Abstract No. 227 External Validation of Different Renal Tumor Ablation Risk Stratification Scoring Systems in Predicting Complication Rates and Post-Ablative Residual Disease Rates in Heat and Cold Ablative Modalities

To compare and externally validate multiple proposed renal tumor ablation risk-stratification algorithms on prediction of complication and post-ablative residual disease following renal cell carcinoma (RCC) ablation with both heat and cold modalities. Retrospective study performed on 100 patients (mean age, 67.3±10.4 years) and 126 biopsy-confirmed RCC who underwent renal ablation (tumor diameter 2.5±1.0 cm; left:right kidney 56:70, T1a:T1b 121:5). Mean follow-up duration was 67.3±37.2 months. Ablative modality included 75 radiofrequency ablations (RFA), 50 cryoablations, and 1 microwave ablation (MWA). Complications were classified using the Society of Interventional Radiology adverse event classification criteria. RENAL, mRENAL, (MC)2, P-RAC, and P-RENAL were validated with area under the curve (AUC) values of ROC curves against complications and incomplete ablation. Out of 126 ablations, 11 mild (8.7%) and 2 moderate (1.6%) complications occurred. 3 of the mild complications occurred with cryoablation and 8 with RFA. Both moderate complications occurred following MWA. No severe complications were observed. T1a tumors included 10 mild complications and 2 moderate complications. T1b tumors included 1 mild complication. The most common complications were bleeding and pain. 97 ablations achieved complete ablation on first follow up, while 29 did not achieve complete ablation. Mean scores of each of the predictive algorithms are depicted in Table 227.1. Within the context of all ablative modalities, none of the algorithms were strong predictors for complications given the AUC values. When isolated to the context of cryoablations only, P-RAC (0.89), mRENAL (0.86), and RENAL (0.81) were strong predictions for complications. None of the algorithms demonstrated strong prediction for incomplete tumor ablation. Comparative evaluation suggests that the five proposed risk-scoring systems do not correlate well to either complication risk or incomplete ablation for RCC ablation with all ablative modalities; however, P-RAC, mRENAL, RENAL, and P-RENAL all achieve reasonable AUC values in the context of predicting complications using cryoablation only.