An analysis of dose-effect relationship in the radiotherapy of malignant gliomas.

The relationship between increasing survival and increasing doses of radiotherapy has been examined in 621 patients who were entered into three successive Brain Tumor Study Group protocols between 1966 and 1975. These patients were operated upon and had histologically proven malignant gliomas. The median survival of patients who received no radiotherapy was 18.0 weeks; for those who had ?4500 rad, it was 13.5 weeks (p = .346); those who received 5000 rad had a median survival of 28 weeks (p < .001), 5500 rad - 36.0 weeks (p < .001), and 6000 rad - 42.0 weeks (p < .001). The specific relationship between 5000 and 6000 rad indicates a 1.3 times increase in median life span associated with the higher dose (p = .004). A detailed analysis of specific factors which might have significantly biased results indicated that patients who received less than 4500 rad were not comparable to other groups of patients because they had a poorer initial performance status and a greater number died before completion of radiotherapy. Factors of importance including parameters of radiotherapy, pathology distribution, the influence of corticosteroids, and the effect of age, sex, and initial performance status were all comparable within the various other subgroups. No other treatment characteristics or selective factors which might have a direct effect on survival were identified. It was concluded that radiotherapy had a significant influence on the survival of patients with malignant glioma and that a clear-cut dose-effect relationship exists.