No abstract

transarterial chemoembolization (cTACE) vs. yttrium-90 radioembolization (Y90) for Hepatocellular Carcinoma (HCC). Materials: With IRB approval, 45 patients with unresectable HCC were prospectively randomized to cTACE vs. Y90 [PREMIERE TRIAL (1)]. From this cohort, the fluoroscopy times and doses were recorded for all of these patients' procedures. Statistical comparison between both cohorts was conducted using independent sample t-test assuming unequal variances. Statistical significance was set at po0.05.Results: Twenty-one patients received cTACE (32 treatment sessions) and 24 patients received Y90 (29 treatment sessions) with glass microspheres. All Y90 patients had a separate planning angiogram prior to radioembolization. Mean fluoroscopic time (95%CI) for the planning angiography prior to Y90 was 11.6 minutes (9.3-14) and mean fluoroscopic dose was 1611 mGy (969-2253). Mean fluoroscopy time (95%CI) for Y90 treatment procedures was 13.3 minutes (10-16.5) and 18.3 minutes (15.5-21) for cTACE procedures (P ¼ 0.021). Mean fluoroscopic dose (95%CI) for Y90 treatment was 1154 mGy (712-1596) and 1834 mGy (1356-2312) for cTACE (P ¼ 0.035). Mean accumulative fluoroscopic time/dose for Y90 included the dose administered on planning angiography and on treatment day. Mean accumulative fluoroscopy time (95%CI) for cTACE patients after multiple treatments was 29 minutes (21-38) and 28 minutes for Y90 patients (24-33) (P ¼ 0.8213). Mean fluoroscopic dose (95%CI) for Y90 patients was 2978 mGy and 3941 mGy (1830-6053) for cTACE patients (P ¼ 0.4). Conclusions: Y90 radioembolization exposes patients to significantly less fluoroscopy time and dose compared to cTACE per treatment session. When combined with planning angiography, there is no statistically significant difference in fluoroscopy time and dose between Y90 and cTACE.

Total patients (n) 46 151 Deceased patients (n) 26 104 OS (P = 0.48) 263 ± 200 days 230 ± 211 days RECIST response CR 0 (0%) 4 (4%) PR 10 (32%) 21 (21%) SD 13 (42%) 40 (41%) PD 8 (26%) 33 (34%) No follow-up 15 53 Average delivered dose (P = 0.46) 34.74 ±14.57 mCi 36.81 ± 12.26 mCi Average prescribed dose (P = 0.31) 42.02 ± 17.03 mCi 38.89 ± 13.03 mCi

Total patients (n) 46 151 Deceased patients (n) 26 104 OS (P = 0.48) 263 ± 200 days 230 ± 211 days RECIST response CR 0 (0%) 4 (4%) PR 10 (32%) 21 (21%) SD 13 (42%) 40 (41%) PD 8 (26%) 33 (34%) No follow-up 15 53 Average delivered dose (P = 0.46) 34.74 ±14.57 mCi 36.81 ± 12.26 mCi Average prescribed dose (P = 0.31) 42.02 ± 17.03 mCi 38.89 ± 13.03 mCi

Segmentation was performed in 3DSlicer (v4.5). For HV and PV segmentation, fiducial seeds were placed and a semiautomated simple region growing algorithm was applied, followed by marching cubes to generate a mesh. For liver segmentation, slices were contoured, followed by a robust statistics segmenter and mesh generation by marching cubes. The final model for 3D printing was generated by performing a boolean subtraction of vasculature, and exported as a stereolithography (.STL) file. The model was then printed on a ProJet 3500 HD Max in VisiJet Crystal, a translucent acrylate polymer. To better visualize HVs and PVs, the interior walls were painted in vibrant colors. Results: The resultant liver model features parenchyma formed with translucent acrylate polymer, enabling visualization of the separately colored HVs and PVs. The HVs and PVs are hollow, allowing evaluation of the geometric relationships between venous systems and "dry run" catheter interrogation. Selection of a post-printing method for coloring the HVs and PVs was important as it greatly reduces the complexity and expense of the 3D print itself, when compared to a multi-color, multi-material approach. Conclusions: As 3D printers become ubiquitous tools of modern medical care, routine 3D printing for TIPS planning is feasible. Using on-site segmentation, a physically transparent liver model may be produced containing hollowed, color coded HVs and PVs to facilitate TIPS planning and potentially decrease morbidity.