Quantification of Liver Fibrosis with MRI and Deep Learning - Project Summary/Abstract Chronic liver disease (CLD) is a common cause of morbidity and mortality in the U.S. and throughout the world. In 2017, CLD had an age-adjusted death rate of 10.9/100,000 total population and an estimated lifetime cost of fatty liver disease alone in the U.S. of ~$222 billion. Liver fibrosis (LF) is the most important and only histologic feature known to predict outcomes from CLD. The current standard for assessing LF is biopsy, which is costly, prone to sampling error, and invasive with poor patient acceptance. Thus, there is an urgent unmet need for noninvasive, highly accurate and precise diagnostic technologies for detection and quantification of LF. Our overarching objective is to apply Deep Learning (DL) methods using conventional non-elastographic magnetic resonance (MR) images, MR elastography (MRE), and clinical data to accurately detect and measure LF in children and adults with CLD, using biopsy-derived histologic data as the reference standard. In this project, we will dedicate our efforts to accomplishing the following specific aims. In Aim 1, we will develop and validate a DL framework to accurately segment liver and spleen in order to extract radiomic (gray-scale signal intensity distribution, shape and morphology, volumetry, and inter-voxel signal intensity pattern and texture) and deep features (complex abstractions of patterns non-linearly constructed throughout the transformation estimated by data-driven DL training procedures) from conventional multiparametric MRI. These features allow detection of liver and spleen structural abnormalities/tissue aberrations. In Aim 2, we will develop and validate an “ensemble” DL model (LFNet) to predict biopsy-derived LF stage and LF percentage using the integration of conventional multimodal MRI radiomic and deep features, MRE data, as well as clinical data. In Aim 3, we will develop and validate a DL model (LSNet) to quantify MRE-derived liver stiffness (LS) using conventional multiparametric MRI radiomic and deep features as well as clinical data. The proposed models will help physicians to more accurately detect and follow CLD by 1) quantifying LS from conventional MR imaging without the need for MRE; and, more importantly, 2) predicting histologic LF stage and LF percentage without the need for biopsy, while avoiding inter- radiologist variability, reducing radiologist workload, and ultimately reducing healthcare costs. We will validate the models using both internal and independent external data from various scanners and sites. The techniques we develop are expected to improve medical diagnosis and prognostication in the same way as DL has revolutionized other fields. This study will significantly impact public health because it will allow physicians and researchers to more accurately diagnose and quantify CLD and LF as well as permit more frequent assessments in a noninvasive, patient-centric manner, thus potentially improving patient outcomes while lowering healthcare costs. The techniques we develop also can be readily extended for the prediction of other important liver-related clinical outcomes, including impending complications such as portal hypertension, time to liver transplant/transplant listing, and mortality risk, among others.
