Toward perioperative, numerically assisted irreversible electroporation for hepatocellular carcinoma: clinical outcomes informed by numerical simulations.

[OBJECTIVES] Numerical evaluation of irreversible electroporation (IRE) holds significant potential to assist practitioners in planning and guiding these complex procedures. We aimed to establish a correlation between the results of retrospective numerical simulations of IRE and clinical outcomes in the treatment of hepatocellular carcinoma (HCC).

[MATERIALS AND METHODS] IRE procedures conducted at our center for HCCs < 5 cm were reconstructed. Per-procedural imaging was used to generate a 3D geometric model of the tumor and its surroundings. The spatial coordinates of electrodes and treatment parameters were used to retrospectively simulate the EF using the static linear model. The proportion of tumor volume encompassed by 3D mappings of different EF magnitudes was correlated with clinical efficacy using a logistic regression algorithm.

[RESULTS] 31 IRE procedures were included. HCCs had a median diameter of 20 mm (range: 7-45 mm). At 1 month, 29/31 tumors (93.5%) showed complete ablation, but 10 of these recurred later (median delay: 7 months). The percentages of tumor coverage by EF isolines were significantly lower in the 12 cases of local IRE failure compared to the 19 cases of IRE success: 97.7% vs. 100% at 300 V/cm (p < 0.001); 89.7% vs. 100% at 400 V/cm (p < 0.001) and 82.8% vs. 95.4% at 500 V/cm (p < 0.01) with 400 V/cm being the most discriminative threshold dose (ROC-AUC: 0.904). Tumor coverage < 95% by the 400 V/cm isoline was associated with IRE failure (2-year incidence: 89% vs. 21.7%, p < 0.0001).

[CONCLUSION] Numerical simulations of IRE correlate with local treatment outcomes. This study suggests that, in HCC, inadequate tumor coverage by the 400 V/cm isoline predicts and localizes recurrence.

[KEY POINTS] Question Can the clinical outcomes of irreversible electroporation for hepatocellular carcinoma be predicted using numerical simulation of the electric field (EF) distribution? Findings Inadequate tumor coverage by simulated EF maps predicts and localizes recurrence, with 400 V/cm being the most discriminative isodose threshold (ROC-AUC > 0.9). Clinical relevance Numerical simulations of the electric field correlate with local treatment outcomes of irreversible electroporation for hepatocellular carcinoma within a computational framework compatible with clinical use. Real-time identification of insufficient tumor coverage can enable operators to adapt and optimize treatment delivery.