Multicenter and inter-software evaluation of ablative margins after thermal ablation of colorectal liver metastases.

Abstract

To assess the association between minimal ablative margin (MAM) and local tumor progression (LTP) following CT-guided thermal ablation of colorectal liver metastases (CRLM) in a multicenter cohort and across two confirmation software. This multicenter retrospective study included patients who underwent CT-guided radiofrequency or microwave ablation for CRLM between 2009 and 2021 in three institutions. Three-dimensional (3D) MAM was retrospectively assessed using dedicated ablation confirmation software by automatic non-rigid (Ablation-fit) or semi-automatic rigid co-registration (SAFIR) of intraprocedural pre- and post-ablation contrast-enhanced CT scans by two independent reader teams blinded to patient outcomes. LTP was assessed on a per-tumor basis. Factors associated with LTP-free survival were assessed using multivariable Cox regression analysis. Overall, 113 patients (mean age: 67 ± 10 years; 78 men) who underwent thermal ablation for 189 CRLM (mean diameter: 1.9 ± 1.1 cm) met the inclusion criteria. 173/189 (92%) CRLM could be successfully analyzed using both software. Over a median follow-up of 31 months (IQR: 22-47), 21 of 173 CRLM (12.1%) developed LTP. On multivariable analysis, 3D MAM was independently associated with LTP in both software (Ablation-fit: HR 0.47, 95% CI: 0.36-0.61, p < 0.001; SAFIR: HR 0.42, 95% CI: 0.32-0.55, p < 0.001). No LTP was observed in CRLM ablated with MAM ≥ 4 mm (Ablation-fit) and ≥ 5 mm (SAFIR). The per-tumor median absolute difference in MAM quantification between both software was 2 mm (IQR: 1-3). MAM was independently associated with LTP after thermal ablation of CRLM across multicenter data and two confirmation software. Ablations achieving a MAM ≥ 5 mm were associated with local control in both software. MAMs from intraprocedural contrast-enhanced CT were independently associated with LTP after thermal ablation of CRLM across multicenter data and two confirmation software, with a margin ≥ 5 mm associated with local control in both software. Sufficient ablative margins are critical for local control following thermal ablation of CRLM. Intraprocedural CT-derived MAM was the only independent factor associated with LTP across two confirmation software. No LTP was observed in CRLM ablated with a MAM ≥ 5 mm.