Quantification of patient-reported outcome measures of radiation-induced skin reactions for use in clinical trial design.

Abstract

Skin toxicity is a common effect from radiotherapy, although difficult to predict on an individual basis, and there is little evidence-based management. This study aimed to quantify inter-patient variation in patient-reported outcome measures for radiation-induced skin reactions (RISR) to enable the determination of the number of patients required for adequate power in a comparative trial of RISR management strategies. The study included 154 patients scheduled to receive breast cancer radiotherapy. Patients filled in a weekly questionnaire during and up to 4weeks following the end of radiotherapy scoring five aspects of their experience of RISR: skin redness, and bother from redness like itching, burning sensation and tenderness/pain. Assessment of patients' reported experience of their RISR was shown to be feasible, with 91% of patients returning at least two questionnaires. The mean score increase between weeks 1 and 4 was 25 points (p value <0.0001, 95% CI 21-29), and the estimated standard deviation at 4weeks was 18 (95% CI 16-21). Patients' assessment of their reaction was not predicted on the basis of treatment and patient-related characteristics. Based on the observed variance in scores at 4weeks, we could calculate the sample size required for a comparative study of two RISR management policies would be 200 patients to have statistical power to detect a clinically significant difference in patient-rated scores of their skin reactions. A trial employing this tool would help provide an evidence base to guide policy in advising patients how to manage their RISR.