Encrypted login | home

Program Information

Interfraction Bladder Contour Inconsistencies Lead to Miscalculations of D2cc in Cervical Cancer Brachytherapy

no image available
D Park

D Park*, A Medina-Palomo , D Scanderbeg , L Mell , L Cervino , UC San Diego, La Jolla, CA

Presentations

SU-I-GPD-T-39 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall


Purpose: To investigate how bladder contour inconsistencies affect the cumulative D2cc in cervical cancer brachytherapy when using inter-fraction deformable image registration (DIR) for dose accumulation, as compared to direct DVH-based addition.

Methods: This study involves data from three anonymized patients who underwent 3-4 fractions of cervical cancer brachytherapy. The applicator regions were segmented based on Zhen et al. PMB 2015, and removed from the images prior to DIR. To study how differences in contouring affect the accumulated D2cc, we used three different bladder contours in each fraction: one manually segmented (Bi), one obtained by contour-propagation from the reference image (Bi*) and a slight manual modification of Bi* (Bi*’). Finally, we registered each image and dose maps to the reference fraction to accumulate the doses. The total D2cc is calculated considering all the deformed contours. Intensity-based DIR and D2cc calculations were done in MIM. Three different values of D2cc were calculated based on the contours Bi, Bi*, and Bi*’.

Results: The DIR-based cumulative dose D2cc was found to be up to 19.87% larger than the direct DVH addition D2ccclinic. Total D2cc using contours Bi* and Bi*’ were more similar to D2ccB1 when the Dice coefficient between them and B1 approaches 1. Upon visual inspection of images, it was found that D2cc discrepancies were due to both DIR inaccuracies and inter-fraction bladder contour discrepancies. The larger D2cc difference found in this study was mostly due to contour discrepancies.

Conclusion: Besides incorrect DIR, inconsistent contouring of the bladder near the applicator might cause miscalculation of cumulative D2cc. This can cause a D2cc to exceed the clinical worst-case value currently used in clinics. We recommend to use deformed contours, or to pay close attention to consistency in contouring in different image sets. Care should be taken when using contour-based DIR in ensuring contour consistency.

Funding Support, Disclosures, and Conflict of Interest: This project was funded with a UCSD CTRMI seed grant.


Contact Email: