Progression free survival was defined as the duration from the date of randomization to the first date of documented disease progression or death from any cause. Tumor assessments were performed every three 21-day cycles during induction and crossover. Progression free survival was censored at the date of the last follow-up visit for participants who were still alive and who had not progressed. Results are presented as a comparison between the two study treatment sequences (induction therapy followed by elective consolidation or crossover therapy) rather than the two induction therapies.

Intent to treat (ITT) population, which includes all randomized participants. 3 participants in the Gemcitabine/Carboplatin treatment sequence were excluded due to missing data, and 1 participant in the Paclitaxel/Carboplatin treatment sequence was excluded for missing data. G/C Arm: 292 events, 122 censored. P/C Arm: 279 events, 134 censored.

Response rate (RR) = proportion of participants with best overall Complete Response (CR: disappearance of all target lesions [TL]) or Partial Response (PR: 30% decrease in sum of longest diameter of TL). Induction therapy RR = number of participants with CR or PR during induction divided by number of participants with measurable disease at baseline (TL measurement during screening). Crossover therapy RR = number of participants with CR or PR during crossover divided by number of participants with measurable disease at baseline (latest TL measurement by first dose date of crossover therapy).

All ITT participants with measurable disease at baseline (screening) for induction period; all ITT participants who crossed over to single agent therapy and had measurable disease before or at crossover for crossover period. Participants in consolidation therapy achieved CR during induction therapy and were not included in analysis.

Time to treatment failure was defined as the duration from date of randomization to the date of the first of the following events: early discontinuation of study therapy; progression of disease, or death due to any cause. Time to treatment failure will be censored at the date of the last follow-up visit for participants who did not discontinue early, who are still alive, and who have not progressed. Results are presented as a comparison between the two study treatment sequences (induction therapy followed by elective consolidation or crossover therapy) rather than the two induction therapies.

Intent to treat (ITT) population, which includes all randomized participants. 3 participants in the Gemcitabine/Carboplatin treatment sequence were excluded due to missing data, and 1 participant in the Paclitaxel/Carboplatin treatment sequence was excluded for missing data. G/C Arm: 336 events, 78 censored. P/C Arm: 330 events, 83 censored.

Overall survival is defined as the duration from baseline to death. For participants who are still alive at the data cut-off date, survival will be censored at the last contact date. Results are presented as a comparison between the two study treatment sequences (induction therapy followed by elective consolidation or crossover therapy) rather than the two induction therapies.

Intent to treat (ITT) population, which includes all randomized participants. 3 participants in the Gemcitabine/Carboplatin treatment sequence were excluded due to missing data, and 1 participant in the Paclitaxel/Carboplatin treatment sequence was excluded for missing data. G/C Arm: 194 events, 220 censored. P/C Arm: 159 events, 254 censored.