Cytotoxic Chemotherapy in Advanced Non-Small Cell Lung Cancer: A Review of Standard Treatment Paradigms

In 1995, the Non-Small Cell Lung Cancer (NSCLC) Collaborative Group published a meta-analysis examining the impact of chemotherapy in all stages of NSCLC (1). In advanced, metastatic disease, platinum-based chemotherapy was shown to provide a survival benefit compared with best supportive care. This observation was based on nine trials, which included 1190 patients (Table 1). Since then, several randomized trials using either platinum-based combination regimens (2, 3) or selected new single agents (4, 5, 6) have confirmed this observation (Table 1). Although the survival advantage is modest, it is both clinically and statistically significant (7). In addition, those studies using a quality-of-life measure (6, 8) have all suggested improved quality-of-life for those patients receiving chemotherapy versus best supportive care.

Over the past decade or so, several new agents with significant single-agent activity in advanced NSCLC have been developed (9). These agents include paclitaxel, docetaxel, vinorelbine, gemcitabine, and irinotecan. Two recent meta-analyses (10, 11) have suggested that regimens including these newer agents improve outcomes of efficacy compared with older regimes, although, again, the benefits are modest. Several randomized trials (12, 13, 14, 15, 16, 17, 18) have evaluated these modern platinum-based doublets and suggested that no one combination is superior when using survival as the primary measure of outcome (Table 2). Taken as a whole, these Phase III trials support the contention that there is not one standard regimen but several standard options for use in the treatment of advanced NSCLC.

The optimal number of agents to use in this setting has been the focus of several recent randomized trials (3, 17, 19) as well as two meta-analyses (20, 21). Table 3 shows the results of the two recently reported meta-analyses performed independently of each other. Delbaldo et al. (20) evaluated all of the trials performed since 1980, whereas Baggstrom et al. (21) evaluated only trials using the new agents. Both meta-analyses are consistent in their conclusions that two drugs yield superior response and survival outcomes compared with one drug, whereas three-drug regimens may improve response rates but do not improve survival outcomes in advanced NSCLC. These results support the use of two-drug combinations in advanced NSCLC.

Platinum-based combinations were the first regimens to convincingly have an impact on survival and have been the standard of care in NSCLC (22). Given the single-agent activity and favorable toxicity profiles of the new agents, several nonplatinum-based doublets have been developed. The rationale for this development was the possibility that survival could be enhanced with the more active agents in combination without an increase in the risk of severe toxicity (or perhaps even less toxicity if the platinums were deleted). Several randomized Phase III trials have been reported comparing platinum-based to nonplatinum-based doublets (14, 23, 24, 25). No survival benefits have been observed for the nonplatinum-based doublets. In fact, concern for decreased survival with the nonplatinum-based doublets has been raised in two of these trials (14, 25). Also, toxicity profiles for the nonplatinum-based regimens have not been substantially different from the platinum-based regimens. The experience reported thus far does not provide a compelling reason for the routine use of nonplatinum-based doublets.

If we accept the fact that platinum-based doublets remain the standard of care, is there a “preferred” platinum? The two choices include cisplatin or carboplatin, although others (i.e., oxaliplatin) also appear active (26). Many of the trials shown in Table 2 have included both cisplatin and carboplatin-based regimens. As can be seen, no significant survival differences are evident in the majority of the trials. Two recent trials (18, 27) have suggested superiority of cisplatin doublets over carboplatin doublets, but this has not been a consistent finding in the palliative setting.

Because the role of chemotherapy in advanced NSCLC is to prolong survival and palliate symptoms, the optimal duration of therapy in the first-line setting is particularly pertinent as chemotherapy has toxicity (acute and cumulative), costs (financial and personal), and a degree of intrusiveness associated with it. Four recent trials have addressed this issue either directly (28, 29, 30) or indirectly (Ref. 31; Table 4). All of the trials compared either 3 or 4 cycles to longer durations of therapy defined as either 6 cycles (28, 30), continuous therapy (29), or “maintenance” vinorelbine (31). None of the trials suggested any benefit for prolonging therapy beyond 3 or 4 cycles, and all showed increased rates of cumulative toxicities.

The elderly (typically defined as either ≥65 or 70 years of age) have traditionally been considered a “special interest” group, although age-related analysis of several recent platinum-based trials have suggested equivalent response and survival rates for the elderly compared with the younger patients (32, 33, 34). In general, the elderly have been underrepresented in clinical trials. Currently, 30–40% of patients diagnosed with lung cancer are ≥70 years of age, whereas recent clinical trials have included only 15–29% of patients 70 years and older (35). A reasonable conclusion from the data gathered thus far is that the “fit” elderly (defined as “robust” enough for a platinum-based clinical trial) have efficacy outcomes similar to younger patients. However, there is significant heterogeneity among the elderly, and many are “unfit” due to comorbidities or the aging process itself (36). Both the Elderly Lung Cancer-Vinovelbine Italian Study and Multicenter Italian Lung Cancer Elderly Study trials support a monotherapy approach using nonplatinum agents (6, 37). Patient selection issues make it difficult to reconcile these observations with the previously noted ones suggesting elderly patients enjoy similar benefits from platinum-based combinations (17, 32, 33, 34), compared with their younger counterparts. Additional investigation in this area is needed and should include evaluating the influence of comorbidities on treatment selection, tolerance, and outcomes, as well as the benefits of therapy in the extremely aged (>80 years of age).

Performance status (PS) 2 patients also pose a challenge, because it is very clear that their survival is significantly shorter than PS 0–1 patients (38). A median survival time of ∼4 months for PS 2 patients has been demonstrated in two recent Phase III trials (38, 39). It is unclear how much survival is impacted in this patient population by chemotherapy. There has not been a randomized trial specific to this population comparing some form of treatment to best supportive care. The recent trial reported by Lilenbaum et al. (17) suggested a survival benefit to carboplatin plus paclitaxel versus paclitaxel alone in a planned prospective subset analysis of CALGB 9730. The median survival time and 1-year survival rate for the PS 2 patients treated with paclitaxel alone were 2.4 months and 10% compared with 4.7 months and 18% for the PS 2 patients treated with the combination of carboplatin and paclitaxel (log rank P < 0.0177) Other trials have documented a significant palliative effect of treatment for PS 2 patients (2). The optimal management strategy for PS 2 patients remains controversial.

Because metastatic disease is not curable with current treatment regimens, all of the patients will eventually have disease progression unless they succumb to a disease-related or comorbid complication. Because many patients retain a good PS at the time of progression, second-line treatment is an option and supported by two Phase III trials comparing docetaxel to either best supportive care (40) or a choice of vinorelbine or ifosfamide (41). In these trials, improved survival and disease palliation was observed for docetaxel 75 mg/m² every 3 weeks. On the basis of these observations, docetaxel was approved by the Food and Drug Administration and remains the only approved option in this setting. Weekly administration compared with the every-3-week schedules of docetaxel has been evaluated in two recent randomized trials (42, 43). Although the data are still preliminary, response and survival outcomes appear similar with toxicity profiles slightly in favor of weekly administration. Other agents are also active (44), and a recent Phase III trial comparing docetaxel to pemetrexed (45) revealed similar efficacy outcomes but less hematological toxicity with pemetrexed (Table 5). Gefitinib and erlotinib, both novel oral tyrosine kinase inhibitors of the epidermal growth factor receptor, also have activity in refractory NSCLC (46, 47, 48). Gefitinib was approved recently for use in the third-line setting based on convincing evidence that objective responses and symptom palliation could be realized. The availability of these options for patients with refractory NSCLC raises several issues including the appropriate surveillance strategies after first-line therapy, the optimal timing, choice, and sequence of agents, as well as identification of prognostic and predictive markers in this setting.

In summary, well-conducted clinical trials have provided several treatment options for patients with advanced NSCLC in the first-, second-, and third-line setting. Unfortunately, the disease in the advanced metastatic setting remains incurable, and our current armamentarium of cytotoxic agents has achieved a survival plateau unlikely to be raised in future trials using current therapeutics. Future research efforts should focus on identifying prognostic and predictive markers of benefit not only for standard cytotoxic agents but also for the new “targeted” agents currently under evaluation in this disease. Continued efforts directed at understanding the biology of lung cancer will hopefully provide novel targets for new therapeutics. The ability to better understand which genetic marker or critical pathways are vital to the cancer of a particular patient will hopefully allow an individualized therapeutic approach for patients in the future.

Dr. Thomas Lynch: Your trial showed that 4 cycles of therapy appear to be optimal. In considering a platform to which to add novel agents, do you think you can make a statement that we should limit it to 4 cycles?

Dr. Socinski: Not only our trial, but three other trials that have addressed this issue either directly or indirectly suggest little benefit in going beyond 3 or 4 cycles. The median number of cycles delivered is typically about 4; the data suggest that you don’t get any further benefit, that what you buy is cumulative toxicity, which can be an issue when you add a third agent.

Dr. David Gandara: From a statistical standpoint, we’ve analyzed our last three trials in the SWOG database and with the paclitaxel/carboplatin combination, we know the percentage of patients with nonprogression after 4 cycles at 68%.

Dr. Lynch: Do people feel that we should be looking at nonplatinum based doublets as the basis for integration with novel agents? Are there advantages to doing nonplatinum-based doublets as opposed to platinum based doublets?

Dr. Socinski: My argument would be only if there are compelling preclinical synergistic data or something along those lines.

Dr. Giorgio Scagliotti: I don’t see any advantage to nonplatinum-based doublets. Even with toxicity, the data available are not consistent enough to replace the cisplatin in doublets or combination chemotherapy.

Dr. Lynch: What should be considered the standard of care for comparing second-line agents as we go forward? Will it be weekly pemetrexed? Will it be docetaxel every 3 weeks?

Dr. Socinski: My bias is that it will be pemetrexed. The trial was very large and showed equivalent efficacy to the standard with substantially two advantages. One was toxicity, and the other was ease of administration. I was not overly enthralled with the data that was presented with weekly docetaxel, and certainly making these patients come once a week versus once every 3 weeks with pemetrexed doesn’t register well. The other, more pertinent issue is the EGFR drugs and how we integrate them in the third-line setting. There are going to be confirmatory studies in this population of patients comparing to our perceived standard, so we need to think that docetaxel q. 3 weeks may not be what everyone is doing a year from now.

Dr. Lynch: What would impress you in terms of outcome from a metastatic disease trial and would make you think this is something we need to take to Phase III? What median survival target?

Dr. Socinski: We were burned by the Isis data with LY900003, and the median survival looked so good in that trial–16 months and 54%. But we didn’t account for the fact that they were the best patients, those who could travel to Stanford and get on Phase I trial wearing a pump for 2 weeks–we missed that on the radar screen. We all know that you can make a mediocre drug look good if you have good patients, and you can make probably a good drug look bad if you pick bad patients. The second-line literature is full of those experiences. If you could control for patient selection, I’d like to see at least a 12-month survival in a very robust Phase II trial as an intermediary point to go forward in Phase III.

Dr. Alan Sandler: I think that the concept of the randomized Phase II trial in that setting is helpful, even though the statisticians say it is not supposed to be used for that analysis and you are not supposed to make comparisons. If you do a Phase II study and get a 12-month median survival and you have a control arm, you can at least say something about the potential.