Liquid biopsy and radiomics for liver cancer surveillance

ABSTRACT Hepatocellular carcinoma (HCC) is the fastest growing cause of cancer death in the United States. While prevention efforts are paramount, most patients succumb to advanced HCC disease. Thus, enrollment of at- risk patients (e.g., cirrhosis of any etiology) in early detection programs is recommended in clinical practice

guidelines. Longstanding challenges to improving early-stage HCC detection are suboptimal performance of the recommended surveillance tools [i.e., abdominal ultrasound and serum alpha-fetoprotein (AFP)] and the low implementation rate of surveillance programs (as low as 25% in the United States). Various studies have

tried to utilize tumor nucleic acids released to the bloodstream (i.e., “liquid biopsy”) as novel early HCC detection tools, but its role in this clinical setting is largely unexplored. Up to 18% of patients with cirrhosis have indeterminate nodules detected during surveillance. In these patients, imaging is inconclusive, and patients

require either a biopsy of the suspicious nodule and/or close imaging follow-up. Our project is designed to overcome these problems by using new blood-based liquid biopsy biomarkers and magnetic resonance imaging (MRI)-based radiomics. We have assembled a multi-institutional Translational Research Center including leading academic centers in NYC (Mount Sinai, Columbia, Cornell, and Montefiore).

We plan to collect blood, clinical and imaging data from a multiracial cohort of 2,560 patients (early HCC cases and controls at high risk). In Aim 1, we will determine the clinical role of new liquid biopsy technologies (i.e., cell-free DNA fragment analysis and a 3-small RNA signatures from extracellular vesicles in plasma) as a

novel surveillance approach in HCC. In Aim 2, we will integrate MRI-based radiomics models with our liquid biopsy technologies to better characterize indeterminate nodules in cirrhosis. Our project is timely and uniquely poised to respond to the imperative of developing noninvasive biomarkers of early HCC detection.