Valencia Lozano Ingrid, Buss Elizabeth, Spina Catherine S, Horowitz David P, Kachnic Lisa A, Price Michael, Wang Yi-Fang, Munbodh Reshma
Department of Radiation Oncology, Columbia University Irving Medical Center, New York, New York, USA.
J Appl Clin Med Phys. 2025 Jun;26(6):e70074. doi: 10.1002/acm2.70074. Epub 2025 Mar 21.
This study quantitatively evaluates bladder changes and their dosimetric impact during the on-couch adaptive process on a commercial CBCT-based online adaptive radiotherapy (CT-gART) platform.
Data from 183 fractions of ten patients receiving online ART for pelvic cancers were analyzed retrospectively. Bladder contours were automatically generated and revised by an expert for each pair of planning and verification CBCTs. Bladder shape changes were assessed using geometric and boundary distance metrics. A deformable image registration (DIR) workflow was implemented to obtain spatial motion characteristics, validated by the dice similarity coefficient between bladder contours. Dosimetric parameters were quantified by warping 'intended' dose distributions to the verification CBCT anatomy using DIR to evaluate coverage and OAR objectives.
Bladder volume changed noticeably during the on-couch adaptation process (19.7 ± 3.3 min). Day-to-day bladder expansion showed an average increase of 3.4 cc/min ± 1.5 cc/min for the full bladder and 0.8 cc/min ± 0.3 cc/min for empty bladder protocols. Deformation occurred mainly in the superior region and was more pronounced for the full bladder protocol. Displacements over 5 mm in cranial-caudal and anterior-posterior directions averaged 16% and 5% of the volume for full bladders and 5% and 4% for empty bladders, respectively. CTV coverage (V100%) was maintained when the bladder was the target, but PTV V95% was reduced by an average of 7%. For non-bladder treatments, bladder constraints increased slightly for supine subjects (0.5 Gy/fx), with prone subjects almost unaffected.
A framework using auto-segmentation and DIR was developed to evaluate the intra-fractional motion of the bladder during CTgART. Results suggest that reducing the isotropic PTV margin to less than 7 mm may be feasible for oART, allowing patient-specific anisotropic margins while maintaining the quality of the adaptive plan.
本研究定量评估在基于商用CBCT的在线自适应放疗（CT-gART）平台上进行的治疗中自适应过程期间膀胱的变化及其剂量学影响。
回顾性分析了10例接受盆腔癌在线自适应放疗的患者的183个分次的数据。由专家针对每对计划和验证CBCT自动生成并修正膀胱轮廓。使用几何和边界距离度量评估膀胱形状变化。实施了可变形图像配准（DIR）工作流程以获得空间运动特征，并通过膀胱轮廓之间的骰子相似系数进行验证。通过使用DIR将“预期”剂量分布扭曲到验证CBCT解剖结构来量化剂量学参数，以评估靶区覆盖和危及器官目标。
在治疗中自适应过程（19.7±3.3分钟）期间膀胱体积有明显变化。日常膀胱扩张显示，对于全膀胱方案，全膀胱平均增加3.4 cc/分钟±1.5 cc/分钟，空膀胱方案为0.8 cc/分钟±0.3 cc/分钟。变形主要发生在上部区域，对于全膀胱方案更为明显。全膀胱在头脚和前后方向上超过5毫米的位移分别平均为体积的16%和5%，空膀胱分别为5%和4%。当膀胱为靶区时，CTV覆盖率（V100%）得以维持，但PTV V95%平均降低了7%。对于非膀胱治疗，仰卧位患者的膀胱限制略有增加（0.5 Gy/分次），俯卧位患者几乎未受影响。
开发了一个使用自动分割和DIR的框架来评估CTgART期间膀胱的分次内运动。结果表明，对于在线自适应放疗，将各向同性PTV边界减小到小于7毫米可能是可行的，在保持自适应计划质量的同时允许根据患者情况设置各向异性边界。
