24 March 2016 A B-spline image registration based CAD scheme to evaluate drug treatment response of ovarian cancer patients
Author Affiliations +
Abstract
Ovarian cancer is the second most common cancer amongst gynecologic malignancies, and has the highest death rate. Since the majority of ovarian cancer patients (>75%) are diagnosed in the advanced stage with tumor metastasis, chemotherapy is often required after surgery to remove the primary ovarian tumors. In order to quickly assess patient response to the chemotherapy in the clinical trials, two sets of CT examinations are taken pre- and post-therapy (e.g., after 6 weeks). Treatment efficacy is then evaluated based on Response Evaluation Criteria in Solid Tumors (RECIST) guideline, whereby tumor size is measured by the longest diameter on one CT image slice and only a subset of selected tumors are tracked. However, this criterion cannot fully represent the volumetric changes of the tumors and might miss potentially problematic unmarked tumors. Thus, we developed a new CAD approach to measure and analyze volumetric tumor growth/shrinkage using a cubic B-spline deformable image registration method. In this initial study, on 14 sets of pre- and post-treatment CT scans, we registered the two consecutive scans using cubic B-spline registration in a multiresolution (from coarse to fine) framework. We used Mattes mutual information metric as the similarity criterion and the L-BFGS-B optimizer. The results show that our method can quantify volumetric changes in the tumors more accurately than RECIST, and also detect (highlight) potentially problematic regions that were not originally targeted by radiologists. Despite the encouraging results of this preliminary study, further validation of scheme performance is required using large and diverse datasets in future.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Maxine Tan, Maxine Tan, Zheng Li, Zheng Li, Kathleen Moore, Kathleen Moore, Theresa Thai, Theresa Thai, Kai Ding, Kai Ding, Hong Liu, Hong Liu, Bin Zheng, Bin Zheng, "A B-spline image registration based CAD scheme to evaluate drug treatment response of ovarian cancer patients", Proc. SPIE 9785, Medical Imaging 2016: Computer-Aided Diagnosis, 97853D (24 March 2016); doi: 10.1117/12.2216303; https://doi.org/10.1117/12.2216303
PROCEEDINGS
7 PAGES


SHARE
Back to Top