Paclitaxel added to the cisplatin/etoposide regimen in extensive-stage small cell lung cancer -- the use of complete response rate as the primary endpoint in phase II trials.

Obtaining a complete response (CR) is the most powerful predictor of survival in extensive-stage small cell lung cancer (SCLC). Improvements in long-term survival in extensive-stage SCLC can be made if the proportion of complete responders to induction therapy can be increased. We performed a phase II trial of the feasibility of adding paclitaxel to standard cisplatin/etoposide (EP regimen) in extensive-stage SCLC. The primary endpoint for this trial is the proportion of patients (pts) obtaining a CR rather than overall response. The null hypothesis for this trial consists of the absence of a CR rate >20%. Paclitaxel was given at doses of 135 (3 pts) or 170 mg/m(2) i.v. over 3 h on day 1. Cisplatin 60 mg/m(2) was given on day 1. On days 1-3 etoposide 80 mg/m(2) per day i.v. was given. G-CSF was used from days 5 to 14 of each cycle. Cycles were repeated q21 days. A two-stage design was used for patient accrual, based on the occurrence of complete responses. Initially, 16 patients were to be accrued. If more than three complete responses were to occur, a further 20 patients would be accrued to the study (Simon's optimal two stage design). Sixteen patients were enrolled. Two patients had a CR (13%) and nine patients had a partial response (56%) for an overall response rate of 69%. The trial was suspended due to the low CR rate. Review of the literature for paclitaxel based front-line treatment combined with EP therapy, in extensive stage SCLC, consistently shows a CR rate <20% but high overall response rate is maintained (thus most responses are partial). As virtually all long-term survivors in extensive-disease SCLC have had a CR to induction therapy and CR remains the strongest predictor of survival for this disease, this may suggest that paclitaxel added to standard EP may improve progression-free survival (and possibly median survival) but is unlikely to significantly improve long-term survival. Initial randomized phase III data confirm the absence of impact on survival for this triple-drug regimen compared to EP therapy alone. Furthermore, other regimens comparing favorably to the EP regimen have all shown consistent CR rates >20% in the phase II setting. In conclusion, consideration should be given to the use of CR rate as a phase II endpoint to determine if a particular regimen should be compared to the standard in a phase III setting for extensive-stage SCLC. A two-stage phase II design based on a minimum required completed responses for further patient accrual is recommended.