OBJECTIVETo analyze the prognostic impact of preoperative (18)F-fluorodeoxyglucose (FDG) PET-CT on postoperative recurrence in patients with completely resected stage I non-small cell lung cancer (NSCLC).

METHODSThe clinic data of 182 patients with stage I NSCLC who underwent (18)F-FDG PET-CT scan before surgical resection between June 2005 and June 2012 were reviewed retrospectively. There were 121 male and 61 female patients, with an average age of 68 years (range from 34 to 85 years). The pathological stage was I A in 98 patients, I B in 84 patients; the histology were adenocarcinoma in 137 patients, squamous cell carcinoma in 35 patients, and others in 10 patients. Clinicopathological factors including gender, age, smoking history, SUV(max), surgical procedure, pathological features and adjuvant chemotherapy were evaluated to identify the independent factors predicting postoperative recurrences by univariate and multivariate analysis. The survivals were calculated by the Kaplan-Meier method and differences in variables were analyzed by the Log-rank test.

RESULTSThe postoperative recurrence rate was 15.9%. The univariate analysis identified that the SUV(max) (t=3.278, P<0.001), p-stage (χ² =5.204, P=0.026), blood vessel invasion (χ² =5.333, P=0.027) and visceral pleural invasion (χ² =7.697, P=0.009) are factors for predicting postoperative recurrence. Only SUV(max) was found to be a significant independent factor according to multivariate analysis (HR=1.068, 95%CI: 1.015 to 1.123, P=0.001). The study population was stratified into three groups by SUV(max), patients with SUV(max) > 5.0 had significantly higher risk of recurrence (23.9%) than those with 2.5 < SUV(max) ≤ 5.0 (15.0%) and SUV(max) ≤ 2.5 (7.3%) (P=0.043); patients with SUV(max) ≤ 2.5 had significantly better 5-year recurrence-free survival rate (90.9%) than those with 2.5 < SUV(max) ≤ 5.0 (82.7%) and SUV(max) ≤ 2.5 (71.0%) (P=0.030). There was a trend toward higher probability of blood vessel invasion (χ² =20.267, P < 0.001), visceral pleural invasion (χ² =6.185, P=0.045) and pathological stage I B (χ² =13.589, P=0.001) with increased SUV(max).

CONCLUSIONSPreoperative SUV(max) of primary tumor is a predictor of postoperative relapse for stage I NSCLC after surgical resection. Therefore, it can contribute to the risk stratification for patients with the same pathological stage and selecting the optimal postoperative follow-up and therapeutic strategy.