Utility of pre-procedural [^99mTc]TcMAA SPECT/CT Multicompartment Dosimetry for Treatment Planning of ⁹⁰Y Glass microspheres in patients with Hepatocellular Carcinoma: comparison of anatomic versus [^99mTc]TcMAA-based Segmentation

Purpose: Pre-treatment [^99mTc]TcMAA-based radioembolization treatment planning using multicompartment dosimetry involves the definition of the tumor and normal tissue compartments and calculation of the prescribed absorbed doses. The aim was to compare the real-world utility of anatomic and [^99mTc]TcMAA-based segmentation of tumor and normal tissue compartments. Materials and methods: Included patients had HCC treated by glass [⁹⁰Y]yttrium microspheres, ≥ 1 tumor, ≥ 3 cm diameter and [^99mTc]TcMAA SPECT/CT imaging before treatment. Segmentation was performed retrospectively using dedicated dosimetry software: (1) anatomic (diagnostic CT/MRI-based), and (2) [^99mTc]TcMAA threshold-based (i.e., using an activity-isocontour threshold). CT/MRI was co-registered with [^99mTc]TcMAA SPECT/CT. Logistic regression and Cox regression, respectively, were used to evaluate relationships between total perfused tumor absorbed dose (TAD) and objective response rate (ORR) and overall survival (OS). In a subset-analysis pre- and post-treatment dosimetry were compared using Bland-Altman analysis and Pearson’s correlation coefficient. Results: A total of 209 patients were enrolled. Total perfused tumor and normal tissue volumes were larger when using anatomic versus [^99mTc]TcMAA threshold segmentation, resulting in lower absorbed doses. mRECIST ORR was higher with increasing total perfused TAD (odds ratio per 100 Gy TAD increase was 1.22 (95% CI: 1.01–1.49; p = 0.044) for anatomic and 1.19 (95% CI: 1.04–1.37; p = 0.012) for [^99mTc]TcMAA threshold segmentation. Higher total perfused TAD was associated with improved OS (hazard ratio per 100 Gy TAD increase was 0.826 (95% CI: 0.714–0.954; p = 0.009) and 0.847 (95% CI: 0.765–0.936; p = 0.001) for anatomic and [^99mTc]TcMAA threshold segmentation, respectively). For pre- vs. post-treatment dosimetry comparison, the average bias for total perfused TAD was + 11.5 Gy (95% limits of agreement: -227.0 to 250.0) with a strong positive correlation (Pearson’s correlation coefficient = 0.80). Conclusion: Real-world data support [^99mTc]TcMAA imaging to estimate absorbed doses prior to treatment of HCC with glass [⁹⁰Y]yttrium microspheres. Both anatomic and [^99mTc]TcMAA threshold methods were suitable for treatment planning. Trial registration number: NCT03295006.