4:03 PMAbstract No. 244 - Radioembolization improves survival in intrahepatic cholangiocarcinoma: a SEER-Medicare population study

Abstract

Purpose: It is well known that planning 99mTc-MAA (TMAA) scans are not consistently reliable for predicting the therapeutic 90Y microsphere (Y90MS) distributions as determined by either SPECT or PET. In this work, we quantify the differences between the planning TMAA distribution and quantitative post-therapy 90YMS bremsstrahlung SPECT/CT by evaluating differences in tumor-to-normal ratios (TNR) for individual tumors. Materials: Between 1/2011 and 12/2012, 11 patients that underwent both TMAA SPECT/CT and post-therapy 90YMS SPECT/CT imaging were retrospectively selected. In all patients, the TMAA and the Y90MS were administered from the common hepatic artery using a novel balloon occlusion technique to induce flow reversal of the gastroduodenal artery and thereby maintaining arterial flow to the liver. Up to three tumors per patient were segmented by an interventional radiologist on a diagnostic CT or MRI. The diagnostic scans were rigidly registered to both the TMAA and 90YMS SPECT/CTs in MiM Maestro. TNR was calculated for each tumor (N1⁄430). Paired t-test with unequal variance was performed to assess for differences in population. Correlation of TNR between TMAA and Y90MS were evaluated. The mean paired 90YMS to TMAA TNR ratio and its variance was also computed. Results: The mean (sd) population TNR for TMAA and 90YMS was 2.4 (1.6) and 2.0 (1.3), respectively. The paired ttest did not show statistically different means between the TMAA and 90YMS population TNR. The mean (sd) paired ratio of Y90 to MAA TNR was 0.99 (0.51) demonstrating equivalence albeit with large variability. A statistically significant Pearson’s correlation coefficient of 0.54 was observed (p1⁄40.002) between 90YMS and TMAA TNR. Conclusions: A well-controlled catheter localization technique reproduces the TMAA TNR on post-therapy 90YMS scans; however, large variances may still be present for individual tumors. Some of these differences may be due to registration errors in reporting TNR and non-ideal image quality for both TMAA and Y90MS scans. The implication of TNR differences for predictive tumor dosimetry is currently under investigation.