Deep learning-based prediction of post-pancreaticoduodenectomy pancreatic fistula.

Postoperative pancreatic fistula is a life-threatening complication with an unmet need for accurate prediction. This study was aimed to develop preoperative artificial intelligence-based prediction models. Patients who underwent pancreaticoduodenectomy were enrolled and stratified into model development and validation sets by surgery between 2016 and 2017 or in 2018, respectively. Machine learning models based on clinical and body composition data, and deep learning models based on computed tomographic data, were developed, combined by ensemble voting, and final models were selected comparison with earlier model. Among the 1333 participants (training, n = 881; test, n = 452), postoperative pancreatic fistula occurred in 421 (47.8%) and 134 (31.8%) and clinically relevant postoperative pancreatic fistula occurred in 59 (6.7%) and 27 (6.0%) participants in the training and test datasets, respectively. In the test dataset, the area under the receiver operating curve [AUC (95% confidence interval)] of the selected preoperative model for predicting all and clinically relevant postoperative pancreatic fistula was 0.75 (0.71-0.80) and 0.68 (0.58-0.78). The ensemble model showed better predictive performance than the individual ML and DL models.

Prognostic value of CT-based radiomics in grade 1-2 pancreatic neuroendocrine tumors.

BACKGROUND: Surgically resected grade 1-2 (G1-2) pancreatic neuroendocrine tumors (PanNETs) exhibit diverse clinical outcomes, highlighting the need for reliable prognostic biomarkers. Our study aimed to develop and validate CT-based radiomics model for predicting postsurgical outcome in patients with G1-2 PanNETs, and to compare its performance with the current clinical staging system. METHODS: This multicenter retrospective study included patients who underwent dynamic CT and subsequent curative resection for G1-2 PanNETs. A radiomics-based model (R-score) for predicting recurrence-free survival (RFS) was developed from a development set (441 patients from one institution) using least absolute shrinkage and selection operator-Cox regression analysis. A clinical model (C-model) consisting of age and tumor stage according to the 8th American Joint Committee on Cancer staging system was built, and an integrative model combining the C-model and the R-score (CR-model) was developed using multivariable Cox regression analysis. Using an external test set (159 patients from another institution), the models' performance for predicting RFS and overall survival (OS) was evaluated using Harrell's C-index. The incremental value of adding the R-score to the C-model was evaluated using net reclassification improvement (NRI) and integrated discrimination improvement (IDI). RESULTS: The median follow-up periods were 68.3 and 59.7 months in the development and test sets, respectively. In the development set, 58 patients (13.2%) experienced recurrence and 35 (7.9%) died. In the test set, tumors recurred in 14 patients (8.8%) and 12 (7.5%) died. In the test set, the R-score had a C-index of 0.716 for RFS and 0.674 for OS. Compared with the C-model, the CR-model showed higher C-index (RFS, 0.734 vs. 0.662, p = 0.012; OS, 0.781 vs. 0.675, p = 0.043). CR-model also showed improved classification (NRI, 0.330, p < 0.001) and discrimination (IDI, 0.071, p < 0.001) for prediction of 3-year RFS. CONCLUSIONS: Our CR-model outperformed the current clinical staging system in prediction of the prognosis for G1-2 PanNETs and added incremental value for predicting postoperative recurrence. The CR-model enables precise identification of high-risk patients, guiding personalized treatment planning to improve outcomes in surgically resected grade 1-2 PanNETs.

Large Duct Pancreatic Ductal Adenocarcinoma: A Morphological Variant of Pancreatic Ductal Adenocarcinoma With Distinct CT and MRI Characteristics.

OBJECTIVE: To investigate the imaging characteristics of large duct pancreatic ductal adenocarcinoma (LD-PDAC) on computed tomography (CT) and magnetic resonance imaging (MRI). MATERIALS AND METHODS: Thirty-five patients with LD-PDAC (63.2 ± 9.7 years) were retrospectively evaluated. Tumor morphology on CT and MRI (predominantly solid mass vs. solid mass with prominent cysts vs. predominantly cystic mass) was evaluated. Additionally, the visibility, quantity, shape (oval vs. branching vs. irregular), and MRI signal intensity of neoplastic cysts within the LD-PDAC were investigated. The radiological diagnoses rendered for LD-PDAC in radiology reports were reviewed. RESULTS: LD-PDAC was more commonly observed as a solid mass with prominent cysts (45.7% [16/35] on CT and 37.1% [13/35] on MRI) or a predominantly cystic mass (20.0% [7/35] on CT and 40.0% [14/35] on MRI) and less commonly as a predominantly solid mass on CT (34.3% [12/35]) and MRI (22.9% [8/35]). The tumor morphology on imaging was significantly associated with the size of the cancer gland on histopathological examination (P = 0.020 [CT] and 0.013 [MRI]). Neoplastic cysts were visible in 88.6% (31/35) and 91.4% (32/35) of the LD-PDAC cases on CT and MRI, respectively. The cysts appeared as branching (51.6% [16/35] on CT and 59.4% [19/35] on MRI) or oval shapes (45.2% [14/35] on CT and 31.2% [10/35] on MRI) with fluid-like MRI signal intensity. In the radiology reports, 10 LD-PDAC cases (28.6%) were misinterpreted as diseases other than typical PDAC, particularly intraductal papillary mucinous neoplasms. CONCLUSION: LD-PDAC frequently appears as a solid mass with prominent cysts or as a predominantly cystic mass on CT and MRI. Radiologists should be familiar with the imaging features of LD-PDAC to avoid misdiagnosis.

Prognostic value of tumor-to-parenchymal contrast enhancement ratio on portal venous-phase CT in pancreatic neuroendocrine neoplasms.

OBJECTIVES: We aimed to evaluate the prognostic value of tumor-to-parenchymal contrast enhancement ratio on portal venous-phase CT (CER on PVP) and compare its prognostic performance to prevailing grading and staging systems in pancreatic neuroendocrine neoplasms (PanNENs). METHODS: In this retrospective study, data on 465 patients (development cohort) who underwent upfront curative-intent resection for PanNEN were used to assess the performance of CER on PVP and tumor size measured by CT (CT-Size) in predicting recurrence-free survival (RFS) using Harrell's C-index and to determine their optimal cutoffs to stratify RFS using a multi-way partitioning algorithm. External data on 184 patients (test cohort) were used to validate the performance of CER on PVP in predicting RFS and overall survival (OS) and compare its predictive performance with those of CT-Size, 2019 World Health Organization classification system (WHO), and the 8th American Joint Committee on Cancer staging system (AJCC). RESULTS: In the test cohort, CER on PVP showed C-indexes of 0.83 (95% confidence interval [CI], 0.74-0.91) and 0.84 (95% CI, 0.73-0.95) for predicting RFS and OS, respectively, which were higher than those for the WHO (C-index: 0.73 for RFS [p = .002] and 0.72 for OS [p = .004]) and AJCC (C-index, 0.67 for RFS [p = .002] and 0.58 for OS [p = .002]). CT-Size obtained C-indexes of 0.71 for RFS and 0.61 for OS. CONCLUSIONS: CER on PVP showed superior predictive performance on postoperative survival in PanNEN than current grading and staging systems, indicating its potential as a noninvasive preoperative prognostic tool. KEY POINTS: • In pancreatic neuroendocrine neoplasms, the tumor-to-parenchymal enhancement ratio on portal venous-phase CT (CER on PVP) showed acceptable predictive performance of postoperative outcomes. • CER on PVP showed superior predictive performance of postoperative survival over the current WHO classification and AJCC staging system.

Deep Learning-based Detection of Solid and Cystic Pancreatic Neoplasms at Contrast-enhanced CT.

Background Deep learning (DL) may facilitate the diagnosis of various pancreatic lesions at imaging. Purpose To develop and validate a DL-based approach for automatic identification of patients with various solid and cystic pancreatic neoplasms at abdominal CT and compare its diagnostic performance with that of radiologists. Materials and Methods In this retrospective study, a three-dimensional nnU-Net-based DL model was trained using the CT data of patients who underwent resection for pancreatic lesions between January 2014 and March 2015 and a subset of patients without pancreatic abnormality who underwent CT in 2014. Performance of the DL-based approach to identify patients with pancreatic lesions was evaluated in a temporally independent cohort (test set 1) and a temporally and spatially independent cohort (test set 2) and was compared with that of two board-certified radiologists. Performance was assessed using receiver operating characteristic analysis. Results The study included 852 patients in the training set (median age, 60 years [range, 19-85 years]; 462 men), 603 patients in test set 1 (median age, 58 years [range, 18-82 years]; 376 men), and 589 patients in test set 2 (median age, 63 years [range, 18-99 years]; 343 men). In test set 1, the DL-based approach had an area under the receiver operating characteristic curve (AUC) of 0.91 (95% CI: 0.89, 0.94) and showed slightly worse performance in test set 2 (AUC, 0.87 [95% CI: 0.84, 0.89]). The DL-based approach showed high sensitivity in identifying patients with solid lesions of any size (98%-100%) or cystic lesions measuring 1.0 cm or larger (92%-93%), which was comparable with the radiologists (95%-100% for solid lesions [P = .51 to P > .99]; 93%-98% for cystic lesions ≥1.0 cm [P = .38 to P > .99]). Conclusion The deep learning-based approach demonstrated high performance in identifying patients with various solid and cystic pancreatic lesions at CT. © RSNA, 2022 Online supplemental material is available for this article.

Preoperative data-based deep learning model for predicting postoperative survival in pancreatic cancer patients.

BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) has a poor prognosis even after curative resection. A deep learning-based stratification of postoperative survival in the preoperative setting may aid the treatment decisions for improving prognosis. This study was aimed to develop a deep learning model based on preoperative data for predicting postoperative survival. METHODS: The patients who underwent surgery for PDAC between January 2014 and May 2015. Clinical data-based machine learning models and computed tomography (CT) data-based deep learning models were developed separately, and ensemble learning was utilized to combine two models. The primary outcomes were the prediction of 2-year overall survival (OS) and 1-year recurrence-free survival (RFS). The model's performance was measured by area under the receiver operating curve (AUC) and was compared with that of American Joint Committee on Cancer (AJCC) 8th stage. RESULTS: The median OS and RFS were 23 and 10 months in training dataset (n = 229), and 22 and 11 months in test dataset (n = 53), respectively. The AUC of the ensemble model for predicting 2-year OS and 1-year RFS in the test dataset was 0.76 and 0.74, respectively. The performance of the ensemble model was comparable to that of the AJCC in predicting 2-year OS (AUC, 0.67; P = 0.35) and superior to the AJCC in predicting 1-year RFS (AUC, 0.54; P = 0.049). CONCLUSION: Our ensemble model based on routine preoperative variables showed good performance for predicting prognosis for PDAC patients after surgery.

Prediction of margin-negative resection of pancreatic ductal adenocarcinoma following neoadjuvant therapy: Diagnostic performance of NCCN criteria for resection vs CT-determined resectability.

BACKGROUND: Accurate assessment of pancreatic ductal adenocarcinoma (PDAC) resectability after neoadjuvant therapy (NAT) is crucial. Recently, the NCCN introduced criteria for resection of PDAC following NAT. METHODS: We analyzed 127 patients who underwent NAT and pancreatectomy for PDAC between January 2010 and March 2020. CT-determined resectability according to the NCCN guideline and CA 19-9 level was evaluated before and after NAT. Diagnostic performance of the NCCN criteria for margin-negative (R0) resection was investigated and compared with CT alone. RESULTS: R0 resection was achieved in 104 (81.9%) patients. After NAT, there were 30 (23.6%) resectable, 90 (70.9%) borderline resectable, and seven (5.5%) locally advanced tumors. Significantly decreased or stable CA 19-9 levels were noted in 114 (89.8%) patients. The sensitivity and specificity of the NCCN criteria were 87.5% (91/104) and 21.7% (5/23), respectively, which were significantly different from CT including only resectable PDAC (26.9% [28/104] and 91.3% [21/23]; P < .001), but less prominently different from CT including resectable and borderline resectable PDAC (95.2% [99/104]; P = .022 and 8.7% [2/23]; P = .375). CONCLUSIONS: The NCCN criteria for resection following NAT showed high sensitivity and low specificity for predicting R0 resection. It had supplementary benefit over CT alone, mainly in preventing underestimation of R0 resection.

MRI features for differentiation of autoimmune pancreatitis from pancreatic ductal adenocarcinoma: A systematic review and meta-analysis.

BACKGROUND AND AIMS: The accurate differential diagnosis between autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDAC) is clinically important. We aimed to determine significant MRI features for differentiating AIP from PDAC, including assessment of diffusion-weighted imaging (DWI). METHODS: We performed a systematic search using three databases. The pooled diagnostic odds ratio was calculated using a bivariate random effects model to determine significant MRI features for differentiating AIP from PDAC. The pooled sensitivity and specificity were calculated. The qualitative systematic review for DWI assessment was performed. RESULTS: Of nine studies (775 patients), multiple main pancreatic duct (MPD) strictures, absence of upstream marked MPD dilatation, peripancreatic rim, and duct penetration sign were significant MRI features for differentiating AIP from PDAC. Absence of MPD dilatation had the highest pooled sensitivity (87%, 95% CI=68-96%), whereas peripancreatic rim had the highest pooled specificity (100%, 95% CI=88-100%). Of 12 studies evaluating DWI, seven reported statistically significant differences in apparent diffusion coefficient (ADC) values between AIP and PDAC; however, four reported lower ADC values in AIP than in PDAC, but three reported the opposite result. CONCLUSION: The four significant MRI features can be useful to differentiate AIP from PDAC, but DWI assessment might be limited.

Combined Hepatocellular-Cholangiocarcinoma: Magnetic Resonance Imaging Features and Prognosis According to Risk Factors for Hepatocellular Carcinoma.

BACKGROUND: Combined hepatocellular-cholangiocarcinoma (cHCC-CCA) can develop in patients with and without risk factors for hepatocellular carcinoma (HCC). PURPOSE: To compare the clinical and magnetic resonance imaging (MRI) characteristics of cHCC-CCA in patients with and without risk factors for HCC, and to assess the influence of risk factors on patient prognosis. STUDY TYPE: Retrospective. POPULATION: A total of 152 patients with surgically confirmed cHCC-CCA. FIELD STRENGTH/SEQUENCE: 1.5-T and 3-T/T1-weighted dual gradient-echo in- and opposed-phase, T2-weighted turbo-spin-echo, diffusion-weighted single-shot spin-echo echo-planar, and T1-weighted three-dimensional gradient-echo contrast-enhanced sequences. ASSESSMENT: MRI features according to the Liver Imaging Reporting and Data System (LI-RADS) and pathologic findings based on revised classification were compared between patients with and without risk factors for HCC. Overall survival (OS) and recurrence-free survival (RFS) were also compared between the two groups, and factors associated with survival were evaluated. STATISTICAL TESTS: The clinico-pathologic and MRI features of the two groups were compared using Student's t-tests, Mann-Whitney U-tests, and chi-square tests. OS and RFS were evaluated by the Kaplan-Meier method, and factors associated with survival were evaluated by Cox proportional hazard model. RESULTS: cHCC-CCA in patients with risk factors were more frequently classified as LI-RADS category 4 or 5 (LR-4/5; probably or definitely HCC) (48.7%), whereas those without risk factors were more frequently classified as category M (LR-M; probably malignant, not specific for HCC) (63.6%). RFS and OS did not differ significantly according to risk factors (P = 0.63 and 0.83). Multivariable analysis showed that pathologic tumor type (hazard ratio 2.02; P < 0.05) and LI-RADS category (hazard ratio 2.19; P < 0.05) were significantly associated with RFS and OS, respectively. DATA CONCLUSION: Although MRI features of cHCC-CCA differed significantly between patients with and without risk factors for HCC, postsurgical prognosis did not. LI-RADS category and pathologic tumor type were independently correlated with postsurgical prognosis in patients with cHCC-CCA. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY STAGE: 2.

CT-determined resectability of borderline resectable and unresectable pancreatic adenocarcinoma following FOLFIRINOX therapy.

OBJECTIVES: We aimed to assess the ability of CT-determined resectability, as defined by a recent version of NCCN criteria, and associated CT findings to predict margin-negative (R0) resection in patients with PDAC after neoadjuvant FOLFIRINOX chemotherapy. METHODS: Sixty-four patients (36 men and 28 women; mean age, 58.8 years) with borderline resectable or unresectable PDAC who received neoadjuvant FOLFIRINOX were evaluated retrospectively. CT findings were independently assessed by two abdominal radiologists according to NCCN criteria (version 3. 2019). Tumor resectability was classified as resectable, borderline resectable, or unresectable, and change in resectability was classified as regression, stability, or progression. The associations of R0 resection rate with CT-determined resectability and change in resectability categories were evaluated, as were the sensitivity and specificity of NCCN criteria for R0 resection. Factors associated with R0 resection were identified by logistic regression analysis. RESULTS: R0 resection rate did not differ significantly among the resectable, borderline resectable, or unresectable PDAC (67-73%, p = 0.95) or among PDAC with regression, stability, or progression (56-77%, p = 0.39). The sensitivity and specificity for R0 resection were 67% and 37%, respectively, for resectability (resectable/borderline vs. unresectable) and 80% and 21%, respectively, for changes in resectability (regression/stable vs. progression). Low-contrast enhancement of soft tissue contacting artery (≤ 46.4 HU) was independently associated with R0 resection (p = 0.01). CONCLUSION: CT-determined resectability after neoadjuvant FOLFIRINOX chemotherapy was relatively insensitive and non-specific for predicting R0 resection. Low-contrast enhancement of soft tissue contacting artery may increase the ability of CT to predict R0 resection. KEY POINTS: • Margin-negative resection rate of pancreatic cancer following FOLFIRINOX therapy did not differ among each resectability (67-73%, p = 0.95) based on NCCN criteria or changes in resectability categories (56-77%, p = 0.39). • The sensitivity and specificity for margin-negative resection were 67% and 37% for resectability (resectable/borderline vs. unresectable) and 80% and 21% for changes in resectability (regression/stable vs. progression). • Low-contrast enhancement of soft tissue contacting artery (≤ 46.4 HU) was independently associated with margin-negative resection (p = 0.01).

Radiologic Evaluation and Structured Reporting Form for Extrahepatic Bile Duct Cancer: 2019 Consensus Recommendations from the Korean Society of Abdominal Radiology.

Radiologic imaging is important for evaluating extrahepatic bile duct (EHD) cancers; it is used for staging tumors and evaluating the suitability of surgical resection, as surgery may be contraindicated in some cases regardless of tumor stage. However, the published general recommendations for EHD cancer and recommendations guided by the perspectives of radiologists are limited. The Korean Society of Abdominal Radiology (KSAR) study group for EHD cancer developed key questions and corresponding recommendations for the radiologic evaluation of EHD cancer and organized them into 4 sections: nomenclature and definition, imaging technique, cancer evaluation, and tumor response. A structured reporting form was also developed to allow the progressive accumulation of standardized data, which will facilitate multicenter studies and contribute more evidence for the development of recommendations.

CT in the prediction of margin-negative resection in pancreatic cancer following neoadjuvant treatment: a systematic review and meta-analysis.

OBJECTIVES: We aimed to systematically evaluate the diagnostic accuracy of CT-determined resectability following neoadjuvant treatment for predicting margin-negative resection (R0 resection) in patients with pancreatic ductal adenocarcinoma (PDAC). METHODS: Original studies with sufficient details to obtain the sensitivity and specificity of CT-determined resectability following neoadjuvant treatment, with a reference on the pathological margin status, were identified in PubMed, EMBASE, and Cochrane databases until February 24, 2020. The identified studies were divided into two groups based on the criteria of R0 resectable tumor (ordinary criterion: resectable PDAC alone; extended criterion: resectable and borderline resectable PDAC). The meta-analytic summary of the sensitivity and specificity for each criterion was estimated separately using a bivariate random-effect model. Summary results of the two criteria were compared using a joint-model bivariate meta-regression. RESULTS: Of 739 studies initially searched, 6 studies (6 with ordinary criterion and 5 with extended criterion) were included for analysis. The meta-analytic summary of sensitivity and specificity was 45% (95% confidence interval [CI], 19-73%; I2 = 88.3%) and 85% (95% CI, 65-94%; I2 = 60.5%) for the ordinary criterion, and 81% (95% CI, 71-87%; I2 = 0.0%) and 42% (95% CI, 28-57%; I2 = 6.2%) for the extended criterion, respectively. The diagnostic accuracy significantly differed between the two criteria (p = 0.02). CONCLUSIONS: For determining resectability on CT, the ordinary criterion might be highly specific but insensitive for predicting R0 resection, whereas the extended criterion increased sensitivity but would decrease specificity. Further investigations using quantitative parameters may improve the identification of R0 resection. KEY POINTS: • CT-determined resectability of PDAC after neoadjuvant treatment using the ordinary criterion shows low sensitivity and high specificity in predicting R0 resection. • With the extended criterion, CT-determined resectability shows higher sensitivity but lower specificity than with the ordinary criterion. • CT-determined resectability with both criteria achieved suboptimal diagnostic performances, suggesting that care should be taken while selecting surgical candidates and when determining the surgical extent after neoadjuvant treatment in patients with PDAC.

Multiparametric MRI for prediction of treatment response to neoadjuvant FOLFIRINOX therapy in borderline resectable or locally advanced pancreatic cancer.

OBJECTIVES: To identify multiparametric MRI biomarkers to predict the tumor response to neoadjuvant FOLFIRINOX therapy in patients with borderline resectable (BR) or locally advanced (LA) pancreatic ductal adenocarcinoma (PDAC). METHODS: From May 2016 to March 2018, adult patients with BR or LA PDAC were prospectively enrolled in this study. They received eight cycles of FOLFIRINOX therapy and underwent multiparametric MRI twice (at baseline and after the second cycle). MRI evaluations included dynamic contrast-enhanced MRI, intravoxel incoherent motion diffusion-weighted imaging, and assessment of T2* relaxivity (R2*) and the change in T1 relaxivity (ΔR1, equilibrium phase R1 minus non-enhanced R1) of the tumors. Factors to predict the responders determined by the best overall response during FOLFIRINOX therapy and those to predict progression-free survival (PFS) and overall survival (OS) were evaluated using multivariable logistic regression and the Cox proportional hazard model. RESULTS: Forty-one patients (mean age, 60.3 years ± 9.3; 24 men) were included. Among the clinical and MRI factors, the baseline ΔR1 (adjusted odds ratio, 31.07; p = 0.008) was the only independent predictor for tumor response. The baseline ΔR1 was also an independent predictor for PFS (adjusted hazard ratio, 0.40; p = 0.033) along with R0 resection. The use of a cutoff ΔR1 value of ≥ 1.31 s-1 enabled prognostic stratification (median PFS, 16.0 months vs.10.0 months; p = 0.029; median OS, 34.9 months vs. 16.6 months; p = 0 .023, respectively). CONCLUSIONS: The baseline tumor ΔR1 value may be useful to predict tumor response and survival in patients with BR or LA PDAC receiving FOLFIRINOX neoadjuvant therapy. KEY POINTS: • Baseline ΔR1 was an independent predictor for tumor response (adjusted odds ratio, 31.07; p = 0.008) and progression-free survival (adjusted hazard ratio, 0.40; p = 0.033) in patients with borderline resectable or locally advanced pancreatic ductal adenocarcinoma receiving neoadjuvant FOLFIRINOX therapy. • The criterion of baseline ΔR1 value ≥ 1.31 s-1 allowed for the prediction of favorable tumor response and survival outcome after neoadjuvant FOLFIRINOX therapy.

Meta-analysis of CT and MRI for differentiation of autoimmune pancreatitis from pancreatic adenocarcinoma.

OBJECTIVES: To systematically determine the diagnostic performance of computed tomography (CT) and magnetic resonance imaging (MRI) for differentiating autoimmune pancreatitis (AIP) from pancreatic ductal adenocarcinoma (PDAC), with a comparison between the two imaging modalities. METHODS: Literature search was conducted using PubMed and EMBASE databases to identify original articles published between 2009 and 2019 reporting the diagnostic performance of CT and MRI for differentiating AIP from PDAC. The meta-analytic sensitivity and specificity of CT and MRI were calculated, and compared using a bivariate random effects model. Subgroup analysis for differentiating focal AIP from PDAC was performed. RESULTS: Of the 856 articles screened, 11 eligible articles are remained, i.e., five studies for CT, four for MRI, and two for both. The meta-analytic summary sensitivity and specificity of CT were 59% (95% confidence interval [CI], 41-75%) and 99% (95% CI, 88-100%), respectively, while those of MRI were 84% (95% CI, 68-93%) and 97% (95% CI, 87-99%). MRI had a significantly higher meta-analytic summary sensitivity than CT (84% vs. 59%, p = 0.02) but a similar specificity (97% vs. 99%, p = 0.18). In the subgroup analysis for focal AIP, the sensitivity for distinguishing between focal AIP and PDAC was lower than that for the overall analysis. MRI had a higher sensitivity than CT (76% vs. 50%, p = 0.28) but a similar specificity (97% vs. 98%, p = 0.07). CONCLUSION: MRI might be clinically more useful to evaluate patients with AIP, particularly for differentiating AIP from PDAC. KEY POINTS: • MRI had an overall good diagnostic performance to differentiate AIP from PDAC with a meta-analytic summary estimate of 83% for sensitivity and of 97% for specificity. • CT had a very high specificity (99%), but a suboptimal sensitivity (59%) for differentiating AIP from PDAC. • Compared with CT, MRI had a higher sensitivity, but a similar specificity.

Lymph node size and its association with nodal metastasis in ductal adenocarcinoma of the pancreas.

BACKGROUND: Although lymph node metastasis is a poor prognostic factor in patients with pancreatic ductal adenocarcinoma (PDAC), our understanding of lymph node size in association with PDAC is limited. Increased nodal size in preoperative imaging has been used to detect node metastasis. We evaluated whether lymph node size can be used as a surrogate preoperative marker of lymph node metastasis. METHODS: We assessed nodal size and compared it to the nodal metastatic status of 200 patients with surgically resected PDAC. The size of all lymph nodes and metastatic nodal foci were measured along the long and short axis, and the relationships between nodal size and metastatic status were compared at six cutoff points. RESULTS: A total of 4,525 lymph nodes were examined, 9.1% of which were metastatic. The mean size of the metastatic nodes (long axis, 6.9±5.0 mm; short axis, 4.3±3.1 mm) was significantly larger than that of the non-metastatic nodes (long axis, 5.0±4.0 mm; short axis, 3.0±2.0 mm; all p<.001). Using a 10 mm cutoff, the sensitivity, specificity, positive predictive value, overall accuracy, and area under curve was 24.8%, 88.0%, 17.1%, 82.3%, and 0.60 for the long axis and 7.0%, 99.0%, 40.3%, 90.6%, and 0.61 for the short axis, respectively. CONCLUSIONS: The metastatic nodes are larger than the non-metastatic nodes in PDAC patients. However, the difference in nodal size was too small to be identified with preoperative imaging. The performance of preoperative radiologic imaging to predict lymph nodal metastasis was not good. Therefore, nodal size cannot be used a surrogate preoperative marker of lymph node metastasis.

Estimating Recurrence after Upfront Surgery in Patients with Resectable Pancreatic Ductal Adenocarcinoma by Using Pancreatic CT: Development and Validation of a Risk Score.

Background No preoperative model is available for predicting postsurgical prognosis of patients with resectable pancreatic ductal adenocarcinoma (PDAC). Purpose To develop and validate a preoperative risk scoring system using clinical and CT variables to predict recurrence-free survival (RFS) after upfront surgery in patients with resectable PDAC. Materials and Methods In this retrospective study, consecutive patients with resectable PDAC underwent upfront surgery from January 2014 to December 2015 (development set) and from January 2016 to January 2017 (test set). In the development set, multivariable Cox proportional hazard modeling with bootstrapping was used to select clinical and CT variables associated with RFS and to construct a risk scoring system. The discrimination capability of the risk score was assessed by using the Harrell C-index and compared with that of pathologic American Joint Committee on Cancer tumor stage. The risk score was validated in the test set. Results A total of 395 patients were evaluated, including 262 patients (mean age ± standard deviation, 64 years ± 10; 155 men) in the development set and 133 (mean age, 64 years ± 9; 79 men) in the test set. Five independent variables predicted risk of recurrence or death: tumor size (hazard ratio [HR], 1.23; 95% confidence interval [CI]: 1.05, 1.44; P = .009), hypodense tumor in the portal venous phase (HR, 1.66; 95% CI: 1.01, 2.73; P = .04), tumor necrosis (HR, 2.04; 95% CI: 1.38, 3.03; P < .001), peripancreatic tumor infiltration (HR, 1.50; 95% CI: 1.07, 2.11; P = .02), and suspicious metastatic lymph nodes (HR, 1.94; 95% CI: 1.38, 2.72; P < .001). In the test set, the risk score showed good discrimination capability (C-index of 0.68; 95% CI: 0.63, 0.74) and outperformed the pathologic tumor stage (C-index of 0.60; 95% CI: 0.55, 0.66; P = .03). Patients were categorized into favorable, intermediate, and poor prognosis groups with 1-year RFS of 0.87, 0.58, and 0.26, respectively. Conclusion The presented preoperative risk score can predict recurrence-free survival after upfront surgery in patients with resectable pancreatic ductal adenocarcinoma. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Pandharipande and Anderson in this issue.

Comparison between neuroendocrine carcinomas and well-differentiated neuroendocrine tumors of the pancreas using dynamic enhanced CT.

OBJECTIVES: To identify CT features distinguishing neuroendocrine carcinomas (NECs) of pancreas from well-differentiated neuroendocrine tumors (NETs) according to the World Health Organization 2017 and 2019 classification systems. METHODS: This retrospective study included 69 patients with pathologically confirmed pancreatic neuroendocrine neoplasms who underwent dynamic CT (17, 17, 18, and 17 patients for well-differentiated grade 1, 2, 3 NET and NEC, respectively). CT was used to perform qualitative analysis (component, homogeneity, calcification, peripancreatic infiltration, main pancreatic ductal dilatation, bile duct dilatation, intraductal extension, and vascular invasion) and quantitative analysis (interface between tumor and parenchyma [delta], arterial enhancement ratio [AER], portal enhancement ratio [PER], and dynamic enhancement pattern). Uni- and multivariate logistic regression analyses were performed to identify features indicating NEC. Optimal cutoff values for enhancement ratios were determined. RESULTS: NECs demonstrated significantly higher frequencies of main pancreatic ductal dilatation, bile duct dilatation, vascular invasion, and significantly lower delta (i.e., lower conspicuity), AER, and PER than well-differentiated NET (p < 0.05). On multivariate analysis, PER was the only independent factor selected by the model for differentiation of NEC from well-differentiated NET (odds ratio, < 0.001; 95% confidence interval [CI], < 0.001-0.012). PER < 0.8 showed the sensitivity of 94.1% (95% CI, 71.3-99.9) and the specificity of 88.5% (95% CI, 76.6-95.6). When three significant CT features were combined, the sensitivity and specificity for diagnosing NEC were 88.2% and 88.5%, respectively. CONCLUSIONS: Tumor-parenchyma enhancement ratio in portal phase is a useful CT feature to distinguish NECs from well-differentiated NETs. Combining qualitative and quantitative CT features may aid in achieving good diagnostic accuracy in the differentiation between NEC and well-differentiated NET. KEY POINTS: • Neuroendocrine carcinoma of the pancreas should be distinguished from well-differentiated neuroendocrine tumor in line with the revised grading and staging system. • Neuroendocrine carcinoma of the pancreas can be differentiated from well-differentiated neuroendocrine tumor on dynamic CT based on assessment of the portal enhancement ratio, arterial enhancement ratio, tumor conspicuity, dilatation of the main pancreatic duct or bile duct, and vascular invasion. • Tumor-parenchyma enhancement ratio in portal phase of dynamic CT is a useful feature, which may help to distinguish neuroendocrine carcinoma from well-differentiated neuroendocrine tumor of the pancreas.

Value of apparent diffusion coefficient for differentiating peripancreatic tuberculous lymphadenopathy from metastatic lymphadenopathy.

PURPOSE: To evaluate effectiveness of the apparent diffusion coefficient (ADC) values of the peripancreatic lymphadenopathy to differentiate tuberculous lymphadenopathy from metastatic lymphadenopathy. MATERIALS AND METHODS: Twenty-nine patients with 65 peripancreatic necrotic tuberculous lymphadenopathy and 31 patients with 47 peripancreatic necrotic metastatic lymphadenopathy from pancreatic ductal adenocarcinoma, who underwent magnetic resonance imaging (MRI), were included in this study. MRI features in the T1-weighted image (WI), T2WI, and diffusion-weighted image were analyzed. The ADC values of necrotic and non-necrotic portions of the lymph nodes were measured and compared using t test. Receiver operating characteristic analysis was performed to obtain the optimal ADC threshold value and diagnostic accuracy for differentiating tuberculous lymphadenopathy from metastatic lymphadenopathy. RESULTS: On T2WI, the signal intensity of necrotic portions was variable in tuberculous lymphadenopathy, but was mostly high in metastatic lymphadenopathy. The mean ADCs of necrotic portions of tuberculous lymphadenopathy were significantly lower than those of metastatic lymphadenopathy ([0.919 ± 0.272] × 10-3 mm2/s vs. [1.553 ± 0.406] × 10-3 mm2/s, p < 0.001). Receiver operating characteristic analysis for differentiating tuberculous from metastatic lymphadenopathy demonstrated an area under the curve for the ADC values of necrotic portions of 0.929 (95% CI, 0.865-0.969) with an ADC threshold of 1.022. The sensitivity and specificity for the differentiation of tuberculous from metastatic lymphadenopathy were 80.0% and 97.8%, respectively. CONCLUSION: The ADC values of necrotic portions of peripancreatic lymphadenopathy may be useful for differentiating tuberculous from metastatic lymphadenopathy.

Smooth Muscle Distribution Patterns of Choledochal Cysts and Their Implications for Pathogenesis and Postoperative Complications.

OBJECTIVES: Histopathologic characteristics of choledochal cysts and their clinical implications have not been previously comprehensively studied. METHODS: Smooth muscle distribution patterns and other histologic findings (inflammation, metaplasia, dysplasia, and heterotopia) in 233 surgically resected choledochal cysts were evaluated. RESULTS: Mean patient age was 23.3 ±â€…19.8 years, with male:female ratio of 0.3. Most cases were Todani type I (175 cases, 75.1%) or IVa (56 cases, 24.1%). Choledochal cysts with thin scattered/no muscle fiber (175 cases, 75.1%) were the predominant pattern and were associated with more frequent postoperative biliary stricture (P = .031), less frequent pyloric metaplasia (P = .016), and mucosal smooth muscle aggregates (P < .001) compared to cysts with thick muscle bundles. Severe chronic cholangitis (P = .049), pyloric metaplasia (P = .019), mucosal smooth muscle aggregates (P < .001), biliary intraepithelial neoplasia (P = .021), and associated bile duct (P = .021) and gallbladder carcinomas (P = .03) were more common in adults (age >20 years vs ≤20 years), suggesting that chronic irritation in association with developmental anomalies involves tumorigenesis from choledochal cysts. CONCLUSION: Smooth muscle distribution pattern of choledochal cyst may predict postoperative complication, raising clinical implications of smooth muscle patterns in postoperative management of choledochal cysts.

Differentiating focal autoimmune pancreatitis and pancreatic ductal adenocarcinoma: contrast-enhanced MRI with special emphasis on the arterial phase.

OBJECTIVES: To compare focal-type autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDA) using contrast-enhanced MR imaging (CE-MRI), and to assess diagnostic performance of the lesion contrast at arterial phase (AP) (ContrastAP) for differentiating between the two diseases. METHODS: Thirty-six patients with focal-type AIP and 72 patients with PDA were included. All included patients underwent CE-MRI with triple phases. The signal intensity (SI) of the mass and normal pancreas was measured at each phase, and the lesion contrast (SIpancreas/SImass) was compared between AIP and PDA groups. The sensitivity and specificity of ContrastAP using an optimal cutoff point were compared with those of key imaging features specific to AIP and PDA. RESULTS: The lesion contrast differed significantly between AIP and PDA groups at all phases of CE-MRI; the maximum difference was observed at AP. For AIP, the sensitivity (94.4%) and specificity (87.5%) of ContrastAP (cutoff ≤ 1.41) were comparable or significantly higher than those of all key imaging features (sensitivity, 38.9-88.9%; specificity, 48.6-95.8%), except for the halo sign. For PDA, the sensitivity (87.5%) and specificity (94.4%) of ContrastAP (cutoff > 1.41) were comparable or significantly higher than those of all key imaging features (sensitivity, 40.3-68.1%; specificity, 72.2-94.4%), except for the discrete mass. CONCLUSIONS: Quantitative analysis of the lesion contrast using CE-MRI, particularly at AP, was helpful to differentiate focal-type AIP from PDA. The diagnostic performance of ContrastAP was mostly comparable or higher than those of the key imaging features. KEY POINTS: • Diagnosis of focal-type AIP vs. PDA using imaging techniques is extremely challenging. • Lesion contrast in the arterial-phase MRI differs significantly between focal-type AIP and PDA. • Quantitative analysis of lesion contrast using CE-MRI, particularly at the arterial phase, is helpful to differentiate focal-type AIP from PDA.