Gynecological Cancer

PV QA 4 - Poster Viewing Q&A 4

TU_14_3455 - Validation of Pre-Brachytherapy MRI-guided, CT-Based Intracavitary High Dose Rate Treatment of Locally Advanced Cervical Cancer Using Deformable Image Registration

Tuesday, October 23
2:45 PM - 4:15 PM
Location: Innovation Hub, Exhibit Hall 3

Validation of Pre-Brachytherapy MRI-guided, CT-Based Intracavitary High Dose Rate Treatment of Locally Advanced Cervical Cancer Using Deformable Image Registration
B. A. Dyer1, Z. Yuan2, J. S. Mayadev3, J. Qiu4, S. H. Benedict1, R. K. Valicenti1, and Y. Rong1; 1University of California Davis Comprehensive Cancer Center, Sacramento, CA, 2Hubei Cancer Hospital, Wuhan, China, 3University of California San Diego, La Jolla, CA, 4Taishan Medical University, Taian, China

Purpose/Objective(s): To establish a feasible clinical workflow for using pre-brachytherapy implantation MRI in guiding CT-based high dose rate (HDR) brachytherapy treatment of locally advanced cervical cancer (LACC) with an advanced deformable image registration (DIR) technique. The accuracy and clinical utility of MRI-to-CT DIR related to significant applicator-induced anatomic distortion is validated.

Materials/Methods: 26 patients with LACC had pre-brachytherapy implantation MRI (pre-MRI) within 2 weeks of the first brachytherapy treatment. CT imaging was used for high risk target volume (HR-CTV) delineation after brachytherapy implantation (post-CT). Pre-MRI was registered and deformed to post-CT images using an anatomically constrained deformation algorithm (ANACONDA). Cervix or cervix plus 10 mm margin was used as a controlling structure (ROI) for deformations. The MRI-defined HR-CTV was propagated to post-CT images and was evaluated against prospectively drawn, CT-defined HR-CTV brachytherapy targets. Clinical utility was defined using a qualitative rubric from 0 (low performing) to 4 (high performing). Quantitative deformation metrics included Dice index, distance to agreement (DTA), center of mass (COM) displacement, and cervical volume change. Various statistical methods helped to identify predictors of clinical utility and deformation modes of failure.

Results: Using the cervix as a controlling ROI, the ANACONDA algorithm achieved clinical utility > 3 in 65% patients. If cervix plus margin was used as the controlling ROI clinical utility improved to 81%. Total COM displacement (cervix plus uterus) had the highest sensitivity in predicting low or high deformation performance, and a displacement threshold for sensitivity was identified. Deformation failure (low clinical utility) resulted from high COM displacement, large cervical volume change, and retroverted uterine anatomy. Quantitative predictive threshold values for the various modes of failure were identified and were strongly correlated with clinical utility. Deformation failure could be overcome by propagating MRI-defined gross tumor volume to the post-CT images with additional margin after deformation.

Conclusion: We established a feasible MRI-to-CT DIR workflow for HDR brachytherapy treatment of LACC. The ANACONDA deformation algorithm produces accurate and precise MRI-to-CT deformations. Several deformation modes of failure were identified and alternative deformation workflows were successfully employed. Deformation failures and DIR solutions warrant further study and prospective validation.

Author Disclosure: B.A. Dyer: None. Z. Yuan: None. J.S. Mayadev: None. S.H. Benedict: None. R.K. Valicenti: None. Y. Rong: Section Editor; JACMP. Associate Editor; Medical Physics, Medical Dosimetry.

Send Email for Brandon Dyer


Assets

TU_14_3455 - Validation of Pre-Brachytherapy MRI-guided, CT-Based Intracavitary High Dose Rate Treatment of Locally Advanced Cervical Cancer Using Deformable Image Registration



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Validation of Pre-Brachytherapy MRI-guided, CT-Based Intracavitary High Dose Rate Treatment of Locally Advanced Cervical Cancer Using Deformable Image Registration