MLH1 single-nucleotide variant in circulating tumor DNA predicts overall survival of patients with hepatocellular carcinoma.

Liquid biopsy can provide a strong basis for precision medicine. We aimed to identify novel single-nucleotide variants (SNVs) in circulating tumor DNA (ctDNA) in patients with hepatocellular carcinoma (HCC). Deep sequencing of plasma-derived ctDNA from 59 patients with HCC was performed using a panel of 2924 SNVs in 69 genes. In 55.9% of the patients, at least one somatic mutation was detected. Among 25 SNVs in 12 genes, four frequently observed SNVs, MLH1 (13%), STK11 (13%), PTEN (9%), and CTNNB1 (4%), were validated using droplet digital polymerase chain reaction with ctDNA from 62 patients with HCC. Three candidate SNVs were detected in 35.5% of the patients, with a frequency of 19% for MLH1 chr3:37025749T>A, 11% for STK11 chr19:1223126C>G, and 8% for PTEN chr10:87864461C>G. The MLH1 and STK11 SNVs were also confirmed in HCC tissues. The presence of the MLH1 SNV, in combination with an increased ctDNA level, predicted poor overall survival among 107 patients. MLH1 chr3:37025749T>A SNV detection in ctDNA is feasible, and thus, ctDNA can be used to detect somatic mutations in HCC. Furthermore, the presence or absence of the MLH1 SNV in ctDNA, combined with the ctDNA level, can predict the prognosis of patients with HCC.