Radiotherapy dose-fractionations and outcomes in cancer patients

Radiotherapy cures many cancers, but the optimum total doses and fractionations used to treat different cancer types remain uncertain. While conventional fractionation (≈2 Gy per fraction) is common in many countries, UK practice has been highly variable. This thesis compared different curative-intent radiotherapy dose-fractionations used in non-small cell lung and breast cancer. These two cancers together make up over a quarter of UK cancer incidence and mortality, and radiotherapy can increase cure rates of both cancers. Two studies were conducted: (A) A meta-analysis of randomised radiotherapy trials in non-small cell lung cancer and (B) A cohort study of non-small cell lung and breast cancer radiotherapy in the Thames Valley. For the meta-analysis, a systematic search was conducted. Eligible studies were randomised comparisons of two or more radiotherapy regimens. Median survival ratios were calculated for each comparison and pooled. 3,795 patients in 25 randomised comparisons of radiotherapy dose were studied. When radiotherapy was given alone, the higher dose within-trial resulted in increased survival (median survival ratio 1.13, 95% confidence interval 1.04-1.22). When radiotherapy was given with concurrent chemotherapy, the higher dose within-trial resulted in decreased survival (median survival ratio 0.83, 95% confidence interval 0.71-0.97). For the cohort study, multiple Public Health England data sources were combined to obtain information on radiotherapy, patient characteristics, and outcomes. Multivariable Cox regressions were conducted separately by cancer site. 324 non-small cell lung, 8,879 invasive breast, and 477 ductal carcinoma in situ patients were studied. In analyses of both non-small cell lung and invasive breast cancer, increasing radiotherapy dose was associated with improved survival in some treatment centres, while in other centres the opposite was true. These opposite trends by treatment centre were unlikely to be explained by chance, and they suggest that differences in patient selection were driving results. There were insufficient events among ductal carcinoma in situ patients to assess associations. Findings from the meta-analysis support consideration of further radiotherapy dose escalation trials, making use of modern methods to reduce toxicity. Findings from the cohort study suggest that it is not possible to use observational studies to examine causal effects of radiotherapy dose-fractionation. This thesis therefore shows the continued importance of conducting sufficiently large randomised trials to ascertain optimal dose-fractionation in radiotherapy.