INTRODUCTION

Although histology has not consistently been associated with treatment outcome in advanced non-small cell lung cancer, a recent phase III trial comparing pemetrexed plus cisplatin and gemcitabine plus cisplatin (GC) demonstrated better efficacy for pemetrexed plus cisplatin in nonsquamous (adenocarcinoma and large cell carcinoma) carcinoma than in squamous cell carcinoma. Herein, retrospective analysis is used to explore the potential predictive and prognostic role of non-small cell lung cancer histology in patients treated with three first-line, platinum-based regimens.

METHODS

Survival and time to progression (TTP) data from a phase III trial comparing paclitaxel plus carboplatin (PCb), GC, and vinorelbine plus cisplatin (VC) were analyzed. Using Cox multiple regression, factors for one model included treatment (PCb, GC, and VC), histology (squamous, adenocarcinoma, large cell, and other), gender, Eastern Cooperative Oncology Group performance status (0/1 and 2), stage (IIIB and IV), number of metastatic sites (< or = 1 and >1), and smoking history (yes or no). In another model, histology was simply considered as squamous versus nonsquamous. An interaction value of p < 0.10 was considered significant.

RESULTS

Baseline patient and disease characteristics for the 607 treated patients were balanced among the arms. No significant treatment-by-histology interaction was seen in either model for either end point. Nevertheless, histology was a significant prognostic factor for survival in the first model (p = 0.0183) and marginally significant for TTP (p = 0.0783). Subsequent pairwise comparisons of histology groups demonstrated a survival advantage for squamous cell carcinoma over adenocarcinoma (p = 0.0021).

CONCLUSIONS

Histology was not predictive of PCb, GC, or VC treatment effect for either survival or TTP. Histology was prognostic for survival, with better outcomes associated with squamous cell carcinoma.