Biological modeling of pelvic radiotherapy: Potential gains from conformal techniques

Abstract

Models have been developed to describe the dose and volume dependences of various long-term rectal complications of radiotherapy; assumptions underlying the models are consistent with clinical and experimental descriptions of complication pathogenesis. In particular, rectal bleeding—perhaps the most common complication of modern external beam prostate radiotherapy, and which might be viewed as its principal dose-limiting toxicity—has been modeled as a parallel-type complication. Rectal dose-surface-histograms have been calculated for 79 patients treated, in the course of the Royal Marsden trial of pelvic conformal radiotherapy, for prostate cancer using conformal or conventional techniques; rectal bleeding data are also available for these patients. A maximum-likelihood fit of the parallel bleeding model to the dose-surface-histograms and complication data shows that the complication status of the patients analyzed (most of whom received reference point doses of 64 Gy) was significantly dependent on, and almost linearly proportional to, the volume of highly dosed rectal wall. Over the range of highly-dosed fractional wall volumes available for analysis (2.5%–6.5%), the rate of RTOG grade 1,2,3 bleeding falls by about 1.1% [95% confidence interval (0.04%, 2.2%)] for each 1% (absolute) reduction in the fraction of rectal wall receiving a dose of 57 Gy or more. The parallel model fit to the bleeding data is only marginally biased by uncertainties in the calculated dose-surface-histograms (due to setup errors, rectal wall movement, and absolute rectal surface area variability), causing the gradient of the observed volume-response curve to be slightly lower than that which would be seen in the absence of these uncertainties. An analysis of published complication data supports these single-center findings and indicates that the reductions in highly dosed rectal wall volumes obtainable using conformal radiotherapy techniques can be exploited to allow escalation of the dose delivered to the prostate target volume, the level of dose-escalation achievable being quite sensitive to the exact conformal technique used. © 2000, American Association of Physicists in Medicine. All rights reserved.