AlphaMissense for Identifying Pathogenic Missense Mutations in DNA Damage Repair Genes in Cancer

Abstract

Purpose: AlphaMissense is a new artificial intelligence-based approach to predicting the pathogenicity of missense variants. However, whether its predictions can be directly applied to clinical decision making remains unclear. This study aimed to evaluate the accuracy of AlphaMissense predictions for DNA damage repair (DDR) genes using genomic and clinical characteristics.

Methods: Sequencing data from 56,965 patients with cancer who underwent Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) testing between April 2015 and March 2023 and data from The Cancer Genome Atlas (TCGA)/Pan-Cancer Analysis of Whole Genomes (PCAWG) were analyzed. AlphaMissense pathogenicity was evaluated in six commonly mutated DDR genes. Missense mutations were classified into three categories on the basis of AlphaMissense and OncoKB: known pathogenic, newly identified pathogenic by AlphaMissense, or benign.

Results: In the MSK-IMPACT cohort, 1,182 (17.5%) of 6,743 unique DDR gene missense mutations were newly identified as pathogenic. In breast, ovarian, pancreatic, and prostate cancers, homologous recombination (HR)-deficiency signatures were more common in tumors with new pathogenic missense mutations in BRCA1/2, PALB2, and RAD51C than in benign missense mutations (66.7% v 35.2%, P = .021). In the TCGA/PCAWG data set, HR-deficiency signatures were also more frequent in tumors with new pathogenic mutations than in wild-type (46.2% v 19.2%, P = .036). For tumors with POLE missense mutations, there were no significant differences in tumor mutation burden and POLE-associated signatures between new pathogenic and benign mutations. Patients with new pathogenic ATM missense mutations had fewer TP53 mutations (30.5% v 54.6%, P < .001) and showed improved irradiated tumor control (hazard ratio, 0.58 [95% CI, 0.35 to 0.95]; P = .03) compared with those with benign missense mutations.

Conclusion: Our findings suggest that AlphaMissense can help identify previously unknown pathogenic DDR gene mutations, but its accuracy is gene-dependent. AlphaMissense prediction still requires additional confirmation with clinical and functional validation.

FIG 1.

Frequency of missense mutations in DDR genes. Frequency of missense mutations in selected DDR genes identified in MSK-IMPACT cohort stratified according to pathogenicity classification. DDR, DNA damage repair; MSK-IMPACT, Memorial Sloan Kettering–Integrated Mutation Profiling of Actionable Cancer Targets.

FIG 2.

Genomic characteristics in tumors with missense mutations in core-HR genes. Association of signature three positivity and biallelic mutations in core-HR genes (BRCA1/2, PALB2, and RAD51C) in BOPP tumors in (A) the MSK-IMPACT cohort and (B) the TCGA/PCAWG cohort (P values calculated using the chi-square test; ns, nonsignificant, ≥.05). Biallelic missense mutation locations in (C) BRCA1 and (D) BRCA2 in the MSK-IMPACT cohorts stratified by signature three positivity (top Sig3-positive; bottom Sig3-negative). BOPP, breast, ovary, pancreatic, and prostate; BRCT, BRCA1 C terminus; DBD, DNA-binding domain, HR, homologous recombination; MSK-IMPACT, Memorial Sloan Kettering–Integrated Mutation Profiling of Actionable Cancer Targets; TCGA/PCAWG, The Cancer Genome Atlas/Pan-Cancer Analysis of Whole Genomes.

FIG 3.

Genomic characteristics in tumors with POLE missense mutations. Association of (A) TMB, (B) POLE signature, and POLE ED missense mutations stratified by pathogenicity (P values calculated using the Wilcoxon signed rank test; ns, nonsignificant, ≥.05; ****, <.0001). (C) Mutation sites in POLE missense mutations in the MSK-IMPACT cohort. ^aA456P was known pathogenic and A456T was new pathogenic; ^bM471V was benign and M471I was new pathogenic. AA, amino acid; ED, exonuclease domain; MSK-IMPACT, Memorial Sloan Kettering–Integrated Mutation Profiling of Actionable Cancer Targets; TMB, tumor mutation burden.

FIG 4.

Genomic and clinical characteristics in tumors with ATM missense mutations. Frequency of (A) FAT/kinase domain mutations and (B) co-occurrent TP53 mutations in known pathogenic versus new pathogenic versus benign ATM missense mutations in MSK-IMPACT cohort (P values calculated using the chi-square test; ns, nonsignificant, ≥.05). (C) Cumulative incidence of irradiated tumor progression by pathogenicity of ATM missense mutations. (D) Exceptional responder to RT in a patient with urothelial cancer with ATM new pathogenic mutation (p.E158K). Abdominal MRI and PET scan of gastric metastases showing near-complete response 7 months after RT. The red circles highlight a target lesion. FAT, FRAP, ATM and TRRAP; MRI, magnetic resonance imaging; MSK-IMPACT, Memorial Sloan Kettering–Integrated Mutation Profiling of Actionable Cancer Targets; PET, positron emission tomography; ref, reference; RT, radiotherapy.