Accumulated bladder wall dose and urinary toxicity 143 Discussion Our study suggests a strong relationship between Dmean as well as low-mediumdoses (V10-35Gy) to the bladder and bladder wall and patient-reported acute urinary toxicity in prostate cancer patients treated with MRI-guided SBRT. These correlations persisted after correction for multiple baseline characteristics. This is the first study to identify such a relationship for the bladder wall using the accumulated dose over the five treatment fractions. Based on our results, we suggest the use of constraints for bladder wall V25Gy (cm3) and bladder (wall) Dmean for treatment planning for MRIguided prostate cancer SBRT. Prospective validation of the suggested (soft) constraints is warranted. Furthermore, future research should focus on determination of the optimal constraints, feasibility of these constraints in treatment planning, and finally the clinical effects with respect to acute urinary toxicity. While some of the bladder dose parameters were associated with the outcome, these associations seemed generally weaker (lower AUC values) compared to those for the bladder wall. This suggests higher accuracy of the bladder wall dose in predicting toxicity and is in line with the hypothesis that irradiation of the bladder wall – and not bladder content – induces toxicity. For the relative dose parameters (V10-37Gy in % and Dmean), correlations between the bladder and bladder wall were high, whereas for absolute V10-30Gy (cm3), correlations between the bladder and bladder wall dose were generally weaker except for the highest dose-volumes (Supplementary C). This probably is caused by the fact that the lower absolute bladder dose volumes also include large parts of bladder content. Together with the lack of significant correlations for the absolute bladder dose parameters, this suggests that the absolute bladder volume receiving X Gy is not a sufficient proxy for the bladder wall. Although the bladder wall parameters showed higher predictive value (based on AUC), the relative bladder dose parameters might be a sufficient proxy, especially for Dmean. Absolute volumetric constraints are more practical because they do not require (online) delineation of the entire organ. However, concerning absolute dose parameters, we only found significant correlations for the bladder wall, and therefore we suggest delineating the bladder wall to be able to use for example V25Gy (cm3) in treatment planning. Especially for the bladder wall, absolute volumetric constraints can be used, since bladder wall volume does not change over the course of treatment. Since Dmean seems important in the prediction of acute urinary toxicity, the bladder size during treatment could contribute to the risk of toxicity. Mean bladder volume was larger (but not statistically significant) in the non-toxicity group compared to the group of patients with toxicity (Table 1). No strict bladder filling protocol was applied and therefore bladder size varied significantly between fractions and patients. A bladder filling protocol aimed at a stable, (comfortably) filled bladder could potentially reduce the mean dose and thereby reduce the risk of toxicity.26 In addition, improvements in delivery accuracy, i.e. smaller PTV margins, are warranted to significantly reduce the dose to the entire bladder (wall) while providing adequate target coverage. To achieve this, fast, online adaptive techniques are needed to counteract intrafraction motion.27,28 7
RkJQdWJsaXNoZXIy MTk4NDMw