Propensity score matching demonstrates similar results for radiofrequency ablation compared to surgical resection in colorectal liver metastases

Journal: European Journal of Surgical Oncology
Year of publication: 2022
Page: S0748-7983(22)00009-9

T.W. van de Geest, M.J. van Amerongen, P.M.H. Nierop, D.J. Höppener, D.J. Grünhagen, A. Moelker, J.J. Fütterer, C. Verhoef & J.H.W. de Wilt

PURPOSE: Minimally invasive ablative treatments, such as radiofrequency ablation (RFA), are increasingly used in the curative treatment of patients with colorectal liver metastases (CRLM). Selection bias plays an important role in the evaluation of early and late results between RFA and surgery. The purpose of this study was to evaluate recurrences and oncological survival following these two treatment modalities using single pair propensity score matching.

METHODS: Between 2000 and 2018, patients curatively treated for CRLM were included in a multicentre database. Patients were excluded when receiving two-staged treatment, synchronous treatment with primary tumor or combination of modalities. Propensity score matching was used to minimize influence of known covariates, i.e., age, ASA, FONG CRS, location and T-stage of the primary tumor.

RESULTS: Before matching, the RFA group contained 39 patients and the surgery group 982 patients, after matching both groups contained 36 patients. After matching, mean age was 69 years (53-86) for RFA and 68 (50-86) for surgery, with a mean tumor size of respectively 2.5 cm (0.8-6.5) and 3.4 cm (1-7.5). Both groups showed similar complication rate according to Clavien-Dindo (17vs.33%; p = 0.18), recurrence rate (58vs.64%; p = 0.09) without significant differences in 5-year DFS and OS (RFA compared to surgery respectively 25vs.37%; p = 0.09 and 42vs.53%; p = 0.09).

CONCLUSION: After propensity score matching, RFA showed lower complications and similar oncological survival compared to surgical resection. In patients who are suboptimal candidates for surgery, RFA seems to be a good and safe alternative.