Why recent studies relating normal tissue response to individual radiosensitivity might have failed and how new studies should be performed

Abstract

Purpose: New insights into the kinetics of late complications occurring after radiation therapy indicated that all patients have a constant risk of developing late tissue complications. These observations might have a great impact on studies relating normal tissue complications to individual radiosensitivity.

Methods and materials: Data previously published by Peacock et al. were used for analysis. In this study, 39 breast cancer patients with severe reactions (responders) were compared with 65 matched patients showing no reactions (nonresponders). Cellular radiosensitivity as measured in vitro in terms of D(0.01) did not show significant differences between the two groups, both for high-dose-rate (5.84 +/- 0.06 vs. 5.85 +/- 0.07 Gy) and low-dose-rate (7.44 +/- 0.10 vs. 7.56 +/- 0.09 Gy) irradiation.

Results: A theoretical distribution was calculated for the individual radiosensitivity of patients with Grade <or= 1, Grade 2, or Grade 3 reactions under the following assumptions: (1). The variation of the individual radiosensitivity is described by a normal distribution. (2). All patients and not only a subgroup have a risk of developing late complications. Based on the normal distribution of low-dose-rate data (mean value [MV] = 7.56 Gy, standard deviation [SD] = 0.5 Gy), a total of 200 hypothetical patients were divided into three groups: a resistant group with a sensitivity >or=(MV + SD), a normal group with a sensitivity between MV - SD and MV + SD, and a sensitive group <or=(MV - SD), the relative fractions being 16%, 68%, and 16%, respectively. It was assumed that these groups differed in the risk of developing late complication; for Grade 3 the annual incidence rate was set at 1%, 2%, and 4% and for Grade 2 at 5%, 10%, and 20% per year, respectively. It was shown that the mean cellular sensitivity calculated for Grade 3 (7.39 +/- 0.10 Gy) or Grade 2 patients (7.46 +/- 0.06 Gy) was slightly but not significantly lower than that of Grade <or= 1 patients (7.65 +/- 0.04 Gy). This result demonstrated that even if the risk was assumed to depend clearly on the individual radiosensitivity, significant differences in the mean cellular sensitivity between responders and nonresponders were not expected, just as found by Peacock et al. It was shown that a significant correlation between these two parameters could be detected only when the risk was analyzed separately for each group of patients.

Conclusion: Most data published so far aiming at establishing a relationship between cellular radiosensitivity and the risk of late complications might need to be reevaluated, because the questions asked up to now were inadequate to arrive at a meaningful answer.