Assessment of delivered rectal dose in prostate cancer radiotherapy

Abstract

Prostate radiotherapy gives excellent long-term disease control; however, morbidity from treatment can negatively impact quality of life. To minimise the risk of toxicity a treatment plan is created, which meets dose volume constraints for organs at risk (OAR) such as the rectum. These constraints have been informed by dose response studies which have attempted to establish a relationship between dosimetric parameters and the incidence of toxicity. The position and shape of the surrounding OAR vary over the course of treatment, and as a consequence, the treatment plan may not represent the actual delivered dose. Precise knowledge of the relationship between dose to the OAR and toxicity, essential for improving the radiotherapeutic index, may be limited if derived using the treatment plan. The aim of this PhD was to determine if the accumulated dose could be used to more accurately predict toxicity than using the planned dose. Delivered dose to the rectum was calculated using daily cone beam computed tomography (CBCT) scans to describe the patient anatomy at each treatment fraction in 86 patients. Deformable image registration was used to establish spatial correspondence between the CBCTs, enabling dose to be accumulated. Dose was also accumulated using dose surface maps which described the spatial distribution of dose to the rectal wall. Differences between planned and accumulated dose were compared and correlated with toxicity data, collected using established toxicity scoring measures. This study quantified uncertainties in the dose accumulation methodology including the effect of missing data. It was found that the delivered dose to the rectum accumulated using a subset of CBCTs gave a comparable result to the dose accumulated using a CBCT from all fractions. Comparison of both planned and accumulated rectal dose-volume parameters and spatial metrics showed that planned dose is a good approximation of accumulated rectal dose when state-of-the-art daily image guidance is used to correct for geometric uncertainty in prostate position. The small differences between planned and accumulated dose meant neither was found to be a better predictor of toxicity. Although not statistically significant, the results suggest that dose surface metrics may have had a stronger association with late toxicity than dose-volume metrics.