Optimal timing of two-stage pulmonary metastasectomy for synchronous bilateral colorectal pulmonary metastases: A time-dependent survival analysis.

BACKGROUND: Synchronous bilateral colorectal pulmonary metastases (CPM) present a surgically complex scenario in which two-stage video-assisted thoracoscopic surgery (VATS) is increasingly practised. The optimal inter-VATS interval remains undefined, and conventional analyses of this question are vulnerable to both selection bias and immortal time bias. We aimed to characterize the continuous, non-linear relationship between the inter-VATS interval and overall survival (OS) using three complementary bias-corrected statistical approaches. METHODS: From a single-center retrospective cohort of 385 CPM patients (January 2009 - December 2023), 152 with synchronous bilateral disease were analyzed. Level 1: Inverse probability of treatment weighting (IPTW) addressed selection bias in the comparison of second VATS versus no second VATS (n = 152). Level 2, restricted to patients who completed both procedures (n = 76; 27 deaths): (a) restricted cubic spline (RCS) Cox regression modeled the continuous interval-outcome dose-response; (b) time-dependent Cox regression in counting-process format eliminated immortal time bias by modeling the second VATS as a time-varying event; (c) landmark analysis at 3 months provided a bias-free dichotomous confirmatory comparison (early ≤3 months vs delayed >3 months). RESULTS: After IPTW adjustment, a second VATS was associated with a non-significant directional survival benefit (HR = 0.67, 95% CI: 0.39-1.12, p = 0.124), consistent with an effect contingent on appropriate patient selection rather than universal benefit. Among patients with two VATS procedures, the RCS model demonstrated a statistically significant non-linear dose-response between the inter-VATS interval and mortality hazard (model likelihood-ratio p = 0.006). The hazard ratio followed a U-shaped trajectory: intervals shorter than 3 months were associated with substantially elevated risk compared to the nadir; risk reached its minimum at 5-8 months and rose again beyond approximately 15 months. Time-dependent Cox regression confirmed the non-linear interval effect and identified extrapulmonary metastasis as the sole independent adverse prognostic factor (HR = 2.95, 95% CI: 1.11-7.81, p = 0.030; model concordance = 0.725). Landmark analysis at 3 months demonstrated that patients who underwent early second VATS (≤3 months, n = 50) had a three-fold higher risk of death compared to those who waited (HR = 3.08, 95% CI: 1.20-7.92, p = 0.020). All three analytical approaches converged on a consistent clinical message. CONCLUSION: A two-stage pulmonary metastasectomy interval of 5-8 months is associated with optimal survival in patients with synchronous bilateral CPM. Sequential surgery within 3 months significantly worsens prognosis and should be avoided unless mandated by clinical urgency. Selective two-stage metastasectomy - guided by multidisciplinary team consensus and timed at 5-8 months - minimizes surgical morbidity while preserving the survival benefit of repeat resection. These findings provide an evidence-based framework for individualized surgical planning and call for prospective validation.