We propose to develop, implement, and test in human studies the use of a computerized tool - an interventionalist's 'cyber-aide'for monitoring MRI-guided cryoablation. The overall purpose of this cyber-aide is to improve the safety and effectiveness of tumor ablations. It will accomplish this purpose by improving both the monitoring and controlling aspects of the procedure. Using MRI's ability to depict the ablation iceball as a well circumscribed high contrast to noise ratio object in all tissues, the cyber aide will provide 3D volume renderings of the tumor, cryoablation probes, and iceball. It will produce continuous, real-time metrics quantifying the 3D comparison of iceball volume versus the tumor volume and adjacent critical structures. As a result, the interventionalist will be able to control the treatment to maximize tumor coverage and minimize risk of injury to adjacent structures. The unique aspect of this proposed software product is its capabilities that extend beyond human limits to quantify the progress of the therapy and monitor its safe application in real time. In our proposed software, the progress of the therapy will be succinctly displayed to the interventionalist in the procedure room. The display will contain: 3D/ 2D display for visualization of the segmented and rendered tumor, cryoablation probes and iceball as of the latest image acquisition during the therapy;a prediction of near future iceball coverage computed from optical flow;and, current and predicted tumor coverage metrics. The cyber-aide's user interface will also allow the interventionalist to initially 'mark'nearby critical structures (e.g., bowel wall) that are to be protected. An alarm system will be provided that will 'warn'the interventionalist of an upcoming possibility that a critical structure may be affected. The software will do this by comparing the optical flow prediction of iceball volume at the next imaging time-point to the segmented protected structures - a critical feature for safety and effectiveness of image-guided thermal therapies. After off-line validation tests, the cyber-aide will be implemented at two BWH interventional MRI suites (70 cm wide-bore Siemens Verio 3.0T MRI) to clinically test its use in cryoablation of renal cell carcinoma in 59 human subjects performed by our Tumor Ablation Program.
Our Specific Aims i nclude: (1) technical development of our 'cyber-aide'software, including the designing, programming and testing of the segmentation, non-rigid registration, optical flow, statistical analysis, and user interface software;(2) optimization and validation of the accuracy of each software component using off-line testing;and (3) validation of the safety and efficacy in human subject cryoablation of renal cell carcinomas using post-procedural MRI findings as gold standard and by comparing two groups of procedures performed before and after the implementation of the software. Results of this study are expected to have an important positive impact, not only for MRI-guided kidney tumor cryoablations, but also for all other image-guided ablations in need of quantitative real-time monitoring and control, like MRI- guided liver, soft-tissue, bone and CT-guided lung, liver, soft-tissue and bone tumor ablations.
The proposed study is important because it will test the clinical validity of a novel computerized ablation monitoring and control tool that automates, combines, and displays all the information that has to be continuously evaluated and acted upon during these complex procedures, such as how well the tumor is treated and should the treatment plan be changed, and whether surrounding critical structures will be safe as the procedure continues and when should the ablation be stopped. Once proven, it can have a positive impact to public health since this cyber-tool can be adapted and used during all other type of image-guided ablations of tumors.
| Liu, Xinyang; Tuncali, Kemal; Wells, William M et al. (2015) Automatic iceball segmentation with adapted shape priors for MRI-guided cryoablation. J Magn Reson Imaging 41:517-24 |
| Liu, Xinyang; Tuncali, Kemal; Wells 3rd, William M et al. (2012) Fully automatic 3D segmentation of iceball for image-guided cryoablation. Conf Proc IEEE Eng Med Biol Soc 2012:2327-30 |
