The differentiation of autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDAC) poses a relevant diagnostic challenge and can lead to misdiagnosis and consequently poor patient outcome. Recent studies have shown that radiomics-based models can achieve high sensitivity and specificity in predicting both entities. However, radiomic features can only capture low level representations of the input image. In contrast, convolutional neural networks (CNNs) can learn and extract more complex representations which have been used for image classification to great success. In our retrospective observational study, we performed a deep learning-based feature extraction using CT-scans of both entities and compared the predictive value against traditional radiomic features. In total, 86 patients, 44 with AIP and 42 with PDACs, were analyzed. Whole pancreas segmentation was automatically performed on CT-scans during the portal venous phase. The segmentation masks were manually checked and corrected if necessary. In total, 1411 radiomic features were extracted using PyRadiomics and 256 features (deep features) were extracted using an intermediate layer of a convolutional neural network (CNN). After feature selection and normalization, an extremely randomized trees algorithm was trained and tested using a two-fold shuffle-split cross-validation with a test sample of 20% (n = 18) to discriminate between AIP or PDAC. Feature maps were plotted and visual difference was noted. The machine learning (ML) model achieved a sensitivity, specificity, and ROC-AUC of 0.89 ± 0.11, 0.83 ± 0.06, and 0.90 ± 0.02 for the deep features and 0.72 ± 0.11, 0.78 ± 0.06, and 0.80 ± 0.01 for the radiomic features. Visualization of feature maps indicated different activation patterns for AIP and PDAC. We successfully trained a machine learning model using deep feature extraction from CT-images to differentiate between AIP and PDAC. In comparison to traditional radiomic features, deep features achieved a higher sensitivity, specificity, and ROC-AUC. Visualization of deep features could further improve the diagnostic accuracy of non-invasive differentiation of AIP and PDAC.

Deep Convolutional Neural Network-Assisted Feature Extraction for Diagnostic Discrimination and Feature Visualization in Pancreatic Ductal Adenocarcinoma (PDAC) versus Autoimmune Pancreatitis (AIP)

BACKGROUND
Detection of hepatic metastases during EUS is an important component of tumor staging.


OBJECTIVE
To describe our experience with EUS-guided FNA (EUS-FNA) of solid hepatic masses and derive and validate criteria to help distinguish between benign and malignant hepatic masses.


DESIGN
Retrospective study, survey.


SETTING
Single, tertiary-care referral center.


PATIENTS
Medical records were reviewed for all patients undergoing EUS-FNA of solid hepatic masses over a 12-year period.


INTERVENTIONS
EUS-FNA of solid hepatic masses.


MAIN OUTCOME MEASUREMENTS
Masses were deemed benign or malignant according to predetermined criteria. EUS images from 200 patients were used to create derivation and validation cohorts of 100 cases each, matched by cytopathologic diagnosis. Ten expert endosonographers blindly rated 15 initial endosonographic features of each of the 100 images in the derivation cohort. These data were used to derive an EUS scoring system that was then validated by using the validation cohort by the expert endosonographer with the highest diagnostic accuracy.


RESULTS
A total of 332 patients underwent EUS-FNA of a hepatic mass. Interobserver agreement regarding the initial endosonographic features among the expert endosonographers was fair to moderate, with a mean diagnostic accuracy of 73% (standard deviation 5.6). A scoring system incorporating 7 EUS features was developed to distinguish benign from malignant hepatic masses by using the derivation cohort with an area under the receiver operating curve (AUC) of 0.92; when applied to the validation cohort, performance was similar (AUC 0.86). The combined positive predictive value of both cohorts was 88%.


LIMITATIONS
Single center, retrospective, only one expert endosonographer deriving and validating the EUS criteria.


CONCLUSION
An EUS scoring system was developed that helps distinguish benign from malignant hepatic masses. Further study is required to determine the impact of these EUS criteria among endosonographers of all experience.

Iconographies supplémentaires de l'article : EUS-derived criteria for distinguishing benign from malignant metastatic solid hepatic masses

The development of artificial intelligence (AI) has increased dramatically in the last 20 years, with clinical applications progressively being explored for most of the medical specialties. The field of gastroenterology and hepatology, substantially reliant on vast amounts of imaging studies, is not an exception. The clinical applications of AI systems in this field include the identification of premalignant or malignant lesions (e.g., identification of dysplasia or esophageal adenocarcinoma in Barrett's esophagus, pancreatic malignancies), detection of lesions (e.g., polyp identification and classification, small-bowel bleeding lesion on capsule endoscopy, pancreatic cystic lesions), development of objective scoring systems for risk stratification, predicting disease prognosis or treatment response [e.g., determining survival in patients post-resection of hepatocellular carcinoma), determining which patients with inflammatory bowel disease (IBD) will benefit from biologic therapy], or evaluation of metrics such as bowel preparation score or quality of endoscopic examination. The objective of this comprehensive review is to analyze the available AI-related studies pertaining to the entirety of the gastrointestinal tract, including the upper, middle and lower tracts; IBD; the hepatobiliary system; and the pancreas, discussing the findings and clinical applications, as well as outlining the current limitations and future directions in this field.

Artificial intelligence in gastroenterology: A state-of-the-art review.

Pancreatic ductal adenocarcinoma (PDAC) carries a deadly diagnosis, due in large part to delayed presentation when the disease is already at an advanced stage. CA19-9 is currently the most commonly utilized biomarker for PDAC; however, it lacks the necessary accuracy to detect precursor lesions or stage I PDAC. Novel biomarkers that could detect this malignancy with improved sensitivity (SN) and specificity (SP) would likely result in more curative resections and more effective therapeutic interventions, changing thus the present dismal survival figures. The aim of this study was to systematically and comprehensively review the scientific literature on non-invasive biomarkers in biofluids such as blood, urine and saliva that were attempting earlier PDAC detection. The search performed covered a period of 10 years (January 2010-August 2020). Data were extracted using keywords search in the three databases: MEDLINE, Web of Science and Embase. The Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool was applied for study selection based on establishing the risk of bias and applicability concerns in Patient Selection, Index test (biomarker assay) and Reference Standard (standard-of-care diagnostic test). Out of initially over 4000 published reports, 49 relevant studies were selected and reviewed in more detail. In addition, we discuss the present challenges and complexities in the path of translating the discovered biomarkers into the clinical setting. Our systematic review highlighted several promising biomarkers that could, either alone or in combination with CA19-9, potentially improve earlier detection of PDAC. Overall, reviewed biomarker studies should aim to improve methodological and reporting quality, and novel candidate biomarkers should be investigated further in order to demonstrate their clinical usefulness. However, challenges and complexities in the path of translating the discovered biomarkers from the research laboratory to the clinical setting remain and would have to be addressed before a more realistic breakthrough in earlier detection of PDAC is achieved.

Non-Invasive Biomarkers for Earlier Detection of Pancreatic Cancer-A Comprehensive Review.

Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) is a first-line procedure for definitive tissue diagnosis of pancreatic cancer because of its high accuracy and low complication rate. The overall sensitivity of EUS-FNA for the diagnosis of pancreatic cancer is approximately 90%; however, its diagnostic ability for small lesions (<10 mm) remains limited. Although EUS-FNA is a relatively safe procedure, with an overall morbidity of ≤2%, it should be noted that needle tract seeding (NTS) can occur after EUS-FNA, which may affect the patient prognosis negatively. In patients with resected pancreatic tumors, preoperative EUS-FNA is not associated with an increased risk of postoperative recurrence. However, NTS after EUS-FNA for resectable pancreatic body and tail cancer has been highlighted recently, particularly by Japanese endoscopists. Thus, the use of preoperative EUS-FNA for the diagnosis of resectable pancreatic body and tail cancer has become a clinical dilemma that challenges gastroenterologists and must be carefully considered on a case-by-case basis by weighing the benefits and risks. This review summarizes the pros and cons of performing EUS-FNA in patients with resectable pancreatic body and tail cancer and provides valuable insight for gastroenterologists treating pancreatic cancer.

https://onlinelibrary.wiley.com/doi/pdf/10.1111/den.14120

Clinical dilemma of endoscopic ultrasound-guided fine needle aspiration for resectable pancreatic body and tail cancer.

Objective The diagnosis of autoimmune pancreatitis (AIP) is challenging. Sonographic and cross-sectional imaging findings of AIP closely mimic pancreatic ductal adenocarcinoma (PDAC) and techniques for tissue sampling of AIP are suboptimal. These limitations often result in delayed or failed diagnosis, which negatively impact patient management and outcomes. This study aimed to create an endoscopic ultrasound (EUS)-based convolutional neural network (CNN) model trained to differentiate AIP from PDAC, chronic pancreatitis (CP) and normal pancreas (NP), with sufficient performance to analyse EUS video in real time. Design A database of still image and video data obtained from EUS examinations of cases of AIP, PDAC, CP and NP was used to develop a CNN. Occlusion heatmap analysis was used to identify sonographic features the CNN valued when differentiating AIP from PDAC. Results From 583 patients (146 AIP, 292 PDAC, 72 CP and 73 NP), a total of 1 174 461 unique EUS images were extracted. For video data, the CNN processed 955 EUS frames per second and was: 99% sensitive, 98% specific for distinguishing AIP from NP; 94% sensitive, 71% specific for distinguishing AIP from CP; 90% sensitive, 93% specific for distinguishing AIP from PDAC; and 90% sensitive, 85% specific for distinguishing AIP from all studied conditions (ie, PDAC, CP and NP). Conclusion The developed EUS-CNN model accurately differentiated AIP from PDAC and benign pancreatic conditions, thereby offering the capability of earlier and more accurate diagnosis. Use of this model offers the potential for more timely and appropriate patient care and improved outcome.

https://gut.bmj.com/content/gutjnl/70/7/1335.full.pdf

Utilisation of artificial intelligence for the development of an EUS-convolutional neural network model trained to enhance the diagnosis of autoimmune pancreatitis.

Application of artificial intelligence using a novel EUS-based convolutional neural network model to identify and distinguish benign and malignant hepatic masses.

Zaiyang Long

Papers

Utilisation of artificial intelligence for the development of an EUS-convolutional neural network model trained to enhance the diagnosis of autoimmune pancreatitis.

Application of artificial intelligence using a novel EUS-based convolutional neural network model to identify and distinguish benign and malignant hepatic masses.