Incidence, Correlates, and Prognostic Significance of Mixed Response in Advanced Non-small Cell Lung Cancer

Abstract

Abstract Background Mixed response (MR), a scenario featuring discordant tumor changes, has been reported primarily with targeted therapies or immunotherapy. We determined the incidence and prognostic significance of MR in advanced non–small cell lung cancer (NSCLC) treated with cytotoxic chemotherapy. Patients and Methods We analyzed patient-level data from ECOG-ACRIN E5508 (carboplatin-paclitaxel + bevacizumab induction followed by randomization to maintenance therapy regimens). For patients with at least 2 target lesions and available measurements after cycle 2, we characterized response as homogeneous response (HR, similar behavior of all lesions), MR (similar behavior but >30% difference in magnitude of best and least responding lesions), or true mixed response (TMR, best and least responding lesions showing different behavior: ≥10% growth versus ≥10% shrinkage). We compared category characteristics using Mann-Whitney U and Chi-square tests, and overall survival (OS) using log-rank test and Cox models. Results Among 965 evaluable patients, HR occurred in 609 patients (63%), MR in 208 (22%), and TMR in 148 (15%). Median OS was 13.6 months for HR, 12.0 months for MR, and 7.6 months for TMR (P < .001). Compared to HR, TMR had inferior OS among stable disease cases (HR 1.62; 95% CI, 1.23-2.12; P < .001) and a trend toward inferior OS among progressive disease cases (HR 1.39; 95% CI, 0.83-2.33; P = .2). In multivariate analysis, TMR was associated with worse OS (HR 1.48; 95% CI, 1.22-1.79; P < .001). Conclusion True mixed response occurs in a substantial minority of lung cancer cases treated with chemotherapy and independently confers poor prognosis.