Aims: Serial photographs have been collected prospectively to evaluate the effect of radiotherapy on normal tissues in the breast. The aim of this study was to compare two methods of scoring radiation-induced changes.
Materials and methods: Five-year photographs of 400 patients randomised to receive either 42.9 or 39 Gy in 13 fractions to the whole breast after tumour excision of early breast cancer were compared with a post-surgery baseline and scored for change in breast appearance on a three-point graded scale. Two alternative methods of scoring using three observers were compared: (a) scores allocated independently, with independent resolution of discrepancies, and (b) scores allocated by consensus.
Results: Treatment effects estimated from the consensus and independent scores were very similar (odds ratio 1.89, 95% confidence interval 1.21-2.96 vs 2.28, 95% confidence interval 1.50-3.47, respectively). Agreement between the scores obtained from each method was reasonable, and the repeatability of the consensus method was good.
Conclusions: The consensus method of scoring photographic change in breast appearance seems to be no less sensitive to randomised dose as the independent method of assessment, but is much quicker to administer. The consensus method has been used to score over 3000 sets of photographs in the National Cancer Research Institute Standardisation of Breast Radiotherapy trial.