David Thomas Glidden BS, Presenter: Nothing to Disclose

Grayson L. Baird MS, Abstract Co-Author: Nothing to Disclose

Derek Merck, Abstract Co-Author: Nothing to Disclose

Damian E. Dupuy MD, Abstract Co-Author: Research Grant, NeuWave Medical Inc Board of Directors, BSD Medical Corporation Stockholder, BSD Medical Corporation Speaker, Educational Symposia

To propose the foundation of a quantitative method for validation of thermal ablations.

24 patients (M:F= 10:14) with solitary lung tumors underwent microwave ablation under CT-guidance. Each tumor was treated with one of four MW applicators (BSD Medical, Salt Lake City, UT, Neuwave Medical, Madison, WI) for 5-15 minutes according to the manufacturers’ specifications. Each case included a CT scan pre- intra- and post-procedure. Tumor volumes were manually segmented from pre-scans and ablation volumes from post-scans using the ground glass halo surrounding the tumor. Pre-scans were fused onto post-scans using two algorithms—a rigid registration, and a rigid plus deformable registration. Volume overlap resulting from both algorithms were calculated. Bland-Altman plots and Deming regression were used to identify possible differences in these image fusion techniques.

The volume overlap between tumors and ablation zones increased proportional to tumor size when deformable registration was applied (p < 0.001). Deming regression showed a significant deviation from perfect concordance between rigid and deformable registration (95 % CI: [1.13, 1.39]) in which more volume overlap was attributable to deformable registration.

Quantitative validation of thermal ablation margin analysis remains challenging due to inherent tumor position and morphology changes after ablation. Rigid registration techniques rarely reflect how an ablation zone covers the tumor and margin because of movement (e.g. respiratory, tumor displacement, patient position). The addition of deformable registration may more accurately reflect how the tumor and ablation zone overlap, thus improving local control outcomes.

Improved fusion between pre- and post-scans using deformable registration will provide a basis for quantitative validation of thermal ablations by correcting for anatomical movement.

Glidden, D, Baird, G, Merck, D, Dupuy, D, Quantitative Validation of Thermal Ablation: An Improved Image Fusion Algorithm to Reflect Treatment Coverage.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14011262.html