Precision Medicine in Cancer: Genomic Testing Explained

Precision medicine cancer is an approach to oncology that matches treatment to the specific molecular characteristics of each patient’s individual tumor — rather than treating all patients with the same cancer type the same way. In traditional oncology, treatment was determined by where the tumor arose and what it looked like under the microscope: its histology and anatomy. In precision oncology, the primary treatment determinant is the molecular alteration driving the tumor — and that alteration may matter more than the organ of origin.

The clearest proof of this shift: in November 2018, the FDA approved larotrectinib for any solid tumor with an NTRK gene fusion, regardless of where it originated. Across 12 different histologies in a single basket trial, the drug produced an overall response rate of 75%. Eight months earlier, pembrolizumab received the same type of tissue-agnostic approval for any solid tumor with mismatch repair deficiency (MSI-H/dMMR) — the first tissue-agnostic cancer drug approval in history. Both approvals confirmed the central premise of precision medicine: identify the molecular target, select the drug that hits it, and the organ of origin becomes secondary. For the targeted drugs that precision medicine selects, see the targeted therapy cancer guide. For immunotherapy biomarkers like MSI-H and TMB, see cancer immunotherapy.

75% ORR of larotrectinib in NTRK fusion-positive solid tumors across 12 histologies — first tissue-agnostic targeted therapy (FDA Nov 2018, NEJM 2018)

~40% ORR of pembrolizumab in MSI-H/dMMR tumors regardless of origin — first tissue-agnostic cancer drug approval (FDA May 2017)

10+ FDA tissue-agnostic cancer drug approvals to date — covering MSI-H/dMMR, NTRK, TMB-H, BRAF V600E, RET, KRAS G12C, and dMMR

25–40% Proportion of patients with advanced cancer who have an actionable genomic alteration with an approved targeted therapy

From Histology to Genomics

For most of the 20th century, cancer was classified by where it arose and what it looked like under a microscope. Treatment followed that classification: all NSCLC patients received platinum-doublet chemotherapy; all metastatic colon cancer patients received FOLFOX or FOLFIRI; all breast cancer patients received anthracycline-based regimens. Clinical trials were organized by tumor type.

This worked when treatments were non-selective cytotoxic agents — chemotherapy kills all rapidly dividing cells, whether the tumor has any particular mutation or not. But molecularly targeted drugs changed the equation: they only work in tumors that carry the specific alteration the drug was designed to hit. Imatinib works in CML with BCR-ABL translocation — not in other leukemias. Trastuzumab works in HER2-amplified breast cancer — not in HER2-normal tumors. Osimertinib works in EGFR-mutated NSCLC — not in EGFR wild-type disease.

This created the precision oncology framework: before selecting treatment, test the tumor for the molecular alteration the treatment targets. The evolution continued toward its logical conclusion — some molecular alterations matter more than the tissue of origin. NTRK fusions in salivary gland, thyroid, and colon cancers all respond similarly to larotrectinib. MSI-H colorectal, endometrial, and gastric cancers all respond to pembrolizumab. Tissue-agnostic approvals formalized this conclusion: if the target is present, the drug can work, regardless of where the tumor is located.

Biomarker Testing Technologies

No single test identifies everything. Test selection is guided by the clinical question, cancer type, stage, and tissue availability. Understanding what each platform detects — and what it misses — is essential for ordering the right test at the right time.

IHC

Detects protein expression (ER/PR, HER2, PD-L1, MMR proteins, ALK). Rapid (1–3 days), inexpensive, universal availability. Misses mutations and novel rearrangements.

FISH

Detects gene amplification (HER2) and translocations (ALK, ROS1 fusions). Gold standard for HER2 confirmation (IHC 2+). Targeted — one gene per test.

PCR / RT-PCR

Detects known hotspot mutations (EGFR, KRAS, BRAF V600E, MSI by fragment analysis). Fast (1–3 days). Misses novel or atypical mutations outside the panel.

NGS Gene Panel

50–500 genes; detects somatic SNVs, indels, CNVs, selected fusions. Clinical workhorse. Misses alterations outside the panel. 7–14 day turnaround.

CGP (Comprehensive Genomic Profiling)

Near-genome-wide; all alteration types + TMB + MSI. FoundationOne CDx (324 genes), Tempus xT (648 genes). FDA-approved companion diagnostics. 2–4 week turnaround.

RNA Sequencing

Detects expressed fusion transcripts — NTRK, RET, ALK, ROS1, FGFR — with exceptional sensitivity including novel fusion partners missed by DNA panels. Increasingly combined with DNA-NGS.

Comprehensive Genomic Profiling (CGP) in Practice

CGP using platforms like FoundationOne CDx (Foundation Medicine/Roche, 324 genes, FDA-approved companion diagnostic for 30+ drug-tumor combinations) or Tempus xT (648 genes + RNA) is increasingly the preferred first-line test for patients with advanced solid tumors. CGP detects all somatic alteration types in a single test — point mutations, small indels, copy number changes, gene fusions, MSI status, and TMB — and can reveal unexpected findings (an NTRK3 fusion in a colon cancer patient, an RET fusion in a pancreatic cancer patient) that would be missed by targeted single-gene PCR tests or small hotspot panels.

NCCN guidelines now recommend broad molecular profiling (CGP) for patients with advanced cancers across most tumor types. Medicare covers CGP for advanced solid tumors under Coverage with Evidence Development (CED).

Liquid Biopsy and Circulating Tumor DNA

Liquid biopsy analyzes cell-free DNA (cfDNA) in plasma — a fraction of which originates from tumor cells (circulating tumor DNA, or ctDNA). Because blood draws are non-invasive and repeatable, liquid biopsy solves problems that tissue biopsy cannot: it captures tumor heterogeneity across all metastatic sites simultaneously, tracks tumor evolution over time without repeat invasive procedures, and provides access to tumor DNA when tissue biopsy is not feasible.

FDA-Approved Liquid Biopsy Platforms

FoundationOne Liquid CDx (Foundation Medicine): 324-gene plasma panel; FDA companion diagnostic for olaparib (BRCA breast/ovarian/prostate), pembrolizumab (MSI-H), alpelisib (PIK3CA breast), and others
Guardant360 CDx (Guardant Health): 74-gene plasma panel; FDA companion diagnostic for sotorasib (KRAS G12C NSCLC), olaparib (BRCA prostate), and others

Clinical Applications

Therapy selection: Plasma EGFR T790M detection guides osimertinib after first/second-generation EGFR TKI resistance. Plasma ESR1 mutation testing identifies patients with HR+/HER2- metastatic breast cancer for elacestrant after aromatase inhibitor progression. Plasma KRAS G12C confirms sotorasib/adagrasib eligibility.

Minimal residual disease (MRD) detection: ctDNA detected in blood after curative-intent surgery indicates residual microscopic disease weeks to months before imaging shows relapse. The DYNAMIC trial (Tie J, NEJM Evidence 2022) in stage II colon cancer was landmark: patients whose ctDNA was undetectable after resection could safely forgo adjuvant chemotherapy with equivalent 2-year recurrence-free survival — the first randomized trial to guide treatment de-escalation by ctDNA MRD. This opened the door to ctDNA-guided adjuvant therapy decisions across multiple cancer types.

Resistance mechanism identification: When targeted therapy fails, liquid biopsy can identify how resistance developed — EGFR C797S after osimertinib; MET amplification or KRAS G12C as bypass mechanisms after EGFR TKI; ESR1 mutations after aromatase inhibitors — enabling rational selection of next treatment or clinical trial.

precision medicine cancer genomic testing NGS liquid biopsy tumor profiling — Precision medicine cancer relies on genomic profiling — from tumor tissue NGS to liquid biopsy ctDNA — to match each patient’s treatment to the specific molecular alteration driving their tumor.

Limitations

Sensitivity ~70–90%: Tumors that shed little ctDNA — CNS primaries, peritoneal-only disease, some indolent histologies, early-stage cancers — may be liquid biopsy-negative despite harboring detectable tissue mutations
Clonal hematopoiesis (CHIP): Age-related somatic mutations in blood stem cells (DNMT3A, TET2, TP53) appear in plasma and can mimic ctDNA. Affects ~10% of adults over 65. Requires paired tissue confirmation to resolve
Does not replace tissue biopsy: Histologic confirmation, PD-L1 IHC, and some companion diagnostic requirements still mandate tissue for initial diagnosis and regulatory compliance

Companion Diagnostics and Molecular Tumor Boards

Companion Diagnostics

A companion diagnostic (CDx) is a laboratory test required by the FDA to determine whether a specific drug should be used in a specific patient. The CDx and the drug are co-approved — without the specific test, the drug cannot be prescribed per its regulatory label in that indication.

Drug	Biomarker	Required CDx
Pembrolizumab (NSCLC monotherapy)	PD-L1 TPS ≥50%	22C3 pharmDx IHC
Pembrolizumab (TMB-H any tumor)	TMB ≥10 mut/Mb	FoundationOne CDx
Osimertinib (EGFR NSCLC)	EGFR exon 19 del / L858R	cobas EGFR or FoundationOne CDx
Olaparib (germline BRCA breast)	Germline BRCA1/2 mutation	BRACAnalysis CDx (Myriad)
Sotorasib (KRAS G12C NSCLC)	KRAS G12C	Guardant360 CDx (plasma)
Elacestrant (ESR1 breast)	ESR1 mutation	Guardant360 CDx (ctDNA)
Larotrectinib (NTRK fusions)	NTRK1/2/3 fusion	FoundationOne CDx (multiple)
Trastuzumab (HER2+ breast)	HER2 amplification/overexpression	IHC + FISH (PATHWAY HER2)

Molecular Tumor Boards (MTB) — Interpreting Complex Genomic Results

When a patient’s CGP reveals alterations beyond standard approved indications — common with comprehensive profiling — a molecular tumor board provides multidisciplinary interpretation. Teams include molecular oncologists, pathologists, clinical genomicists, bioinformaticians, and clinical trial coordinators. MTB classifies variants by tier (Tier I: FDA-approved; Tier II: investigational or cross-tumor evidence; Tier III: variant of uncertain significance; Tier IV: benign), identifies clinical trial matches for Tier II alterations, and interprets co-occurring mutations that modify drug sensitivity. Studies show MTB participation correlates with higher rates of matched therapy use and improved outcomes vs. standard histology-based treatment.

Basket Trials and Tissue-Agnostic Oncology

A basket trial enrolls patients across multiple different cancer types based on a shared molecular alteration — testing whether a drug works across histologies sharing the same molecular target. Basket trials are the clinical engine of tissue-agnostic oncology.

Drug	Biomarker	Key Trial / Result	FDA Approval
Pembrolizumab	MSI-H / dMMR any tumor	KEYNOTE-158/051: ORR ~40%	May 2017 — first tissue-agnostic ever
Larotrectinib	NTRK fusion any solid tumor	LOXO-TRK/NAVIGATE: ORR 75%, CR 22%, 12 histologies (NEJM 2018)	Nov 2018
Entrectinib	NTRK / ROS1 fusions	STARTRK-2: NTRK ORR 57%; ROS1+ NSCLC ORR 77%	Aug 2019
Pembrolizumab	TMB-H ≥10 mut/Mb any tumor	KEYNOTE-158: ORR ~29% (10 tumor types)	Jun 2020
Dostarlimab	dMMR any solid tumor	GARNET: ORR ~42% across tumor types	Apr 2021
Dabrafenib + trametinib	BRAF V600E any solid tumor (exc. CRC)	BRF117019 + multiple: ORR across anaplastic thyroid, biliary, other	Jun 2022
Selpercatinib	RET fusion any solid tumor	LIBRETTO-001: ORR 44% non-thyroid tumors	Sep 2022

The SHIVA Caveat: Genomic Match ≠ Guaranteed Benefit

The SHIVA trial (Le Tourneau C, Lancet Oncol 2015) — the first randomized basket trial — found no PFS benefit of MTB-matched therapy versus physician’s choice (2.3 vs. 2.0 months). The lesson: molecular presence of an alteration does not guarantee functional dependency. A clonal, truncal alteration driving every tumor cell is a better drug target than a subclonal alteration in a minority of cells. Co-occurring mutations can blunt efficacy (KRAS mutation negates EGFR inhibitor benefit in CRC even when EGFR-expressing). Context — clonality, co-mutations, tumor type — matters as much as the alteration itself.

Germline vs. Somatic Genomic Testing

Somatic testing identifies mutations acquired during tumor development — present only in cancer cells, not in normal tissue. This is the vast majority of precision oncology testing: EGFR mutations in NSCLC, KRAS in CRC, BRAF V600E in melanoma. Somatic results have no implications for family members.

Germline testing identifies inherited mutations present in every cell of the body from birth — and potentially inheritable by children. In cancer care, germline testing is indicated when the tumor may reflect a hereditary cancer syndrome:

BRCA1/2 germline mutations: ~70% lifetime breast cancer risk (BRCA1); ~45% ovarian cancer risk (BRCA1); elevated prostate and pancreatic cancer risk (BRCA2). Germline BRCA1/2 confers eligibility for olaparib adjuvant therapy after early HER2-negative breast cancer (OlympiA trial: iDFS HR 0.58) and PARP inhibitors in other settings
Lynch syndrome (MLH1/MSH2/MSH6/PMS2/EPCAM germline): ~50–70% lifetime CRC risk for MLH1/MSH2 carriers; elevated endometrial, ovarian, gastric, and urologic cancer risk. When dMMR/MSI-H is found in tumor, germline MMR testing is now standard
Other hereditary syndromes: CDH1 (diffuse gastric cancer); TP53 (Li-Fraumeni — pan-cancer); PALB2 (breast/pancreatic); ATM (breast/prostate); APC (familial adenomatous polyposis)

Somatic BRCA vs. Germline BRCA — Not the Same

Tumor sequencing may detect a BRCA1/2 mutation that is somatic-only (present only in the cancer cells) or may reflect an underlying germline mutation present in all cells. Tumor-only sequencing cannot definitively distinguish the two — finding BRCA in tumor does not automatically mean the patient has hereditary BRCA syndrome. Patients with BRCA detected by tumor NGS should be referred to a genetic counselor for germline testing, particularly when PARP inhibitor eligibility or family implications are at stake.

Limitations, Access, and What’s Next

Precision medicine cancer is powerful but not universal. Approximately 25–40% of patients with advanced cancer have an actionable genomic alteration with an approved matched targeted therapy. For the remaining majority, genomic profiling may identify clinical trial options (Tier II alterations) or confirm no currently actionable target — in which case standard chemotherapy or immunotherapy remains the standard of care.

Additional limitations:

Tumor heterogeneity: A biopsy captures one site at one time; the driver alteration in a primary tumor may differ from a metastasis; clonal vs. subclonal distinction matters for predicting drug response
Turnaround time: CGP takes 2–4 weeks; for rapidly progressing patients, empiric treatment may need to start before results return
Resistance is inevitable: Most TKIs develop resistance within 12–24 months; serial re-profiling (liquid biopsy or repeat tissue biopsy) at progression is essential to guide next-line therapy selection
Access and cost: CGP costs $3,000–$6,000; liquid biopsy $1,000–$3,000; coverage improving but still uneven across payers and geographies; both Foundation Medicine and Guardant offer patient assistance programs

The next frontier is multi-omic profiling — combining genomic (DNA), transcriptomic (RNA), proteomic, and epigenomic data — and AI-driven models that predict response, resistance, and optimal drug combinations across data streams. The NCI-MATCH trial, ASCO TAPUR registry, and ESMO SCALE consortium continue to build the evidence base. For questions to ask your oncologist about genomic testing options, see the cancer diagnosis questions guide.

Frequently Asked Questions

What is precision medicine in cancer?

Precision medicine in cancer (also called precision oncology) is an approach that matches treatment to the specific molecular, genetic, or genomic characteristics of each patient’s individual tumor — rather than treating all patients with the same cancer type identically. Traditional oncology used histology and anatomy (lung cancer, breast cancer) to determine treatment; precision oncology adds a molecular layer: does this lung cancer have an EGFR mutation? Is this breast cancer HER2-amplified? Does this colon cancer have MSI-H/dMMR? The answer determines which drug is most likely to work — or whether a tissue-agnostic drug (larotrectinib for NTRK fusions, pembrolizumab for MSI-H any tumor) is appropriate regardless of cancer origin. According to the National Cancer Institute, genomics-based precision medicine uses information about a tumor’s genes to guide treatment decisions — an approach now standard for advanced cancer care at comprehensive cancer centers.

What is comprehensive genomic profiling (CGP)?

Comprehensive genomic profiling (CGP) is a broad-based next-generation sequencing (NGS) test that analyzes hundreds of cancer-relevant genes simultaneously in tumor tissue — detecting all types of somatic alterations: point mutations (SNVs), insertions/deletions (indels), copy number variations (amplifications/deletions), gene fusions/rearrangements, MSI status, and tumor mutational burden (TMB) — all in a single test. FDA-approved CGP platforms include FoundationOne CDx (Foundation Medicine, 324 genes) and Tempus xT (648 genes + RNA). Unlike targeted PCR tests that only detect pre-specified mutations, CGP can find novel or unexpected alterations — a NTRK3 fusion in a colon cancer patient, an RET fusion in a pancreatic cancer patient — that would be missed by small panels. NCCN guidelines recommend broad molecular profiling for advanced solid tumors across most cancer types. Medicare covers CGP for patients with advanced solid tumors under Coverage with Evidence Development (CED).

What is a liquid biopsy and what does it test for?

A liquid biopsy is a blood test that analyzes circulating tumor DNA (ctDNA) — DNA fragments shed by cancer cells into the bloodstream. FDA-approved platforms (FoundationOne Liquid CDx, Guardant360 CDx) detect cancer-driving mutations, fusions, and copy number changes from a blood draw without invasive tumor sampling. Clinical uses include: selecting targeted therapies (ESR1 mutations guide elacestrant use in metastatic breast cancer; EGFR T790M confirms osimertinib eligibility after first-generation TKI failure; KRAS G12C confirms sotorasib eligibility); detecting minimal residual disease (MRD) after surgery to predict recurrence before imaging — the DYNAMIC trial demonstrated ctDNA-guided de-escalation of adjuvant chemotherapy in stage II colon cancer was non-inferior to giving all patients chemo; and identifying resistance mechanisms when targeted therapy stops working. Limitations include lower sensitivity in low-shedding tumors (brain tumors, peritoneal disease) and the possibility of detecting clonal hematopoiesis (CHIP) — blood cell mutations unrelated to cancer that can appear as false positives.

What is a tissue-agnostic cancer drug?

A tissue-agnostic (tumor-agnostic) cancer drug is FDA-approved to treat any solid tumor with a specific molecular alteration — regardless of the organ of origin. Traditional drug approvals are site-specific (approved for “breast cancer” or “lung cancer”); tissue-agnostic approvals are alteration-specific. Examples: larotrectinib and entrectinib for any NTRK fusion-positive tumor (ORR 75% across 12 histologies); pembrolizumab for any MSI-H/dMMR solid tumor (first tissue-agnostic approval, May 2017) and for any TMB-H (≥10 mut/Mb) solid tumor; dabrafenib + trametinib for any BRAF V600E solid tumor (excluding CRC); selpercatinib for any RET fusion-positive solid tumor. These approvals mean a patient with an NTRK-fusion-positive colon cancer is eligible for larotrectinib even though colon cancer is not in the drug’s name — the fusion is the eligibility criterion. Tissue-agnostic drugs require comprehensive molecular testing (CGP or RNA sequencing) to detect the qualifying alteration, since small PCR panels may miss novel fusion variants. More information on approved tissue-agnostic indications is available at the NCI-MATCH trial page.

What is a companion diagnostic test?

A companion diagnostic (CDx) is a laboratory test required to determine whether a specific drug should be used in a specific patient. The FDA requires CDx co-approval with the drug when biomarker testing is essential for safe and effective use — the CDx and drug are inseparable from a regulatory standpoint. Examples: the 22C3 pharmDx PD-L1 IHC test is required for pembrolizumab monotherapy eligibility in NSCLC (TPS ≥50%); FoundationOne CDx is the required companion diagnostic for pembrolizumab in TMB-H settings; BRACAnalysis CDx (Myriad Genetics) or FoundationOne CDx determines germline or somatic BRCA1/2 status for olaparib eligibility in early breast cancer; Guardant360 CDx detects KRAS G12C in plasma for sotorasib/adagrasib eligibility. Importantly, different drugs targeting the same biomarker (e.g., multiple PD-1/PD-L1 inhibitors) may have different required CDx assays — using an unapproved assay (even one that detects the same protein) may produce discordant or regulatory non-compliant results. When ordering CGP or targeted testing for treatment selection, confirm the specific CDx assay required for the targeted drug.

Should every cancer patient get genomic testing?

Genomic testing recommendations depend on cancer type, stage, and clinical question. For patients with advanced solid tumors (stage III–IV), comprehensive genomic profiling is recommended by NCCN and ESMO guidelines for most cancer types where actionable alterations frequently influence first-line treatment: NSCLC (EGFR, ALK, ROS1, BRAF, RET, MET, KRAS G12C, NTRK); colorectal cancer (MSI-H, RAS/RAF, HER2, NTRK); breast cancer (HER2, HR status, BRCA, PIK3CA, ESR1 in metastatic); gastric/GEJ (HER2, MSI-H, PD-L1); cholangiocarcinoma (FGFR2, IDH1); endometrial (MSI-H, POLE). For early-stage patients, germline genetic testing (BRCA1/2, Lynch syndrome) should be considered based on personal and family history — because germline results affect treatment decisions (adjuvant olaparib in early BRCA-positive breast cancer) and family surveillance programs. For some cancers where all patients receive the same first-line regimen regardless of genomic profile (e.g., small cell lung cancer), upfront CGP has lower yield but repeat profiling at progression is valuable. For questions about specific testing recommendations for your cancer, see the cancer diagnosis questions guide and the AACR GENIE project for real-world genomic landscape data.

Drilon A et al. — Larotrectinib NTRK basket trial; NEJM 2018
Marabelle A et al. — KEYNOTE-158 TMB-H tissue-agnostic; Lancet Oncol 2020
Le Tourneau C et al. — SHIVA randomized basket trial; Lancet Oncol 2015
Tie J et al. — DYNAMIC ctDNA-guided de-escalation colon cancer; NEJM Evidence 2022
Tutt ANJ et al. — OlympiA (olaparib adjuvant germline BRCA breast); NEJM 2021
Schwaederle M et al. — MTB-guided therapy outcomes; JCO 2016
Jaiswal S / Genovese G et al. — CHIP prevalence in older adults; NEJM 2014
André T et al. — KEYNOTE-177 (pembro MSI-H mCRC); NEJM 2020
National Cancer Institute Genomics — cancer.gov/research/areas/treatment/genomics
NCI-MATCH — NCI-MATCH Precision Oncology Trial
AACR GENIE — AACR Project GENIE

This article is for educational purposes only and does not constitute medical advice. Discuss all cancer genomic testing and treatment decisions with your oncology care team.