Detecting likely germline variants during tumor-based molecular profiling

Abstract

As the use of molecular profiling of tumors expands, cancer diagnosis, prognosis, and treatment planning increasingly rely on the information it provides. Although primarily designed to detect somatic variants, next-generation sequencing (NGS) tumor-based profiling also identifies germline DNA alterations, necessitating careful clinical interpretation of the data. Traditionally, germline risk testing has depended on prioritizing individuals based on physical exam findings consistent with known hereditary cancer syndromes, tumor-specific features, age at diagnosis, personal history, and family history. As NGS-based molecular profiling is used increasingly to diagnose, prognosticate, and follow cancer progression, DNA variants that are likely to be of germline origin are identified with increased frequency. Because pathogenic/likely pathogenic germline variants are critical biomarkers for risk stratification and treatment planning, consensus guidelines are expanding to recommend comprehensive germline testing for more cancer patients. This Review highlights the nuances of identifying DNA variants of potential germline origin incidentally at the time of NGS-based molecular profiling and emphasizes key differences between comprehensive germline versus tumor-based platforms, sample types, and analytical methodologies. In the growing era of precision oncology, clinicians should be adept at navigating these distinctions to optimize testing strategies and leverage insights regarding germline cancer risk surveillance and management for all people with cancer.

Figure 1. DNA repair pathways and additional mechanisms underlying hereditary cancer risk.

(A) Defects in the homologous recombination repair (HRR) genes BRCA1 and BRCA2 impair the accurate repair of double-strand DNA breaks, resulting in error-prone mechanisms like single-strand annealing, alternative end joining, and non-homologous end joining, leading to increased genomic instability and the accumulation of somatic variants. HRR deficiency is associated with several tumor types, including breast, ovarian, prostate, and pancreatic cancers. (B) Defects in the mismatch repair (MMR) genes MLH1, MSH2, MSH6, and PMS2 impair the repair of DNA replication errors, leading to microsatellite instability and genome-wide hypermutation. MMR deficiency is commonly associated with Lynch syndrome, predisposing individuals to a variety of cancers, including colorectal, endometrial, and ovarian cancers. Similar to HRR defects, MMR defects result in a mutator phenotype that drives tumorigenesis by allowing the accumulation of somatic variants. (C) Additional pathways implicated in hereditary cancer risk are shown, highlighting commonly altered cancer susceptibility genes recommended for further germline evaluation by the ACMG and the ESMO PMWG when identified on tumor-based profiling. Corresponding clinical phenotypes and penetrance estimates are detailed in Table 1.

Figure 2. Proposed algorithm to aid the identification of likely germline variants during tumor-based profiling.

In a patient with a probable malignancy (i.e., before pathologic confirmation), a thorough clinical history should be performed to determine risk. High-risk patients are prioritized for paired tumor-normal sequencing to differentiate between somatic and germline alterations. For patients not identified as high-risk who undergo tumor-based profiling, variant allele frequency (VAF) thresholds and pathogenicity classifications should guide confirmatory germline testing. In patients with a VAF less than 0.3 who experience a change in disease status, clinicians should evaluate the appropriateness of additional germline testing based on the clinical context. Confirmed germline variants should prompt targeted clinical action. Shared decision-making is essential, particularly when prognosis is limited or results are unlikely to influence management. In the future (indicated by the dashed line), we anticipate standardized integration of simultaneous germline testing and tumor-based profiling at the time of diagnosis. ^AGuidelines from ASCO and NCCN recommend universal germline testing in patients with epithelial ovarian cancer, pancreatic adenocarcinoma, metastatic or high-risk prostate cancer, pleural mesothelioma, adrenocortical carcinoma, pheochromocytomas, or paragangliomas, regardless of age, family history, or personal history (, , –109). ^BGene selection for multi-gene panels should adhere to ASCO guidelines, which consider clinical relevance, actionability, penetrance, and associations with hereditary cancer syndromes (106). ^CFor paired tumor-normal testing, best practice entails using cultured skin for germline analysis. ^DThis VAF threshold is derived from ESMO PMWG, which designated a cutoff of ≥0.3 for single-nucleotide variants and ≥0.2 for indels (48). Notably, VAF alone, particularly when derived from a tumor sample at a single time point, remains insufficient to determine germline origin. ^EVAF may fluctuate across time points because of technical variation, tumor heterogeneity, or biological phenomena such as allelic loss or subclonal architecture. Thus, redemonstration of the same P/LP variant on repeat tumor sequencing, independent of VAF, warrants consideration for germline evaluation (66, 106).

Figure 3. Comparison of germline, tumor-based, and paired tumor-normal sequencing.

Paired tumor-normal sequencing, the gold standard for confirming incidental germline findings, involves simultaneous sequencing of tumor and matched normal, non-malignant tissue from the same patient. This approach identifies inherited variants to generate a germline report. Tumor-based profiling analyzes DNA from tumor cells, reporting both somatic alterations and potential germline variants. Germline sequencing should be performed using non-hematopoietic tissue, with cultured skin fibroblasts considered the gold standard. Alternative sources are depicted in descending order based on DNA quality, ease of collection, and interpretive reliability. By subtracting germline variants from the tumor sequence, paired testing differentiates between somatic and germline variants, improving specificity and reducing false-positive somatic calls. Unlike tumor-based profiling, paired tumor-normal sequencing provides direct discrimination between somatic and germline variants.