Memorial Sloan Kettering Cancer Center New York, NY
N. Hassan Rezaeian1, C. Gui2, L. Bosma1, S. Madhavan1, S. Burleson3, N. Cote4, V. S. Brennan2, and N. Tyagi3; 1Memorial Sloan Kettering Cancer Center, New York, NY, 2Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, 3Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, NY, 4Memorial Sloan Kettering Cancer Center, New York, NY, United States
Purpose/Objective(s): To perform deformable dose accumulation on prostate cancer patients undergoing MR-guided radiotherapy (MRgRT). Materials/
Methods: Over 100 intermediate risk prostate cancer patients were treated with daily online MR-guided adaptive replanning on the 1.5 Tesla MR-LINAC system and a prescribed dose of 40Gy in 5 fractions with a boost of 45Gy to the dominant intraprostatic lesion (DIL). Target and normal organ contours were generated daily on pre-beam images and an adaptive plan was created. A workflow was developed inside MIMTM to register and deform pre- and post-treatment MR for each fraction to obtain intra-fraction displacement-vector-fields (DVF) A multimodality deformable image registration method with normalization and smoothness factor of 0.5 was used in MIM. Inter-fraction deformation was then performed between each daily fraction’s images. QA of deformable image registration (DIR) and deformed contours was performed by calculating Hausdorff Distance at the 95th percentile (HD95), Jacobian determinant (JD), Curls, and DICE coefficient. A ground truth prostate dataset with empty bladder was created based on biomechanical properties of pelvis organs and deformed to another state (full bladder). Ground truth contours and DVFs were compared with the MIM DIR method. Accumulated Dmax, D10%, and D50% metrics for bladder, urethra and rectum were extracted and compared with our institutional constraints (bladder - Dmax<42Gy, D10%<36Gy and D50%<20Gy; urethra - Dmax<42Gy; rectum: Dmax<41.2Gy, D1cc<38.5Gy). GU toxicity was scored using an IPSS increase of more than 10 points and CTCAE for events occurring within 180 days of the first adaptive RT fraction (acute) and after 180 days (late). Results: Dose accumulation was completed on the first 15 patients. JD between MIM DIR and ground truth dataset for bladder, rectum and prostate was 1.04 vs 0.94; 1.02 vs 0.96; 0.93 vs 1.01. For first 15 cases, HD95, JD, Curl and DICE for prostate, rectum and bladder was (1.42, 1.08, 0.15, 0.89); (1.33, 1.1, 0.19, 0.86) and (1.8, 0.88, 0.21 and 0.89) (Fig1). Urethra being a very small structure yielded the smallest value for DICE, Curl and HD95.The average accumulated Dmax doses for Bladder, Urethra, and Rectum were 4182±101, 4278±68, and 4038±176 cGy. The average accumulated D10% and D50% doses for bladder were 3417±311cGy and 1695±402cGy. The average accumulated D1cc for the rectum was 3462±270cGy. Figure 2 shows the spread in accumulated bladder and urethra DVHs for these cases. Acute urinary toxicity was observed in two patients while late toxicity was observed in one patient. Urethra Dmax was 100cGy and 190cGy above our institutional constraints on two patients. One patient exceeded bladder Dmax by 20cGy from ourconstraint.
Conclusion: Dmax for urethra exceeded our constraints for 12 out of 15 patients. Future work will include reevaluating our institutional constraints based on accumulated doses rather than planned dose.
