Predictors of post-recurrence survival in patients with non-small-cell lung cancer initially completely resected.

OBJECTIVES: Despite recent progress in diagnostic technology and therapeutic approaches to non-small-cell lung cancer (NSCLC), 30–75% of patients develop tumour recurrence after resection. However, the details of post-recurrence survival (PRS) are not well understood. We aimed to investigate the predictors of PRS in patients with NSCLC initially completely resected. METHODS: A series of 568 NSCLC patients who had undergone complete resection between 2000 and 2009 were evaluated retrospectively. Patients who had developed recurrent NSCLC after complete resection were subjected to the current analysis. We examined PRS using the Kaplan–Meier method and multivariate Cox regression analyses. RESULTS: Of the 568 patients, 138 (24.3%) were identified as having disease recurrence. The 2-year and 5-year PRS rates were 44.6 and 25.9%, respectively, while the median PRS time was 22.5 months. Non-adenocarcinoma histology [hazard ratio (HR) = 2.825, 95% confidence interval (CI): 1.825–4.367, P< 0.001], serum carcinoembryonic antigen (CEA) at recurrence ≥5.0 mg/dl (HR = 2.205, 95% CI: 1.453– 3.344, P< 0.001) and no systemic chemotherapy (HR = 2.137, 95% CI: 1.304–3.247, P= 0.002) were independent prognostic factors for PRS. CONCLUSIONS: The current results showed that non-adenocarcinoma histology, elevated serum CEA at recurrence and no systemic chemotherapy were independent unfavourable post-recurrence prognostic factors. The current data can be informative for patient followup after complete resection and further clinical investigation may give us more information about PRS and accurate treatment strategy for recurrent NSCLC after initial complete resection.