Noninvasive Detection of Hepatocellular Carcinoma with Circulating Tumor DNA Features and α-Fetoprotein

Liver cancer is the fifth-most common cancer worldwide, with the third-highest rate of cancer-related mortality. Hepatocellular carcinoma (HCC) is the leading pathologic subtype, contributing 85% to 90% of cases of primary liver cancer. Most HCC patients are diagnosed at an advanced stage at which treatment is not curative. This study assessed the performance of a newly developed blood-based assay that utilizes genomic features and protein markers for the early detection of HCC. Two cancer-associated hallmarks, copy-number aberrations (CNA) and fragment size (FS), were characterized by shallow whole-genome sequencing of cell-free DNA and utilized to differentiate cancer patients from healthy subjects. As a clinically implemented biomarker of HCC, plasma α-fetoprotein (AFP) was also used with the genomic surrogates to optimize the detection of HCCs. The sensitivity of AFP ≥20.0 μg/L in detecting HCC was 57.9%. The combined genomic classifier CNA + FS via cell-free DNA shallow whole-genome sequencing identified nearly half of AFP-negative HCC patients (43.8%). By integrating CNA, FS as well as AFP (HCCseek), 75.0% sensitivity was achieved at 98.0% specificity, resulting in 92.6% accuracy, with 58.6% sensitivity in stage I HCC. The quantitative output of HCCseek was correlated with the severity of the disease (tumor size, stage, and recurrence-free survival). In summary, this study describes an efficient, noninvasive, and cost-effective method to detect HCC.