Background & Aims
Tumour mutational burden (TMB) predicts improved response and survival to immunotherapy. In this pilot study, we optimized targeted next‐generation sequencing (tNGS) to estimate TMB in hepatocellular carcinoma (HCC).

Methods
We sequenced 48 non‐paired samples (21 fresh‐frozen [FF] and 27 paraffin‐embedded [FFPE]), among which 11 FFPE samples were pretreated with uracil‐DNA glycosylase (UDG). Thirty samples satisfied post‐sequencing quality control. High/low TMB was defined by median number of mutations/Mb (Mut/Mb), across different minimum allele frequency (MAF) thresholds (≥0.05, ≥0.1 and ≥0.2).

Results
Eligible patients (n = 29) were cirrhotic (84%) with TNM stage I‐II HCC (75%). FFPE samples had higher TMB (median 958.39 vs 2.51 Mut/Mb, P < .0001), estimated deamination counts (median 1335.50 vs 0, P < .0001) and C > T transitions at CpG sites (median 60.3% vs 9.1%, P = .002) compared to FF. UDG‐treated samples had lower TMB (median 4019.92 vs 353 Mut/Mb, P = .041) and deamination counts (median 6393.5 vs 328.5, P = .041) vs untreated FFPE. At 0.2 MAF threshold with UDG treatment, median TMB was 5.48 (range 1.68‐16.07) and did not correlate with salient pathologic features of HCC, including survival.

Conclusion
While tNGS on fresh HCC samples appears to be the optimal source of tumour DNA, the low median TMB values observed may limit the role of TMB as a predictor of response to immunotherapy in HCC.