Residency · Residency · Medical Genetics Genomics

Tumor Genomic Profiling and Somatic Variant Interpretation

Overview

Tumor genomic profiling involves sequencing tumor tissue to identify somatic genetic alterations that drive cancer and may be therapeutically targetable. The shift from single-gene testing to comprehensive genomic profiling (CGP) has transformed oncology practice. Key applications include identifying actionable mutations for targeted therapy, predicting immunotherapy response, prognostication, and clinical trial matching. The distinction between somatic-only and tumor-normal paired analysis has important implications for detecting germline findings.

Principles of Somatic Variant Calling

Tumor Sample Requirements

Formalin-fixed paraffin-embedded (FFPE) tissue is the standard sample type for tumor genomic profiling. Reliable variant detection typically requires a minimum tumor cellularity (tumor purity) exceeding 20%. Pre-analytical factors including fixation time, tissue age, necrosis, and DNA quality can affect results. Liquid biopsy using circulating tumor DNA (ctDNA) is increasingly employed when tissue is unavailable or for longitudinal monitoring.

Somatic vs. Germline Distinction

Somatic mutations are acquired in tumor cells and are not inherited. Without a matched normal sample, distinguishing somatic from germline variants requires bioinformatic filtering using population databases such as gnomAD. Tumor-only sequencing carries a higher risk of misclassifying germline variants as somatic, especially in patients from underrepresented populations whose variants are less well-characterized in existing databases. Paired tumor-normal sequencing subtracts germline variants to yield true somatic calls and simultaneously identifies germline pathogenic variants. ACMG and CGC recommend paired analysis when possible.

Variant Calling Pipeline

Reads are aligned to the reference genome (hg38), and variant calling algorithms designed for somatic mutations account for low allele frequencies and tumor heterogeneity. Variant allele frequency (VAF) interpretation requires clinical context: a high VAF near 50% may indicate a germline variant or a clonal somatic event in a high-purity tumor, while a low VAF below 10% suggests a subclonal somatic event. VAF must always be interpreted in the context of tumor purity and ploidy. Dedicated algorithms detect structural variants, copy number alterations, and gene fusions.

<image>Schematic of tumor genomic profiling workflow from tissue acquisition through DNA extraction, library preparation, sequencing, bioinformatic analysis, and clinical reporting with key quality checkpoints labeled</image>

Key Biomarkers in Tumor Genomic Profiling

Tumor Mutational Burden (TMB)

TMB is defined as the number of somatic mutations per megabase of sequenced DNA. TMB-High (10 or more mutations per megabase) is an FDA-approved biomarker for pembrolizumab across solid tumors under a tumor-agnostic approval. It correlates with neoantigen load and immunotherapy response in many tumor types. However, TMB thresholds vary by assay, tumor type-specific cutoffs may be more appropriate, and TMB alone is not perfectly predictive. Extremely high TMB may indicate mismatch repair deficiency or POLE/POLD1 proofreading domain mutations.

Microsatellite Instability (MSI)

MSI-High status, caused by deficient mismatch repair (dMMR), results from either somatic mechanisms (MLH1 promoter hypermethylation) or germline causes (Lynch syndrome). MSI-H/dMMR is an FDA-approved tumor-agnostic biomarker for pembrolizumab in solid tumors. Detection methods include PCR-based testing (Bethesda markers), immunohistochemistry for MMR proteins (MLH1, MSH2, MSH6, PMS2), and NGS-based MSI calling, which is increasingly incorporated into standard genomic panels.

PD-L1 Expression

PD-L1 expression is assessed by immunohistochemistry rather than genomic profiling per se, though it is often integrated into tumor profiling reports. Different scoring systems (tumor proportion score, combined positive score) apply to different tumor types and antibody clones. PD-L1 is complementary to but distinct from TMB and MSI as an immunotherapy biomarker.

Actionable Alterations and Targeted Therapy

Oncogene-Driven Cancers

Major actionable alterations include EGFR mutations in non-small cell lung cancer (exon 19 deletions, L858R, treated with osimertinib), ALK fusions (treated with alectinib or lorlatinib), BRAF V600E across multiple tumor types (treated with BRAF/MEK inhibitor combinations), HER2 amplification or mutations (trastuzumab, T-DXd), KRAS G12C (sotorasib, adagrasib -- historically considered "undruggable"), NTRK fusions (tumor-agnostic approval for larotrectinib and entrectinib), RET fusions and mutations (selpercatinib, pralsetinib), ROS1 fusions (crizotinib, entrectinib), PIK3CA mutations in breast cancer (alpelisib), and FGFR alterations in urothelial carcinoma and cholangiocarcinoma (erdafitinib, pemigatinib).

Tumor Suppressor Alterations

Tumor suppressor losses are often not directly targetable but inform prognosis and indirect therapeutic strategies. Somatic BRCA1/2 mutations confer PARP inhibitor sensitivity in ovarian, breast, prostate, and pancreatic cancers. TP53 mutations are prognostic in many tumor types but have limited direct therapeutic options. RB1 loss has implications for CDK4/6 inhibitor resistance in breast cancer. PTEN loss activates the PI3K/AKT pathway and may predict sensitivity to pathway inhibitors.

<image>Table of FDA-approved tumor-agnostic biomarkers and therapies including MSI-H/dMMR with pembrolizumab, NTRK fusions with larotrectinib and entrectinib, TMB-High with pembrolizumab, and RET fusions with selpercatinib</image>

Molecular Tumor Boards

Structure and Function

Molecular tumor boards are multidisciplinary teams comprising medical oncologists, molecular pathologists, clinical geneticists, bioinformaticians, genetic counselors, and pharmacists. They review complex cases where genomic findings require expert interpretation, match patients to appropriate targeted therapies, immunotherapies, or clinical trials, discuss variants of uncertain significance in the somatic context, and address incidental germline findings with appropriate referral pathways.

Evidence Tiering for Somatic Variants

Knowledgebases such as OncoKB and CIViC provide evidence-based annotation of somatic variants. The AMP/ASCO/CAP tier classification system categorizes variants as Tier I (strong clinical significance with FDA-approved therapy or guideline inclusion), Tier II (potential clinical significance with clinical trial evidence or off-label use), Tier III (unknown clinical significance), or Tier IV (benign or likely benign). The distinction between companion diagnostic variants (required for drug prescription) and complementary diagnostic variants (informative but not required) is clinically important.

Comprehensive Genomic Profiling Platforms

Commercial Platforms

FoundationOne CDx covers 324 genes and holds FDA approval for multiple companion diagnostic indications. Tempus xT/xF offers a large gene panel with matched RNA sequencing. MSK-IMPACT covers 505 genes, is FDA-authorized, and includes a matched normal sample. Guardant360 CDx is a liquid biopsy platform covering 74 genes with FDA approval for EGFR, BRCA, ALK, and KRAS G12C. Caris Molecular Intelligence provides multi-omic profiling including IHC, ISH, and NGS.

Tissue vs. Liquid Biopsy

Tissue remains the gold standard, capturing the full genomic landscape including amplifications and structural variants. Liquid biopsy (ctDNA) offers advantages of non-invasiveness, capture of tumor heterogeneity, and the ability for serial monitoring. Its limitations include lower sensitivity for early-stage or CNS tumors with low ctDNA shedding and less reliable detection of amplifications. Applications include treatment response monitoring, minimal residual disease detection, and resistance mechanism identification. A negative ctDNA result does not rule out actionable mutations, and tissue testing may still be needed.

<image>Comparison diagram of tumor-only versus paired tumor-normal sequencing approaches showing differences in variant classification accuracy, germline detection capability, and clinical implications</image>

Challenges in Somatic Variant Interpretation

Clonal Hematopoiesis of Indeterminate Potential (CHIP)

CHIP involves age-related somatic mutations in hematopoietic cells (commonly in DNMT3A, TET2, ASXL1, TP53, PPM1D) that can be detected in blood-derived "normal" samples, complicating paired tumor-normal analysis. CHIP can also appear in tumor tissue if contaminated with blood or if the tumor is hematopoietic in origin. Prevalence increases with age, reaching approximately 10% by age 70. These mutations must be distinguished from true tumor-derived alterations.

Tumor Heterogeneity

Intratumoral heterogeneity means that different subclones within a tumor may carry different mutations. Spatial heterogeneity limits what a single biopsy can capture of the full mutational landscape. Temporal heterogeneity reflects tumor evolution under therapy, leading to resistance mutations. Multi-region sequencing or liquid biopsy may better capture this complexity.

Interpretation in Diverse Populations

Population-specific germline variants may be misclassified as somatic in tumor-only testing. Underrepresentation in variant databases creates filtering gaps, and higher false-positive somatic call rates occur in non-European populations when using tumor-only approaches.

Clinical Pearls

Tumor-only sequencing can misclassify germline pathogenic variants as somatic -- germline testing should always be considered when a known cancer predisposition gene variant is detected at approximately 50% VAF. TMB and MSI are related but not identical: MSI-H tumors usually have high TMB, but high TMB can occur without MSI (such as in POLE-mutated tumors). Somatic BRCA1/2 mutations in ovarian and prostate cancer confer PARP inhibitor sensitivity similar to germline mutations. KRAS mutations in colorectal cancer predict resistance to anti-EGFR therapy (cetuximab, panitumumab), making expanded RAS testing standard of care. A "negative" comprehensive genomic profile does not mean the tumor has no driver, as drivers may reside in non-sequenced regions, involve epigenetic mechanisms, or reflect complex structural variants. CHIP-associated TP53 mutations detected in blood should not be interpreted as tumor-derived without careful VAF analysis.

References

• Frampton GM, Fichtenholtz A, Otto GA, et al. "Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing." Nature Biotechnology. 2013;31(11):1023-1031.
• Li MM, Datto M, Duncavage EJ, et al. "Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the AMP, ASCO, and CAP." Journal of Molecular Diagnostics. 2017;19(1):4-23.
• Chakravarty D, Gao J, Phillips SM, et al. "OncoKB: a precision oncology knowledge base." JCO Precision Oncology. 2017;1:1-16.
• Meric-Bernstam F, Brusco L, Shaw K, et al. "Feasibility of large-scale genomic testing to facilitate enrollment onto genomically matched clinical trials." Journal of Clinical Oncology. 2015;33(25):2753-2762.
• Mosele F, Remon J, Mateo J, et al. "Recommendations for the use of next-generation sequencing (NGS) for patients with metastatic cancers: a report from the ESMO Precision Medicine Working Group." Annals of Oncology. 2020;31(11):1491-1505.

Read this lecture as Markdown