Precision Medicine in Oncology

Note: This page is educational and reflects the state of the literature in 2025. It does not replace medical advice.

TL;DR

Precision oncology matches treatment to a tumor's molecular profile rather than only its organ of origin. The core idea: cancers are now classified by driver alterations (single mutations, fusions, copy-number changes, expression signatures) — and many of those drivers have specific drugs. The infrastructure: NGS panels, liquid biopsy, molecular tumor boards, clinical decision support, and basket/umbrella trial designs. The hard parts are no longer the sequencing — they are variant interpretation, access and reimbursement, equity, and operationalization at scale.

1. From organ-based to molecular-based oncology

Historically, oncology classified tumors by tissue of origin (lung, breast, colon). Precision oncology adds a second axis — molecular alterations — which sometimes matters more than organ:

Tissue-agnostic approvals — pembrolizumab in MSI-H/dMMR tumors; larotrectinib and entrectinib in NTRK fusions; dostarlimab in dMMR; selpercatinib in RET fusions across tumor types.
Driver-specific therapies — EGFR (osimertinib), ALK (alectinib), KRAS G12C (sotorasib, adagrasib), BRAF V600E (dabrafenib + trametinib), HER2 (trastuzumab, T-DXd), BRCA1/2 (PARP inhibitors), FGFR (erdafitinib).
Histology-agnostic biomarkers — TMB-high (≥10 mut/Mb), MSI-H, dMMR, NTRK fusion.

Endometrial cancer is illustrative: 2025 management is now organized around TCGA molecular classes (POLE-ultramutated, MSI-H/dMMR, copy-number low, copy-number high) — each with distinct prognosis and therapy. Sources: [1]

2. The molecular profiling toolkit

Tool	What it tells you	Notes
Tumor DNA sequencing (panel/WES/WGS)	Driver mutations, indels, structural variants, TMB	Panel = focused (~500 genes); WES = exome; WGS = whole genome
Tumor RNA sequencing	Fusions (especially), expression signatures	Best detection of fusion variants
IHC	Protein expression (HER2, PD-L1, ER/PR)	Cheap, fast, widely available
FISH	Specific rearrangements / amplifications	Falling out of favor where NGS available
MSI / MMR testing	Mismatch-repair status	PCR or NGS; immunotherapy decision
Liquid biopsy (ctDNA)	Circulating tumor mutations	MRD, monitoring, when biopsy unavailable
Methylation profiling	Epigenetic class (CNS tumor classifier)	CNS tumors especially
Functional assays	Cell-line / organoid drug response	Research / personalized drug discovery

For data engineering background, see From FASTQ to variants and Biomarkers & companion diagnostics.

3. Variant interpretation — the hard part

Sequencing is now commoditized; interpretation is the bottleneck. Frameworks in clinical use:

AMP/ASCO/CAP tier system (Tier I–IV) — clinical actionability classification.
OncoKB (MSKCC) — public knowledge base; drug-variant evidence levels (1, 2, 3A/3B, 4, R1/R2).
CIViC — community-curated variant interpretation.
ClinVar / dbSNP — germline classification (for hereditary cancer questions).
VICC meta-knowledgebase — harmonization across sources.

Each panel and EHR vendor has its own pipeline: call → annotate → filter → tier → review → report. Common interpretive pitfalls:

VUS overload — variants of uncertain significance dominate report length but rarely action.
Germline contamination — somatic-only panels may flag germline variants without confirmation.
Clonal hematopoiesis (CHIP) — DNMT3A/TET2 in blood-contaminating samples mimic somatic findings.
Tumor purity — low purity makes VAFs unreliable.
Copy-number from panels — methodologically tricky; performance varies a lot.

4. The molecular tumor board (MTB)

The MTB is where precision oncology becomes practice. A typical case discussion includes:

Clinical context — diagnosis, stage, prior therapies, performance status.
Pathology — diagnosis, IHC, tumor purity.
Molecular results — variants, signatures, expression, methylation if available.
Interpretation — which alterations are pathogenic? Which are actionable?
Recommendation — approved drug, off-label, clinical trial, observation.
Documentation — variant calls, evidence levels, decision rationale.

Quality MTBs are multidisciplinary: oncologist, pathologist, geneticist, bioinformatician, pharmacist, clinical research coordinator. Without that, cases get superficial reads. Operationally, MTBs are a major target for software automation — case prep, trial matching, structured note generation, outcome capture.

5. Liquid biopsy: when tissue isn't enough

When tumor tissue is unavailable, insufficient, or temporally outdated, circulating tumor DNA (ctDNA) offers an alternative:

First-line genotyping — when biopsy isn't feasible.
Resistance monitoring — track emerging resistance mutations (e.g., EGFR T790M, C797S; ESR1 in breast).
MRD — residual disease detection post-curative therapy (CRC, breast, others).
Multi-cancer early detection (MCED) — screening healthy populations (still investigational, see Emerging therapies).

Sensitivity is the limiting factor — low-shedding tumors (some brain, kidney, prostate) may not have detectable ctDNA. Tumor-informed assays (designed against the patient's tumor variants) are more sensitive than tumor-naïve.

6. Trials matched to precision medicine

Three designs that became standard for precision oncology:

Basket — one drug, multiple tumor types sharing a biomarker.
Umbrella — one tumor type, multiple drugs matched to biomarkers.
Platform — long-running infrastructure, drugs entering and exiting on evidence.

See Clinical trials for design details. NCI-MATCH (US), TAPUR (ASCO), Drug Rediscovery Protocol DRUP (Netherlands), and similar studies generated much of the modern evidence for off-label and tissue-agnostic targeting.

7. Equity, access, and Brazilian context

Precision oncology amplifies existing disparities unless deliberately countered:

Sequencing access — high-cost panels not universally reimbursed; small / public hospitals lag.
Targeted drug access — expensive, sometimes restricted to private insurance or special programs.
Knowledge bases — disproportionately built on European-ancestry data; performance gaps for Latin American, African, and Asian populations.
Operational capacity — interpretation expertise concentrated in large centers; pipelines for community oncology immature.

Brazilian-specific dynamics:

SUS has been incorporating selected targeted therapies and biomarker tests; coverage varies by state and CONITEC decisions.
A.C. Camargo Cancer Center, Hospital Sírio-Libanês, Hospital Israelita Albert Einstein, and academic networks have built MTBs and panels.
Biological diversity of Brazilian populations is under-represented in public databases — local cohorts (e.g., ABRACE) help close that gap.
Liquid biopsy is increasingly used in private oncology; access via SUS still limited.

For ethics and regulatory dimensions, see Regulatory & ethics.

8. What the technology side actually looks like

For an engineer building precision-oncology infrastructure:

Sequencing pipelines — fastq → BAM → VCF → annotated TSV → JSON for downstream — versioned, reproducible, validated. Sources: [2]
Knowledge integration — synchronize OncoKB, CIViC, ClinVar; manage update cadence; reconcile conflicts.
Variant tiering automation — implement AMP tiers; surface evidence concisely.
Trial matching — use eligibility criteria (structured + NLP) to find candidate trials per case.
MTB software — case agenda, evidence collation, outcome tracking.
Outcome registry — close the loop on what worked, for whom, why.
Federated analytics — let centers learn from each other without moving PHI.

For trade-offs and pitfalls, see ML in oncology pitfalls and Data governance & LGPD.

References

Corr BR, Erickson BK, Barber EL, Fisher CM, Slomovitz B. Advances in the management of endometrial cancer. BMJ 2025;388:e080978. PMID 40044230. https://doi.org/10.1136/bmj-2024-080978 (model of modern molecularly classified disease management)
Meyer ML, Fitzgerald BG, Paz-Ares L, et al. New promises and challenges in the treatment of advanced non-small-cell lung cancer. Lancet 2024;404:803-822. PMID 39121882. https://doi.org/10.1016/S0140-6736(24)01029-8
Dumontet C, Reichert JM, Senter PD, Lambert JM, Beck A. Antibody-drug conjugates come of age in oncology. Nat Rev Drug Discov 2023;22:641-661. PMID 37308581. https://doi.org/10.1038/s41573-023-00709-2
U.S. National Cancer Institute. Precision medicine in cancer treatment. https://www.cancer.gov/about-cancer/treatment/types/precision-medicine
American Cancer Society. Cancer A-Z. https://www.cancer.org/cancer.html
Cleveland Clinic. Cancer (overview). https://my.clevelandclinic.org/health/diseases/12194-cancer
A.C. Camargo Cancer Center. https://accamargo.org.br
Fundação do Câncer (Brasil). https://www.cancer.org.br/
Ministério da Saúde / BVS. ABC do câncer. https://bvsms.saude.gov.br/bvs/publicacoes/abc_do_cancer.pdf
OncoKB (Memorial Sloan Kettering). https://www.oncokb.org
CIViC. https://civicdb.org

Precision Medicine in Oncology ​

TL;DR ​

1. From organ-based to molecular-based oncology ​

2. The molecular profiling toolkit ​

3. Variant interpretation — the hard part ​

4. The molecular tumor board (MTB) ​

5. Liquid biopsy: when tissue isn't enough ​

6. Trials matched to precision medicine ​

7. Equity, access, and Brazilian context ​

8. What the technology side actually looks like ​

See also ​

References ​