Radiogenomics and Toxicity Prediction in Cancer Therapy

Radiogenomics integrates genomic profiling with radiation oncology to understand how inherited genetic variants influence individual responses to radiotherapy. By linking single nucleotide polymorphisms and gene expression patterns to clinical outcomes, this field seeks to predict normal tissue toxicity and guide personalised treatment plans. Predictive biomarkers range from functional assays such as radiation-induced lymphocyte apoptosis to polygenic risk scores derived from genome-wide association studies. Advances in computational methods and machine learning have enhanced the accuracy of risk stratification, enabling clinicians to identify patients at high or low risk of adverse effects. Ultimately, radiogenomics aims to optimise therapeutic efficacy while minimising side effects, improving quality of life for cancer survivors and informing dose‐modification strategies across diverse tumour types.

Research from Nature Portfolio

Recent studies have refined functional and computational approaches for toxicity prediction. A multicentre investigation in prostate cancer patients demonstrated that pre-treatment measurements of radiation-induced lymphocyte apoptosis reliably predict late urinary side effects. Patients were stratified by apoptotic response, with lower apoptosis rates correlating with higher symptom scores, and models incorporating apoptosis outperformed those based solely on clinical variables. Another study introduced a novel machine learning framework to construct polygenic risk models from hundreds of GWAS-derived variants. Applied to prostate radiotherapy endpoints such as rectal bleeding and erectile dysfunction, this pre-conditioned random forest regression approach yielded improved predictive performance over traditional methods and highlighted key DNA repair and immune pathways implicated in normal tissue damage.

Radiogenomics and Toxicity Prediction in Cancer Therapy publication trend

The graph below shows the total number of articles in radiogenomics and toxicity prediction in cancer therapy across all publications each year (not limited to Nature Index journals).

Technical terms

Radiogenomics: The study of how genetic variation affects individual responses to radiation therapy.

Single nucleotide polymorphism (SNP): A DNA sequence variation involving a single base pair, used as a genetic marker in association studies.

Genome-wide association study (GWAS): An approach that scans the genome for SNPs to identify genetic variants associated with traits or treatment outcomes.

Radiation-induced lymphocyte apoptosis (RILA): A functional assay measuring programmed cell death in lymphocytes after radiation exposure, indicative of radiosensitivity.

Polygenic risk model: A predictive tool combining multiple genetic variants to estimate an individual’s risk of a specific outcome, such as toxicity.

References

1. A two-stage genome-wide association study identifies novel germline genetic variations in CACNA2D3 associated with radiotherapy response in nasopharyngeal carcinoma. Journal of Translational Medicine (2023).
2. Genome-wide association study of treatment-related toxicity two years following radiotherapy for breast cancer. Radiotherapy and Oncology (2023).
3. Assessment of the predictive power the radiation-induced lymphocyte apoptosis method in prostate cancer patients. Scientific Reports (2025).
4. Radiogenomics Consortium Genome-Wide Association Study Meta-Analysis of Late Toxicity After Prostate Cancer Radiotherapy. Journal of the National Cancer Institute (2019).
5. Computational methods using genome-wide association studies to predict radiotherapy complications and to identify correlative molecular processes. Scientific Reports (2017).