By Wolfgang Hackl, MD, CEO OncoGenomX, Allschwil, Switzerland
Precision oncology is entering a new phase.
Over the next 10-15 years, it is likely to evolve from a discipline focused mainly on molecular matching into one that is dynamic, predictive, and deeply integrated with tumour biology, causal reasoning, and advanced AI.
For HealthTech innovators, clinicians, researchers, payers, and regulators, this is not just a technical shift—it is an opportunity to build a more adaptive and more patient-centred cancer care ecosystem.
A more complete view of tumour biology
The next generation of precision oncology will be built on a richer biological understanding of cancer.
Instead of relying on a single biopsy or one biomarker category at one time point, future systems will increasingly integrate genomics, transcriptomics, proteomics, spatial biology, immune context, and longitudinal clinical data.
That will make it possible to capture how tumours evolve, how resistance emerges, and how the microenvironment shapes treatment response.
This matters because tumours are not static. They change over time, often under the selective pressure of therapy.
Better biological models will help stakeholders move from reacting to resistance toward anticipating it.
From correlation to causality
A major step forward will be the shift from descriptive analytics to causal understanding.
Today, many models are excellent at identifying patterns, but the next wave of progress will come from systems that can better estimate what actually changes outcomes.
That includes predicting which therapy is likely to work, for whom, and why.
This is especially important in oncology, where treatment decisions are high-stakes and patient populations are heterogeneous.
Causal methods can help bridge the gap between clinical trial evidence and real-world practice, support individualised treatment selection, and improve confidence in decision-making across different care settings.
AI as a clinical copilot
Artificial intelligence will likely become a core layer of the precision oncology workflow.
Not as a replacement for clinical judgment, but as an assistant that can synthesize complex data, flag uncertainty, suggest scenarios, and support faster and more informed decisions.
Deep learning will continue to be valuable for pattern recognition in imaging, pathology, and multi-omics data, while newer approaches will increasingly combine prediction with interpretability and uncertainty estimation.
The strongest systems will not be the opaquest. They will be the ones clinicians can understand, validate, and trust.
That makes explainability, calibration, and workflow integration essential design principles.
Prediction models that evolve with the patient
A key opportunity lies in moving from one-time risk scores to continuously updated patient models.
As more longitudinal data become available, prediction systems can be refreshed over time to reflect tumour evolution, treatment response, toxicity risk, and changing clinical context.
This opens the door to more proactive care: earlier intervention, smarter sequencing of therapies, and better support for shared decision-making.
In practical terms, this means precision oncology will become less like a snapshot and more like a live map.
That shift could improve outcomes while also making care more efficient and better aligned with patient needs.
Learning from physics and other disciplines
One of the most promising developments is the growing exchange between oncology and fields such as physics, mathematics, systems engineering, and complexity science.
Physics-inspired thinking encourages multi-scale modeling, dynamic systems analysis, and the use of constraints to improve robustness.
These ideas are highly relevant to cancer, where processes unfold across molecular, cellular, tissue, and patient levels.
Cross-disciplinary learning can help the field design models that are not only accurate, but also stable, interpretable, and transferable.
For a mixed HealthTech audience, this is a useful reminder: major innovation often happens when disciplines borrow wisely from one another.
What stakeholders can do now
To ensure this future moves in the right direction, the ecosystem should act early and deliberately.
- Researchers should prioritise longitudinal, multi-modal, and mechanistically rich datasets that support both prediction and causal inference.
- Clinicians should help define clinically meaningful endpoints, validate tools in real workflows, and keep usability central.
- Industry players should build AI systems that are transparent, interoperable, and designed for continuous improvement rather than one-time deployment.
- Payers and health systems should support evidence frameworks that recognise adaptive models, real-world validation, and value-based outcomes.
- Regulators and policymakers should encourage standards for safety, fairness, traceability, and post-deployment monitoring.
- Patients and advocates should be included early, so innovation reflects real priorities such as quality of life, access, and informed choice.
The common thread is collaboration.
Precision oncology will only reach its full potential if the entire value chain works together to create data systems, evidence standards, and care pathways that are both innovative and trustworthy.
A positive path forward
The next decade could make cancer care more personalised, more predictive, and more humane.
If the field invests in biological depth, causal reasoning, responsible AI, and cross-disciplinary learning, precision oncology can become a model for how HealthTech innovation translates complexity into better care.
The opportunity is not only to treat cancer more effectively, but to build an ecosystem that learns continuously and improves with every patient experience.
That is the direction worth pursuing.
