Inefficient DNA-DSB repair via the homologous recombination pathway in prostate cancer patients with late radiation toxicity

Download

Purpose: Severe late normal tissue damage limits radiotherapy treatment regimens. This study aims to validate γ-H2AX foci decay ratios and induced expression levels of DNA double strand break (DSB) repair genes, found in a retrospective study, as possible predictors for late radiation toxicity. Methods and Materials: Prospectively, decay ratios (initial/residual γ-H2AX foci numbers) and genome-wide expression profiles were examined in ex vivo irradiated lymphocytes of 198 prostate cancer patients. All patients were followed ≥2 years after radiotherapy, clinical characteristics were assembled and toxicity was recorded using the Common Terminology Criteria (CTCAE) v4.0. Results: No clinical factors were correlated with late radiation toxicity. Analysis of γ-H2AX foci uncovered a negative correlation between the foci decay ratio and toxicity grade. Significantly smaller decay ratios were found in grade≥3 compared to grade 0 patients (p=0.02), indicating less efficient DNA-DSB repair in radio-sensitive patients. Moreover, utilizing a foci decay ratio threshold determined in our previous retrospective study correctly classified 23 of the 28 grade≥3 patients (sensitivity, 82%) and 9 of the 14 grade 0 patients (specificity, 64%). Grade of toxicity also correlated with a reduced induction of the homologous recombination (HR) repair gene-set. The difference in average fold induction of the HR gene-set was most pronounced between grade 0 and grade≥3 patients (p=0.008). Conclusions: Reduced responsiveness of HR repair genes to irradiation and inefficient DSB repair correlate with an increased risk of late radiation toxicity. Using a decay ratio classifier, we could correctly classify 82% of the patients with grade≥3 toxicity. Additional studies are required to further optimize and validate the foci decay assay and to assess its predictive value for late radiation toxicity in patients prostate cancer

Cancer Type: Prostate Cancer

GPL: 13158

Number of Genes with Data: 978

Number Observations: 440

Number Outputs: 2

Imbalanced Ratio: 1

Default Task: Classification