Circulating Biomarkers and Hepatocellular Carcinoma in a High-risk Population

Description

Circulating cell-free DNA (ccfDNA) shed from solid tumors provides an opportunity to detect earlystage cancer non-invasively. Several ccfDNA signatures have been studied for cancer detection, such as somatic mutation, nucleosome footprint, genome-wide fragmentation profile, fragment end sequence, and methylation. However, few ccfDNA biomarkers have been validated in a longitudinal, populationbased setting. Furthermore, lessons learnt from other screening strategies, such as prostate-specificantigen testing, indicate the importance of differentiating aggressive tumors from non-consequential lesions to minimize overtreatment. Telomeres, located at the terminal of linear chromosomes, affect genome integrity, cell immortalization, and cancer development. Shortening of telomeres is closely related to the number of cell divisions and occurs during the early development of cancer. We hypothesize that accelerated proliferation of the cancer genome and accompanying telomere shortening and shedding may increase the abundance of circulating telomere. we have developed a high-preservation technology for ccfDNA analysis.Hepatocellular carcinoma (HCC) is the third most common cause of cancer mortality globally, estimated to cause approximately 830,000 deaths in 2020. More than 50% of HCC cases worldwide are attributable to chronic hepatitis B virus (HBV) infection. Current international guidelines recommend HCC screening among high-risk individuals, including patients with chronic HBV infection or cirrhosis. However, the effectiveness of HCC screening is hampered by the lack of highly sensitive, specific, andefficient tools.We aim to first identify biomarkers using a cross-sectional case-control study on hospital HCC patients and apply a machine learning approach to model circulating ccfDNA biomarkers including telomere traces detected by our high-preservation technology – a decade of work on using trace amount of starting DNA materials for highly sensitive sequencing. Then, we will evaluate the performance of the model in an independent, prospective, population-based cohort, with over 8 years of follow-up of nearly 3,000 high-risk individuals who were positive for HBV infection at baseline. Positive predictive value (PPV) and negative predictive value (NPV) are key in evaluating the effectiveness of a screening program. We will set a high specificity value at least 98% to emphasize PPV that is important for a population screening program to be cost-effective. Furthermore, we aim to evaluate the predictive power of the potential biomarker model in HCC patient survival, to understand whether the biomarker not only can detect cancer early but, more importantly, able to detect those tumors destine to be aggressive HCC for timely action and leaving non-consequential ones alone.