Physician MD Anderson Cancer Center, United States
Purpose: Characterization of hepatocellular carcinoma (HCC) tumor responses for single-arm single-center prospective 90Y-radioembolization clinical trial (n=40) that used patient-specific voxel-dosimetry-based treatment planning (NCT03896646).
Materials and Methods: Eligibility criteria included adult HCC patients with non-infiltrative tumor diameters ≥3 cm otherwise eligible for 90Y-glass radioembolization. All patients underwent hepatic angiography CT (AngioCT), received 99mTc-MAA injections at all sites of treatment, and underwent SPECT/CT imaging. Fused AngioCT and 99mTc-MAA SPECT/CT was used to plan tumor mean-voxel dose ≥200 Gy and normal liver mean-voxel dose ≤100 Gy. Post-treatment dose verification was performed using 90Y-SPECT/CT and 90Y-PET/CT. Localized mRECIST tumor responses were determined at 6-month follow-up. We characterize tumor response relationships, based on 37 patients 59 tumors who completed 6-month follow-up, with size and dose-coverage metrics.
Results: At 6 months, the DCR [=(CR+PR+SD)/(CR+PR+SD+PD)] was 100%, with 97% ORR [=(CR+PR)/(CR+PR+SD+PD)], and 73% CRR [=(CR)/(CR+PR+SD+PD)]. The median (range) tumor effective diameter was 5.4 (3-15.4) cm; with 40% 3-5cm, 17% 5-7cm, and 25% >7cm. The median (range) tumor voxel-dose delivered was 361 (208-981) Gy; with 100% >200Gy, 83% >300Gy, and 65% >400Gy. The Welch's t-test reported non-significant differences (p=0.68) in tumor mean-dose for CR (420Gy) versus others (396Gy); but significant differences (p=0.0003) in size for CR (4.8cm) versus others (8.4cm). Non-significant differences were also observed when assessing other dose-coverage metrics: D50, D70, D90, D70-50_gradient, D90-50_gradient. Multi-variate logistic regression showed tumor size was prognostic of CR with odd’s ratio 0.39 (95%CI:0.19-0.62) and ROC curve AUC 0.85 (95%CI:0.73-0.98, p< 0.0001). The threshold size for CR was estimated as < 5.7cm with sensitivity of 82% and specificity of 93%. Dose heterogeneity was not correlated with tumor size.
Conclusion: The feasibility and efficacy of prospective, patient-specific, MAA-based voxel dosimetry treatment planning for 90Y-radioembilization has been demonstrated. Mean tumor doses alone do not account for tumor response for tumors larger than 5cm. Other factors in larger tumors may be the reason for lack of CR.
