Predictors of Loco-Regional Recurrence After Pulmonary Non-Anatomical Metastasectomy: A Multicentre Retrospective Analysis

OBJECTIVES: This study aimed at identifying predictors of loco-regional recurrence in patients who underwent non-anatomical (wedge) resection(s) of lung metastases at 15 European centres. METHODS: Multicentre retrospective analysis of patients ≥18 years who underwent curative-intent wedge resection(s) of lung metastases (January 2010-December 2018). Exclusion criteria were: previous metastasectomy, non-curative intent, incomplete (R1/R2) resection, and lack of data concerning recurrence. Loco-regional recurrence was defined as any recurrence occurring in the lungs, hilar-mediastinal lymph nodes, and/or pleurae. Subset analyses were conducted on patients with a solitary metastasis. RESULTS: A total of 588 patients were included (56.1% adenocarcinoma). Five-year overall survival was 63.9% (95% CI = 60.43; 67.18). Five-year loco-regional recurrence rate was 47.7% (95% CI = 42.8; 52.4). Mean resection margin width was 6.8 mm (IQR = 2.0-10.0). 422 patients (71.8%) underwent a single wedge resection. In the total sample, univariable Cox regression showed that primary tumour site (P = .0003), primary tumour histology (P = .0027), resection margin width (P = .0060), log(margin-to-tumour size ratio) (P = .0022), and number of metastases (P < .0001) were significantly associated with loco-regional recurrence. In patients with a solitary metastasis, univariable analyses showed that primary tumour site (P = .0150), primary tumour histology (P = .0248), and log(margin-to-tumour size ratio) (P = .0355) were significantly associated with loco-regional recurrence. In multivariable analyses (solitary metastasis group), primary squamous-cell carcinoma histology was significantly associated with loco-regional recurrence (P = .0023). Resection margin width and log(margin-to-tumour size ratio) did not significantly affect loco-regional recurrence. CONCLUSIONS: Loco-regional recurrence after pulmonary metastasectomy seems to be correlated with primary tumour histology and the number of metastases. Further studies are needed to clarify the role of size-adjusted margin parameters in preventing recurrence.