ABSTRACT
Radiotherapy schedules maximize tumour control probability and minimize normal tissue control probability (NTCP), but are population-based. Variation in the human genome can take several different forms including single nucleotide polymorphisms, rare variants, small insertions and deletions, copy number variants and epigenetic modifications. NTCP models aim to describe the probability of complication in normal tissue by modelling dose-response relationships. There is both empirical and theoretical evidence supporting the notion that incorporation of genetic or other biologic data into NTCP models can improve their sensitivity and specificity for prediction of toxicity. The inclusion of genetic variables in NTCP models requires identifying sufficient variants to have a clinically useful test. It is the common variants that are of interest given rare variants such as homozygous mutations in the ATM gene can be associated with syndromes that are phenotypically obvious.
