Using Artificial Intelligence to Predict Cirrhosis From Computed Tomography Scans

Abstract

Background: Undiagnosed cirrhosis remains a significant problem. In this study, we developed and tested an automated liver segmentation tool to predict the presence of cirrhosis in a population of patients with paired liver biopsy and CT scans. Methods: We utilized a cohort of 1590 CT scans within the Morphomics database to train an automated liver segmentation model using 3D-U-Net and Google’s DeeplLabv3+. Imaging features were then automatically calculated from an external test cohort of patients with chronic liver disease who had a paired liver biopsy and CT within 6 months of each other in Jan 2004-2012. Using gradient boosting decision trees, we developed multivariate models to predict presence of histologic cirrhosis and evaluated with five fold cross validated c-statistic. Results: Our cohort had 351 patients, 96 patients had cirrhosis. Of the total cohort, 72 were post liver transplant. Both FIB-4 and liver morphomics alone performed equally well with AUROC of 0.76 (95% CI 0.70-0.81) and 0.71 (95% CI 0.65-0.76), respectively (p = 0.2). However, the combination of liver morphomics with laboratory values or liver morphomics with laboratory and demographic data resulted in significant improved performance with AUROC of 0.84 (0.80-0.89) and 0.85 (0.81-0.90) compared to FIB-4 alone (p <0.001). In a subgroup analysis, we also examined performance in patients without liver transplantation and saw similar augmentation of FIB-4. Conclusion: This proof of principle study demonstrates that automatically extracted features within CT scans can be combined with classic electronic medical record data to improve prediction of cirrhosis in patients with liver disease. This tool may be utilized in both pre and post-transplant patients and has the potential to improve our ability to detect undiagnosed cirrhosis.

The Use of Artificial Intelligence in Gastroenterology: A Glimpse Into the Present

“…Another exciting application of AI is its ability to improve population health screening. In this month's issue, you'll read about machine learning's utility in searching through electronic health records to predict who may develop incident Barrett's esophagus and esophageal adenocarcinoma ( 6 ) and to review cross-sectional imaging to identify patients with cirrhosis ( 7 ). In addition, the promise of AI to save us time in our daily activities is also addressed.…”