High performance in risk stratification of intraductal papillary mucinous neoplasms by confocal laser endomicroscopy image analysis with convolutional neural networks (with video)

Deep Learning and Automatic Differentiation of Pancreatic Lesions in Endoscopic Ultrasound: A Transatlantic Study

INTRODUCTION

Endoscopic ultrasound (EUS) allows for characterization and biopsy of pancreatic lesions. Pancreatic cystic neoplasms (PCN) include mucinous (M-PCN) and nonmucinous lesions (NM-PCN). Pancreatic ductal adenocarcinoma (P-DAC) is the commonest pancreatic solid lesion (PSL), followed by pancreatic neuroendocrine tumor (P-NET). Although EUS is preferred for pancreatic lesion evaluation, its diagnostic accuracy is suboptimal. This multicentric study aims to develop a convolutional neural network (CNN) for detecting and distinguishing PCN (namely M-PCN and NM-PCN) and PSL (particularly P-DAC and P-NET).

METHODS

A CNN was developed with 378 EUS examinations from 4 international reference centers (Centro Hospitalar Universitário São João, Hospital Universitario Puerta de Hierro Majadahonda, New York University Hospitals, Hospital das Clínicas Faculdade de Medicina da Universidade de São Paulo). About 126.000 images were obtained-19.528 M-PCN, 8.175 NM-PCN, 64.286 P-DAC, 29.153 P-NET, and 4.858 normal pancreas images. A trinary CNN differentiated normal pancreas tissue from M-PCN and NM-PCN. A binary CNN distinguished P-DAC from P-NET. The total data set was divided into a training and testing data set (used for model's evaluation) in a 90/10% ratio. The model was evaluated through its sensitivity, specificity, positive and negative predictive values, and accuracy.

RESULTS

The CNN had 99.1% accuracy for identifying normal pancreatic tissue, 99.0% and 99.8% for M-PCN and NM-PCN, respectively. P-DAC and P-NET were distinguished with 94.0% accuracy.

DISCUSSION

Our group developed the first worldwide CNN capable of detecting and differentiating the commonest PCN and PSL in EUS images, using examinations from 4 centers in 2 continents, minimizing the impact of the demographic bias. Larger multicentric studies are needed for technology implementation.

Endoscopic ultrasound-guided radiofrequency ablation of pancreatic tumors

PURPOSE OF REVIEW

Surgery is a cornerstone in the management of pancreatic cancer and precancerous pancreatic lesions. However, many patients are not suitable candidates for surgery at the time of diagnosis for various reasons. Endoscopic ultrasound-guided radiofrequency ablation (EUS-RFA) appears to be a promising treatment option for patients who are ineligible for surgery for management of pancreatic adenocarcinoma (PDAC), and pancreatic neuroendocrine tumors (PNETs), and pancreatic cystic lesions (PCLs).

RECENT FINDINGS

EUS-RFA may serve as an adjunct to chemotherapy or palliative measures for inoperable cases of PDAC. Given its feasibility and efficacy, EUS-RFA has an evolving niche as a minimally invasive and potentially definitive treatment for PNETs and high-risk PCLs such as intraductal papillary mucinous neoplasms (IPMNs). EUS-RFA is a generally well tolerated procedure, with abdominal pain and acute pancreatitis being the most common adverse effects, though the risk can be mitigated through prophylactic measures.

SUMMARY

There is an increasing body of evidence to support the use of EUS-RFA in managing pancreatic lesions, either as definitive, adjunctive, or palliative treatment, depending on lesion type.

Endoscopic ultrasound-guided needle-based confocal laser endomicroscopy for pancreatic cystic lesions: current status and future prospects

Pancreatic cystic lesions (PCLs) have increased in prevalence due to the increased usage and advancements in cross-sectional abdominal imaging. Current diagnostic techniques cannot distinguish between PCLs requiring surgery, close surveillance, or expectant management. This has increased the morbidity and healthcare costs from inappropriately aggressive and conservative management strategies. Endoscopic ultrasound (EUS) needle-based confocal laser endomicroscopy (nCLE) allows for microscopic examination and delineation of the surface epithelium of PCLs. Landmark studies have identified characteristics distinguishing various types of PCLs, confirmed the high diagnostic yield of EUS-nCLE (especially for PCLs with an equivocal diagnosis), and shown that EUS-nCLE helps to change management and reduce healthcare costs. Refining procedure technique and reducing procedure length have improved the safety of EUS-nCLE. The utilization of artificial intelligence and its combination with other EUS-based advanced diagnostic techniques would further improve the results of EUS-based PCL diagnosis. A structured training program and device improvements to allow more complete mapping of the pancreas cyst epithelium will be crucial for the widespread adoption of this promising technology.

Advancements in biliopancreatic endoscopy: a comprehensive review of artificial intelligence in EUS and ERCP

The development and implementation of artificial intelligence (AI), particularly deep learning (DL) models, has generated significant interest across various fields of gastroenterology. While research in luminal endoscopy has seen rapid translation to clinical practice with approved AI devices, its potential extends far beyond, offering promising benefits for biliopancreatic endoscopy like optical characterization of strictures during cholangioscopy or detection and classification of pancreatic lesions during diagnostic endoscopic ultrasound (EUS). This narrative review provides an up-to-date of the latest literature and available studies in this field. Serving as a comprehensive guide to the current landscape of AI in biliopancreatic endoscopy, emphasizing technological advancements, main applications, ethical considerations, and future directions for research and clinical implementation.

Basic Principles and Role of Endoscopic Ultrasound in Diagnosis and Differentiation of Pancreatic Cancer from Other Pancreatic Lesions: A Comprehensive Review of Endoscopic Ultrasound for Pancreatic Cancer

Pancreatic cancer is one of the leading causes of cancer-related deaths worldwide. Pancreatic lesions consist of both neoplastic and non-neoplastic lesions and often pose a diagnostic and therapeutic challenge due to similar clinical and radiological features. In recent years, pancreatic lesions have been discovered more frequently as incidental findings due to the increased utilization and widespread availability of abdominal cross-sectional imaging. Therefore, it becomes imperative to establish an early and appropriate diagnosis with meticulous differentiation in an attempt to balance unnecessary treatment of benign pancreatic lesions and missing the opportunity for early intervention in malignant lesions. Endoscopic ultrasound (EUS) has become an important diagnostic modality for the identification and risk stratification of pancreatic lesions due to its ability to provide detailed imaging and acquisition of tissue samples for analysis with the help of fine-needle aspiration/biopsy. The recent development of EUS-based technology, including contrast-enhanced endoscopic ultrasound, real-time elastography-endoscopic ultrasound, miniature probe ultrasound, confocal laser endomicroscopy, and the application of artificial intelligence has significantly augmented the diagnostic accuracy of EUS as it enables better evaluation of the number, location, dimension, wall thickness, and contents of these lesions. This article provides a comprehensive overview of the role of the different types of EUS available for the diagnosis and differentiation of pancreatic cancer from other pancreatic lesions while discussing their key strengths and important limitations.

Current status of artificial intelligence analysis for the treatment of pancreaticobiliary diseases using endoscopic ultrasonography and endoscopic retrograde cholangiopancreatography

Pancreatic and biliary diseases encompass a range of conditions requiring accurate diagnosis for appropriate treatment strategies. This diagnosis relies heavily on imaging techniques like endoscopic ultrasonography and endoscopic retrograde cholangiopancreatography. Artificial intelligence (AI), including machine learning and deep learning, is becoming integral in medical imaging and diagnostics, such as the detection of colorectal polyps. AI shows great potential in diagnosing pancreatobiliary diseases. Unlike machine learning, which requires feature extraction and selection, deep learning can utilize images directly as input. Accurate evaluation of AI performance is a complex task due to varied terminologies, evaluation methods, and development stages. Essential aspects of AI evaluation involve defining the AI's purpose, choosing appropriate gold standards, deciding on the validation phase, and selecting reliable validation methods. AI, particularly deep learning, is increasingly employed in endoscopic ultrasonography and endoscopic retrograde cholangiopancreatography diagnostics, achieving high accuracy levels in detecting and classifying various pancreatobiliary diseases. The AI often performs better than doctors, even in tasks like differentiating benign from malignant pancreatic tumors, cysts, and subepithelial lesions, identifying gallbladder lesions, assessing endoscopic retrograde cholangiopancreatography difficulty, and evaluating the biliary strictures. The potential for AI in diagnosing pancreatobiliary diseases, especially where other modalities have limitations, is considerable. However, a crucial constraint is the need for extensive, high-quality annotated data for AI training. Future advances in AI, such as large language models, promise further applications in the medical field.

Endoscopic Ultrasound-Guided Ablation of Premalignant Pancreatic Cysts and Pancreatic Cancer

Pancreatic cancer is on the rise and expected to become the second leading cause of cancer-related death by 2030. Up to a one-fifth of pancreatic cancers may arise from mucinous pancreatic cysts, which are frequently present in the general population. Currently, surgical resection is the only curative approach for pancreatic cancer and its cystic precursors. However, only a dismal proportion of patients are eligible for surgery. Therefore, novel treatment approaches to treat pancreatic cancer and precancerous pancreatic cysts are needed. Endoscopic ultrasound (EUS)-guided ablation is an emerging minimally invasive method to treat pancreatic cancer and premalignant pancreatic cysts. Different ablative modalities have been used including alcohol, chemotherapy agents, and radiofrequency ablation. Cumulative data over the past two decades have shown that endoscopic ablation of mucinous pancreatic cysts can lead to cyst resolution in a significant proportion of the treated cysts. Furthermore, novel data are emerging about the ability to endoscopically ablate early and locally advanced pancreatic cancer. In this review, we aim to summarize the available data on the efficacy and safety of the different EUS-ablation modalities for the management of premalignant pancreatic cysts and pancreatic cancer.

Structured training program on confocal laser endomicroscopy for pancreatic cystic lesions: a multicenter prospective study among early-career endosonographers (with video)

BACKGROUND AND AIMS

Data on how to teach endosonographers needle-based confocal laser endomicroscopy (nCLE)-guided histologic diagnosis of pancreatic cystic lesions (PCLs) are limited. Hence, we developed and tested a structured educational program to train early-career endosonographers in nCLE-guided diagnosis of PCLs.

METHODS

Twenty-one early-career nCLE-naïve endosonographers watched a teaching module outlining nCLE criteria for diagnosing PCLs. Participants then reviewed 80 high-yield nCLE videos, recorded diagnoses, and received expert feedback (phase 1). Observers were then randomized to a refresher feedback session or self-learning at 4 weeks. Eight weeks after training, participants independently assessed the same 80 nCLE videos without feedback and provided histologic predictions (phase 2). Diagnostic performance of nCLE to differentiate mucinous versus nonmucinous PCLs and to diagnose specific subtypes were analyzed using histopathology as the criterion standard. Learning curves were determined using cumulative sum analysis.

RESULTS

Accuracy and diagnostic confidence for differentiating mucinous versus nonmucinous PCLs improved as endosonographers progressed through nCLE videos in phase 1 (P < .001). Similar trends were observed with the diagnosis of PCL subtypes. Most participants achieved competency interpreting nCLE, requiring a median of 38 assessments (range, 9-67). During phase 2, participants independently differentiated PCLs with high accuracy (89%), high confidence (83%), and substantial interobserver agreement (κ = .63). Accuracy for nCLE-guided PCL subtype diagnoses ranged from 82% to 96%. The learned nCLE skills did not deteriorate at 8 weeks and were not impacted by a refresher session.

CONCLUSIONS

We developed a practical, effective, and durable educational intervention to train early-career endosonographers in nCLE-guided diagnosis of PCLs.

Improving Pancreatic Cyst Management: Artificial Intelligence-Powered Prediction of Advanced Neoplasms through Endoscopic Ultrasound-Guided Confocal Endomicroscopy

Despite the increasing rate of detection of incidental pancreatic cystic lesions (PCLs), current standard-of-care methods for their diagnosis and risk stratification remain inadequate. Intraductal papillary mucinous neoplasms (IPMNs) are the most prevalent PCLs. The existing modalities, including endoscopic ultrasound and cyst fluid analysis, only achieve accuracy rates of 65-75% in identifying carcinoma or high-grade dysplasia in IPMNs. Furthermore, surgical resection of PCLs reveals that up to half exhibit only low-grade dysplastic changes or benign neoplasms. To reduce unnecessary and high-risk pancreatic surgeries, more precise diagnostic techniques are necessary. A promising approach involves integrating existing data, such as clinical features, cyst morphology, and data from cyst fluid analysis, with confocal endomicroscopy and radiomics to enhance the prediction of advanced neoplasms in PCLs. Artificial intelligence and machine learning modalities can play a crucial role in achieving this goal. In this review, we explore current and future techniques to leverage these advanced technologies to improve diagnostic accuracy in the context of PCLs.

Applications and Prospects of Artificial Intelligence-Assisted Endoscopic Ultrasound in Digestive System Diseases

Endoscopic ultrasound (EUS) has emerged as a widely utilized tool in the diagnosis of digestive diseases. In recent years, the potential of artificial intelligence (AI) in healthcare has been gradually recognized, and its superiority in the field of EUS is becoming apparent. Machine learning (ML) and deep learning (DL) are the two main AI algorithms. This paper aims to outline the applications and prospects of artificial intelligence-assisted endoscopic ultrasound (EUS-AI) in digestive diseases over the past decade. The results demonstrated that EUS-AI has shown superiority or at least equivalence to traditional methods in the diagnosis, prognosis, and quality control of subepithelial lesions, early esophageal cancer, early gastric cancer, and pancreatic diseases including pancreatic cystic lesions, autoimmune pancreatitis, and pancreatic cancer. The implementation of EUS-AI has opened up new avenues for individualized precision medicine and has introduced novel diagnostic and treatment approaches for digestive diseases.

Using Endoscopy in the Diagnosis of Pancreato-Biliary Cancers

Pancreatic cancer and cholangiocarcinoma are life threatening oncological conditions with poor prognosis and outcome. Pancreatic cystic lesions are considered precursors of pancreatic cancer as some of them have the potential to progress to malignancy. Therefore, accurate identification and classification of these lesions is important to prevent the development of invasive cancer. In the biliary tract, the accurate characterization of biliary strictures is essential for providing appropriate management and avoiding unnecessary surgery. Techniques have been developed to improve the diagnosis, risk stratification, and management of pancreato-biliary lesions. Endoscopic ultrasound (EUS) and associated techniques, such as elastography, contrasted-enhanced EUS, and EUS-guided needle confocal laser endomicroscopy, may improve diagnostic accuracy. In addition, intraductal techniques applied during endoscopic retrograde cholangiopancreatography (ERCP), such as new generation cholangioscopy and in vivo cellular evaluation through probe-based confocal laser endomicroscopy, can increase the diagnostic yield in characterizing indeterminate biliary strictures. Both EUS-guided and intraductal approaches can provide the possibility for tissue sampling with new tools, such as needles, biopsies forceps, and brushes. At the molecular level, novel biomarkers have been explored that provide new insights into diagnosis, risk stratification, and management of these lesions.

Intraductal Papillary Mucinous Neoplasm of the Pancreas: A Challenging Diagnosis

Intraductal papillary mucinous neoplasm of the pancreas (IPMN) was classified as a distinct entity from mucinous cystic neoplasm by the WHO in 1995. It represents a mucin-producing tumor that originates from the ductal epithelium and can evolve from slight dysplasia to invasive carcinoma. In addition, different aspects of tumor progression may be seen in the same lesion. Three types are recognized, the branch duct variant, the main duct variant, which shows a much higher prevalence for malignancy, and the mixed-type variant, which combines branch and main duct characteristics. Advances in cross-sectional imaging have led to an increased rate of IPMN detection. The main imaging characteristic of IPMN is the dilatation of the pancreatic duct without the presence of an obstructing lesion. The diagnosis of a branch duct IPMN is based on the proof of its communication with the main pancreatic duct on MRI-MRCP examination. Early identification by imaging of the so-called worrisome features or predictors for malignancy is an important and challenging task. In this review, we will present recent imaging advances in the diagnosis and characterization of different types of IPMNs, as well as imaging tools available for early recognition of worrisome features for malignancy. A critical appraisal of current IPMN management guidelines from both a radiologist's and surgeon's perspective will be made. Special mention is made of complications that might arise during the course of IPMNs as well as concomitant pancreatic neoplasms including pancreatic adenocarcinoma and pancreatic endocrine neoplasms. Finally, recent research on prognostic and predictive biomarkers including radiomics will be discussed.

Artificial Intelligence in the Diagnosis and Treatment of Pancreatic Cystic Lesions and Adenocarcinoma

Pancreatic cancer is projected to become the second leading cause of cancer-related mortality in the United States by 2030. This is in part due to the paucity of reliable screening and diagnostic options for early detection. Amongst known pre-malignant pancreatic lesions, pancreatic intraepithelial neoplasia (PanIN) and intraductal papillary mucinous neoplasms (IPMNs) are the most prevalent. The current standard of care for the diagnosis and classification of pancreatic cystic lesions (PCLs) involves cross-sectional imaging studies and endoscopic ultrasound (EUS) and, when indicated, EUS-guided fine needle aspiration and cyst fluid analysis. However, this is suboptimal for the identification and risk stratification of PCLs, with accuracy of only 65-75% for detecting mucinous PCLs. Artificial intelligence (AI) is a promising tool that has been applied to improve accuracy in screening for solid tumors, including breast, lung, cervical, and colon cancer. More recently, it has shown promise in diagnosing pancreatic cancer by identifying high-risk populations, risk-stratifying premalignant lesions, and predicting the progression of IPMNs to adenocarcinoma. This review summarizes the available literature on artificial intelligence in the screening and prognostication of precancerous lesions in the pancreas, and streamlining the diagnosis of pancreatic cancer.

Deep-Learning-Enabled Computer-Aided Diagnosis in the Classification of Pancreatic Cystic Lesions on Confocal Laser Endomicroscopy

Accurate classification of pancreatic cystic lesions (PCLs) is important to facilitate proper treatment and to improve patient outcomes. We utilized the convolutional neural network (CNN) of VGG19 to develop a computer-aided diagnosis (CAD) system in the classification of subtypes of PCLs in endoscopic ultrasound-guided needle-based confocal laser endomicroscopy (nCLE). From a retrospectively collected 22,424 nCLE video frames (50 videos) as the training/validation set and 11,047 nCLE video frames (18 videos) as the test set, we developed and compared the diagnostic performance of three CNNs with distinct methods of designating the region of interest. The diagnostic accuracy for subtypes of PCLs by CNNs with manual, maximal rectangular, and U-Net algorithm-designated ROIs was 100%, 38.9%, and 66.7% on a per-video basis and 88.99%, 73.94%, and 76.12% on a per-frame basis, respectively. Our per-frame analysis suggested differential levels of diagnostic accuracy among the five subtypes of PCLs, where non-mucinous PCLs (serous cystic neoplasm: 93.11%, cystic neuroendocrine tumor: 84.31%, and pseudocyst: 98%) had higher diagnostic accuracy than mucinous PCLs (intraductal papillary mucinous neoplasm: 84.43% and mucinous cystic neoplasm: 86.1%). Our CNN demonstrated superior specificity compared to the state-of-the-art for the classification of mucinous PCLs (IPMN and MCN), with high specificity (94.3% and 92.8%, respectively) but low sensitivity (46% and 45.2%, respectively). This suggests the complimentary role of CNN-enabled CAD systems, especially for clinically suspected mucinous PCLs.

Deep learning for pancreatic diseases based on endoscopic ultrasound: A systematic review

BACKGROUND AND AIMS

Endoscopic ultrasonography (EUS) is one of the main examinations in pancreatic diseases. A series of the studies reported the application of deep learning (DL)-assisted EUS in the diagnosis of pancreatic diseases. This systematic review is to evaluate the role of DL algorithms in assisting EUS diagnosis of pancreatic diseases.

METHODS

Literature search were conducted in PubMed and Semantic Scholar databases. Studies that developed DL models for pancreatic diseases based on EUS were eligible for inclusion. This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and quality assessment of the included studies was performed according to the IJMEDI checklist.

RESULTS

A total of 23 studies were enrolled into this systematic review, which could be categorized into three groups according to computer vision tasks: classification, detection and segmentation. Seventeen studies focused on the classification task, among which five studies developed simple neural network (NN) models while twelve studies constructed convolutional NN (CNN) models. Three studies were concerned the detection task and five studies were the segmentation task, all based on CNN architectures. All models presented in the studies performed well based on EUS images, videos or voice. According to the IJMEDI checklist, six studies were recognized as high-grade quality, with scores beyond 35 points.

CONCLUSIONS

DL algorithms show great potential in EUS images/videos/voice for pancreatic diseases. However, there is room for improvement such as sample sizes, multi-center cooperation, data preprocessing, model interpretability, and code sharing.

The Role of Genetic, Metabolic, Inflammatory, and Immunologic Mediators in the Progression of Intraductal Papillary Mucinous Neoplasms to Pancreatic Adenocarcinoma

Intraductal papillary mucinous neoplasms (IPMN) have the potential to progress to pancreatic ductal adenocarcinoma (PDAC). As with any progression to malignancy, there are a variety of genetic and metabolic changes, as well as other disruptions to the cellular microenvironment including immune alterations and inflammation, that can contribute to tumorigenesis. Previous studies further characterized these alterations, revealing changes in lipid and glucose metabolism, and signaling pathways that mediate the progression of IPMN to PDAC. With the increased diagnosis of IPMNs and pancreatic cysts on imaging, the opportunity to attenuate risk with the removal of high-risk lesions is possible with the understanding of what factors accelerate malignant progression and how they can be clinically utilized to determine the level of dysplasia and stratify the risk of progression. Here, we reviewed the genetic, metabolic, inflammatory, and immunologic pathways regulating the progression of IPMN to PDAC.

Role of Endoscopic Ultrasound in the Evaluation of Pancreatic Cystic Neoplasms: A Concise Review

Pancreatic cystic lesions are being discovered as incidental lesions during cross-sectional imaging studies of the abdomen with increasing frequency. Endoscopic ultrasound is an important diagnostic modality for managing pancreatic cystic lesions. There are various types of pancreatic cystic lesions, from benign to malignant. Endoscopic ultrasound has a multifactorial role in delineating the morphology of pancreatic cystic lesions, ranging from fluid and tissue acquisition for analysis-fine needle aspiration and through-the-needle biopsy, respectively-to advanced imaging techniques, such as contrast-harmonic mode endoscopic ultrasound and EUS-guided needle-based confocal laser endomicroscopy. In this review, we will summarize and provide an update on the specific role of EUS in the management of pancreatic cystic lesions.

A Review of Endoscopic Ultrasound-Guided Chemoablative Techniques for Pancreatic Cystic Lesions

Pancreatic cystic lesions (PCLs) are known precursors to pancreatic cancer, one of the deadliest types of cancer worldwide. Surgical removal or pancreatectomies remain the central approach to managing precancerous high-risk PCLs. Endoscopic ultrasound (EUS)-guided therapeutic management of PCLs is a novel management strategy for patients with prohibitive surgical risks. Various ablation techniques have been explored in previous studies utilizing EUS-guided fine needle injection (FNI) of alcohol and chemotherapeutic agents. This review article focuses on EUS-FNI and chemoablation, encompassing the evolution of chemoablation, pancreatic cyst selection, chemotherapy drug selection, including novel agents, and a discussion of its safety and efficacy.

Pancreatic Cystic Lesions: A Focused Review on Cyst Clinicopathological Features and Advanced Diagnostics

Macroscopic, endomicroscopic, and histologic findings and correlation are an integral part of the diagnostic evaluation of pancreatic cystic lesions (PCLs), as complementing morphologic features seen by different specialties are combined to contribute to a final diagnosis. However, malignancy risk stratification of PCLs with worrisome features can still be challenging even after endoscopic ultrasound guided-fine needle aspiration (EUS-FNA) with cytological evaluation. This review aims to summarize cyst clinicopathological features from the pathologists' perspective, coupled with knowledge from advanced diagnostics-confocal laser endomicroscopy and cyst fluid molecular analysis, to demonstrate the state-of-art risk stratification of PCLs. This review includes illustrative photos of surgical specimens, endomicroscopic and histologic images, and a summary of cyst fluid molecular markers.

Historical perspective on needle development: From the past to the future

With the introduction of EUS, endoscopy was no longer limited to luminal indications. However, the method was unable to distinguish malignant from benign lesions. Consequently, needles designed for tissue acquisition under EUS-guidance was designed. Initially, the needles were designed for fine needle aspiration (FNA); nevertheless, with increased requirement for the precured tissue in terms of quality and quantity, newly design needles aimed at obtaining tissue cores for histological assessment were developed. Recent studies demonstrate superiority of these fine needle biopsy needles (FNB) compared to FNA needles.

Deep Learning for Automatic Differentiation of Mucinous versus Non-Mucinous Pancreatic Cystic Lesions: A Pilot Study

Endoscopic ultrasound (EUS) morphology can aid in the discrimination between mucinous and non-mucinous pancreatic cystic lesions (PCLs) but has several limitations that can be overcome by artificial intelligence. We developed a convolutional neural network (CNN) algorithm for the automatic diagnosis of mucinous PCLs. Images retrieved from videos of EUS examinations for PCL characterization were used for the development, training, and validation of a CNN for mucinous cyst diagnosis. The performance of the CNN was measured calculating the area under the receiving operator characteristic curve (AUC), sensitivity, specificity, and positive and negative predictive values. A total of 5505 images from 28 pancreatic cysts were used (3725 from mucinous lesions and 1780 from non-mucinous cysts). The model had an overall accuracy of 98.5%, sensitivity of 98.3%, specificity of 98.9% and AUC of 1. The image processing speed of the CNN was 7.2 ms per frame. We developed a deep learning algorithm that differentiated mucinous and non-mucinous cysts with high accuracy. The present CNN may constitute an important tool to help risk stratify PCLs.

MRI-Based Pancreatic Atrophy Is Associated With Malignancy or Invasive Carcinoma in Intraductal Papillary Mucinous Neoplasm

Background

Abrupt change in the caliber of the main pancreatic duct (MPD) with distal pancreatic atrophy (PA) was considered as one of worrisome features in the International Association of Pancreatology guideline and American College of Gastroenterology guideline for the management of intraductal papillary mucinous neoplasms (IPMNs). However, this feature was not included in other guidelines. Moreover, the association between PA alone and malignancy in IPMNs has not been fully evaluated. In the present study, we investigated the role of image-based PA in identifying malignant IPMNs or invasive carcinoma.

Methods

A total of 186 patients with IPMNs were included for analysis. The tumor size, location, MPD diameter, presence of a mural nodule (MN), and PA were evaluated using magnetic resonance imaging. Demographic information and serum carbohydrate antigen 19-9 and carcinoembryonic antigen (CEA) levels were also collected. IPMNs with high-grade dysplasia and associated invasive carcinoma were regarded as malignant IPMNs.

Results

PA was observed in 34 cases (18.3%). The occurrence of malignant IPMNs or invasive carcinoma in patients with PA were significantly higher than in those without PA (52.9% vs. 22.3%; 44.1% vs. 8.9%, all P < 0.01). Multivariate logistic regression analysis showed that PA was an independently associated factor for malignant IPMNs [odds ratio (OR) = 2.69, 95% confidence interval (CI): 1.07-6.78] or invasive carcinoma (OR = 7.78, 95%CI: 2.62-23.10) after modified with confounders. Subgroup analysis in MPD-involved IPMNs also indicated that PA was an independently associated factor for invasive carcinoma (OR = 9.72, 95%CI: 2.43-38.88). PA had a similar performance with MPD plus MN [the area under the curve (AUC) was both 0.71] in identifying malignancy. PA had a higher performance in identifying invasive carcinoma in MPD-involved IPMNs than MN (AUC = 0.71 vs. 0.65, P = 0.02).

Conclusion

Our data showed that imaging-based PA was associated with malignancy or invasive carcinoma regardless of abrupt change in the caliber of MPD in IPMNs. PA had an acceptable performance in identifying malignant IPMNs.

Application of Artificial Intelligence in the Management of Pancreatic Cystic Lesions

The rate of incidentally detected pancreatic cystic lesions (PCLs) has increased over the past decade and was recently reported at 8%. These lesions pose a unique challenge, as each subtype of PCL carries a different risk of malignant transformation, ranging from 0% (pancreatic pseudocyst) to 34–68% (main duct intraductal papillary mucinous neoplasm). It is imperative to correctly risk-stratify the malignant potential of these lesions in order to provide the correct care course for the patient, ranging from monitoring to surgical intervention. Even with the multiplicity of guidelines (i.e., the American Gastroenterology Association guidelines and Fukuoka/International Consensus guidelines) and multitude of diagnostic information, risk stratification of PCLs falls short. Studies have reported that 25–64% of patients undergoing PCL resection have pancreatic cysts with no malignant potential, and up to 78% of mucin-producing cysts resected harbor no malignant potential on pathological evaluation. Clinicians are now incorporating artificial intelligence technology to aid in the management of these difficult lesions. This review article focuses on advancements in artificial intelligence within digital pathomics, radiomics, and genomics as they apply to the diagnosis and risk stratification of PCLs.

Artificial intelligence implementation in pancreaticobiliary endoscopy

INTRODUCTION

Artificial intelligence has been rapidly deployed in gastroenterology and endoscopy. The acceleration of deep convolutional neural networks along with hardware development has allowed implementation of artificial intelligence algorithms into real-time endoscopy, particularly colonoscopy. However, artificial intelligence implementation in pancreaticobiliary endoscopy is nascent.

AREAS COVERED

Initial studies have been conducted in endoscopic retrograde pancreatography (ERCP), endoscopic ultrasound (EUS), and digital single operator cholangioscopy (DSOC). Machine learning has been implemented in identifying significant landmarks, including the ampulla on ERCP, and the bile duct, pancreas, and portal confluence on EUS. Moreover, artificial intelligence algorithms have been deployed in differentiating pathology including pancreas cancer, autoimmune pancreatitis, pancreatic cystic lesions, and biliary strictures.

EXPERT OPINION

There have been relatively few studies with limited sample sizes in developing these machine learning algorithms. Despite the early successful demonstration of artificial intelligence in pancreaticobiliary endoscopy, additional research needs to be conducted with larger data sets to improve generalizability and assessed in real-time endoscopy before clinical implementation. However, pancreaticobiliary endoscopy remains a promising avenue of artificial intelligence application with the potential to improve clinical practice and outcomes.

State-of-the-Art Update of Pancreatic Cysts

Andrew Canakis.

The development and clinical application of microscopic endoscopy for in vivo optical biopsies: Endocytoscopy and confocal laser endomicroscopy

Endoscopies are crucial for detecting and diagnosing diseases in gastroenterology, pulmonology, urology, and other fields. To accurately diagnose diseases, sample biopsies are indispensable and are currently considered the gold standard. However, random 4-quadrant biopsies have sampling errors and time delays. To provide intraoperative real-time microscopic images of suspicious lesions, microscopic endoscopy for in vivo optical biopsy has been developed, including endocytoscopy and confocal laser endomicroscopy. This article reviews recent advances in technology and clinical applications, as well as their shortcomings and future directions.

The role of artificial intelligence in pancreatic surgery: a systematic review

Artificial intelligence (AI), including machine learning (ML), is being slowly incorporated in medical practice, to provide a more precise and personalized approach. Pancreatic surgery is an evolving field, which offers the only curative option for patients with pancreatic cancer. Increasing amounts of data are available in medicine: AI and ML can help incorporate large amounts of information in clinical practice. We conducted a systematic review, based on PRISMA criteria, of studies that explored the use of AI or ML algorithms in pancreatic surgery. To our knowledge, this is the first systematic review on this topic. Twenty-five eligible studies were included in this review; 12 studies with implications in the preoperative diagnosis, while 13 studies had implications in patient evolution. Preoperative diagnosis, such as predicting the malignancy of IPMNs, differential diagnosis between pancreatic cystic lesions, classification of different pancreatic tumours, and establishment of the correct management for each of these lesions, can be facilitated through different AI or ML algorithms. Postoperative evolution can also be predicted, and some studies reported prediction models for complications, including postoperative pancreatic fistula, while other studies have analysed the implications for prognosis evaluation (from predicting a textbook outcome, the risk of metastasis or relapse, or the mortality rate and survival). One study discussed the possibility of predicting an intraoperative complication-massive intraoperative bleeding. Artificial intelligence and machine learning models have promising applications in pancreatic surgery, in the preoperative period (high-accuracy diagnosis) and postoperative setting (prognosis evaluation and complication prediction), and the intraoperative applications have been less explored.

Recent advances in the diagnostic evaluation of pancreatic cystic lesions

Pancreatic cystic lesions (PCLs) are becoming more prevalent due to more frequent abdominal imaging and the increasing age of the general population. It has become crucial to identify these PCLs and subsequently risk stratify them to guide management. Given the high morbidity associated with pancreatic surgery, only those PCLs at high risk for malignancy should undergo such treatment. However, current diagnostic testing is suboptimal at accurately diagnosing and risk stratifying PCLs. Therefore, research has focused on developing new techniques for differentiating mucinous from non-mucinous PCLs and identifying high risk lesions for malignancy. Cross sectional imaging radiomics can potentially improve the predictive accuracy of primary risk stratification of PCLs at the time of detection to guide invasive testing. While cyst fluid glucose has reemerged as a potential biomarker, cyst fluid molecular markers have improved accuracy for identifying specific types of PCLs. Endoscopic ultrasound guided approaches such as confocal laser endomicroscopy and through the needle microforceps biopsy have shown a good correlation with histopathological findings and are evolving techniques for identifying and risk stratifying PCLs. While most of these recent diagnostics are only practiced at selective tertiary care centers, they hold a promise that management of PCLs will only get better in the future.

Risk Stratification of Pancreatic Cysts With Confocal Laser Endomicroscopy

In the modern era of high-quality cross-sectional imaging, pancreatic cysts (PCs) are a common finding. The prevalence of incidental PCs detected on cross-sectional abdominal imaging (such as CT scan) is 3%-14% which increases with age, up to 8% in those 70 years or older. Although PCs can be precursors of future pancreatic adenocarcinoma, imaging modalities such as CT scan, MRI, or endoscopic ultrasound with fine-needle aspiration (EUS-FNA) are suboptimal at risk stratifying the malignant potential of individual cysts. An inaccurate diagnosis could potentially overlook premalignant lesions, which can lead to missed lesions, lead to unnecessary surveillance, or cause significant long-term surgical morbidity from unwarranted removal of benign lesions. Although current guidelines recommend an EUS or MRI for surveillance, they lack the sensitivity to risk stratify and guide management decisions. Needle-based confocal laser endomicroscopy (nCLE) with EUS-FNA can be a superior diagnostic modality for PCs with sensitivity and accuracy exceeding 90%. Despite this, a significant challenge to the widespread use of nCLE is the lack of adequate exposure and training among gastroenterologists for the real-time interpretation of images. Better understanding, training, and familiarization with this novel technique and the imaging characteristics can overcome the limitations of nCLE use, improving clinical care of patients with PCs. Here, we aim to review the types of CLE in luminal and nonluminal gastrointestinal disorders with particular attention to the evaluation of PCs. Furthermore, we discuss the adverse events and safety of CLE.

Efficacy and safety of EUS-guided through-the-needle microforceps biopsy sampling in categorizing the type of pancreatic cystic lesions

BACKGROUND AND AIMS

EUS-guided through-the-needle biopsy sampling (EUS-TTNB) using microbiopsy forceps is performed for the accurate diagnosis of pancreatic cystic lesions (PCLs). However, there are no standardized protocols for this procedure, and the amount of data on its efficacy is limited. Here, we evaluated the feasibility, efficacy, and safety of EUS-TTNB in categorizing the types of PCLs and identified the factors associated with diagnostic failure.

METHODS

The prospectively collected and maintained EUS-TTNB database at Asan Medical Center was reviewed to identify patients with PCLs who underwent EUS-TTNB between January 2019 and January 2021. The primary outcomes were technical success, diagnostic yield, and adverse events. Factors contributing to diagnostic failure and the discrepancies in the diagnosis made by conventional modalities (ie, EUS morphology, cross-sectional imaging, and cystic fluid analysis) were also evaluated.

RESULTS

Forty-five patients were analyzed. EUS-TTNB was successfully performed in all patients (technical success, 100%). Histologic diagnosis of PCLs was made in 37 patients (diagnostic yield, 82%). When comparing EUS-TTNB with a presumptive diagnosis, EUS-TTNB changed the diagnosis in 10 patients in terms of the categorization of the types of PCLs. The diagnostic yield was significantly higher in those who had 4 or more visible biopsy specimens per session (93%) than in those with fewer than 4 visible biopsy specimens per session (67%; P = .045). During follow-up, 3 patients (7%) experienced adverse events (2 acute pancreatitis, 1 intracystic bleeding), and no life-threatening adverse event occurred.

CONCLUSIONS

EUS-TTNB showed high technical feasibility, diagnostic yield, and good safety profile. EUS-TTNB may improve the categorization of the types of PCLs. Studies with standardized procedure protocols are needed to reduce the diagnostic failure for the types of PCLs.

Application of artificial intelligence to pancreatic adenocarcinoma

BACKGROUND AND OBJECTIVES

Pancreatic cancer (PC) is one of the deadliest cancers worldwide although substantial advancement has been made in its comprehensive treatment. The development of artificial intelligence (AI) technology has allowed its clinical applications to expand remarkably in recent years. Diverse methods and algorithms are employed by AI to extrapolate new data from clinical records to aid in the treatment of PC. In this review, we will summarize AI's use in several aspects of PC diagnosis and therapy, as well as its limits and potential future research avenues.

METHODS

We examine the most recent research on the use of AI in PC. The articles are categorized and examined according to the medical task of their algorithm. Two search engines, PubMed and Google Scholar, were used to screen the articles.

RESULTS

Overall, 66 papers published in 2001 and after were selected. Of the four medical tasks (risk assessment, diagnosis, treatment, and prognosis prediction), diagnosis was the most frequently researched, and retrospective single-center studies were the most prevalent. We found that the different medical tasks and algorithms included in the reviewed studies caused the performance of their models to vary greatly. Deep learning algorithms, on the other hand, produced excellent results in all of the subdivisions studied.

CONCLUSIONS

AI is a promising tool for helping PC patients and may contribute to improved patient outcomes. The integration of humans and AI in clinical medicine is still in its infancy and requires the in-depth cooperation of multidisciplinary personnel.

Management of pancreatic cysts and guidelines: what the gastroenterologist needs to know

The prevalence of pancreatic cysts has increased significantly over the last decade, partly secondary to increased quality and frequency of cross-sectional imaging. While the majority never progress to cancer, a small number will and need to be followed. The management of pancreatic cysts can be both confusing and intimidating due to the multiple guidelines with varying recommendations. Despite the differences in the specifics of the guidelines, they all agree on several high-risk features that should get the attention of any clinician when assessing a pancreatic cyst: presence of a mural nodule or solid component, dilation of the main pancreatic duct (or presence of main duct intraductal papillary mucinous neoplasm), pancreatic cyst size ⩾3–4 cm, or positive cytology on pancreatic cyst fluid aspiration. Other important criteria to consider include rapid cyst growth (⩾5 mm/year), elevated serum carbohydrate antigen 19-9 levels, new-onset diabetes mellitus, or acute pancreatitis thought to be related to the cystic lesion.

Early detection of pancreatic cancer: current state and future opportunities

PURPOSE OF REVIEW

Pancreatic ductal adenocarcinoma (PDAC) is third leading cause of cancer death in the United States, a lethal disease with no screening strategy. Although diagnosis at an early stage is associated with improved survival, clinical detection of PDAC is typically at an advanced symptomatic stage when best in class therapies have limited impact on survival.

RECENT FINDINGS

In recent years this status quo has been challenged by the identification of novel risk factors, molecular markers of early-stage disease and innovations in pancreatic imaging. There is now expert consensus that screening may be pursued in a cohort of individuals with increased likelihood of developing PDAC based on genetic and familial risk.

SUMMARY

The current review summarizes the known risk factors of PDAC, current knowledge and recent observations pertinent to early detection of PDAC in these risk groups and outlines future approaches that will potentially advance the field.

Artificial intelligence for pancreatic cancer detection: Recent development and future direction

Artificial intelligence (AI) has been increasingly utilized in medical applications, especially in the field of gastroenterology. AI can assist gastroenterologists in imaging-based testing and prediction of clinical diagnosis, for examples, detecting polyps during colonoscopy, identifying small bowel lesions using capsule endoscopy images, and predicting liver diseases based on clinical parameters. With its high mortality rate, pancreatic cancer can highly benefit from AI since the early detection of small lesion is difficult with conventional imaging techniques and current biomarkers. Endoscopic ultrasound (EUS) is a main diagnostic tool with high sensitivity for pancreatic adenocarcinoma and pancreatic cystic lesion. The standard tumor markers have not been effective for diagnosis. There have been recent research studies in AI application in EUS and novel biomarkers to early detect and differentiate malignant pancreatic lesions. The findings are impressive compared to the available traditional methods. Herein, we aim to explore the utility of AI in EUS and novel serum and cyst fluid biomarkers for pancreatic cancer detection.

A Review of the Diagnosis and Management of Premalignant Pancreatic Cystic Lesions

Pancreatic cystic lesions are an increasingly common clinical finding. They represent a heterogeneous group of lesions that include two of the three known precursors of pancreatic cancer, intraductal papillary mucinous neoplasms (IPMN) and mucinous cystic neoplasms (MCN). Given that approximately 8% of pancreatic cancers arise from these lesions, careful surveillance and timely surgery offers an opportunity for early curative resection in a disease with a dismal prognosis. This review summarizes the current evidence and guidelines for the diagnosis and management of IPMN/MCN. Current pre-operative diagnostic tests in pancreatic cysts are imperfect and a proportion of patients continue to undergo unnecessary surgical resection annually. Balancing cancer prevention while preventing surgical overtreatment, continues to be challenging when managing pancreatic cysts. Cyst fluid molecular markers, such as KRAS, GNAS, VHL, PIK3CA, SMAD4 and TP53, as well as emerging endoscopic technologies such as needle-based confocal laser endomicroscopy and through the needle microbiopsy forceps demonstrate improved diagnostic accuracy. Differences in management and areas of uncertainty between the guidelines are also discussed, including indications for surgery, surveillance protocols and if and when surveillance can be discontinued.

Pre-Operative Imaging and Pathological Diagnosis of Localized High-Grade Pancreatic Intra-Epithelial Neoplasia without Invasive Carcinoma

Pancreatic ductal adenocarcinoma (PDAC) arises from precursor lesions, such as pancreatic intra-epithelial neoplasia (PanIN) and intraductal papillary mucinous neoplasm (IPMN). The prognosis of high-grade precancerous lesions, including high-grade PanIN and high-grade IPMN, without invasive carcinoma is good, despite the overall poor prognosis of PDAC. High-grade PanIN, as a lesion preceding invasive PDAC, is therefore a primary target for intervention. However, detection of localized high-grade PanIN is difficult when using standard radiological approaches. Therefore, most studies of high-grade PanIN have been conducted using specimens that harbor invasive PDAC. Recently, imaging characteristics of high-grade PanIN have been revealed. Obstruction of the pancreatic duct due to high-grade PanIN may induce a loss of acinar cells replaced by fibrosis and lobular parenchymal atrophy. These changes and additional inflammation around the branch pancreatic ducts (BPDs) result in main pancreatic duct (MPD) stenosis, dilation, retention cysts (BPD dilation), focal pancreatic parenchymal atrophy, and/or hypoechoic changes around the MPD. These indirect imaging findings have become important clues for localized, high-grade PanIN detection. To obtain pre-operative histopathological confirmation of suspected cases, serial pancreatic-juice aspiration cytologic examination is effective. In this review, we outline current knowledge on imaging characteristics of high-grade PanIN.