Adaptive-Predictive Organ Localization Using Cone-Beam Computed Tomography for Improved Accuracy in External Beam Radiotherapy for Bladder Cancer

Abstract

To examine patterns of bladder wall motion during high-dose hypofractionated bladder radiotherapy and to validate a novel adaptive planning method, A-POLO, to prevent subsequent geographic miss. Patterns of individual bladder filling were obtained with repeat computed tomography planning scans at 0, 15, and 30minutes after voiding. A series of patient-specific plans corresponding to these time-displacement points was created. Pretreatment cone-beam computed tomography was performed before each fraction and assessed retrospectively for adaptive intervention. In fractions that would have required intervention, the most appropriate plan was chosen from the patient's "library," and the resulting target coverage was reassessed with repeat cone-beam computed tomography. A large variation in patterns of bladder filling and interfraction displacement was seen. During radiotherapy, predominant translations occurred cranially (maximum 2.5 cm) and anteriorly (maximum 1.75 cm). No apparent explanation was found for this variation using pretreatment patient factors. A need for adaptive planning was demonstrated by 51% of fractions, and 73% of fractions would have been delivered correctly using A-POLO. The adaptive strategy improved target coverage and was able to account for intrafraction motion also. Bladder volume variation will result in geographic miss in a high proportion of delivered bladder radiotherapy treatments. The A-POLO strategy can be used to correct for this and can be implemented from the first fraction of radiotherapy; thus, it is particularly suited to hypofractionated bladder radiotherapy regimens.