Barbara Burtness, MD of the Yale Cancer Center discusses the ASCO 2020 Abstract Correlation of tumor mutational burden (TMB) with CDKN2A and TP53 mutation in HPV-negative head and neck squamous cell carcinoma (HNSCC).
Context:
In HNSCC, the TP53 and CDKN2A tumor suppressors are usually mutated or lost, impairing G1 checkpoints. This decreases the ability to repair hypoxia-related DNA damage, replication stress, and mutagen exposure, thereby increasing TMB, a possible predictor of immunotherapy profit. TP53 mutations, with or without dominant-negative (DNE) involvement, gain-of-function (GOF), and benign, can be categorized as loss-of-function (LOF). We investigated whether unique TP53 mutation groups were correlated with increased TMB and whether they cooperated with elevated TMB CDKN2A mutations.
Methodology:
From 2015 to 2019, we analyzed 1010 HPV- HNSCC tumor samples (246 females) profiled by Caris Life Sciences with a 592-gene panel. Oral cavity (285), oropharynx (225), and larynx (153) were the predominant subsites. All somatic nonsynonymous missense mutations detected were reflected by TMB. The mean TMB per megabase (MB) is registered. The pathogenicity of mutations in TP53 and CDKN2A was determined in accordance with the guidelines of the American College of Medical Genetics (ACMG). Based on protein structure review, public databases (IARC, ClinVar, InterVar), and publications (PMID: 25108461 and others) evaluating structure-function relations, four alternative methods of characterizing TP53 mutations were also used.
Outcomes:
By ACMG guidelines, 60 percent of cases had TP53 mutations (TP53mut) designated pathogenic. Among the alternative classification methods, frequency estimates of LOF / DNE mutations ranged from 30-42.8 percent of cases. In 20 percent, damaging CDKN2A mutations were present. The mean TMB per MB in oral cavity cancers ranged from 8.2/8.6 (female/male) to 26.5/27.7 (female/male) in lip cancer. In the presence of damaging LOF / DNE TP53 mutations or CDKN2A mutations, the mean TMB was usually higher, but not TP53 GOF mutations. Based on ACMG, TMB was 8.03 for TP53 and CDKN2A wild type (WT) tumors, 9.82 for CDKN2Amut-only, 10.56 for TP53mut-only, and 17.6 for TP53 mut / CDKN2A mut / CDKN2A (p < 0.001). The mean TMB for destructive TP53mut (Poeta algorithm) was 8.67 for WT / WT, 11.31 for TP53mut, 17.9 for CDKN2Amut, and 15.83 for TP53mut / CDKN2A mut (p < 0.001).
Findings:
TP53 and/or CDKN2A mutations are associated with increased mean TMB relative to WT; mean TMB was highest in both genes for tumors carrying harmful mutations. The mutation of GOF TP53 was not specifically correlated with elevated TMB. Relevant TP53 / CDKN2A mutational status should also be evaluated, as TMB is evaluated as a predictive biomarker in HNSCC immunotherapy.