Artificial intelligence in endoscopic imaging for detection of malignant biliary strictures and cholangiocarcinoma: a systematic review.

Background: Artificial intelligence (AI), when applied to computer vision using a convolutional neural network (CNN), is a promising tool in "difficult-to-diagnose" conditions such as malignant biliary strictures and cholangiocarcinoma (CCA). The aim of this systematic review is to summarize and review the available data on the diagnostic utility of endoscopic AI-based imaging for malignant biliary strictures and CCA. Methods: In this systematic review, PubMed, Scopus and Web of Science databases were reviewed for studies published from January 2000 to June 2022. Extracted data included type of endoscopic imaging modality, AI classifiers, and performance measures. Results: The search yielded 5 studies involving 1465 patients. Of the 5 included studies, 4 (n=934; 3,775,819 images) used CNN in combination with cholangioscopy, while one study (n=531; 13,210 images) used CNN with endoscopic ultrasound (EUS). The average image processing speed of CNN with cholangioscopy was 7-15 msec per frame while that of CNN with EUS was 200-300 msec per frame. The highest performance metrics were observed with CNN-cholangioscopy (accuracy 94.9%, sensitivity 94.7%, and specificity 92.1%). CNN-EUS was associated with the greatest clinical performance application, providing station recognition and bile duct segmentation; thus reducing procedure length and providing real-time feedback to the endoscopist. Conclusions: Our results suggest that there is increasing evidence to support a role for AI in the diagnosis of malignant biliary strictures and CCA. CNN-based machine leaning of cholangioscopy images appears to be the most promising, while CNN-EUS has the best clinical performance application.

Artificial intelligence in medical imaging for cholangiocarcinoma diagnosis: A systematic review with scientometric analysis.

INTRODUCTION: Artificial intelligence (AI), by means of computer vision in machine learning, is a promising tool for cholangiocarcinoma (CCA) diagnosis. The aim of this study was to provide a comprehensive overview of AI in medical imaging for CCA diagnosis. METHODS: A systematic review with scientometric analysis was conducted to analyze and visualize the state-of-the-art of medical imaging to diagnosis CCA. RESULTS: Fifty relevant articles, published by 232 authors and affiliated with 68 organizations and 10 countries, were reviewed in depth. The country with the highest number of publications was China, followed by the United States. Collaboration was noted for 51 (22.0%) of the 232 authors forming five clusters. Deep learning algorithms with convolutional neural networks (CNN) were the most frequently used classifiers. The highest performance metrics were observed with CNN-cholangioscopy for diagnosis of extrahepatic CCA (accuracy 94.9%; sensitivity 94.7%; and specificity 92.1%). However, some of the values for CNN in CT imaging for diagnosis of intrahepatic CCA were low (AUC 0.72 and sensitivity 44%). CONCLUSION: Our results suggest that there is increasing evidence to support the role of AI in the diagnosis of CCA. CNN-based computer vision of cholangioscopy images appears to be the most promising modality for extrahepatic CCA diagnosis. Our social network analysis highlighted an Asian and American predominance in the research relational network of AI in CCA diagnosis. This discrepancy presents an opportunity for coordination and increased collaboration, especially with institutions located in high CCA burdened countries.

A cell type-aware framework for nominating non-coding variants in Mendelian regulatory disorders.

Unsolved Mendelian cases often lack obvious pathogenic coding variants, suggesting potential non-coding etiologies. Here, we present a single cell multi-omic framework integrating embryonic mouse chromatin accessibility, histone modification, and gene expression assays to discover cranial motor neuron (cMN) cis-regulatory elements and subsequently nominate candidate non-coding variants in the congenital cranial dysinnervation disorders (CCDDs), a set of Mendelian disorders altering cMN development. We generated single cell epigenomic profiles for ~86,000 cMNs and related cell types, identifying ~250,000 accessible regulatory elements with cognate gene predictions for ~145,000 putative enhancers. Seventy-five percent of elements (44 of 59) validated in an in vivo transgenic reporter assay, demonstrating that single cell accessibility is a strong predictor of enhancer activity. Applying our cMN atlas to 899 whole genome sequences from 270 genetically unsolved CCDD pedigrees, we achieved significant reduction in our variant search space and nominated candidate variants predicted to regulate known CCDD disease genes MAFB, PHOX2A, CHN1, and EBF3 - as well as new candidates in recurrently mutated enhancers through peak- and gene-centric allelic aggregation. This work provides novel non-coding variant discoveries of relevance to CCDDs and a generalizable framework for nominating non-coding variants of potentially high functional impact in other Mendelian disorders.