RESUMEN
PURPOSE: A confirmatory analysis was performed to determine the prognostic value of metabolic response during induction chemotherapy followed by bimodality/trimodality treatment of patients with operable locally advanced non-small-cell lung cancer. PATIENTS AND METHODS: Patients with potentially operable stage IIIA(N2) or selected stage IIIB non-small-cell lung cancer received three cycles of cisplatin/paclitaxel (induction chemotherapy) followed by neoadjuvant radiochemotherapy (RCT) to 45 Gy (1.5 Gy twice per day concurrent cisplatin/vinorelbine) within the ESPATUE (Phase III Study of Surgery Versus Definitive Concurrent Chemoradiotherapy Boost in Patients With Resectable Stage IIIA[N2] and Selected IIIB Non-Small-Cell Lung Cancer After Induction Chemotherapy and Concurrent Chemoradiotherapy) trial. Positron emission tomography scans were recommended before (t0) and after (t2) induction chemotherapy. Patients who were eligible for surgery after neoadjuvant RCT were randomly assigned to definitive RCT or surgery. The prognostic value of percentage of maximum standardized uptake value (%SUVmax) remaining in the primary tumor after induction chemotherapy-%SUVremaining = SUVmax(t2)/SUVmax(t0)-was assessed by proportional hazard analysis and receiver operating characteristic analysis. RESULTS: Overall, 161 patients were randomly assigned (155 from the Essen and Tübingen centers), and 124 of these received positron emission tomography scans at t0 and t2. %SUVremaining as a continuous variable was prognostic for the three end points of overall survival, progression-free survival, and freedom from extracerebral progression in univariable and multivariable analysis (P < .016). The respective hazard ratios per 50% increase in %SUVremaining from multivariable analysis were 2.3 (95% CI, 1.5 to 3.4; P < .001), 1.8 (95% CI, 1.3 to 2.5; P < .001), and 1.8 (95% CI, 1.2 to 2.7; P = .006) for the three end points. %SUVremaining dichotomized at a cut point of maximum sum of sensitivity and specificity from receiver operating characteristic analysis at 36 months was also prognostic. Exploratory analysis revealed that %SUVremaining was likewise prognostic for overall survival in both treatment arms and was more closely associated with extracerebral distant metastases (P = .016) than with isolated locoregional relapses (P = .97). CONCLUSION: %SUVremaining is a predictor for survival and other end points after multimodality treatment and can serve as a parameter for treatment stratification after induction chemotherapy or for evaluation of adjuvant new systemic treatment options for high-risk patients.