Abstract Germline and somatic genetic testing have become cornerstones for assessing patient risk and tailoring cancer treatment, respectively, relying on the ability to distinguish pathogenic from benign variants. Typically, clinical reporting from these tests is conducted independently, serving distinct clinical objectives and audiences. However, integrating the underlying genomic data from both germline and somatic tests can enhance interpretative clarity, particularly in resolving variants of unknown significance (VUS) that lack sufficient evidence to determine pathogenicity. This ambiguity is especially problematic for non-White populations and those of non-European ancestry, where VUSs are disproportionately found, thereby limiting the scope of precision medicine and perpetuating health disparities. To assess the potential utility of an integrated model, we examined a prospective cohort of 88 self-identified Hispanic patients with colorectal cancer (CRC), performing both germline and somatic genomic analyses. We found that at least one germline VUS was identified in 44% (39/88) of participants, compared to a frequently reported rate of below 15%. By integrating tumor transcriptome data, mutational signatures, and copy number variations with traditional germline testing, we aimed to clarify VUSs and identify driver mutations. Integrating data modalities of the tumor transcriptome, mutational signatures, and copy number variations alongside traditional germline testing can help clarify VUSs and identify driver mutations. We constructed a model whereby the assessment of VUSs was based on the putative functional impact on either the DNA mutational patterns or allelic expression observed in tumor RNA. Using our integrated model, we deemed the oncogene AXIN2 VUS of one participant was likely biologically significant due to mutant-allele specific expression observed in the tumor RNA. Furthermore, we showed that 26% (10/39) of participants harbored a VUS that was very likely benign based on the tumor mutational profile and RNA expression. These included variants associated with homologous recombination in PALB2 and BRCA2 as well as genes associated with DNA repair like MLH1, PMS2, POLE and POLD1. Among patients with VUSs in genes associated with homologous recombination deficiency (HRD), none exhibited HRD via copy number analyses. Conversely, of the six patients that did exhibit HRD, only one had a pathogenic variant in a gene associated with HRD (ATM). Additionally, we identified tumors with a DNA mismatch repair deficiency phenotype (high tumor mutational burden, high microsatellite instability, DNA mismatch repair deficiency mutational signature) but without a pathogenic variant in a Lynch syndrome-associated gene. Methylation analysis revealed hypermethylation of the MLH1 promoter in these cases, explaining the tumor phenotype. This finding indicates that our integrative method can accurately predict DNA mismatch repair deficiency even without a genetic alteration in a DNA mismatch repair gene. Citation Format: Jonathan Amzaleg, Julie O. Culve, Charité N. Ricker, Natalia Gutierrez, Yuxin Jin, Brigette Waldrup, Carmen Chaves, Lucia Enriquez, Joel Sanchez-Mendez, Daisy Hernandez, Bodour Salhia, Lourdes Baezconde-Garbanati, Mariana C. Stern, Enrique Velazquez Villarreal, John D. Carpten, Heinz-Joseph Lenz, David W. Craig. Integrating germline and somatic data to resolve variants of uncertain significance (VUS) in colorectal cancer (CRC) patients who self-identified as Hispanic/Latino [abstract]. In: Proceedings of the 17th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2024 Sep 21-24; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2024;33(9 Suppl):Abstract nr C010.
Read full abstract