This is an R01 grant application for five years of funding to apply novel advanced image analysis techniques and introducing technology that would improve the targeting in CT guided interventions. CT is currently in the United States the most common imaging modality used to guide biopsy and tumor ablation. The majority of liver tumors such as hepatocellular carcinomas are visible on contrast-enhanced CT or MRI obtained prior to the procedure. Yet, these tumors may not be seen or may have poor margin conspicuity on unenhanced CT images acquired during the procedure. This may increase the procedure time, and/or lead to non-diagnostic cytopathologic assessment, requiring repeat biopsy or sub-optimal ablation applicator placement.
We aim to enhance the current information available to the interventional radiologist, by registering the high-resolution pre-procedural images with low quality intra-procedural images. Anatomical details (including the tumor margin) visible on the pre-procedural images will be superimposed on the intra- procedural images. The precise position of ablation applicator and biopsy needle would be thus estimated with respect to the real contour of the tumor as appearing on the pre-procedural scan. Non-rigid registration is demonstrated as a technology to achieve alignment of images with good accuracy, even in the presence of organ motion. However, up to date it has not been used for fusing pre- and intra- procedural data during CT guided interventions in a clinical suite, since it requires significant computational infrastructure and often these methods are not sufficient robust. We have developed a new non-rigid registration method based on biomechanical model, validated in a prospective study in the area of image-guide neurosurgery. Recently, this new technique is also validated on retrospective data from RFA patients. Pre-procedural MRI and unenhanced intra-procedural CT images are aligned within 5 minutes, with 2mm accuracy. This proposal aims to demonstrate the feasibility of this technology in an interventional radiology suite during image-guided abdominal interventions, without impact on the medical decision. Our predictions for biopsy and ablation will be compared with the histological findings and ablated area on post- procedural MRI. Both mathematical methods and visual inspection by two independent radiologists of results will be employed in assessing the results. The success will be determined by the accuracy, robustness, execution time of the non-rigid registration algorithms. This proposal will benefit public health by developing and assessing key technologies to enable enhanced navigation during the image guided biopsy and ablation. The capacity to visualize the tumor and tumor margin throughout the procedure will better enable the interventional radiologists to achieve better diagnosis, and coagulation necrosis for the entire tumor mass, without unexpected site effects.
This work will lead to new technologies for precise targeting in CT guided interventions. These technologies will facilitate better diagnosis and treatment of liver neoplasms, by enabling the interventional radiologists to more accurately place (i) the biopsy needle and (ii) the ablation applicator, and so, they can (i) improve the specificity in detecting even small tumors and (ii) achieve complete tumor ablation with less risk of removing healthy liver tissue located adjacent to the tumor.
Showing the most recent 10 out of 17 publications
