The central obstacle, the authors explain, is heterogeneity. Oesophageal adenocarcinoma is driven in large part by chromosomal instability, a process that generates large-scale genomic chaos rather than the tidy, single-gene mutations seen in some other cancers. This instability produces profound differences not only between patients but also between different regions of the same tumour and between the primary tumour and its metastases. Two cells sitting centimetres apart within one patient’s oesophagus may carry different copy-number landscapes, different mutational burdens and different vulnerabilities. A therapy that eradicates one subclone may simply clear the way for another, which is why responses to treatment are so variable and why resistance so often emerges. Any model that smooths over this complexity risks giving clinicians a misleading picture of how a real tumour will behave.

Precision oncology promises to match each treatment to the biology of each tumour, but the review makes clear that in oesophageal cancer this promise has been constrained by history. Conventional two-dimensional cell lines—the workhorses of cancer biology for decades—grow quickly, are cheap and are easy to manipulate genetically, yet decades of passaging in plastic have driven them far from the tumours they originally came from. They lack the three-dimensional architecture of real tissue, they have lost most of the stromal and immune cells that surround a tumour in the body, and their genomes often no longer reflect the patient’s disease. They remain useful for dissecting mechanisms, the authors concede, but as avatars of an individual patient they fall short of what translational medicine now demands.

The models that have attracted the most excitement in recent years are patient-derived organoids: miniature, self-organising tumour fragments grown from fresh biopsy or surgical tissue in a supportive extracellular matrix. Because they are established directly from a patient and expanded for only a limited number of passages, organoids preserve much of the genotype and phenotype of the parent tumour, including the copy-number aberrations that dominate oesophageal adenocarcinoma. Crucially, they can be grown in multi-well formats, meaning dozens of drugs and drug combinations can be tested against a patient’s own tumour cells within days to weeks—a time horizon that can genuinely inform clinical decision-making. Studies across multiple cancer types have shown that organoid drug responses can predict patient responses with encouraging accuracy, and the review highlights their potential as functional biomarkers for treatment selection in oesophageal cancer specifically.

Yet organoids have an inherent limitation: they usually contain only the epithelial cancer cells. The tumour microenvironment—the fibroblasts, immune cells, blood vessels and signalling molecules that bathe a tumour in vivo—is largely absent, and it is this microenvironment that determines whether immunotherapies work. To close that gap, researchers are developing co-culture systems that introduce cancer-associated fibroblasts or immune cells into organoid cultures, and the review singles out immune-augmented organoid platforms as one of the most promising frontiers. By embedding tumour organoids with autologous immune cells, laboratories can begin to run functional immunology readouts: measuring whether a patient’s own T cells recognise their tumour, whether immune checkpoint blockade reinvigorates an anti-tumour response, and whether resistance mechanisms are already at play. Such systems offer a glimpse of personalised immunotherapy testing—something barely imaginable a decade ago.

At the other end of the biological fidelity spectrum sit patient-derived xenografts, or PDX models, in which fragments of a patient’s tumour are implanted into immunodeficient mice. These models retain the three-dimensional architecture, stromal interactions and evolutionary dynamics of the original tumour, and because they grow inside a living organism they capture whole-body pharmacology—how a drug is absorbed, distributed, metabolised and cleared—that no dish can replicate. Orthotopic variants, implanted directly into the oesophagus, add anatomical realism, while humanised PDX mice, engrafted with a human immune system, allow immunotherapies to be studied in a living setting. The trade-off, the authors stress, is throughput and time: establishing a PDX line takes months, success rates vary, and the cost and animal requirements limit how many patients can be modelled at scale. PDX models therefore serve best as deep characterisation platforms and for studying evolutionary and pharmacological questions rather than as rapid diagnostic tools.

Between the dish and the mouse lies a class of models that the review treats with particular attention: ex vivo organotypic tissue slice platforms and histocultures. Rather than dissociating a tumour or passaging it, these approaches take fresh slices of the actual surgical specimen—preserving the full cellular ecosystem of cancer cells, stroma, vasculature and immune infiltrate—and keep them alive in culture for days to a few weeks. Because nothing is disrupted, these slices offer what may be the highest fidelity to the parent tumour of any platform, and their short turnaround makes them attractive for clinically aligned endpoints such as predicting a patient’s response to neoadjuvant chemotherapy or radiotherapy before treatment begins. The limitations are equally practical: slice viability is finite, oxygen and nutrient penetration constrain slice thickness, and standardisation across laboratories remains immature. Nonetheless, the authors argue that organotypic cultures, especially when paired with immune readouts, occupy a unique translational niche for short-horizon therapeutic testing.

The review then turns to a rapidly accelerating dimension of cancer modelling that involves no cells at all: computation. In silico inference pipelines now integrate whole-genome sequencing, transcriptomics, epigenetics and imaging data to infer tumour evolutionary history, predict vulnerabilities and stratify patients, while computational histopathology—increasingly powered by deep learning applied to routine pathology slides—can extract prognostic and predictive information at a scale no experimental model can match. Digital approaches offer unlimited scalability and near-instant results, and they can integrate multi-omic and imaging information that fragmented experimental systems capture only in part. But the authors are emphatic about a caveat: algorithms trained on retrospective data are only as good as their validation, and rigorous benchmarking against real patient outcomes and against experimental models is essential before computational predictions can safely guide therapy. The most credible future, they suggest, is not a single winning platform but a triangulation in which genomic inference, organoid and slice-based drug testing, and selective PDX experiments corroborate one another.

What emerges from the survey is a portfolio philosophy. No single model satisfies all the translationally relevant criteria the authors apply—fidelity to the parent tumour, representation of stromal and immune compartments, scalability, time-to-result and suitability for clinically aligned endpoints such as response prediction and resistance evolution. Organoids win on speed and scalability; organotypic slices win on microenvironmental fidelity and clinical turnaround; PDX models win on organism-level pharmacology and evolutionary context; and computational pipelines win on throughput and data integration. Used intelligently and in combination, these platforms could finally give oncologists what oesophageal adenocarcinoma has long denied them: a way to test, in advance and in the laboratory, whether a given therapy will work for a given patient, and to watch resistance evolve before it happens in the clinic.

The stakes could hardly be higher. As immune checkpoint inhibitors reshape frontline treatment of gastro-oesophageal cancers and a growing arsenal of targeted agents waits in the wings, the absence of reliable predictive biomarkers means many patients endure toxic therapies from which they derive little benefit, while potentially effective options go untried. The Birmingham team, whose work was supported by Cancer Research UK and the Sir Arthur Thomson Charitable Trust, frames its review as both a critical appraisal and a call to action: the model-building tools now exist, but the field must invest in head-to-head comparisons, standardisation and prospective validation against patient outcomes. If that work succeeds, the era of treating oesophageal adenocarcinoma by trial and error could give way to one in which a patient’s tumour is first grown, challenged and computationally interrogated in the laboratory—so that the first real experiment happens where it matters most, in the clinic, with the odds stacked in the patient’s favour.

Subject of Research: Preclinical and computational modelling of oesophageal adenocarcinoma for precision oncology and immunotherapy

Article Title: Modelling oesophageal adenocarcinoma for precision oncology and immunotherapy

Article References: Anwar, R., Rose, E., Swirsky, F., Kunene, V., & Contino, G. (2026). Modelling oesophageal adenocarcinoma for precision oncology and immunotherapy. Cancer Immunology, Immunotherapy. https://doi.org/10.1007/s00262-026-04447-3

Image Credits: AI Generated

DOI: 10.1007/s00262-026-04447-3

Keywords: oesophageal adenocarcinoma, tumour heterogeneity, patient-derived organoids, patient-derived xenografts, immunotherapy, precision oncology, drug sensitivity, tumour microenvironment, computational histopathology, chromosomal instability, personalised medicine, cancer models

The decision to centre emerging investigators rather than established luminaries is itself a statement about how science should evolve. The authors argue that researchers early in their careers are uniquely positioned to challenge prevailing assumptions, adopt interdisciplinary approaches and redirect priorities that may have calcified over decades. The resulting collection of perspectives spans an unusually wide technical range, from neoadjuvant immunotherapy in colorectal cancer to somatic mosaicism in healthy tissues, from the nervous system’s role in tumour progression to artificial intelligence models that predict cellular responses to genetic perturbation. Together, the six contributions sketch a research agenda that is more integrated, more prevention-focused and more computationally ambitious than anything the field has attempted before.

One thread running through the article is the remarkable maturation of cancer immunotherapy, particularly when treatment is moved earlier in the disease course. Myriam Chalabi, a medical oncologist and physician scientist at the Netherlands Cancer Institute in Amsterdam, has built her research programme around immunotherapy delivered in the neoadjuvant setting, using novel treatment combinations within innovative trial designs. The clinical evidence underpinning this shift is striking: recent work has demonstrated neoadjuvant immunotherapy in mismatch-repair-proficient colon cancers, while separate research has shown that non-operative management of mismatch repair deficient tumours can produce durable responses, in some cases allowing patients with rectal cancer to avoid surgery entirely. These results suggest that the immune system, when engaged before a tumour has been removed, can eliminate disease that conventional staging would consider established, and they raise the prospect of organ-preserving treatment as a realistic goal rather than an aspirational one.

Yet immunotherapy has also exposed the limits of tumour-centric thinking, and several of the authors argue that the next 25 years must focus on the host as much as the tumour. James L. Reading, an associate professor of cancer immunology at UCL who leads the Pre-cancer Immunology Laboratory, studies T cell-driven cancer interception — the idea of detecting and eliminating tumours before they become clinically invasive. His work builds on the discovery that reservoirs of stem-like CD8-positive T cells in tumour-draining lymph nodes sustain ongoing antitumor immune responses, and that conventional type I dendritic cells maintain pools of proliferative, tumour-antigen-specific TCF1-positive CD8-positive T cells in those same nodes. Understanding how these immune reservoirs are established and maintained during pre-invasive disease, he argues, could transform early detection from passive imaging into active, immune-guided interception, catching malignancy at a stage when cure rates approach certainty.

The genomic dimension of this preventive agenda is developed most fully by Tim H. H. Coorens, a group leader at the European Bioinformatics Institute who studies how somatic mutations accumulate in normal cells. Twenty-five years ago, cancer genomes were largely studied in isolation from the tissues that produced them. Today, it is clear that essentially every cell in the body accrues mutations over a lifetime, and that clones of mutant cells — some harmless, some pre-malignant — expand and compete in otherwise healthy tissue. Coorens contributed to the Somatic Mosaicism Across Human Tissues network, an effort to catalogue this variation systematically, and recent analyses have shown that age itself can distinguish selective clonal expansion from simple mutational causation in cancer genomes. Meanwhile, the real-world clinical utility of tumour whole-genome sequencing in solid cancers has now been demonstrated at scale, suggesting that comprehensive genomic profiling is moving from research luxury to standard of care. The conceptual shift is profound: cancer becomes not a foreign invader but one possible endpoint of a lifelong evolutionary process, and the levers for prevention may lie in the dynamics of normal tissue.

Perhaps the most visually striking frontier is cancer neuroscience. Leanne Li, a group leader at the Francis Crick Institute in London, combines cancer genetics with neurotechnologies to decipher the logic of interactions between tumours and the nervous system in mouse models. The field’s roots reach back more than a century to observations of nerves within tumours made using methylene blue vital staining, but modern cancer neuroscience has exploded in the past decade. Recent single-neuron sequencing has revealed how individual neurons are reprogrammed by pancreatic cancer, and comprehensive reviews have mapped the past, present and future of the discipline. Li also leads InteroCANCEption, a multidisciplinary team funded by Cancer Grand Challenges to tackle the broader question of how interoception — the body’s sensing and regulation of its own internal signals, a concept elaborated in modern neuroscience — shapes tumour initiation, growth and response to therapy. If tumours co-opt neural circuitry the way they co-opt blood vessels, then neuromodulatory drugs already approved for other conditions could become unexpected additions to the oncology arsenal.

Resistance to therapy, the stubborn core of cancer mortality, is the focus of Shensi Shen, associate professor at West China Hospital, Sichuan University. His work centres on drug-tolerant persister cells — a subpopulation of cancer cells that survives initial treatment not through genetic mutation but through reversible shifts in cell state. Reviews have traced the journey of persister cell biology from basic questions to clinical opportunities, and single-cell analyses have shown that genetically homogeneous cancer cells can diverge into multiple distinct clonal fates when exposed to the same drug. Shen’s particular interest is in layered translational control: the regulation of how messenger RNA is decoded into protein, which allows cancer cells to deploy hidden protein functions and switch states under therapeutic pressure. Because these transitions are non-genetic, they are also potentially reversible, which makes the persister state an attractive target for combination strategies designed to block the escape routes that tumours use to survive targeted therapy and immunotherapy alike.

Underpinning all of these biological questions is a computational revolution, examined by Ewa Szczurek, associate professor at the University of Warsaw and director of the Institute of AI for Health at Helmholtz Munich. Szczurek develops artificial intelligence models for molecular biology and medicine, and her perspective is notably sober about the current state of the field. While foundation models promise to predict how cells respond to genetic and pharmacological perturbations — an ambition exemplified by recent preprint work on state-based prediction of cellular responses — independent evaluations have shown that deep-learning-based gene perturbation effect prediction does not yet outperform simple linear baselines in many settings. Her message is that the next 25 years of AI in cancer research will be defined not by model size but by data quality, experimental validation and careful benchmarking. If the field heeds that warning, machine learning could genuinely accelerate target discovery and personalised treatment; if it does not, hype risks outrunning biology.

Woven together, the six perspectives describe a field in mid-revolution. The tumour-as-isolated-entity model is giving way to a systems view in which cancer is embedded in the evolutionary dynamics of normal tissues, the immune landscape of pre-invasive disease, the neural circuitry of the host body and the non-genetic plasticity of individual cells. Clinical paradigms are shifting in parallel: treatment is moving earlier, surgery is sometimes becoming optional, and molecular residual disease monitoring — exemplified by analyses of adjuvant osimertinib in resected EGFR-mutated lung cancer — is becoming a guide for post-operative decisions. The anniversary article also translates premalignant biology into strategies for intercepting non-small-cell lung cancer, illustrating how laboratory insight can be converted directly into prevention trials.

What emerges most clearly is a demand for interdisciplinarity as a structural principle rather than a slogan. The authors themselves embody it: a clinician designing immunotherapy trials, a bioinformatician decoding mutation accumulation, a neuroscientist engineering tools to interrogate tumour-nerve crosstalk, an immunologist chasing T cells before invasion, a molecular biologist tracking protein-level resistance and a computer scientist stress-testing the field’s newest models. Their collective wager is that the major killers of the next quarter-century will not be defeated by any single breakthrough but by the deliberate integration of genomics, immunology, neuroscience, developmental biology and computation — and by the willingness of a new generation to ask questions their predecessors did not think to ask.

Subject of Research: Emerging investigators' perspectives on the future priorities of cancer research over the next 25 years

Article Title: The next 25 years of cancer research: emerging perspectives and priorities

Article References: Chalabi, M., Coorens, T. H. H., Li, L., Reading, J. L., Shen, S., & Szczurek, E. (2026). The next 25 years of cancer research: emerging perspectives and priorities. Nature Reviews Cancer. https://doi.org/10.1038/s41568-026-00975-3

Image Credits: AI Generated

DOI: 10.1038/s41568-026-00975-3

Keywords: cancer research, Nature Reviews Cancer, immunotherapy, neoadjuvant therapy, somatic mosaicism, cancer neuroscience, cancer interception, drug-tolerant persister cells, tumour heterogeneity, whole-genome sequencing, artificial intelligence, early detection

The conceptual foundation dates back to 1976, when Peter Nowell proposed that tumour cell populations evolve through acquired genetic lability, with stepwise selection of variant sublines driving progression. Half a century of bulk sequencing vindicated the model, revealing branched evolution, intratumoural heterogeneity and the selective sweeps that follow treatment. Yet bulk measurements average across millions of cells, obscuring which mutation resides in which cell and, crucially, what that mutation actually does to the cell carrying it. The review’s authors argue that closing this genotype-to-phenotype gap is now the central task of cancer evolutionary biology.

One striking motivation comes from healthy tissue. Landmark studies of normal skin, oesophagus, colon, endometrium, bladder and bronchial epithelium have shown that somatic mutations in canonical cancer driver genes are under pervasive positive selection in tissue that looks entirely normal under the microscope. In sun-exposed skin, roughly a quarter of cells carry cancer-associated mutations, and mutation burdens in some cells rival those of tumours. Clonal haematopoiesis, the age-related expansion of blood cells carrying mutations in genes such as DNMT3A, TET2 and JAK2, likewise demonstrates that genetic mosaicism is a feature of ordinary physiology and ageing, seeding the pre-malignant landscape from which frank cancers emerge.

Mapping this diversity, however, is only the first step. The review emphasises that understanding somatic clonal evolution requires defining the phenotypes that give mutated clones a fitness advantage, whether those phenotypes involve altered differentiation, survival, proliferation or interaction with the microenvironment. This is where genotype-aware single-cell multi-omics enters. Methods such as G&T-seq and its descendants physically split or barcode the genome and transcriptome of an individual cell, while genotyping-of-transcriptomes approaches recover expressed mutations directly from single-cell RNA-sequencing data. Targeted strategies enrich for known mutant loci, and chromatin-focused assays now co-capture mutations alongside single-cell accessibility profiles.

The biological payoffs have been substantial. In clonal haematopoiesis, single-cell multi-omics has shown that the effects of a mutation are often cell-state specific. DNMT3A R882 mutations, for example, were found to perturb early progenitor states through selective hypomethylation, a phenotype invisible to bulk assays. Splicing aberrations in haematopoietic clonal outgrowths display distinct cell-type-specific impacts, and maps linking genotypes to chromatin accessibility profiles reveal how individual mutations reshape regulatory landscapes in a lineage-dependent manner. In myeloproliferative neoplasms, clonally resolved analyses have traced how JAK2 and CALR mutations propagate through differentiation hierarchies, while work in acute myeloid leukaemia has connected RAS-mutant leukaemia stem cells to clinical resistance against the BCL-2 inhibitor venetoclax.

Beyond single time points, the review highlights the power of coupling phylogenetic reconstruction with phenotypic measurement. Endogenous marks such as somatic point mutations, copy-number alterations, mitochondrial DNA mutations, microsatellite shifts and stochastic epimutations each leave heritable traces that allow researchers to infer the ancestral relationships among single cells. Mitochondrial mutations in particular have enabled lineage tracing directly in human samples, and somatic epimutations have recently been used to chart the dynamics of blood ageing. Reconstructed single-cell phylogenies can then be time-calibrated, converting a branching diagram into a chronogram that estimates when a clone originated within a patient’s lifespan.

Such temporal mapping demands careful statistical treatment. The authors describe phylogenetic frameworks that quantify heritability and plasticity of cell states, decoupling genetic inheritance from non-genetic, environmentally driven transitions. Molecular clock models, whether strict or relaxed, permit inference of mutation rates and timing of clonal expansions, and phylodynamic approaches borrowed from pathogen genetics now illuminate how tumour population sizes fluctuate over the course of disease. Applied to colorectal cancer, these tools have revisited the Big Bang model of tumour growth, in which most subclonal diversity is generated in an early expansion rather than through later selective sweeps, and have documented phenotypic plasticity under genetic control during malignant progression.

Spatial context adds a further dimension. Multiclonal invasion patterns in breast tumours, spatially resolved copy-number maps in benign and malignant tissue, and spatial genomics of cancer clones all demonstrate that evolutionary dynamics are constrained by tumour architecture. Mechanical confinement has been shown to govern phenotypic plasticity in melanoma, and harsh microenvironments select for glycolytic phenotypes in early breast cancer. Integrating spatially resolved or lineage-resolved phenotypes with genotype maps is therefore revealing how selection operates not just on mutations but on the cell states and niches in which those mutations find themselves.

Translationally, the authors argue that genotype-to-phenotype mapping can expose therapeutic vulnerabilities for the precision elimination of disease-propagating mutant cells. Because mutant phenotypes are often confined to specific cell states or lineages, vulnerabilities may exist that spare wild-type tissue. Single-cell analyses have identified drug-tolerant persister states, non-genetic determinants of clonal fitness, and epigenetically inherited plasticity that drives drug resistance through one-to-many genotype-to-phenotype relationships. In glioblastoma, recurring cellular states whose abundance is modulated by genetic aberrations suggest combination strategies; in IDH-mutant oligodendroglioma, mutant IDH inhibitors induce lineage differentiation detectable at single-cell resolution. Evolutionary steering, in which treatment is designed to guide tumours toward collateral sensitivities, becomes more tractable when clonal phenotypes can be read directly in patients.

The review closes with a sober assessment of remaining challenges. Whole-genome amplification artefacts, allelic dropout and tissue dissociation biases still limit sensitivity and fidelity, and artifacts in mitochondrial DNA analyses can misinform phylogenetic inference if uncorrected. Computational methods continue to mature, from probabilistic single-cell phylogeny inference that accounts for sequencing error to models that relax the infinite-sites assumption when back mutations and parallel evolution occur. Yet the trajectory is clear: by pairing genotype with phenotype in the same cell and embedding those pairs within time-calibrated phylogenies, researchers can now define, directly in primary human samples, the mechanisms underlying clonal expansion in both healthy and malignant tissues. For a disease that evolves to evade every therapy thrown at it, that may prove the most consequential lens oncology has yet acquired.

Subject of Research: Single-cell multi-omics mapping of genotype and phenotype co-evolution in clonal evolution of healthy and cancerous tissues

Article Title: A single-cell lens into the co-evolution of genotypes and phenotypes in cancer

Article References: Izzo, F., Prieto, T., Potenski, C., & Landau, D. A. (2026). A single-cell lens into the co-evolution of genotypes and phenotypes in cancer. Nature Reviews Cancer. https://doi.org/10.1038/s41568-026-00970-8

Image Credits: AI Generated

DOI: 10.1038/s41568-026-00970-8

Keywords: single-cell multi-omics, clonal evolution, genotype-phenotype mapping, tumour heterogeneity, clonal haematopoiesis, single-cell phylogenetics, lineage tracing, somatic mosaicism, phenotypic plasticity, cancer genomics, epigenetics, therapeutic vulnerabilities

The review, authored by Giulio Caravagna of the University of Trieste and Area Science Park, Trevor A. Graham of the Centre for Evolution and Cancer at The Institute of Cancer Research in London, and Andrea Sottoriva of Human Technopole in Milan, makes a deceptively simple but profound point. Sequencing is a snapshot, whereas evolution is a dynamic process. A single tumour biopsy tells us which mutations are present and at what frequencies, but it does not, by itself, reveal the underlying dynamics that produced them. Bridging this gap requires models. Population genetics, the discipline developed over the past century to understand how allele frequencies change in natural populations under the influence of mutation, selection and drift, provides exactly the framework needed to transform static measurements of variant allele frequencies into quantitative estimates of clonal fitness and evolutionary timings.

At the heart of the framework lies the concept of the site frequency spectrum, the distribution of mutations across different variant allele frequencies within a tumour sample. In a neutrally evolving tumour, one in which no clone enjoys a fitness advantage over its neighbours, mathematical theory predicts a characteristic power-law tail: a predictable excess of mutations at progressively lower frequencies, each arising in expanding lineages as passengers hitchhiking along with the growing clone. This ‘neutral tail’ has become a signature of neutral tumour evolution, first identified across cancer types in work led by the same research groups. Deviations from this expected distribution are the fingerprints of selection. When subclones carrying driver mutations expand faster than neutral expectations, the frequency spectrum distorts in characteristic and quantifiable ways, allowing researchers to estimate the strength of selection rather than merely guess at it.

The authors show how modern computational tools exploit these principles. Subclonal deconvolution, the process of resolving a bulk sequencing sample into its constituent clonal and subclonal populations, has traditionally been treated as a clustering problem. But clustering alone cannot distinguish between a tumour shaped by strong selection and one dominated by spatially constrained growth, where genetic diversity accumulates by neutral drift in separate geographic compartments. By embedding population genetic models directly into inference algorithms, for example approaches combining machine learning with branching process theory, researchers can distinguish genuine selection from the illusions created by tumour architecture and sampling bias. Simulations of mutation, drift and selection can be tuned until the synthetic genomic data they produce match patient samples, converting descriptive allele frequencies into estimates of evolutionary parameters such as selection coefficients and the timing of clonal expansions.

Timing is one of the most clinically valuable outputs of this framework. Clock-like mutational processes, such as the spontaneous deamination of methylated cytosines, accumulate at approximately constant rates, providing a molecular clock against which key events can be dated. Studies of clear cell renal cell cancer have used these principles to time landmark events in tumour evolution, revealing that many chromosomal catastrophes occur astonishingly early, sometimes decades before diagnosis. More recent theoretical work has shown that patient age itself can help distinguish selection from causation in cancer genomes, since a mutation that arises early and confers a growth advantage leaves a different statistical imprint than one that simply accumulates with time. Dating driver events, genome doublings and the origins of metastatic seeds transforms the tumour genome from a parts list into a chronicle.

The review also confronts the limitations and confounders that complicate evolutionary inference from real data. Bulk sequencing averages across millions of cells, obscuring rare subclones and entangling spatial structure with temporal dynamics. Multi-region sequencing and single-cell approaches help, but each introduces its own biases: sampling depth, copy number alterations that distort allele frequencies, and the fundamental fact that a biopsy represents only a fragment of a spatially extended population. Copy number changes in particular must be carefully modelled, since amplifications and deletions shift variant allele frequencies in ways that can mimic or mask selection. The authors emphasise that assumptions embedded in population genetic models, such as well-mixed populations or constant growth rates, must be tested rather than taken for granted, because spatially constrained tumour growth can generate patterns that superficially resemble selection in the absence of any fitness advantage.

Beyond genetics, the framework extends to epigenetic inheritance and phenotypic plasticity, two dimensions of cancer evolution that the standard genetic models handle poorly. Epigenetic states such as DNA methylation patterns are heritable across cell divisions and can be under selection, yet they are reversible and can switch stochastically, creating a one-to-many relationship between genotype and phenotype. Quantitative models adapted from evolutionary theory, including those describing phenotypic plasticity and stochastic switching in fluctuating environments, offer a way to measure the heritability, transition rates and fitness consequences of non-genetic states. This matters enormously for therapy, because drug-tolerant persister cells frequently arise through epigenetic reprogramming rather than genetic mutation, and their dynamics determine whether resistance emerges in weeks or years.

The ecological dimension of tumour evolution receives similar treatment. Cancers are not just populations of competing clones; they are ecosystems in which cells cooperate, cheat and interact with stromal and immune cells. Game theory and eco-evolutionary models capture frequency-dependent selection, in which the fitness of a clone depends on the composition of its neighbourhood, something classical population genetics assumes away. These models have practical consequences. Adaptive therapy strategies, which aim to maintain sensitive clones that suppress resistant ones rather than eradicate the tumour wholesale, draw directly on ecological and game-theoretic reasoning. Similarly, immune selection on neoantigens shapes both tumour antigenicity and response to checkpoint inhibitors, and can be quantified using selection metrics adapted from population genetics.

Ultimately, the review’s central message is one of reframing. Cancer genomes should be read not as catalogues of mutations but as quantitative records of evolutionary processes, in which every allele frequency, every frequency-spectrum distortion and every signature of mutational timing encodes information about the dynamics that produced them. The authors argue that population genetics provides the foundation for understanding and, ultimately, predicting the trajectories of cancer evolution. If the framework fulfils its promise, the implications for precision oncology are substantial: forecasts of relapse timing grounded in measured evolutionary parameters, treatment strategies designed around the predictable dynamics of resistance, and clinical decisions informed not merely by which mutations a tumour carries today, but by the evolutionary forces that will shape what it becomes tomorrow.

The intellectual roots of this framework stretch well beyond oncology. Population genetics matured through the study of natural populations, and its migration into cancer research has been gradual, beginning with early attempts to reconstruct individual tumour histories from genetic data and gaining momentum as sequencing costs fell. A parallel body of work on clonal haematopoiesis and on pre-malignant lesions in normal tissues has reinforced the point that the evolutionary processes described in the review operate long before a tumour is diagnosed. Studies of normal breast tissue, for example, have revealed rare aneuploid epithelial populations and copy number alterations shared with frank cancers, suggesting that the same population genetic tools can illuminate the earliest steps of carcinogenesis across a continuous spectrum from healthy tissue to invasive disease.

Selection in tumours is not exclusively positive. Adapted metrics such as the ratio of synonymous to non-synonymous mutations, borrowed directly from classical genetics, have revealed universal patterns of selection in cancer and somatic tissues, while complementary analyses indicate that negative selection acts on essential cellular functions and on the immunopeptidome, pruning mutations that would compromise basic biology or expose cells to immune attack. Copy number amplifications of wild-type regions may further allow tumours to tolerate otherwise deleterious coding mutations, a reminder that the interplay between different classes of genomic alteration can obscure simple fitness calculations. At population scale, large studies of somatic mutation across many individuals now provide the statistical power to measure these forces with unprecedented precision.

Practical questions of study design also fall within the framework’s remit. How many biopsies are needed to confidently identify truly clonal mutations in a heterogeneous tumour, and how sampling schemes affect the inferred frequency spectrum, are problems that can themselves be solved with evolutionary models rather than ad hoc rules. The review’s figures trace this progression, from conceptual overviews of genomic readouts through the dynamics of variant allele frequencies, the confounding effects of spatial structure, and the distinct evolutionary routes by which tumours respond to therapy and relapse. Together they map a research programme in which the reliability of every evolutionary claim is tied explicitly to the sampling strategy and the model assumptions behind it.

Subject of Research: Application of population genetics models to infer tumour evolutionary dynamics from cancer genome sequencing data.

Article Title: A guide to understanding tumour evolution through the lens of population genetics

Article References: Caravagna, G., Graham, T. A., & Sottoriva, A. (2026). A guide to understanding tumour evolution through the lens of population genetics. Nature Reviews Cancer. https://doi.org/10.1038/s41568-026-00973-5

Image Credits: AI Generated

DOI: 10.1038/s41568-026-00973-5

Keywords: tumour evolution, population genetics, cancer genomics, clonal selection, variant allele frequency, subclonal deconvolution, genetic drift, mutational signatures, phenotypic plasticity, adaptive therapy, tumour heterogeneity, molecular clock