Fractionwise verification of delivered proton dose to prostate cancer patients based on daily in-room CT imaging

Purpose: Retrospective proton dose calculation based on a unique dataset of daily CT images to confirm our prostate patient positioning and immobilization protocol for counterbalancing interfractional motion.
Material/Methods: For 12 prostate cancer patients treated to 74GyE with double-scattered lateral or anterior oblique proton fields, daily (27-37, median 32) in-room control CTs (cCT) were acquired. Patient preparation includes a drink protocol, water-filled endorectal balloon insertion, bony anatomy alignment by orthogonal X-Ray imaging, and CT-based verification of prostate location via implanted fiducial marker positions. Fraction doses were calculated on all manually delineated cCTs and accumulated on the planning CT by a deformable image registration (DIR) in RayStation 5.99. DVH parameters of iCTVs, bladder, rectum, femoral heads, bladder and rectal wall were analyzed fractionwise prior and after DIR and values from the cumulated and planned dose distributions were compared.
Results: Fig.1 shows the fractionwise assessed DVH parameters for one patient. 275 fraction doses were analyzed in total without finding trends for improving or worsening DVH parameters over treatment time. Intended target coverage, D98%(iCTV)>95%, was missed in 29 cCTs (10.5%) due to suboptimal bladder filling, endorectal balloon position or delineation variation. No overdosage was observed (D2%(iCTV)<105%). DIR led partly to notable changes of DVH parameters (Fig.1). No alarming differences exist between planned and cumulated doses (Fig.2), but significant changes (p<0.05, Wilcoxon signed rank test) were found for D2%(iCTV), V75%(bladder) and V30Gy(bladder wall).
Conclusion: Despite some suspicious fractions, the total delivered doses to prostate cancer patients are accurate with the applied positioning and immobilization protocol.