Abstract #2599

Characterization of NAFLD disease evolution with multiparametric MRI/MRE and multi-state Hidden Markov Model

Jingbiao Chen^1,2, Yunyi Kang³, Feng Ju³, Jiahui Li¹, Jie Chen¹, Xin Lu¹, Richard L Ehman¹, Vijay H Shah², and Meng Yin¹

¹Radiology, Mayo Clinic, Rochester, MN, United States, ²Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN, United States, ³School of Computing, Informatics, and Decision Systems Engineering, Arizona State University, Tempe, AZ, United States

Non-invasively diagnosing non-alcoholic fatty liver disease (NAFLD) and predicting disease prognosis in individual patients are two main unmet clinical needs. There are very few longitudinal studies that evaluate imaging, biochemical, and histopathological variables that predict disease progression of NAFLD. This study established a multi-state Hidden Markov model (HMM) of NAFLD evolution in an animal model with three imaging biomarkers: MRI derived proton density fat fraction (PDFF) and MR elastography (MRE) assessed liver stiffness (LS) and loss modulus (LM). Results have shown that a 3-state HMM can well characterize the natural history of NAFLD, and predict disease progression or regression.

This abstract and the presentation materials are available to members only; a login is required.

Join Here