3–9.6 per 100000[ 1 ]. Radical resection is the only effective means for treatment, but fewer than 20% of patients are able to undergo surgery at the time of diagnosis, and early recurrence and metastasis frequently occur after radical resection[ 2 ]. PDAC is characterized by perineural growth, and its incidence is 43.2%-100%[ 3 ]. Perineural invasion (PNI) is related to the dissemination and metastasis of PDAC and is an independent risk factor for patient prognosis[ 4 – 6 ]. A previous study showed that patients who received neoadjuvant therapy had a significantly lower PNI than did those who did not receive neoadjuvant therapy[ 7 ]. Felsenstein et al. demonstrated that adjuvant chemotherapy improved the prognosis in patients with PNI-positive PDAC but not in those with PNI-negative disease[ 8 ]. In addition, the PNI status affects whether the Heidelberg procedure is performed[ 9 ]. The assessment of PNI currently relies on histopathology following surgery; however, preoperative knowledge of PNI status holds clinical significance because it has the potential to aid clinicians in identifying high-risk categories beforehand, formulating personalized treatment plans, and ultimately improving patient outcomes. Contrast-enhanced CT (CECT) is the first-line imaging method for the diagnosis, staging and evaluation of the therapeutic effect of PDAC[ 10 , 11 ]. Previous studies[ 12 – 14 ] have evaluated PNI of PDAC via qualitative methods and established criteria for CECT; however, these studies relied on radiologists’ experience and had low accuracy. Guo et al.[ 15 ] established a quantitative method based on the minimum distance between the tumor boundary and adjacent arteries, but the tumor boundary is difficult to define and may affect the measurement results. Radiomics can noninvasively and rapidly obtain diagnostic, prognostic and treatment information from medical images to support clinical decision-making. This information can be used as a complementary tool to verify clinical and imaging results[ 16 – 18 ]. Radiomics has been applied to evaluate lymph node (LN) metastasis and assess the prognosis of PDAC patients[ 19 – 21 ]. Several CT/MRI-based radiomics models for predicting PNI have been introduced for rectal cancer and gastric cancer patients, and they have achieved satisfactory results[ 22 , 23 ]. However, no study has evaluated PNI preoperatively in patients with PDAC based on CT radiomics. The aim of this study was to develop and validate a nomogram based on CT radiomics features and clinical characteristics for the preoperative prediction of PNI in PDAC patients. Materials and Methods This retrospective study was approved by the institutional review board at West China Hospital, Sichuan, China (IRB number: 2023-0003), and the requirement for written informed consent was waived. It performed in accordance with Declaration of Helsinki. Patients A total of 335 patients with PDAC who underwent CT at our hospital between September 2021 and February 2023 were enrolled in this retrospective study. The inclusion criteria were as follows: (1) underwent radical resection, and preoperative CECT images were available at our institution, (2) primary PDAC and definite PNI were confirmed by histopathology, and (3) complete clinicopathological information was available. The exclusion criteria were as follows: (1) lesions that were too small (less than 1 cm) or of poor image quality that did not meet diagnostic criteria, (2) preoperative neoadjuvant therapy such as radiotherapy or chemotherapy, (3) more than 30 days between the preoperative CT scan and surgery, or (4) other retroperitoneal tumors. Finally, 217 patients with PDAC were enrolled (99 PNI-negative patients and 118 PNI-positive patients), and the patients were randomly divided into a training cohort (n = 151) and a validation cohort (n = 66) at a ratio of 7:3 (Fig. 1 ). Demographic information, including age and sex, was collected. The carbohydrate antigen 19 − 9 (CA19-9) and carcinoembryonic antigen (CEA) levels before surgery were recorded. Pathological PNI diagnosis PNI is defined as a tumor located near a nerve, with tumor cells located in at least 33% of the nerve perimeter or in any of the three layers of the nerve sheath[ 24 ]. CT image acquisition The abdominal CT scanning parameters and contrast agents used are described in detail in supplementary A1. CT image analysis Two radiologists with six and eight years of experience in abdominal imaging who were blinded to the pathologic details reviewed the CT images and evaluated the following features: CTPNI, CTLN, location and size of the tumor, and dilatation of the common bile duct and the main pancreatic duct. Discrepancies between observers were resolved by consensus, and further analysis was performed using consensus interpretation. Interagreement between the two reviewers was evaluated by calculating the intraclass correlation coefficient (ICC) for continuous variables and Cohen’s kappa value for categorical variables. CTPNI was defined as the disappearance of the peripancreatic fat sPDACe or peripancreatic vascular sPDACe (including the common hepatic artery, superior mesenteric artery, superior mesenteric vein, celiac artery and splenic vessels) or the appearance of ribbon-like, reticular soft tissue density shadows or irregular mass shadows[ 25 ]. If any of the following conditions were met, the CTLN was evaluated as positive: the short diameter of the LN was more than 10 mm, the density was uneven, the enhancement was uneven, internal necrosis occurred, the LN was fused, the boundary of the LN was unclear, or the LN invaded adjacent organs or blood vessels[ 26 ]. The location and size of the tumor were recorded based on preoperative CT. The location of the tumor was defined as the head on the right side, the neck in front, or the body or tail on the left side according to the confluence of the portal vein and the superior mesenteric vein. The size of the tumor was measured at the axial level according to the largest cross section of the lesion. Tumor size was calculated as the mean of two measurements for further analysis. A common bile duct (CBD) diameter greater than 10 mm and a main pancreatic duct (MPD) diameter greater than 2 mm were defined as dilated. Tumor segmentation Two experienced radiologists in abdominal imaging (with six and eight years of experience from Reader 1 and Reader 2, respectively) manually and independently and blindly sketched the lesion slice by slice along the edge to the results from 30 randomly selected patients based on the arterial and portal venous images of CECT, avoiding the CBD and vessels. Reader 1 sketched the lesion twice, with an interval of more than 1 week, for calculating intra-agreement. Reader 1 and Reader 2 blindly and independently delineated the lesion to measure the interagreement. Then, the sketch of all patients was completed by reader 1. This process was implemented on the open source software IBEX (β1.0, http://bit.ly/IBEX MDAnderson), which runs on MATLAB 2013a. Due to the variability problems caused by voxel size and gray level dependence, it is unrealistic that all radiomics features achieve satisfactory agreement. The ICC was used to assess intraobserver and interobserver agreement. An ICC greater than 0.75 was considered to indicate good consistency. Radiomics extraction and selection The gray level cooccurrence matrix (GLCM), gray level runlength matrix (GLRLM), intensity histogram (IH), intensity direct (ID) and shape feature groups were extracted from IBEX. A total of 808 radiomics features were extracted from arterial and portal venous CECT images. Resampling was applied for preprocessing to eliminate images with different scanning parameters and slice thicknesses[ 27 ]. The Z score was used to eliminate the effect of the data dimension. The radiomics workflow is shown in Fig. 2 . Two steps were adopted for reducing the dimensions and identifying robust radiomics features in the training cohort. Univariate analysis with an independent samples t test or the Mann‒Whitney U test was first applied to select potentially important features. Subsequently, the least absolute shrinkage and selection operator (LASSO) method was applied using tenfold cross-validation for feature selection. Lambda was selected according to the 1-standard error of the minimum criterion (1-SE criterion, a simpler model). The selected optimal radiomics features were weighted by their respective coefficients and a linear combination to obtain the corresponding radiomics score (Rad-score) in the training and testing cohorts. Nomogram construction The significantly different features were used to construct a nomogram for both the training and testing cohorts. The area under the receiver operating characteristic (ROC) curve (AUC), accuracy, sensitivity and specificity were calculated to assess the performance of the nomogram model. A calibration curve was used to evaluate the agreement between the predicted probability of PNI and the actual probability of PNI. The clinical utility of the model was evaluated by a decisive curve. Statistical analysis Statistical analyses of clinical characteristics were conducted with SPSS (statistics 26). The remaining statistical analyses were implemented in R (version 4.3.1, https://www.r-project.org/ ). A significant difference was considered at P < 0.05. Categorical variables were analyzed using the chi-square test or Fisher’s exact test. Continuous variables were analyzed by the independent samples t test or the Mann‒Whitney U test, depending on the type of data distribution. Differences in clinical information and CT characteristics between patients with and without PNI were compared using univariate analysis, and multivariate logistic regression analysis was used to identify independent predictors that were significantly associated with PNI. The “glmnet” PDACkage was used for LASSO regression analysis. The “rms” PDACkage was applied for nomogram construction and calibration curve plotting. The “pROC” PDACkage and “dca. R” were used for the ROC curve and decision curve plot, respectively. Results Clinical characteristics Table 1 shows the clinical data for patients in the training cohort and testing cohort. The LN status and tumor histopathological grade were significantly different between the two groups in the training cohort (P 0.05). The CTLN and CTPNI were significantly different in both the training and testing cohorts (P < 0.05). Table 1 The patients’ characteristics in the training and testing cohort. Characteristics Training cohort P value Testing cohort P value PNI (+) (n = 82) PNI (-) (n = 69) PNI (+) (n = 36) PNI (-) (n = 30) Age (y) 60.1 ± 11.0 60.4 ± 9.2 0.451 59.3 ± 11.3 57.9 ± 10.1 0.578 Sex 0.225 0.451 Male 48 47 22 21 Female 34 22 14 9 Interval between CT and surgery (d) 5 (3,8) 6 (3, 9) 0.401 4 (2,7) 4 (2, 6) 0.869 Location 0.903 0.794 Head 53 47 30 24 Neck 4 3 3 2 Body or tail 25 19 3 4 Vascular 0.652 0.288 Negative 54 43 26 25 Positive 28 26 10 5 Margin 0.641 0.114 Negative 73 63 30 29 Positive 9 6 6 1 MPD 0.229 0.203 Negative 21 23 7 10 Positive 61 46 29 20 CBD 0.284 0.575 Negative 38 26 12 12 Positive 44 43 24 18 CTLN 0.000* 0.034* Negative 47 61 21 25 Positive 35 8 15 5 CTPNI 0.006* 0.004* Negative 45 53 21 28 Positive 37 16 15 2 Size 2.50 (1.90, 4.05) 2.24 (1.84, 3.09) 0.155 2.55 (1.94, 3.19) 2.42 (2, 3) 0.981 CEA 3.60 (2.21, 5.57) 3.32 (2.27, 5.51) 0.977 3.12 (2.21, 7.4) 3.02 (1.72, 6.2) 0.747 CA19-9 270 (65.89, 795.5) 195. 45 (26.23, 577. 63) 0.375 308.25 (75.0, 699.55) 196.25 (61.15, 387.95) 0.245 LN 0.037* 0.389 Negative 42 47 19 19 Positive 40 22 17 11 Grade 0.049* 0.413 Well-differentiated 1 3 1 2 Moderately differentiated 53 54 23 22 Poorly differentiated 28 12 12 6 Rad-score 0.52 (-0.06, 0.99) -0.13 (-0.85, -0.5) 0.000* 0.46 (-0.14, 0.95) -0.08 (-0.58, 0.6) 0.01* * represents a statistically significant difference. CTPNI: Radiologists evaluated the status of PNI based on CECT; CTLN: The lymph node status determined on CT; CA19-9: Carbohydrate antigen 19 − 9; CEA: Carcinoembryonic antigen; CBD: Common bile duct; LN: Lymph node; MPD: Main pancreatic duct; PNI: Perineural invasion; Rad-score: Radiomics score. There was no significant difference between the training and validation cohorts in terms of the percentage of PNI-positive patients. There was no significant difference between the PNI-positive and PNI-negative groups in age, sex, CEA, CA19-9, tumor size or location, CBD or MPD dilation status, or vascular invasion or margin status in either the training or testing cohort. The ICC of the tumor size was 0.936, indicating good consistency. The kappa values of the CTPNI and CTLN were 0.723 and 0.741, respectively, with moderate consistency. According to the univariate analysis of the training cohort, the CTPNI, CTLN, LN and tumor grade were significantly associated with PNI (Table 2 ). According to multivariate analysis, CT features, including the CTPNI (OR = 0.315 [95% CI: 0.131, 0.761], P = 0.01) and CTLN (OR = 0.169 [95% CI: 0.059, 0.479], P = 0.001), were significantly associated with PNI. Table 2 Univariate and multivariate logistic regression analysis for PNI of PDAC. Characteristics Univariate Analysis Multivariate Analysis Odds Ratio (95% CI) P value Odds Ratio (95% CI) P value CTPNI Present 1 1 Absent 0.367 (0.181, 0.746) 0.006 0.315 (0.131, 0.761) 0.01 CTLN Positive 1 1 Negative 0.176 (0.075, 0.415) < 0.000 0.169 (0.059, 0.479) 0.001 LN Positive 1 1 Negative 0.491 (0.252, 0.957) 0.037 0.456 (0.196, 1.065) 0.07 Grade Poorly differentiated 1 1 Moderately differentiated 0.143 (0.013, 1.516) 0.106 0.435 (0.034, 5.52) 0.521 Well-differentiated 0.421 (0.194, 0.913) 0.029 0.434 (0.169, 1.115) 0.083 Rad-score 2.718 (1.172, 4.316) < 0.001 3.666 (2.069, 6.494) < 0.001 CI: Confidence interval; CTPNI: Radiologists evaluated the status of PNI based on CECT; CTLN: The lymph node status determined on CT; LN: Lymph node; Rad-score: Radiomics score. Radiomics feature selection and model construction The mean ICCs for intraobserver agreement and interobserver agreement were 0.889 (range from 0.103 to 0.995) and 0.843 (range from 0.002–0.993), respectively (Supplementary Fig. 1). Ninety radiomics features were excluded because of intraobserver agreement, and 141 radiomics features were excluded because of interobserver agreement. For features with suboptimal agreement, 68 intraobserver-excluded features were included among 141 interobserver-excluded features. Ultimately, 163 radiomics features were excluded due to inferior reproducibility, and the remaining 645 radiomics features were used for the next analysis. After univariate analysis, 91 radiomics features were significantly different between the PNI-positive and PNI-negative groups in the training cohort. The LASSO regression method with 10-fold cross-validation was applied for the remaining features selected, and 8 optimized features were chosen for constructing the model. The selected radiomics features were quantitatively integrated into the Rad-score in the training and testing cohorts. According to the univariate and multivariate analyses, the Rad-score (OR = 3.666 [95% CI: 2.069, 6.494], P < 0.001) was significantly related to PNI. Nomogram construction The Rad-score was independently associated with PNI. The AUC for the Rad-score in the training cohort (0.720, 95% confidence interval [CI]: [0.639, 0.802]) was close to that in the testing cohort (0.640, 95% CI: 0.499, 0.781). The AUC of the CTPNI was 0.610 (95% CI: 0.536, 0.684) in the training cohort and 0.675 (95% CI: 0.582, 0.768) in the testing cohort. Using the Rad-score combined with the CTLN and CTPNI, a nomogram (Fig. 3 ) for predicting PNI was constructed, and it achieved favorable performance both in the training cohort, with an AUC of 0.846 (95% CI: 0.785, 0.907), and in the testing cohort, with an AUC of 0.778 (95% CI: 0.666, 0.889). Comparing the AUCs of the nomogram model with those of the Rad-score and CTPNI through the DeLong test, the nomogram model achieved the best performance (P < 0.05) (Fig. 4 ). The diagnostic performance of the nomogram model, Rad-score and CTPNI in both the training and testing cohorts is summarized in Table 3 . The calibration plot indicated that the predicted PNI was consistent with the actual PNI probability (Fig. 5 A). The decision curve suggested that the nomogram model outperformed CTPNI at any threshold probability (Fig. 5 B). Figure 6 shows examples of the clinical application of the nomogram. Table 3 The performance of the training and testing cohorts. Model Accuracy Sensitivity Specificity AUC (95% CI) Training cohort CTPNI model 0.596 0.451 0.768 0.610 (0.536, 0.684) Radiomics model 0.636 0.720 0.536 0.720 (0.639, 0.802) nomogram 0.781 0.890 0.768 0.846 (0.785, 0.907) Testing cohort CTPNI model 0.652 0.417 0.882 0.675 (0.582, 0.768) Radiomics model 0.636 0.722 0.533 0.640 (0.499, 0.781) nomogram 0.667 0.806 0.733 0.778 (0.666, 0.889) AUC: Area under the receiver operating characteristic curve; CI: Confidence interval; Rad-clinical: The combined of CTLN, CTPNI and radiomics model. Discussion Preoperative accurate evaluation of PNI in patients with PDAC affects the choice of appropriate treatment. Our retrospective study constructed a nomogram to preoperatively predict PNI based on the rad-score from the arterial and portal venous phases of CECT, CTLN and CTPNI. The nomogram could effectively predict the occurrence of PNI in both the training and testing groups. This study demonstrated that the nomogram was superior to the Rad-score and CTPNI for evaluating the occurrence of PNI. The prevalence of PNI in PDAC patients is fairly high, varying from 42.3%-100% in previous reports[ 3 ], with an incidence of 54.4% in this study. Additionally, we found an interesting association between PNI status and LN status, and the PNI-positive group was more prone to LN metastasis. Previous studies have indicated that cancer cells grow along nerves in contact with LNs, indicating a complex link between LN metastasis and PNI[ 28 , 29 ]. There was also a statistically significant difference in the CTLNs. In addition, patients with PNI were more likely to have poor pathological differentiation. Poorly differentiated PDAC has more aggressive behavior, which is related to poor prognosis, PNI and poor differentiation and reflects the malignant biological behavior of PDAC. We did not find significant differences in tumor size; dilation status of the CBD or MPD; CEA or CA19-9 levels; or resection margin status between the PNI-positive and PNI-negative groups. The relationship between PNI and tumor size is controversial. Crippa et al. reported that the incidence of PNI increased with tumor size[ 30 ]. However, Patel et al. suggested that no evidence was found between tumor size and PNI[ 31 ]. PNI occurs early in PDAC, and even tumors less than 2 cm may develop PNI[ 32 ]. This may be related to the greater probability of PNI due to tumor growth beyond the pancreas, while tumors within the pancreas generally do not develop PNI even if the tumor is large, based on the anatomical structure[ 33 ]. These controversial results suggest that the relationship between tumor size and PNI needs further investigation. CBD and MPD dilatation caused by obstruction may not be associated with PNI. CEA and CA19-9 are nonspecific in PDAC and may be abnormal in a variety of diseases, which may account for the lack of differences in CEA and CA19-9 between the PNI-positive and PNI-negative groups. Eight radiomics features related to PNI were selected for this study. Kulkarni et al.[ 34 ] extracted CT texture features to analyze their association with PNI and did not find any texture features related to PNI. The possible reason is that the texture features were only extracted from the maximum level of the tumor, which included incomplete features. In addition, in this study, poorly vascularized tumors located in the head of the pancreas were selected. In our study, the radiomics features of the whole lesion were extracted at the three-dimensional level, which could help to discover more biological characteristics of tumors. These selected features were integrated into a Rad-score and exhibited moderate performance in preoperatively predicting the PNI status of PDAC in both the training and testing cohorts. Radiomics can improve the prediction performance of medical images by improving analysis and using computer algorithms to extract thousands of quantitative features, and it can mine a large amount of information that is invisible to the naked eye. According to the radiologists’ evaluation, the CTPNI achieved inferior performance, which may be attributed to perivascular inflammation or fibrosis easily mimicking PNI on CT. In addition, PDAC is characterized by lymphatic growth and PNI, which are easily confused with microvessels, LN or fibrosis on CT. This may have caused the unsatisfactory agreement between the two reviewers in assessing the CTPNI and CTLN in our study. A nomogram combining the Rad-score, CTLN and CTPNI achieved the best performance (the AUC in the testing cohort was 0.778) for the preoperative assessment of PNI in patients with PDAC. Several possible reasons may contribute to the good performance of the nomogram. One is that the Rad-score combined with arterial and portal venous phases can provide valuable information. In addition, different CT scan parameters may lead to unsatisfactory reproducibility of radiomics features, which can be maximally alleviated by using resampling as a preprocessing method, which optimizes gray dispersion to maintain the stability of features[ 35 ]. Z score standardization eliminates the effect of different data dimensions. Moreover, the good performance of the nomogram was attributed to feature selection and modeling. Univariate analysis and LASSO regression confirmed that important features were selected for modeling. Tenfold validation was applied to guarantee the robustness of the model. Finally, the nomogram integrates selected radiomics features, the presence of PDAC on CT images and the experience of radiologists and combines the performance of different dimensions to better reflect the characteristics of PDAC. This result suggested that the nomogram has the potential to preoperatively predict PNI status in PDAC patients. The calibration curve revealed that the predicted PNI was in good agreement with the actual PNI probability. The decision curve indicated that the nomogram outperformed radiologists at any threshold probability. This study has several limitations. First, this was a single-center study with a limited sample size and no external validation group. However, the number included in our study was relatively larger than that in previous studies. Therefore, the retrospective nature of the study may have led to biased results. Large sample, multicenter and prospective studies should be conducted to further verify the results. Moreover, although several studies have reported that PNI is associated with PDAC patient prognosis, the relationship between PNI and prognosis was not clarified in this study. Subsequently, the corresponding patients should be followed up on the basis of this study to explore the ability of the nomogram to predict survival. Ultimately, there was not a consistent one-to-one match between CT evaluation and pathology. In conclusion, a nomogram based on the rad-score derived from both the arterial and portal venous phases of CECT combined with the CTLN and CTPNI may serve as a valuable noninvasive tool for the preoperative assessment of PNI in patients with PDAC. This approach offers a practical means to classify PDAC patients before surgery and enhance patient management. Abbreviations AUC area under the receiver operating characteristic curve CECT contrast-enhanced computed tomography CTPNI radiologists evaluated the status of PNI based on CECT CTLN the lymph node status determined on CT CA19-9 carbohydrate antigen 19 − 9 CEA carcinoembryonic antigen CBD common bile duct GLCM gray-level cooccurrences matrix GLRLM gray-level run-length matrix IH intensity histogram ID intensity direct ICC intraclass correlation coefficient LASSO least absolute shrinkage and selection operator LN lymph node MPD main pancreatic duct PNI perineural invasion PDAC pancreatic ductal adenocarcinoma Rad-score radiomics score ROC received operating characteristic Declarations Acknowledgements Not applicable. Author contribution Y.D. collected the data and drafted the main manuscript, HP.Y. made the statistical analysis and part writing. Their contributions to this study are consistent, so they are listed as co- first authors. XP.D. collected and interpreted the data. L.L. collected the data. ZX.H. revised the manuscript and B.S. designed the manuscript. The contributions of ZX.H. and B.S. to this study are consistent, so they are listed as co-corresponding authors. All authors reviewed the manuscript. Funding The authors state that this study was supported by the 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University (Grant number ZYGD22004, ZYJC21012); the development project of Hainan provincial clinical medical center and post-doctoral station development project of Sana (Grant number 23CZ009). Data availability The datasets used during the current study are available from the corresponding author on reasonable request. Ethical approval and consent to participate Written informed consent was not required for this study because this is retrospective study, this study was approved by the institutional ethics committee of West China Hospital, Sichuan, China (IRB number: 2023-0003) and performed in accordance with Declaration of Helsinki. Consent for publication Not applicable. Competing interests All authors declare that there is no conflict of interest. References Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. Ca Cancer J Clin. 2023;73(1):17-48. https://doi.org/10.3322/caac.21763. Suzuki S, Shimoda M, Shimazaki J, Maruyama T, Oshiro Y, Nishida K, et al. Predictive Early Recurrence Factors of Preoperative Clinicophysiological Findings in Pancreatic Cancer. Eur Surg Res. 2018;59(5-6):329-38. https://doi.org/10.1159/000494382. Schorn S, Demir IE, Haller B, Scheufele F, Reyes CM, Tieftrunk E, et al. The influence of neural invasion on survival and tumor recurrence in pancreatic ductal adenocarcinoma - A systematic review and meta-analysis. Surg Oncol. 2017;26(1):105-15. https://doi.org/10.1016/j.suronc.2017.01.007. Jurcak NR, Rucki AA, Muth S, Thompson E, Sharma R, Ding D, et al. Axon Guidance Molecules Promote Perineural Invasion and Metastasis of Orthotopic Pancreatic Tumors in Mice. Gastroenterology. 2019;157(3):838-50. https://doi.org/10.1053/j.gastro.2019.05.065. Huang C, Li Y, Guo Y, Zhang Z, Lian G, Chen Y, et al. MMP1/PAR1/SP/NK1R paracrine loop modulates early perineural invasion of pancreatic cancer cells. Theranostics. 2018;8(11):3074-86. https://doi.org/10.7150/thno.24281. Ceyhan GO, Bergmann F, Kadihasanoglu M, Altintas B, Demir IE, Hinz U, et al. Pancreatic neuropathy and neuropathic pain--a comprehensive pathomorphological study of 546 cases. Gastroenterology. 2009;136(1):177-86. https://doi.org/10.1053/j.gastro.2008.09.029. Chatterjee D, Katz MH, Rashid A, Wang H, Iuga AC, Varadhachary GR, et al. Perineural and intraneural invasion in posttherapy pancreaticoduodenectomy specimens predicts poor prognosis in patients with pancreatic ductal adenocarcinoma. Am J Surg Pathol. 2012;36(3):409-17. https://doi.org/10.1097/PAS.0b013e31824104c5. Felsenstein M, Lindhammer F, Feist M, Hillebrandt KH, Timmermann L, Benzing C, et al. Perineural Invasion in Pancreatic Ductal Adenocarcinoma (PDAC): A Saboteur of Curative Intended Therapies? J Clin Med. 2022;11(9). https://doi.org/10.3390/jcm11092367. Schneider M, Strobel O, Hackert T, Büchler MW. Pancreatic resection for cancer-the Heidelberg technique. Langenbecks Arch Surg. 2019;404(8):1017-22. https://doi.org/10.1007/s00423-019-01839-1. Mizrahi JD, Surana R, Valle JW, Shroff RT. Pancreatic cancer. Lancet. 2020;395(10242):2008-20. https://doi.org/10.1016/S0140-6736(20)30974-0. Zaky AM, Wolfgang CL, Weiss MJ, Javed AA, Fishman EK, Zaheer A. Tumor-Vessel Relationships in Pancreatic Ductal Adenocarcinoma at Multidetector CT: Different Classification Systems and Their Influence on Treatment Planning. Radiographics. 2017;37(1):93-112. https://doi.org/10.1148/rg.2017160054. Patel BN, Olcott E, Jeffrey RB. Extrapancreatic perineural invasion in pancreatic adenocarcinoma. Abdom Radiol (Ny). 2018;43(2):323-31. https://doi.org/10.1007/s00261-017-1343-9. Deshmukh SD, Willmann JK, Jeffrey RB. Pathways of extrapancreatic perineural invasion by pancreatic adenocarcinoma: evaluation with 3D volume-rendered MDCT imaging. Ajr Am J Roentgenol. 2010;194(3):668-74. https://doi.org/10.2214/AJR.09.3285. Chang ST, Jeffrey RB, Patel BN, DiMaio MA, Rosenberg J, Willmann JK, et al. Preoperative Multidetector CT Diagnosis of Extrapancreatic Perineural or Duodenal Invasion Is Associated with Reduced Postoperative Survival after Pancreaticoduodenectomy for Pancreatic Adenocarcinoma: Preliminary Experience and Implications for Patient Care. Radiology. 2016;281(3):816-25. https://doi.org/10.1148/radiol.2016152790. Guo X, Gao S, Yu J, Zhou Y, Gao C, Hao J. The imaging features of extrapancreatic perineural invasion (EPNI) in pancreatic Cancer：A comparative retrospective study. Pancreatology. 2021;21(8):1516-23. https://doi.org/10.1016/j.pan.2021.08.010. Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563-77. https://doi.org/10.1148/radiol.2015151169. Lambin P, Leijenaar R, Deist TM, Peerlings J, de Jong E, van Timmeren J, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62. https://doi.org/10.1038/nrclinonc.2017.141. Tomaszewski MR, Gillies RJ. The Biological Meaning of Radiomic Features. Radiology. 2021;299(2):E256. https://doi.org/10.1148/radiol.2021219005. Li K, Yao Q, Xiao J, Li M, Yang J, Hou W, et al. Contrast-enhanced CT radiomics for predicting lymph node metastasis in pancreatic ductal adenocarcinoma: a pilot study. Cancer Imaging. 2020;20(1):12. https://doi.org/10.1186/s40644-020-0288-3. Yao J, Cao K, Hou Y, Zhou J, Xia Y, Nogues I, et al. Deep Learning for Fully Automated Prediction of Overall Survival in Patients Undergoing Resection for Pancreatic Cancer: A Retrospective Multicenter Study. Ann Surg. 2023;278(1):e68-79. https://doi.org/10.1097/SLA.0000000000005465. Du D, Feng H, Lv W, Ashrafinia S, Yuan Q, Wang Q, et al. Machine Learning Methods for Optimal Radiomics-Based Differentiation Between Recurrence and Inflammation: Application to Nasopharyngeal Carcinoma Post-therapy PET/CT Images. Mol Imaging Biol. 2020;22(3):730-8. https://doi.org/10.1007/s11307-019-01411-9. Zheng H, Zheng Q, Jiang M, Han C, Yi J, Ai Y, et al. Contrast-enhanced CT based radiomics in the preoperative prediction of perineural invasion for patients with gastric cancer. Eur J Radiol. 2022;154:110393. https://doi.org/10.1016/j.ejrad.2022.110393. Chen J, Chen Y, Zheng D, Pang P, Zhang H, Zheng X, et al. Pretreatment MR-based radiomics nomogram as potential imaging biomarker for individualized assessment of perineural invasion status in rectal cancer. Abdom Radiol (Ny). 2021;46(3):847-57. https://doi.org/10.1007/s00261-020-02710-4. Liebig C, Ayala G, Wilks JA, Berger DH, Albo D. Perineural invasion in cancer: a review of the literature. Cancer-Am Cancer Soc. 2009;115(15):3379-91. https://doi.org/10.1002/cncr.24396. Mochizuki K, Gabata T, Kozaka K, Hattori Y, Zen Y, Kitagawa H, et al. MDCT findings of extrapancreatic nerve plexus invasion by pancreas head carcinoma: correlation with en bloc pathological specimens and diagnostic accuracy. Eur Radiol. 2010;20(7):1757-67. https://doi.org/10.1007/s00330-010-1727-5. Bian Y, Zheng Z, Fang X, Jiang H, Zhu M, Yu J, et al. Artificial Intelligence to Predict Lymph Node Metastasis at CT in Pancreatic Ductal Adenocarcinoma. Radiology. 2023;306(1):160-9. https://doi.org/10.1148/radiol.220329. Shafiq-Ul-Hassan M, Zhang GG, Latifi K, Ullah G, Hunt DC, Balagurunathan Y, et al. Intrinsic dependencies of CT radiomic features on voxel size and number of gray levels. Med Phys. 2017;44(3):1050-62. https://doi.org/10.1002/mp.12123. Gasparini G, Pellegatta M, Crippa S, Lena MS, Belfiori G, Doglioni C, et al. Nerves and Pancreatic Cancer: New Insights into a Dangerous Relationship. Cancers (Basel). 2019;11(7). https://doi.org/10.3390/cancers11070893. Kayahara M, Nakagawara H, Kitagawa H, Ohta T. The nature of neural invasion by pancreatic cancer. Pancreas. 2007;35(3):218-23. https://doi.org/10.1097/mpa.0b013e3180619677. Crippa S, Pergolini I, Javed AA, Honselmann KC, Weiss MJ, Di Salvo F, et al. Implications of Perineural Invasion on Disease Recurrence and Survival After Pancreatectomy for Pancreatic Head Ductal Adenocarcinoma. Ann Surg. 2022;276(2):378-85. https://doi.org/10.1097/SLA.0000000000004464. Patel BN, Giacomini C, Jeffrey RB, Willmann JK, Olcott E. Three-dimensional volume-rendered multidetector CT imaging of the posterior inferior pancreaticoduodenal artery: its anatomy and role in diagnosing extrapancreatic perineural invasion. Cancer Imaging. 2013;13(4):580-90. https://doi.org/10.1102/1470-7330.2013.0051. Marchegiani G, Andrianello S, Malleo G, De Gregorio L, Scarpa A, Mino-Kenudson M, et al. Does Size Matter in Pancreatic Cancer?: Reappraisal of Tumour Dimension as a Predictor of Outcome Beyond the TNM. Ann Surg. 2017;266(1):142-8. https://doi.org/10.1097/SLA.0000000000001837. Tu W, Gottumukkala RV, Schieda N, Lavallée L, Adam BA, Silverman SG. Perineural Invasion and Spread in Common Abdominopelvic Diseases: Imaging Diagnosis and Clinical Significance. Radiographics. 2023;43(7):e220148. https://doi.org/10.1148/rg.220148. Kulkarni A, Carrion-Martinez I, Jiang NN, Puttagunta S, Ruo L, Meyers BM, et al. Hypovascular pancreas head adenocarcinoma: CT texture analysis for assessment of resection margin status and high-risk features. Eur Radiol. 2020;30(5):2853-60. https://doi.org/10.1007/s00330-019-06583-0. Larue R, van Timmeren JE, de Jong E, Feliciani G, Leijenaar R, Schreurs W, et al. Influence of gray level discretization on radiomic feature stability for different CT scanners, tube currents and slice thicknesses: a comprehensive phantom study. Acta Oncol. 2017;56(11):1544-53. https://doi.org/10.1080/0284186X.2017.1351624. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial.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-4161245","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":284337527,"identity":"e156a7f6-d9db-470b-8d5c-4c9b983ab224","order_by":0,"name":"Yan Deng","email":"","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Deng","suffix":""},{"id":284337528,"identity":"001bfeb7-c705-4a56-993d-3e359a70551f","order_by":1,"name":"Haopeng Yu","email":"","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Haopeng","middleName":"","lastName":"Yu","suffix":""},{"id":284337529,"identity":"a294ad8c-d8bb-479c-b263-7da109ae0f07","order_by":2,"name":"Xiuping Duan","email":"","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Xiuping","middleName":"","lastName":"Duan","suffix":""},{"id":284337530,"identity":"c03edcb5-a3f8-4f0a-a1c5-9f04c6596e98","order_by":3,"name":"Li Liu","email":"","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Liu","suffix":""},{"id":284337531,"identity":"ce5a8205-9c5b-485e-8902-d903424a3d6d","order_by":4,"name":"Zixing Huang","email":"","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Zixing","middleName":"","lastName":"Huang","suffix":""},{"id":284337532,"identity":"cafe48f3-872c-4ac9-9a5b-835ddcd74117","order_by":5,"name":"Bin Song","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABFUlEQVRIiWNgGAWjYLCCBwwHQJTBAYYCIMXe2PjwAyEtCXAtBkCK53CzsQSxWsCIQSK9TYAHj2r59t7DLxJq7jCYtx/eeOCDgY3dhpsP2xgkGOzkdBuwazE4cy7NIuHYMwaZM2kFB2cYpCVvuJ3Y9qCAIdnY7AAOLRI5ZgYJbIeB5uYYHOYxOJxscDux3UCC4UDiNhxa5GeAtPwDauF/A9Vy82CbBA8eLQw3cowfJLYBtUhAbLEzuMGIX4vBmTNmDIl9z4BanoH9kiB5JhEYyAa4/SLf3mP84cO3O0CHJW/+8KHCxp7v+PGHDz9U2Mnh0gIEbKB4q2+A8hIXgFUa4FQOAswoqcNevgGHulEwCkbBKBixAACfD2gW7PKUPwAAAABJRU5ErkJggg==","orcid":"","institution":"Department of Radiology, West China Hospital of Sichuan University","correspondingAuthor":true,"prefix":"","firstName":"Bin","middleName":"","lastName":"Song","suffix":""}],"badges":[],"createdAt":"2024-03-25 06:58:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4161245/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4161245/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":53754981,"identity":"37b64b52-519f-47f3-a2d8-62931f2beffd","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":203590,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of patient recruitment. CECT: contrast-enhanced computed tomography; PDAC: Pancreatic ductal adenocarcinoma; PNI: Perineural invasion.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/00bad2c6850c77b3dc1bdee8.png"},{"id":53754980,"identity":"e724ae66-cecf-4756-8fd4-5337bc7486e7","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":301066,"visible":true,"origin":"","legend":"\u003cp\u003eRadiomicsworkflow. GLCM: Gray-level cooccurrencesmatrix. GLRLM: Gray-level run-lengthmatrix.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/920ee074531a62143f97e34e.png"},{"id":53754979,"identity":"a3c674bc-bb28-4d57-910e-3e774d5823e8","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":125541,"visible":true,"origin":"","legend":"\u003cp\u003eThe nomogram for preoperatively evaluating thePNI status of PDAC patients. A nomogram combining the rad-score, CTLN, and CTPNI for preoperative evaluation of PNI in PDAC patients. The Rad-score, CTPNI and CTLN are summed to obtain the total points on the scale, and the risk of PNI in the PDAC is the corresponding number on the “probability” axis. CTLN: Lymph node status determined on CT; CTPNI: Radiologists evaluated the status of PNI based on CECT; PNI: Perineural invasion; PDAC: Pancreatic ductal adenocarcinoma.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/f06bbb572be808fe53bb183c.png"},{"id":53754983,"identity":"3d12c653-2c3b-46c9-8b9f-72544af2b2e4","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":176098,"visible":true,"origin":"","legend":"\u003cp\u003eThe ROC curve of the AUC comparison among the CTPNI model, radiomics model and Rad-clinical model. A: The training cohort; B: The testing cohort. ROC: Received operating characteristic; AUC: Area under the receiver operating characteristic curve; CTPNI: Radiologists evaluated the status of PNI based on CECT.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/96a291495d708cd729fa19a8.png"},{"id":53757379,"identity":"68bb9b94-a4e1-4995-afc9-bd21a83578a1","added_by":"auto","created_at":"2024-03-29 19:07:09","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":193170,"visible":true,"origin":"","legend":"\u003cp\u003eCalibrationcurve analysis was used to evaluate the nomogram performance (A), and decision curve analysis was used (B). A: The x-axis represents the predicted PNI,and the y-axis represents the actual PNI. B: The x-axis represents the threshold probability, and the y-axis represents the net benefit to the patient. PNI: Perineural invasion. CTPNI: Radiologists evaluated the status of PNI based on CECT.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/28cf3b412893fd02fc18cbc3.png"},{"id":53754982,"identity":"8fcd555a-f7bb-4de5-866e-86f4fca8641d","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":473728,"visible":true,"origin":"","legend":"\u003cp\u003eCT images of patients with PNI(A, B) and patients without PNI(C). A and B: A 67-year-old man with PDAC underwent preoperative contrast-enhanced axial abdominal CT, and the CTPNI (white arrow) and CTLN (white arrowhead) were positive. Thelesion (* in A) was used to calculate aRad-score of 1.11508, and the total score was 11.75. Based on the nomogram, a probability of PNI positivity greater than 0.95 and PNI positivity was confirmed by histopathology. C: A 73-year-old man with PDAC who underwent preoperativecontrast-enhanced axial abdominal CT, CTPNI and CTLN were negative, and the Rad-score was -1.89701, as calculatedby delineating the lesion (* in C). Thetotal score was 3.6, the probability of PNI positivity was less than 0.05, and PNI negativity wasconfirmed by histopathology. PNI: Perineural invasion; PDAC: Pancreatic ductal adenocarcinoma; CTLN: Lymph node status determined on CT; CTPNI: Radiologists evaluated the status of PNI based on CECT.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/8406f0cd348c8aa7f62f4482.png"},{"id":67273378,"identity":"cf922908-afd2-4cef-b502-2b5e069c1b37","added_by":"auto","created_at":"2024-10-23 07:54:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2688694,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/b941e788-684d-4d11-8361-ad5b9b2f6b47.pdf"},{"id":53754985,"identity":"114e734d-5fed-43db-b020-fc45a87c8624","added_by":"auto","created_at":"2024-03-29 18:59:09","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":216776,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-4161245/v1/437ba2e317b1da5926ad2a7f.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A CT-based radiomics nomogram for the preoperative prediction of perineural invasion in pancreatic ductal adenocarcinoma","fulltext":[{"header":"Key points","content":"\u003cp\u003e1. Preoperative evaluation perineural invasion (PNI) affects the treatment and prognosis of patients with pancreatic ductal adenocarcinoma (PDAC).\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp;This\u0026nbsp;retrospective study established a nomogram based on morphological features,\u0026nbsp;and\u0026nbsp;the\u0026nbsp;quantitative\u0026nbsp;radiomics\u0026nbsp;score from contrast-enhanced CT has the potential to preoperatively predict PNI status in PDAC patients.\u003c/p\u003e\n\u003cp\u003e3. The performance of the nomogram was better than that of radiomics score or radiologists in evaluating PNI status.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003ePancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related death, and its prevalence has slowly increased among men, from 12.1 per 100000 in 2000 to 12.7 per 100000 in 2020, while its prevalence has remained stable in women at 9.3\u0026ndash;9.6 per 100000[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Radical resection is the only effective means for treatment, but fewer than 20% of patients are able to undergo surgery at the time of diagnosis, and early recurrence and metastasis frequently occur after radical resection[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e].\u003c/p\u003e \u003cp\u003ePDAC is characterized by perineural growth, and its incidence is 43.2%-100%[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Perineural invasion (PNI) is related to the dissemination and metastasis of PDAC and is an independent risk factor for patient prognosis[\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. A previous study showed that patients who received neoadjuvant therapy had a significantly lower PNI than did those who did not receive neoadjuvant therapy[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Felsenstein et al. demonstrated that adjuvant chemotherapy improved the prognosis in patients with PNI-positive PDAC but not in those with PNI-negative disease[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. In addition, the PNI status affects whether the Heidelberg procedure is performed[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. The assessment of PNI currently relies on histopathology following surgery; however, preoperative knowledge of PNI status holds clinical significance because it has the potential to aid clinicians in identifying high-risk categories beforehand, formulating personalized treatment plans, and ultimately improving patient outcomes.\u003c/p\u003e \u003cp\u003eContrast-enhanced CT (CECT) is the first-line imaging method for the diagnosis, staging and evaluation of the therapeutic effect of PDAC[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Previous studies[\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] have evaluated PNI of PDAC via qualitative methods and established criteria for CECT; however, these studies relied on radiologists\u0026rsquo; experience and had low accuracy. Guo et al.[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] established a quantitative method based on the minimum distance between the tumor boundary and adjacent arteries, but the tumor boundary is difficult to define and may affect the measurement results.\u003c/p\u003e \u003cp\u003eRadiomics can noninvasively and rapidly obtain diagnostic, prognostic and treatment information from medical images to support clinical decision-making. This information can be used as a complementary tool to verify clinical and imaging results[\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Radiomics has been applied to evaluate lymph node (LN) metastasis and assess the prognosis of PDAC patients[\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Several CT/MRI-based radiomics models for predicting PNI have been introduced for rectal cancer and gastric cancer patients, and they have achieved satisfactory results[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. However, no study has evaluated PNI preoperatively in patients with PDAC based on CT radiomics.\u003c/p\u003e \u003cp\u003eThe aim of this study was to develop and validate a nomogram based on CT radiomics features and clinical characteristics for the preoperative prediction of PNI in PDAC patients.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e This retrospective study was approved by the institutional review board at West China Hospital, Sichuan, China (IRB number: 2023-0003), and the requirement for written informed consent was waived. It performed in accordance with Declaration of Helsinki.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eA total of 335 patients with PDAC who underwent CT at our hospital between September 2021 and February 2023 were enrolled in this retrospective study. The inclusion criteria were as follows: (1) underwent radical resection, and preoperative CECT images were available at our institution, (2) primary PDAC and definite PNI were confirmed by histopathology, and (3) complete clinicopathological information was available. The exclusion criteria were as follows: (1) lesions that were too small (less than 1 cm) or of poor image quality that did not meet diagnostic criteria, (2) preoperative neoadjuvant therapy such as radiotherapy or chemotherapy, (3) more than 30 days between the preoperative CT scan and surgery, or (4) other retroperitoneal tumors. Finally, 217 patients with PDAC were enrolled (99 PNI-negative patients and 118 PNI-positive patients), and the patients were randomly divided into a training cohort (n\u0026thinsp;=\u0026thinsp;151) and a validation cohort (n\u0026thinsp;=\u0026thinsp;66) at a ratio of 7:3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDemographic information, including age and sex, was collected. The carbohydrate antigen 19\u0026thinsp;\u0026minus;\u0026thinsp;9 (CA19-9) and carcinoembryonic antigen (CEA) levels before surgery were recorded.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003ePathological PNI diagnosis\u003c/h2\u003e \u003cp\u003ePNI is defined as a tumor located near a nerve, with tumor cells located in at least 33% of the nerve perimeter or in any of the three layers of the nerve sheath[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eCT image acquisition\u003c/h2\u003e \u003cp\u003eThe abdominal CT scanning parameters and contrast agents used are described in detail in supplementary A1.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eCT image analysis\u003c/h2\u003e \u003cp\u003eTwo radiologists with six and eight years of experience in abdominal imaging who were blinded to the pathologic details reviewed the CT images and evaluated the following features: CTPNI, CTLN, location and size of the tumor, and dilatation of the common bile duct and the main pancreatic duct. Discrepancies between observers were resolved by consensus, and further analysis was performed using consensus interpretation. Interagreement between the two reviewers was evaluated by calculating the intraclass correlation coefficient (ICC) for continuous variables and Cohen\u0026rsquo;s kappa value for categorical variables.\u003c/p\u003e \u003cp\u003eCTPNI was defined as the disappearance of the peripancreatic fat sPDACe or peripancreatic vascular sPDACe (including the common hepatic artery, superior mesenteric artery, superior mesenteric vein, celiac artery and splenic vessels) or the appearance of ribbon-like, reticular soft tissue density shadows or irregular mass shadows[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIf any of the following conditions were met, the CTLN was evaluated as positive: the short diameter of the LN was more than 10 mm, the density was uneven, the enhancement was uneven, internal necrosis occurred, the LN was fused, the boundary of the LN was unclear, or the LN invaded adjacent organs or blood vessels[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe location and size of the tumor were recorded based on preoperative CT. The location of the tumor was defined as the head on the right side, the neck in front, or the body or tail on the left side according to the confluence of the portal vein and the superior mesenteric vein. The size of the tumor was measured at the axial level according to the largest cross section of the lesion. Tumor size was calculated as the mean of two measurements for further analysis.\u003c/p\u003e \u003cp\u003eA common bile duct (CBD) diameter greater than 10 mm and a main pancreatic duct (MPD) diameter greater than 2 mm were defined as dilated.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eTumor segmentation\u003c/h2\u003e \u003cp\u003eTwo experienced radiologists in abdominal imaging (with six and eight years of experience from Reader 1 and Reader 2, respectively) manually and independently and blindly sketched the lesion slice by slice along the edge to the results from 30 randomly selected patients based on the arterial and portal venous images of CECT, avoiding the CBD and vessels. Reader 1 sketched the lesion twice, with an interval of more than 1 week, for calculating intra-agreement. Reader 1 and Reader 2 blindly and independently delineated the lesion to measure the interagreement. Then, the sketch of all patients was completed by reader 1. This process was implemented on the open source software IBEX (β1.0, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://bit.ly/IBEX\u003c/span\u003e\u003cspan address=\"http://bit.ly/IBEX\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e MDAnderson), which runs on MATLAB 2013a. Due to the variability problems caused by voxel size and gray level dependence, it is unrealistic that all radiomics features achieve satisfactory agreement. The ICC was used to assess intraobserver and interobserver agreement. An ICC greater than 0.75 was considered to indicate good consistency.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eRadiomics extraction and selection\u003c/h2\u003e \u003cp\u003eThe gray level cooccurrence matrix (GLCM), gray level runlength matrix (GLRLM), intensity histogram (IH), intensity direct (ID) and shape feature groups were extracted from IBEX. A total of 808 radiomics features were extracted from arterial and portal venous CECT images. Resampling was applied for preprocessing to eliminate images with different scanning parameters and slice thicknesses[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. The Z score was used to eliminate the effect of the data dimension. The radiomics workflow is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTwo steps were adopted for reducing the dimensions and identifying robust radiomics features in the training cohort. Univariate analysis with an independent samples t test or the Mann‒Whitney U test was first applied to select potentially important features. Subsequently, the least absolute shrinkage and selection operator (LASSO) method was applied using tenfold cross-validation for feature selection. Lambda was selected according to the 1-standard error of the minimum criterion (1-SE criterion, a simpler model). The selected optimal radiomics features were weighted by their respective coefficients and a linear combination to obtain the corresponding radiomics score (Rad-score) in the training and testing cohorts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eNomogram construction\u003c/h2\u003e \u003cp\u003eThe significantly different features were used to construct a nomogram for both the training and testing cohorts. The area under the receiver operating characteristic (ROC) curve (AUC), accuracy, sensitivity and specificity were calculated to assess the performance of the nomogram model. A calibration curve was used to evaluate the agreement between the predicted probability of PNI and the actual probability of PNI. The clinical utility of the model was evaluated by a decisive curve.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eStatistical analyses of clinical characteristics were conducted with SPSS (statistics 26). The remaining statistical analyses were implemented in R (version 4.3.1, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.r-project.org/\u003c/span\u003e\u003cspan address=\"https://www.r-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). A significant difference was considered at P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Categorical variables were analyzed using the chi-square test or Fisher\u0026rsquo;s exact test. Continuous variables were analyzed by the independent samples t test or the Mann‒Whitney U test, depending on the type of data distribution. Differences in clinical information and CT characteristics between patients with and without PNI were compared using univariate analysis, and multivariate logistic regression analysis was used to identify independent predictors that were significantly associated with PNI.\u003c/p\u003e \u003cp\u003eThe \u0026ldquo;glmnet\u0026rdquo; PDACkage was used for LASSO regression analysis. The \u0026ldquo;rms\u0026rdquo; PDACkage was applied for nomogram construction and calibration curve plotting. The \u0026ldquo;pROC\u0026rdquo; PDACkage and \u0026ldquo;dca. R\u0026rdquo; were used for the ROC curve and decision curve plot, respectively.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eClinical characteristics\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the clinical data for patients in the training cohort and testing cohort. The LN status and tumor histopathological grade were significantly different between the two groups in the training cohort (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), while there was no significant difference in the testing cohort (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The CTLN and CTPNI were significantly different in both the training and testing cohorts (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\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\u003eThe patients\u0026rsquo; characteristics in the training and testing cohort.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eTraining cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eTesting cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePNI (+)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;82)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePNI (-)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;69)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePNI (+)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;36)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePNI (-)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;30)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (y)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e60.1\u0026thinsp;\u0026plusmn;\u0026thinsp;11.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60.4\u0026thinsp;\u0026plusmn;\u0026thinsp;9.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.451\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e59.3\u0026thinsp;\u0026plusmn;\u0026thinsp;11.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e57.9\u0026thinsp;\u0026plusmn;\u0026thinsp;10.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.578\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.225\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.451\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\u003e48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\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\u003e34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInterval between CT and surgery (d)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (3,8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (3, 9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.401\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (2,7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4 (2, 6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.869\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLocation\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.794\u003c/p\u003e \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\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeck\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody or tail\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVascular\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.288\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMargin\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.641\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMPD\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.203\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCBD\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.284\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.575\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCTLN\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.000*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.034*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCTPNI\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.006*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.004*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.50 (1.90, 4.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.24 (1.84, 3.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.155\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.55 (1.94, 3.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.42 (2, 3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.981\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCEA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.60 (2.21, 5.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.32 (2.27, 5.51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.977\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.12 (2.21, 7.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.02 (1.72, 6.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.747\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCA19-9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e270 (65.89, 795.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e195. 45 (26.23, 577. 63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.375\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e308.25 (75.0, 699.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e196.25 (61.15, 387.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.245\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLN\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.037*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.389\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade\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=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.049*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.413\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWell-differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerately differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoorly differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRad-score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.52 (-0.06, 0.99)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.13 (-0.85, -0.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.000*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.46 (-0.14, 0.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-0.08 (-0.58, 0.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.01*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003e* represents a statistically significant difference.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eCTPNI: Radiologists evaluated the status of PNI based on CECT; CTLN: The lymph node status determined on CT; CA19-9: Carbohydrate antigen 19\u0026thinsp;\u0026minus;\u0026thinsp;9; CEA: Carcinoembryonic antigen; CBD: Common bile duct; LN: Lymph node; MPD: Main pancreatic duct; PNI: Perineural invasion; Rad-score: Radiomics score.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThere was no significant difference between the training and validation cohorts in terms of the percentage of PNI-positive patients. There was no significant difference between the PNI-positive and PNI-negative groups in age, sex, CEA, CA19-9, tumor size or location, CBD or MPD dilation status, or vascular invasion or margin status in either the training or testing cohort. The ICC of the tumor size was 0.936, indicating good consistency. The kappa values of the CTPNI and CTLN were 0.723 and 0.741, respectively, with moderate consistency.\u003c/p\u003e \u003cp\u003eAccording to the univariate analysis of the training cohort, the CTPNI, CTLN, LN and tumor grade were significantly associated with PNI (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). According to multivariate analysis, CT features, including the CTPNI (OR\u0026thinsp;=\u0026thinsp;0.315 [95% CI: 0.131, 0.761], P\u0026thinsp;=\u0026thinsp;0.01) and CTLN (OR\u0026thinsp;=\u0026thinsp;0.169 [95% CI: 0.059, 0.479], P\u0026thinsp;=\u0026thinsp;0.001), were significantly associated with PNI.\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\u003eUnivariate and multivariate logistic regression analysis for PNI of PDAC.\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\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eUnivariate Analysis\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eMultivariate Analysis\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOdds Ratio (95% CI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOdds Ratio (95% CI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCTPNI\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 \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePresent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \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 \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbsent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.367 (0.181, 0.746)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.315 (0.131, 0.761)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCTLN\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 \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \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 \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.176 (0.075, 0.415)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.169 (0.059, 0.479)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLN\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 \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePositive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \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 \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNegative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.491 (0.252, 0.957)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.037\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.456 (0.196, 1.065)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade\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 \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoorly differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \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 \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerately differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.143 (0.013, 1.516)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.106\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.435 (0.034, 5.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.521\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWell-differentiated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.421 (0.194, 0.913)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.434 (0.169, 1.115)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.083\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRad-score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.718 (1.172, 4.316)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.666 (2.069, 6.494)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eCI: Confidence interval; CTPNI: Radiologists evaluated the status of PNI based on CECT; CTLN: The lymph node status determined on CT; LN: Lymph node; Rad-score: Radiomics score.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eRadiomics feature selection and model construction\u003c/h2\u003e \u003cp\u003eThe mean ICCs for intraobserver agreement and interobserver agreement were 0.889 (range from 0.103 to 0.995) and 0.843 (range from 0.002\u0026ndash;0.993), respectively (Supplementary Fig.\u0026nbsp;1). Ninety radiomics features were excluded because of intraobserver agreement, and 141 radiomics features were excluded because of interobserver agreement. For features with suboptimal agreement, 68 intraobserver-excluded features were included among 141 interobserver-excluded features. Ultimately, 163 radiomics features were excluded due to inferior reproducibility, and the remaining 645 radiomics features were used for the next analysis.\u003c/p\u003e \u003cp\u003eAfter univariate analysis, 91 radiomics features were significantly different between the PNI-positive and PNI-negative groups in the training cohort. The LASSO regression method with 10-fold cross-validation was applied for the remaining features selected, and 8 optimized features were chosen for constructing the model. The selected radiomics features were quantitatively integrated into the Rad-score in the training and testing cohorts. According to the univariate and multivariate analyses, the Rad-score (OR\u0026thinsp;=\u0026thinsp;3.666 [95% CI: 2.069, 6.494], P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) was significantly related to PNI.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eNomogram construction\u003c/h2\u003e \u003cp\u003eThe Rad-score was independently associated with PNI. The AUC for the Rad-score in the training cohort (0.720, 95% confidence interval [CI]: [0.639, 0.802]) was close to that in the testing cohort (0.640, 95% CI: 0.499, 0.781). The AUC of the CTPNI was 0.610 (95% CI: 0.536, 0.684) in the training cohort and 0.675 (95% CI: 0.582, 0.768) in the testing cohort. Using the Rad-score combined with the CTLN and CTPNI, a nomogram (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) for predicting PNI was constructed, and it achieved favorable performance both in the training cohort, with an AUC of 0.846 (95% CI: 0.785, 0.907), and in the testing cohort, with an AUC of 0.778 (95% CI: 0.666, 0.889). Comparing the AUCs of the nomogram model with those of the Rad-score and CTPNI through the DeLong test, the nomogram model achieved the best performance (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The diagnostic performance of the nomogram model, Rad-score and CTPNI in both the training and testing cohorts is summarized in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The calibration plot indicated that the predicted PNI was consistent with the actual PNI probability (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). The decision curve suggested that the nomogram model outperformed CTPNI at any threshold probability (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e shows examples of the clinical application of the nomogram.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe performance of the training and testing cohorts.\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAUC (95% CI)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eTraining cohort\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCTPNI model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.596\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.451\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.768\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.610 (0.536, 0.684)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.636\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.720\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.536\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.720 (0.639, 0.802)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003enomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.781\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.890\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.768\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.846 (0.785, 0.907)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eTesting cohort\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCTPNI model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.417\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.882\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.675 (0.582, 0.768)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.636\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.533\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.640 (0.499, 0.781)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003enomogram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.806\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.733\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.778 (0.666, 0.889)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eAUC: Area under the receiver operating characteristic curve; CI: Confidence interval; Rad-clinical: The combined of CTLN, CTPNI and radiomics model.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003ePreoperative accurate evaluation of PNI in patients with PDAC affects the choice of appropriate treatment. Our retrospective study constructed a nomogram to preoperatively predict PNI based on the rad-score from the arterial and portal venous phases of CECT, CTLN and CTPNI. The nomogram could effectively predict the occurrence of PNI in both the training and testing groups. This study demonstrated that the nomogram was superior to the Rad-score and CTPNI for evaluating the occurrence of PNI.\u003c/p\u003e \u003cp\u003eThe prevalence of PNI in PDAC patients is fairly high, varying from 42.3%-100% in previous reports[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], with an incidence of 54.4% in this study. Additionally, we found an interesting association between PNI status and LN status, and the PNI-positive group was more prone to LN metastasis. Previous studies have indicated that cancer cells grow along nerves in contact with LNs, indicating a complex link between LN metastasis and PNI[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. There was also a statistically significant difference in the CTLNs. In addition, patients with PNI were more likely to have poor pathological differentiation. Poorly differentiated PDAC has more aggressive behavior, which is related to poor prognosis, PNI and poor differentiation and reflects the malignant biological behavior of PDAC. We did not find significant differences in tumor size; dilation status of the CBD or MPD; CEA or CA19-9 levels; or resection margin status between the PNI-positive and PNI-negative groups. The relationship between PNI and tumor size is controversial. Crippa et al. reported that the incidence of PNI increased with tumor size[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. However, Patel et al. suggested that no evidence was found between tumor size and PNI[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. PNI occurs early in PDAC, and even tumors less than 2 cm may develop PNI[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. This may be related to the greater probability of PNI due to tumor growth beyond the pancreas, while tumors within the pancreas generally do not develop PNI even if the tumor is large, based on the anatomical structure[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. These controversial results suggest that the relationship between tumor size and PNI needs further investigation. CBD and MPD dilatation caused by obstruction may not be associated with PNI. CEA and CA19-9 are nonspecific in PDAC and may be abnormal in a variety of diseases, which may account for the lack of differences in CEA and CA19-9 between the PNI-positive and PNI-negative groups.\u003c/p\u003e \u003cp\u003eEight radiomics features related to PNI were selected for this study. Kulkarni et al.[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] extracted CT texture features to analyze their association with PNI and did not find any texture features related to PNI. The possible reason is that the texture features were only extracted from the maximum level of the tumor, which included incomplete features. In addition, in this study, poorly vascularized tumors located in the head of the pancreas were selected. In our study, the radiomics features of the whole lesion were extracted at the three-dimensional level, which could help to discover more biological characteristics of tumors. These selected features were integrated into a Rad-score and exhibited moderate performance in preoperatively predicting the PNI status of PDAC in both the training and testing cohorts. Radiomics can improve the prediction performance of medical images by improving analysis and using computer algorithms to extract thousands of quantitative features, and it can mine a large amount of information that is invisible to the naked eye. According to the radiologists\u0026rsquo; evaluation, the CTPNI achieved inferior performance, which may be attributed to perivascular inflammation or fibrosis easily mimicking PNI on CT. In addition, PDAC is characterized by lymphatic growth and PNI, which are easily confused with microvessels, LN or fibrosis on CT. This may have caused the unsatisfactory agreement between the two reviewers in assessing the CTPNI and CTLN in our study.\u003c/p\u003e \u003cp\u003eA nomogram combining the Rad-score, CTLN and CTPNI achieved the best performance (the AUC in the testing cohort was 0.778) for the preoperative assessment of PNI in patients with PDAC. Several possible reasons may contribute to the good performance of the nomogram. One is that the Rad-score combined with arterial and portal venous phases can provide valuable information. In addition, different CT scan parameters may lead to unsatisfactory reproducibility of radiomics features, which can be maximally alleviated by using resampling as a preprocessing method, which optimizes gray dispersion to maintain the stability of features[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Z score standardization eliminates the effect of different data dimensions. Moreover, the good performance of the nomogram was attributed to feature selection and modeling. Univariate analysis and LASSO regression confirmed that important features were selected for modeling. Tenfold validation was applied to guarantee the robustness of the model. Finally, the nomogram integrates selected radiomics features, the presence of PDAC on CT images and the experience of radiologists and combines the performance of different dimensions to better reflect the characteristics of PDAC. This result suggested that the nomogram has the potential to preoperatively predict PNI status in PDAC patients. The calibration curve revealed that the predicted PNI was in good agreement with the actual PNI probability. The decision curve indicated that the nomogram outperformed radiologists at any threshold probability.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, this was a single-center study with a limited sample size and no external validation group. However, the number included in our study was relatively larger than that in previous studies. Therefore, the retrospective nature of the study may have led to biased results. Large sample, multicenter and prospective studies should be conducted to further verify the results. Moreover, although several studies have reported that PNI is associated with PDAC patient prognosis, the relationship between PNI and prognosis was not clarified in this study. Subsequently, the corresponding patients should be followed up on the basis of this study to explore the ability of the nomogram to predict survival. Ultimately, there was not a consistent one-to-one match between CT evaluation and pathology.\u003c/p\u003e \u003cp\u003eIn conclusion, a nomogram based on the rad-score derived from both the arterial and portal venous phases of CECT combined with the CTLN and CTPNI may serve as a valuable noninvasive tool for the preoperative assessment of PNI in patients with PDAC. This approach offers a practical means to classify PDAC patients before surgery and enhance patient management.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003earea under the receiver operating characteristic curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCECT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003econtrast-enhanced computed tomography\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCTPNI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradiologists evaluated the status of PNI based on CECT\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCTLN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ethe lymph node status determined on CT\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCA19-9\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecarbohydrate antigen 19\u0026thinsp;\u0026minus;\u0026thinsp;9\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCEA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecarcinoembryonic antigen\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCBD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecommon bile duct\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGLCM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egray-level cooccurrences matrix\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGLRLM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egray-level run-length matrix\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIH\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eintensity histogram\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eID\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eintensity direct\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eICC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eintraclass correlation coefficient\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLASSO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eleast absolute shrinkage and selection operator\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003elymph node\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMPD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emain pancreatic duct\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePNI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eperineural invasion\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePDAC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epancreatic ductal adenocarcinoma\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRad-score\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradiomics score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ereceived operating characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contribution\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eY.D. collected the data and drafted the main manuscript, HP.Y. made the statistical analysis and part writing. Their contributions to this study are consistent, so they are listed as co- first authors. XP.D. collected and interpreted the data. L.L. collected the data. ZX.H. revised the manuscript and\u0026nbsp;B.S.\u0026nbsp;designed the manuscript. The contributions of ZX.H. and B.S. to this study are consistent, so they are listed as co-corresponding authors. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors state that this study was supported by the 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University (Grant number ZYGD22004, ZYJC21012); the development project of Hainan provincial clinical medical center and post-doctoral station development project of Sana (Grant number 23CZ009).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was not required for this study because this is retrospective\u003c/p\u003e\n\u003cp\u003estudy, this study was approved by the institutional ethics committee of West China Hospital, Sichuan, China (IRB number: 2023-0003) and performed in accordance with Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare that there is no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSiegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. Ca Cancer J Clin. 2023;73(1):17-48. https://doi.org/10.3322/caac.21763.\u003c/li\u003e\n\u003cli\u003eSuzuki S, Shimoda M, Shimazaki J, Maruyama T, Oshiro Y, Nishida K, et al. Predictive Early Recurrence Factors of Preoperative Clinicophysiological Findings in Pancreatic Cancer. Eur Surg Res. 2018;59(5-6):329-38. https://doi.org/10.1159/000494382.\u003c/li\u003e\n\u003cli\u003eSchorn S, Demir IE, Haller B, Scheufele F, Reyes CM, Tieftrunk E, et al. The influence of neural invasion on survival and tumor recurrence in pancreatic ductal adenocarcinoma - A systematic review and meta-analysis. Surg Oncol. 2017;26(1):105-15. https://doi.org/10.1016/j.suronc.2017.01.007.\u003c/li\u003e\n\u003cli\u003eJurcak NR, Rucki AA, Muth S, Thompson E, Sharma R, Ding D, et al. Axon Guidance Molecules Promote Perineural Invasion and Metastasis of Orthotopic Pancreatic Tumors in Mice. Gastroenterology. 2019;157(3):838-50. https://doi.org/10.1053/j.gastro.2019.05.065.\u003c/li\u003e\n\u003cli\u003eHuang C, Li Y, Guo Y, Zhang Z, Lian G, Chen Y, et al. MMP1/PAR1/SP/NK1R paracrine loop modulates early perineural invasion of pancreatic cancer cells. Theranostics. 2018;8(11):3074-86. https://doi.org/10.7150/thno.24281.\u003c/li\u003e\n\u003cli\u003eCeyhan GO, Bergmann F, Kadihasanoglu M, Altintas B, Demir IE, Hinz U, et al. Pancreatic neuropathy and neuropathic pain--a comprehensive pathomorphological study of 546 cases. Gastroenterology. 2009;136(1):177-86. https://doi.org/10.1053/j.gastro.2008.09.029.\u003c/li\u003e\n\u003cli\u003eChatterjee D, Katz MH, Rashid A, Wang H, Iuga AC, Varadhachary GR, et al. Perineural and intraneural invasion in posttherapy pancreaticoduodenectomy specimens predicts poor prognosis in patients with pancreatic ductal adenocarcinoma. Am J Surg Pathol. 2012;36(3):409-17. https://doi.org/10.1097/PAS.0b013e31824104c5.\u003c/li\u003e\n\u003cli\u003eFelsenstein M, Lindhammer F, Feist M, Hillebrandt KH, Timmermann L, Benzing C, et al. Perineural Invasion in Pancreatic Ductal Adenocarcinoma (PDAC): A Saboteur of Curative Intended Therapies? J Clin Med. 2022;11(9). https://doi.org/10.3390/jcm11092367.\u003c/li\u003e\n\u003cli\u003eSchneider M, Strobel O, Hackert T, B\u0026uuml;chler MW. Pancreatic resection for cancer-the Heidelberg technique. Langenbecks Arch Surg. 2019;404(8):1017-22. https://doi.org/10.1007/s00423-019-01839-1.\u003c/li\u003e\n\u003cli\u003eMizrahi JD, Surana R, Valle JW, Shroff RT. Pancreatic cancer. Lancet. 2020;395(10242):2008-20. https://doi.org/10.1016/S0140-6736(20)30974-0.\u003c/li\u003e\n\u003cli\u003eZaky AM, Wolfgang CL, Weiss MJ, Javed AA, Fishman EK, Zaheer A. Tumor-Vessel Relationships in Pancreatic Ductal Adenocarcinoma at Multidetector CT: Different Classification Systems and Their Influence on Treatment Planning. Radiographics. 2017;37(1):93-112. https://doi.org/10.1148/rg.2017160054.\u003c/li\u003e\n\u003cli\u003ePatel BN, Olcott E, Jeffrey RB. Extrapancreatic perineural invasion in pancreatic adenocarcinoma. Abdom Radiol (Ny). 2018;43(2):323-31. https://doi.org/10.1007/s00261-017-1343-9.\u003c/li\u003e\n\u003cli\u003eDeshmukh SD, Willmann JK, Jeffrey RB. Pathways of extrapancreatic perineural invasion by pancreatic adenocarcinoma: evaluation with 3D volume-rendered MDCT imaging. Ajr Am J Roentgenol. 2010;194(3):668-74. https://doi.org/10.2214/AJR.09.3285.\u003c/li\u003e\n\u003cli\u003eChang ST, Jeffrey RB, Patel BN, DiMaio MA, Rosenberg J, Willmann JK, et al. Preoperative Multidetector CT Diagnosis of Extrapancreatic Perineural or Duodenal Invasion Is Associated with Reduced Postoperative Survival after Pancreaticoduodenectomy for Pancreatic Adenocarcinoma: Preliminary Experience and Implications for Patient Care. Radiology. 2016;281(3):816-25. https://doi.org/10.1148/radiol.2016152790.\u003c/li\u003e\n\u003cli\u003eGuo X, Gao S, Yu J, Zhou Y, Gao C, Hao J. The imaging features of extrapancreatic perineural invasion (EPNI) in pancreatic Cancer：A comparative retrospective study. Pancreatology. 2021;21(8):1516-23. https://doi.org/10.1016/j.pan.2021.08.010.\u003c/li\u003e\n\u003cli\u003eGillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563-77. https://doi.org/10.1148/radiol.2015151169.\u003c/li\u003e\n\u003cli\u003eLambin P, Leijenaar R, Deist TM, Peerlings J, de Jong E, van Timmeren J, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62. https://doi.org/10.1038/nrclinonc.2017.141.\u003c/li\u003e\n\u003cli\u003eTomaszewski MR, Gillies RJ. The Biological Meaning of Radiomic Features. Radiology. 2021;299(2):E256. https://doi.org/10.1148/radiol.2021219005.\u003c/li\u003e\n\u003cli\u003eLi K, Yao Q, Xiao J, Li M, Yang J, Hou W, et al. Contrast-enhanced CT radiomics for predicting lymph node metastasis in pancreatic ductal adenocarcinoma: a pilot study. Cancer Imaging. 2020;20(1):12. https://doi.org/10.1186/s40644-020-0288-3.\u003c/li\u003e\n\u003cli\u003eYao J, Cao K, Hou Y, Zhou J, Xia Y, Nogues I, et al. Deep Learning for Fully Automated Prediction of Overall Survival in Patients Undergoing Resection for Pancreatic Cancer: A Retrospective Multicenter Study. Ann Surg. 2023;278(1):e68-79. https://doi.org/10.1097/SLA.0000000000005465.\u003c/li\u003e\n\u003cli\u003eDu D, Feng H, Lv W, Ashrafinia S, Yuan Q, Wang Q, et al. Machine Learning Methods for Optimal Radiomics-Based Differentiation Between Recurrence and Inflammation: Application to Nasopharyngeal Carcinoma Post-therapy PET/CT Images. Mol Imaging Biol. 2020;22(3):730-8. https://doi.org/10.1007/s11307-019-01411-9.\u003c/li\u003e\n\u003cli\u003eZheng H, Zheng Q, Jiang M, Han C, Yi J, Ai Y, et al. Contrast-enhanced CT based radiomics in the preoperative prediction of perineural invasion for patients with gastric cancer. Eur J Radiol. 2022;154:110393. https://doi.org/10.1016/j.ejrad.2022.110393.\u003c/li\u003e\n\u003cli\u003eChen J, Chen Y, Zheng D, Pang P, Zhang H, Zheng X, et al. Pretreatment MR-based radiomics nomogram as potential imaging biomarker for individualized assessment of perineural invasion status in rectal cancer. Abdom Radiol (Ny). 2021;46(3):847-57. https://doi.org/10.1007/s00261-020-02710-4.\u003c/li\u003e\n\u003cli\u003eLiebig C, Ayala G, Wilks JA, Berger DH, Albo D. Perineural invasion in cancer: a review of the literature. Cancer-Am Cancer Soc. 2009;115(15):3379-91. https://doi.org/10.1002/cncr.24396.\u003c/li\u003e\n\u003cli\u003eMochizuki K, Gabata T, Kozaka K, Hattori Y, Zen Y, Kitagawa H, et al. MDCT findings of extrapancreatic nerve plexus invasion by pancreas head carcinoma: correlation with en bloc pathological specimens and diagnostic accuracy. Eur Radiol. 2010;20(7):1757-67. https://doi.org/10.1007/s00330-010-1727-5.\u003c/li\u003e\n\u003cli\u003eBian Y, Zheng Z, Fang X, Jiang H, Zhu M, Yu J, et al. Artificial Intelligence to Predict Lymph Node Metastasis at CT in Pancreatic Ductal Adenocarcinoma. Radiology. 2023;306(1):160-9. https://doi.org/10.1148/radiol.220329.\u003c/li\u003e\n\u003cli\u003eShafiq-Ul-Hassan M, Zhang GG, Latifi K, Ullah G, Hunt DC, Balagurunathan Y, et al. Intrinsic dependencies of CT radiomic features on voxel size and number of gray levels. Med Phys. 2017;44(3):1050-62. https://doi.org/10.1002/mp.12123.\u003c/li\u003e\n\u003cli\u003eGasparini G, Pellegatta M, Crippa S, Lena MS, Belfiori G, Doglioni C, et al. Nerves and Pancreatic Cancer: New Insights into a Dangerous Relationship. Cancers (Basel). 2019;11(7). https://doi.org/10.3390/cancers11070893.\u003c/li\u003e\n\u003cli\u003eKayahara M, Nakagawara H, Kitagawa H, Ohta T. The nature of neural invasion by pancreatic cancer. Pancreas. 2007;35(3):218-23. https://doi.org/10.1097/mpa.0b013e3180619677.\u003c/li\u003e\n\u003cli\u003eCrippa S, Pergolini I, Javed AA, Honselmann KC, Weiss MJ, Di Salvo F, et al. Implications of Perineural Invasion on Disease Recurrence and Survival After Pancreatectomy for Pancreatic Head Ductal Adenocarcinoma. Ann Surg. 2022;276(2):378-85. https://doi.org/10.1097/SLA.0000000000004464.\u003c/li\u003e\n\u003cli\u003ePatel BN, Giacomini C, Jeffrey RB, Willmann JK, Olcott E. Three-dimensional volume-rendered multidetector CT imaging of the posterior inferior pancreaticoduodenal artery: its anatomy and role in diagnosing extrapancreatic perineural invasion. Cancer Imaging. 2013;13(4):580-90. https://doi.org/10.1102/1470-7330.2013.0051.\u003c/li\u003e\n\u003cli\u003eMarchegiani G, Andrianello S, Malleo G, De Gregorio L, Scarpa A, Mino-Kenudson M, et al. Does Size Matter in Pancreatic Cancer?: Reappraisal of Tumour Dimension as a Predictor of Outcome Beyond the TNM. Ann Surg. 2017;266(1):142-8. https://doi.org/10.1097/SLA.0000000000001837.\u003c/li\u003e\n\u003cli\u003eTu W, Gottumukkala RV, Schieda N, Lavall\u0026eacute;e L, Adam BA, Silverman SG. Perineural Invasion and Spread in Common Abdominopelvic Diseases: Imaging Diagnosis and Clinical Significance. Radiographics. 2023;43(7):e220148. https://doi.org/10.1148/rg.220148.\u003c/li\u003e\n\u003cli\u003eKulkarni A, Carrion-Martinez I, Jiang NN, Puttagunta S, Ruo L, Meyers BM, et al. Hypovascular pancreas head adenocarcinoma: CT texture analysis for assessment of resection margin status and high-risk features. Eur Radiol. 2020;30(5):2853-60. https://doi.org/10.1007/s00330-019-06583-0.\u003c/li\u003e\n\u003cli\u003eLarue R, van Timmeren JE, de Jong E, Feliciani G, Leijenaar R, Schreurs W, et al. Influence of gray level discretization on radiomic feature stability for different CT scanners, tube currents and slice thicknesses: a comprehensive phantom study. Acta Oncol. 2017;56(11):1544-53. https://doi.org/10.1080/0284186X.2017.1351624.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"pancreatic ductal adenocarcinoma, perineural invasion, computed tomography, radiomics, nomogram","lastPublishedDoi":"10.21203/rs.3.rs-4161245/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4161245/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003ePreoperative evaluation perineural invasion (PNI) affects the treatment and prognosis of patients with pancreatic ductal adenocarcinoma (PDAC). This study aims to develop a nomogram based on a CT radiomics nomogram for the preoperative prediction of PNI in PDAC patients.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA total of 217 patients with histologically confirmed PDAC were enrolled in this retrospective study. Radiomics features were extracted from the whole tumor. Univariate analysis and least absolute shrinkage and selection operator logistic regression were applied for feature selection and radiomics model construction. Finally, a nomogram combining the radiomics score (Rad-score) and clinical characteristics was established. Receiver operating characteristic curve analysis, calibration curve analysis and decision curve analysis (DCA) were used to evaluate the predictive performance of the radiomics nomogram.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eAccording to multivariate analysis, CT features, including the evaluation of radiologists regarding PNI status based on CECT (CTPNI) (OR\u0026thinsp;=\u0026thinsp;0.315 [95% CI: 0.131, 0.761], P\u0026thinsp;=\u0026thinsp;0.01), the lymph node status determined on CECT (CTLN) (OR\u0026thinsp;=\u0026thinsp;0.169 [95% CI: 0.059, 0.479], P\u0026thinsp;=\u0026thinsp;0.001) and the Rad-score (OR\u0026thinsp;=\u0026thinsp;3.666 [95% CI: 2.069, 6.494], P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), were significantly associated with PNI. The area under the receiver operating characteristic curve (AUC) for the nomogram combined with the Rad-score, CTLN and CTPNI achieved favorable discrimination of PNI status, with AUCs of 0.846 and 0.778 in the training and testing cohorts, respectively, which were superior to those of the Rad-score (AUC of 0.720 in the training cohort and 0.640 in the testing cohort) and CTPNI (AUC of 0.610 in the training cohort and 0.675 in the testing cohort). The calibration plot and decision curve showed good results.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe CT-based radiomics nomogram has the potential to accurately predict PNI in patients with PDAC.\u003c/p\u003e","manuscriptTitle":"A CT-based radiomics nomogram for the preoperative prediction of perineural invasion in pancreatic ductal adenocarcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-29 18:59:04","doi":"10.21203/rs.3.rs-4161245/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":"d738724c-e455-4377-8f8c-231bdfac279a","owner":[],"postedDate":"March 29th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-10-23T07:53:50+00:00","versionOfRecord":[],"versionCreatedAt":"2024-03-29 18:59:04","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4161245","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4161245","identity":"rs-4161245","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}