Cancer Evolution & Clonal Dynamics
Note: This page is educational and reflects the state of the literature in 2025. It does not replace medical advice.
TL;DR
A tumor is not a snapshot — it is an evolving population. Mutations accumulate, the microenvironment selects, and therapy applies a hard selective pressure. The Darwinian framing (variation + selection) explains a lot, but recent evidence shows tumors also evolve via catastrophic, punctuated events (whole-genome doubling, chromothripsis), neutral drift, and age-related somatic mosaicism. To predict relapse and resistance you must read the tumor's history, not just its current state. Sources: [1], [2]
1. The Darwinian backbone
Peter Nowell's 1976 paper formalized cancer as clonal evolution: a normal cell acquires a mutation that gives it a survival or proliferation advantage; its descendants form a clone; further mutations create sub-clones; selective pressures (immune, metabolic, therapeutic) shape which clones expand. The model has three pillars:
- Variation — mutations (SNVs, indels, structural variants), copy-number changes, epigenetic alterations.
- Inheritance — daughter cells inherit alterations.
- Selection — fitter clones outgrow others.
This is the basis for "branching evolution" trees inferred from sequencing data. Sources: [1]
2. Three flavors of evolutionary trajectory
Modern data suggest tumors do not all evolve the same way: Sources: [1], [2]
a) Linear / strong selection
A driver mutation appears, sweeps to fixation, then the next driver appears, etc. Most evident in some hematological malignancies and early-stage tumors.
b) Branching evolution
Multiple sub-clones coexist, each carrying different drivers. The classic "tree" inferred by phylogenetic methods. Common in advanced solid tumors. Implication: a biopsy from one region may miss drivers present in another region (spatial heterogeneity — see Tumor microenvironment).
c) Neutral evolution
A surprisingly large fraction of mutations may be passengers under little selection. The tumor still grows, but as drift, not Darwinian sweeps. Distinguishing neutral from selected requires careful statistics. Sources: [1]
3. Punctuated equilibria — when evolution is not gradual
A growing body of evidence shows that some of the most consequential events in cancer evolution happen catastrophically, not gradually: Sources: [2]
- Whole-genome doubling (WGD) — duplicates the entire genome in one event; common in lung, breast, and esophageal cancers; correlates with poor prognosis.
- Chromothripsis — one or a few catastrophic shattering-and-rejoining events that produce dozens of rearrangements at once.
- Chromoplexy — coordinated rearrangements involving several chromosomes simultaneously; described in prostate cancer.
- Kataegis — localized hypermutation, often APOBEC-driven.
These contradict gradualism — a single instant in cancer's life can scramble its genome more than years of slow mutation. They also create vulnerabilities (e.g., WGD may sensitize to certain spindle-checkpoint inhibitors). Sources: [2]
4. Age-related somatic mosaicism
Even healthy people accumulate somatic mutations with age (sun-exposed skin, esophageal lining, hematopoietic stem cells). Some cells acquire driver mutations and expand into "fields" without becoming cancer. Known as:
- Clonal hematopoiesis of indeterminate potential (CHIP) — DNMT3A/TET2/ASXL1 clones in blood; risk factor for myeloid malignancy and cardiovascular disease.
- Field cancerization — mutated patches in normal tissue (skin, esophagus, head & neck) from which independent tumors can arise.
Clinical implication: cancer is partially a disease of accumulated normal aging. This affects how we interpret variants in liquid biopsy and how we model risk. Sources: [2]
5. Therapy as the strongest selective pressure
Anything that kills a fraction of tumor cells selects for the survivors:
- Targeted therapy resistance — pre-existing minor clones (e.g., T790M in EGFR-mutant lung cancer) expand under drug pressure.
- Chemotherapy resistance — efflux pumps, DNA repair upregulation, metabolic shifts.
- Immunotherapy resistance — loss of HLA expression, β2-microglobulin mutations, antigen loss.
Two distinct mechanisms: Sources: [1]
- Selection of pre-existing clones — the resistance was already there at low frequency.
- De novo mutations under therapy pressure — new alterations arise during treatment.
Telling them apart matters: pre-existing requires earlier or combination therapy; de novo may benefit from intermittent dosing or hypomutator strategies.
6. How we measure tumor evolution
Methods, ordered by resolution: Sources: [1]
| Method | Resolution | What it tells you |
|---|---|---|
| Bulk WES/WGS | Population-level | Driver mutations, clonal architecture (via VAF + phylogeny) |
| Multi-region sequencing | Spatial | Branching trees (e.g., TRACERx in lung cancer) |
| Single-cell DNA-seq | Per cell | Sub-clone structure without averaging artifacts |
| scRNA-seq + scDNA-seq | Per cell | Genotype-to-phenotype linking |
| ctDNA / liquid biopsy | Longitudinal | Track clonal dynamics non-invasively |
| Lineage tracing (preclinical) | Per ancestor | True ancestor-descendant maps in mouse models |
The bioinformatics workhorses include PyClone-VI / SciClone (clonal deconvolution), MEDICC2 / CONIPHER / PhyloWGS (tumor phylogeny), dN/dS in cancer (selection inference). Background pipeline: see From FASTQ to variants.
7. Implications for treatment
If tumors are evolving systems, treatment must account for it: Sources: [1], [2]
- Adaptive therapy — dose to control rather than maximize kill, preserving treatment-sensitive clones to suppress resistant ones (Gatenby et al., prostate cancer).
- Combination therapy — simultaneous attack on multiple vulnerabilities reduces probability of pre-existing resistance to all.
- Liquid biopsy monitoring — detect emerging resistance before clinical relapse.
- Targeting evolutionary vulnerabilities — synthetic lethality with WGD, chromosomal instability, or replication stress.
- Avoid evolutionary dead ends — don't drive the tumor into a more aggressive state by overly aggressive monotherapy (the "evolutionary trap" of MTD chemotherapy).
See also
- Hallmarks of cancer (2011–2022)
- Tumor microenvironment
- Stem cells & cancer
- From FASTQ to variants
- ML pitfalls in oncology
References
- Graham TA, Sottoriva A. Measuring cancer evolution from the genome. J Pathol 2017;241:183-191. PMID 27741350. https://doi.org/10.1002/path.4821
- Vendramin R, Litchfield K, Swanton C. Cancer evolution: Darwin and beyond. EMBO J 2021;40:e108389. PMID 34459009. https://doi.org/10.15252/embj.2021108389
- Nowell PC. The clonal evolution of tumor cell populations. Science 1976;194:23-28. PMID 959840. https://doi.org/10.1126/science.959840 (historical / foundational)
- National Cancer Institute (NCI). What is cancer? https://www.cancer.gov/about-cancer/understanding/what-is-cancer
- 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/