A genome-based model for adjusting radiotherapy dose (GARD): a retrospective, cohort-based study

Area of Science:

• Oncology
• Radiation Oncology
• Genomics

Background:

• Radiotherapy is a common cancer treatment but lacks personalization.
• Current methods do not adjust radiation doses based on biological differences in tumors.
• There is a need for precision medicine approaches in radiotherapy.

Purpose of the Study:

• To assess if a patient-specific molecular signature of radiation sensitivity can determine the optimal radiotherapy dose.
• To introduce and validate the genomic-adjusted radiation dose (GARD) as a predictor of radiotherapy outcomes.

Main Methods:

• Derived the genomic-adjusted radiation dose (GARD) using a gene-expression-based radiation sensitivity index and the linear-quadratic model.
• Calculated GARD for 8271 primary tumors across 20 disease sites from the Total Cancer Care (TCC) protocol.
• Assessed GARD's association with clinical outcomes in five independent patient cohorts using multivariable Cox modeling.

Main Results:

• Calculated GARD for 8271 tissue samples, revealing wide variations across and within tumor types.
• Median GARD values varied significantly by cancer type, being lowest in gliomas/sarcomas and highest in cervical/head and neck cancers.
• GARD independently predicted clinical outcomes in breast, lung, glioblastoma, and pancreatic cancers, with high GARD correlating to better survival in breast cancer patients.

Conclusions:

• A GARD-based clinical model can individualize radiotherapy dosage based on tumor radiosensitivity.
• This approach provides a framework for developing genomically-guided clinical trials in radiation oncology.
• GARD represents a significant step towards precision medicine in cancer radiotherapy.