16.04.2026

Artificial intelligence, genomics and personalised oncology

Jacques Bernier, Chief Science Officer, Swiss Medical Network

The exponential growth in genomic data generated by technologies such as whole-exome sequencing (WES) and whole-genome sequencing (WGS) has opened up unprecedented opportunities for precision medicine, particularly in oncology.

However, the complexity and volume of this data are rapidly outstripping the capabilities of traditional human interpretation. It is in this context that artificial intelligence (AI) is proving transformative, enabling faster, more in-depth and clinically relevant analyses.

AI and genomics: detection, annotation and discovery of signatures

MammaPrint: A clinically validated test authorised by the FDA, using a 70-gene signature to predict the risk of recurrence in early-stage breast cancer, enabling the identification of patients who may be able to avoid chemotherapy.
Oncotype DX Colon Cancer Assay: Analyses the expression of 12 genes to predict the risk of recurrence in stage II colon cancer, helping oncologists decide whether adjuvant chemotherapy is necessary.
EPIC Seq (Epigenetic Profiling in cfDNA): An AI-enhanced non-invasive test using cell-free DNA (cfDNA) to classify different cancers (e.g. NSCLC, DLBCL), predict response to immunotherapy and subtype tumours with over 90% accuracy.

These tools illustrate how AI accelerates the interpretation of variants, automates clinical annotation and detects complex genomic signatures beyond what conventional bioinformatics allows.

AI and personalised cancer treatment: tailoring therapy to the genomic profile

Machine learning predictive models

Deep learning models trained on large patient cohorts and drug response databases can predict tumour sensitivity to certain treatments based on their molecular profile.

Counterfactual AI models in ovarian cancer

New AI models simulate various treatment scenarios and provide personalised recommendations, complete with a confidence score and an explainable rationale, by incorporating genomic and transcriptomic data.

Functional precision oncology

Emerging platforms combine AI with ex vivo testing on live tumour cells to identify, for each patient, the most effective and best-tolerated therapies, even in the absence of known mutations.

AI in histopathology and imaging

Stanford AI predicts gene expression from images

A deep learning model developed at Stanford can deduce gene expression profiles of tumours directly from H&E-stained histopathology slides, offering a rapid and less expensive alternative to sequencing in certain contexts.

Multimodal models in radiation oncology

AI systems now combine imaging, genomic data and clinical information to design personalised radiotherapy plans, whilst quantifying uncertainty using advanced models

Global access and equity through AI: examples

Prostate Cancer Programme – Queen’s University Belfast

This programme combines genomics, clinical data, imaging and health economic analysis to identify new biomarkers and optimise treatments whilst reducing side effects.

The US Cancer Moonshot and AI

AI plays a key role in Moonshot initiatives aimed at reducing inequalities in oncology, improving clinical trial matching and supporting clinical decision-making.

Clinical benefits

Time savings: rapid interpretation of high-throughput genomic data
Improved accuracy: mutation detection, risk stratification and treatment prediction
Truly personalised medicine: treatments tailored to molecular and functional profiles
Therapeutic de-escalation: avoiding unnecessary treatments
Wider access: roll-out of precision medicine tools beyond university centres

Remaining challenges

Rigorous clinical validation of AI models across diverse populations
Explainability: clinicians must understand how AI generates its recommendations
Data harmonisation: integrating multi-omic data whilst protecting confidentiality
Training healthcare professionals in the responsible use of AI
Bias and equity: ensuring that AI operates fairly across different healthcare settings

Implications pour Swiss Medical Network

As part of the strategic collaboration between Swiss Medical Network and the Mayo Clinic, AI-assisted genomics can accelerate innovation in precision oncology by:

incorporating validated tests such as MammaPrint or Oncotype DX into care pathways;
deploying next-generation liquid biopsy platforms such as EPIC Seq for non-invasive tumour monitoring;
using multimodal AI tools to guide treatment decisions based on real-time data;
participating in inter-institutional AI collaborations aligned with Mayo Clinic initiatives (the Atlas platform and fundamental models in pathology).

Conclusion

Artificial intelligence applied to genomics and personalised oncology is no longer a theoretical prospect: it is already transforming the way we diagnose, classify and treat cancer. By combining large-scale data analysis with clinical decision support, AI enables clinicians to move from reactive medicine to truly predictive and personalised medicine.

For institutions such as the Swiss Medical Network and the Mayo Clinic, this convergence of AI and precision medicine represents a major driver for accelerating translational research, broadening patient access to innovation and redefining standards of cancer care on a global scale.

Clinique de Genolier

Hospitals

Centres

For patients

About us

Integrated care