The majority of patients with non-small-cell lung cancer (NSCLC) who are of good performance status will receive chemotherapy at some time in the management of their disease. A metaanalysis of 52 randomized trials, published in 1995, supported the use of cisplatin(Drug information on cisplatin)-based chemotherapy in NSCLC, based on a 1.5-month improvement in median survival and a 10% absolute 1-year survival benefit with chemotherapy vs best supportive care (NSCLC Collaborative Group).[1] Chemotherapy was also associated with improved symptom control and quality of life. More recently, platinum-based doublets containing newer agents including gemcitabine(Drug information on gemcitabine) (Gemzar), the taxanes, vinorelbine (Navelbine), irinotecan(Drug information on irinotecan) (Camptosar) and others have demonstrated improved activity in the firstline treatment of NSCLC, with response rates ≥ 40% in several randomized trials.[2-5] With several potential regimens available, it is worth examining key trials that helped establish current treatment approaches in NSCLC. This paper will focus on first-line chemotherapy for NSCLC, using mainly phase III randomized trials to examine issues such as which new-generation regimen to use, whether a platinum agent should be included, and if so, cisplatin or carboplatin(Drug information on carboplatin) (Paraplatin), how many drugs to combine, and the optimal treatment duration. Trials that will be discussed, primarily in terms of response and survival results, are summarized in Table 1. Which Third-Generation Regimen to Use? Vinorelbine/Platinum Trials
Vinorelbine was the first of the third-generation chemotherapy drugs established for use in lung cancer, and has now accrued more than a decade of experience in this disease. Le Chevalier et al conducted a pivotal trial comparing vinorelbine/cisplatin, vindesine(Drug information on vindesine)/cisplatin, and vinorelbine alone in 612 patients with stage IIIA/B and IV NSCLC.[2] Response rate improved significantly with vinorelbine/ cisplatin compared with the other treatments, (30%, 19% [P = .02], 14% [P < .001], respectively). Notably, this was the first trial demonstrating a 40% 1-year survival rate in advanced NSCLC (in the vinorelbine/ cisplatin patients), whereas 1-year survival rates with previous-generation regimens had ranged around 20% to 25%. In the TAX 326 trial, first-line vi-norelbine/cisplatin was compared with docetaxel(Drug information on docetaxel)/cisplatin and docetaxel/carboplatin in more than 1,200 advanced NSCLC patients, with no significant benefit shown for vinorelbine.[3] Depierre et al also compared vinorelbine/ cisplatin with single-agent vinorelbine. Response rate was significantly higher (43% vs 16%, P = .0001) when cisplatin was added to single-agent vinorelbine, although there were no differences in either median survival (~ 8 mo) or 1-year survival rates (~ 35%) between treatment groups in this study.[6] Paclitaxel/Platinum Trials
An important study conducted by the Eastern Cooperative Oncology Group (ECOG 5592) helped establish the role of paclitaxel(Drug information on paclitaxel) in NSCLC.[7] Patients (N = 599) received first-line etoposide(Drug information on etoposide)/cisplatin or paclitaxel/cisplatin, with paclitaxel administered at two doses (135 mg/m² or 250 mg/m²) infused over 24 hours. The two paclitaxel groups combined had superior 1-year survival rates and median survival times (39% and 9.9 months vs 32% and 7.6 months, P = .048) and significantly improved response rates (~26% vs 12%, P < .001) compared with those achieved with etoposide/ cisplatin, respectively. A European Organization for Research and Treatment of Cancer (EORTC) trial with more than 300 advanced NSCLC patients compared teniposide (Vumon)/cisplatin and paclitaxel/ cisplatin using a more standard 3-hour paclitaxel infusion time. A significant response benefit was seen in the paclitaxel arm (41% vs 28%, P = .018), although 1-year survival rates were similar (~42%) in the two groups.[4] In a study of first-line paclitaxel/cisplatin vs cisplatin alone in 414 patients, Gatzemeier et al reported a significant response improvement (26% vs 17%, P = .028) but no survival benefit with the paclitaxel doublet.[8] Standard initial chemotherapy for advanced NSCLC in the United States is paclitaxel plus carboplatin. Belani et al[9] compared paclitaxel and carboplatin with the second-generation regimen of etoposide and cisplatin, and showed only nonsignificant trends in favor of the paclitaxel regimen. The Southwest Oncology Group (SWOG) compared paclitaxel/carboplatin and vinorelbine/cisplatin (SWOG 9509) and found response rates of 25% in the paclitaxel/carboplatin arm and 28% in the vinorelbine/ cisplatin arm, with similar survival rates (median: 8 months).[10] Patients on the vinorelbine arm had significantly more nausea and vomiting, whereas patients on the paclitaxel arm had significantly more neurotoxicity. More patients receiving vinorelbine/cisplatin discontinued treatment because of toxicity, but quality- of-life analyses identified no differences between the two arms. In a multinational study reported by Rosell et al, 618 chemotherapy-naive advanced NSCLC patients received paclitaxel infused over 3 hours combined with carboplatin or cisplatin, with no differences in response rates (~25%) shown.[11] However, with 22 months of follow-up, updated analysis showed median survivals of 8.2 and 9.8 months, respectively (P = .019), suggesting that cisplatin may be a slightly more active agent than carboplatin when combined with paclitaxel. To compare the most important third-generation regimens in advanced NSCLC (stage IIIB, IV, or recurrent disease), Schiller and colleagues at ECOG undertook a large 1,200-patient trial (ECOG 1594) that compared the best arm of their previous trial (paclitaxel at 135 mg/m² infused over 24 hours and cisplatin at 75 mg/m²) to three other third-generation regimens, including gemcitabine/cisplatin, docetaxel/ cisplatin, and paclitaxel at 225 mg/m² infused over 3 hours with carboplatin (Paraplatin). The results of the paclitaxel/carboplatin arm were somewhat disappointing, with a response rate of only 17% and median and 1-year survival rates of 8.5 months and 34%, respectively.[12] However, these results were similar to those seen in the other three arms, and none of the three study arms was significantly better than the reference arm of paclitaxel and cisplatin. Gemcitabine/Platinum Studies
Gemcitabine has been studied combined with platinum as first-line therapy for NSCLC in several phase III randomized studies. Sandler et al from the Hoosier Oncology Group (HOG) conducted a randomized phase III study (n = 522) comparing cisplatin at 100 mg/m² on day 1 plus gemcitabine at 1,000 mg/m² on days 1, 8, and 15 every 4 weeks to cisplatin at 100 mg/m² on day 1, every 4 weeks.[13] Patients receiving cisplatin/gemcitabine had significantly improved response (30.4% vs 11.1%; P < .0001), median time to progression (5.6 vs 3.7 months; P = .0013), and overall survival (9.1 vs 7.6 months; P = .004). Toxicity was mostly hematologic (more pronounced in the combination arm) with grade 4 neutropenia in 35.3% of patients vs 1.2%; neutropenic fever was < 5% in both arms. Grade 4 thrombocytopenia occurred in 25.4% of patients on the combination arm vs 0.8% of those on the monotherapy arm. In a study conducted by Cardenal et al comparing gemcitabine/cisplatin and the second- generation regimen etoposide/cisplatin in 133 advanced NSCLC patients, the gemcitabine regimen produced a significantly higher overall response rate (41% vs 22%, P = .02) and a significant delay in time to disease progression (6.9 vs 4.3 months, P = .01).[14] Survival was similar in the two arms. The ECOG 1594 study mentioned previously, which compared four third-generation regimens, included gemcitabine/cisplatin as one treatment arm. Response and 1-year survival rates with the gemcitabine combination (22%, 36%) were similar to those achieved with paclitaxel/cisplatin (21%, 31%), and higher than those of docetaxel/cisplatin (17%, 31%) and paclitaxel/carboplatin (17%, 34%) although these differences were not significant.[12] The only significant study finding was that of time to disease progression, favoring the gemcitabine/ cisplatin doublet (4.2 vs 3.4, 3.7, and 3.1 months, respectively, P < .05). The gemcitabine/cisplatin doublet and paclitaxel/cisplatin doublets comprised two out the three arms of an EORTC trial (EORTC 08975) [15] vs MIC (mitomycin [Mutamycin], ifosfamide [Ifex], cisplatin). No significant differences were seen for either response or survival rates. Notwithstanding the excellent response rates achieved when gemcitabine is combined with cisplatin, many physicians prefer to use carboplatin based on tolerability issues. When administered with carboplatin, gemcitabine is generally given on a day 1, day 8 schedule and the carboplatin dose must be reduced to an area under the concentration curve (AUC) of 5 to lessen risk of severe thrombocytopenia. Gemcitabine/carboplatin has only been compared in randomized trials to second-generation regimens. In one study of gemcitabine/carboplatin vs MIC or MVP (mitomycin, vinblastine(Drug information on vinblastine), cisplatin), no significant differences in response and survival outcomes were noted.[16] However, a significant survival advantage favoring gemcitabine/carboplatin over MIP (mitomycin, ifosfamide, cisplatin) was demonstrated in another randomized trial (n = 422) for median survival, 10.0 vs 6.5 mo; 1-year survival rates, 38% vs 28%; P = .0043.[17] Gemcitabine/carboplatin also provided significantly better response (27% vs 15%, P < .05) and mean (11.5 vs 7.9 mo, P = .0001) and 1-year (36% vs 12%, P < .05) survival results than the secondgeneration vinblastine/cisplatin regimen.[ 18] These results thus demonstrate the utility of carboplatin as an alternate to cisplatin in gemcitabine- containing regimens. Docetaxel/Platinum Regimens
The four-arm ECOG 1594 study included the docetaxel/cisplatin regimen, wherein the response rate (17%) and median survival (7.4 months) were lower than those achieved with the paclitaxel/cisplatin or gemcitabine/ cisplatin doublets (P = not significant).[ 12] On the other hand, in the large TAX 326 study, docetaxel/cisplatin had improved response rates (32%, 24%, 25%, P = .029) relative to docetaxel/carboplatin and vinorelbine/cisplatin, respectively.[3] Furthermore, while 2-year survival rates are often not reported, docetaxel/cisplatin treatment resulted in a 7% absolute 2-year survival improvement over that achieved with vinorelbine/ cisplatin (21% vs. 14%). Quality-oflife assessments also favored the docetaxel arms in this trial. In a smaller study from Japan involving more than 300 chemotherapy- naive stage IV NSCLC patients, docetaxel/cisplatin significantly improved response rate (37% vs 21%, P < .01) and median survival (11.3 vs 9.6 mo, P = .014) over that achieved with vindesine/cisplatin, and there was a trend toward improved 1-year survival (48% vs 41%).[19] Irinotecan/Platinum Regimens
Randomized trials of first-line irinotecan in advanced NSCLC have also been carried out. Negoro et al compared irinotecan/cisplatin, vindesine/ cisplatin, and irinotecan alone in approximately 400 patients.[20] Significant improvements in response rates (44%, 32%, 21%, respectively, P<.001) were observed, while median (12.5, 11.5, 11.5 months) and 1-year survival (47%, 38%, 42%) rates were somewhat better with irinotecan/cisplatin vs the other treatments.[20] Survival was significantly better for the subgroup of patients who had stage IV tumors (median survival 50 vs 36.4 weeks, P = .004, for irinotecan/cisplatin vs vindesine/cisplatin; and 42.1 vs 36.4 weeks for irinotecan alone vs vindesine/cisplatin, P = .018). In contrast, preliminary data from a randomized trial comparing irinotecan/ cisplatin with vindesine/cisplatin, paclitaxel/ carboplatin, and gemcitabine/ cisplatin showed similar response rates (~30%) for the four treatment groups, with no survival data yet available.[21] Pemetrexed
Pemetrexed(Drug information on pemetrexed) (Alimta) is a multitargeted antifolate with response rates of ≥ 20% as a single agent and approximately 40% when combined with cisplatin, even in patients with advanced NSCLC.[22,23] This agent has demonstrated its activity in second-line NSCLC, and randomized trials in the first-line setting are awaited to determine further the role of pemetrexed. Possible combination partners under investigation include gemcitabine, vinorelbine, carboplatin, and oxaliplatin(Drug information on oxaliplatin) (Eloxatin). Is Platinum Necessary? Data from several randomized trials of first-line treatment in advanced NSCLC can be used to address whether a platinum agent is necessary. For example, EORTC 08975 included combinations of cisplatin with gemcitabine or paclitaxel and the nonplatinum paclitaxel/gemcitabine doublet.[ 15] No significant differences were noted, although the lowest response rate of 27% and median survival time of 6.9 months occurred in the nonplatinum arm-results lower than one would expect with a thirdgeneration platinum-based regimen. A trial by Gridelli et al undertaken by the National Cancer Institute (NCI) of Italy Naples group and the NCI Canada also found inferior results for the non-platinum regimen (vinorelbine/ gemcitabine) compared with gemcitabine or vinorelbine combined with cisplatin.[24] Time to disease progression was significantly poorer in the nonplatinum arm (P = .004), while response and 1-year survival differences were not statistically significant, although trended in favor of cisplatin use. Toxicity was greater in this study in the cisplatin arm, but patients treated with the cisplatinbased regimens had more improvement in their symptoms. It is also of interest that, despite greater toxicity in the platinum arm, quality of life was similar in both arms. Two other studies involving approximately 900 chemotherapy-naive advanced NSCLC patients demonstrated virtual equivalence of nonplatinum and platinum regimens, in comparisons of docetaxel/cisplatin vs docetaxel/gemcitabine[25], and paclitaxel/ carboplatin vs paclitaxel/gemcitabine.[ 26] Response rates ranged from 28% to 35% with the four regimens, with ~10-month median survival times and ~40% 1-year survival rates. While these study results are somewhat mixed regarding the need for platinum therapy, a recent meta-analysis by D'Addario et al provides support for platinum-containing chemotherapy.[ 27] This meta-analysis compared effects of platinum and nonplatinum regimens administered to more than 7,500 patients with NSCLC participating in 37 trials. Response rates significantly improved with platinum- based chemotherapy (P < .0001) when all trials were considered, and also when only the 1,563 patients treated with third-generation regimens were assessed (P = .0032). Survival also favored cisplatin-based treatment, although the difference was not statistically significant for the third-generation regimens.