Can Current Treatments for Advanced Non–Small-Cell Lung Cancer Be Improved?

Lung cancer is a worldwide epidemic primarily caused by tobacco smoking.¹ Approximately 1 million new cases of lung cancer are diagnosed each year worldwide, resulting in more than 900 000 deaths.² Of these, approximately 175 000 new cases and 160 000 deaths occur annually in the United States.³ Unfortunately, late diagnosis of lung cancer is the rule, and no curative therapies exist for metastatic disease. Nevertheless, perceptible progress in the treatment of this disease has been made as the result of painstaking efforts and decades of clinical investigations.

The role of chemotherapy is now well established in essentially all stages of non–small-cell lung cancer (NSCLC). In advanced NSCLC, the superiority of chemotherapy over best supportive care alone has been repeatedly demonstrated in patients with good performance status.^{4 - 6} If untreated, patients with metastatic NSCLC have a median survival of 4 to 5 months, with a 1-year survival of about 10%.^{4 - 6} The use of chemotherapy yields almost a doubling of the median survival, which is accompanied by an improvement in quality of life.^{4 - 6} Although the improvements in outcome are clinically and statistically significant, the absolute survival benefit with treatment of advanced NSCLC has been small. To improve on these results, over the past several decades investigators have explored various combinations of different chemotherapy drugs.

In this issue of the JAMA, Delbaldo and colleagues⁷ report a meta-analysis of 57 randomized trials that evaluated the addition of an additional cytotoxic drug to a single agent or a double-agent regimen in advanced, incurable NSCLC. These studies spanned 22 years of clinical research, from 1980 to 2001, and involved more than 11 000 patients. The addition of 1 drug to a single agent doubled response rates from 13% to 26% (odds ratio [OR], 0.42; P<.001) and improved median survival (median ratio, 0.83; P<.001) and 1-year survival (OR, 0.80; P<.001). On the other hand, the addition of a third drug to a 2-agent regimen, improved response rates from 23% to 31% (OR, 0.66; P<.001) but failed to improve the 1-year survival (OR, 1.01; P = .88) or the median survival (median ratio, 1.00; P = .97). In a minority of the studies, toxicity data were available. As expected, hematologic toxicities and some nonhematologic toxicities were more frequently seen with 2-agent regimens vs monotherapy and with 3-agent regimens vs 2-agent regimens.

Several methodologic deficiencies can be identified in this study. For instance, the meta-analysis did not use individual patient data. A meta-analysis using individual patient data is laborious but may provide a more valid tool for analysis than the use of published data.⁸ Also, only a fraction of patients had survival data available: 88% and 53% of patients had data available for median survival and 1-year survival, respectively. Despite these relatively minor flaws, the results of the meta-analysis by Delbaldo et al seem conclusive and confirm the belief of most clinicians who treat patients with NSCLC that 2-drug combinations are standard first-line therapy for advanced NSCLC and that 3 drugs are not better than 2. This seemingly algebraic paradox has generated frustration for patients and researchers throughout the world. Contemporary 2-drug combination regimens achieve, at best, a median survival of 8 to 10 months and a 1-year survival rate of 35% to 40%. Yet, the question remains: Why has medical research not advanced the field beyond the results achieved with 2 chemotherapeutics in advanced NSCLC?

A general principle of cancer chemotherapy is to combine cytotoxics with different mechanisms of action and partially nonoverlapping toxic effects with the goal of improving antitumor efficacy.⁹ With a few notable exceptions, such as for the treatment of hairy-cell leukemia, combination chemotherapy has evolved as standard therapy for most neoplasms.⁹ In highly chemosensitive neoplasms, such as germ-cell tumors and aggressive lymphomas, the use of multiagent combinations is potentially curative. On the other hand, the task of showing a survival benefit with the addition of a third cytotoxic to a 2-agent regimen has generally been arduous, and 3-agent combinations have not become standard therapy for most advanced solid tumors. In the meta-analysis reported by Delbaldo et al, a small improvement in response rates by 8% with the use of 3 vs 2 drugs in advanced NSCLC was not sufficient to prolong survival. It is possible that such minute increases in response rates will become more important for patient outcome in early stage NSCLC when eradication of micrometastatic disease is required.

Perhaps the quality rather than the quantity of the drugs should be questioned. Poor ingredients make poor combinations. How powerful is the current armamentarium against lung cancer? Since the time nitrogen mustard was introduced as a promising therapy for bronchogenic carcinoma in 1948,¹⁰ only a handful of agents have been found to be clinically useful against NSCLC. Platinum agents have been the mainstay of lung cancer chemotherapy since a meta-analysis published in 1995 demonstrated a small but statistically and clinically significant survival advantage compared with no therapy.⁶ In the 1990s, 5 newer agents, namely paclitaxel, docetaxel, vinorelbine, gemcitabine, and irinotecan, were added to the armamentarium of cytotoxics and have been combined in 2- or 3-agent regimens, in many instances accompanying cisplatin or carboplatin. The single-agent response rates observed with newer agents, even though better than those observed with old agents, are modest, ranging between 10% to 30%.¹¹ Delbaldo et al showed that the magnitude of improvement in response rates and survival with the addition of a second agent was smaller when the control group was receiving a new drug vs an old drug, which suggests a possible superior outcome with the new agents.⁷ However, their study does not allow for a direct comparison between old and new agents, and definitive conclusions cannot be drawn.

A major impediment to the development of cytotoxic combinations has been the emergence of intolerable toxicities. Adding a cytotoxic drug usually requires a compromise in dose intensity of individual agents due to unwarranted adverse effects. It is possible, but unproved, that dose intensity, or at least a dose threshold, is important for optimal antitumor efficacy in NSCLC. In accordance with this hypothesis, Delbaldo et al showed that trials with planned attenuation of drug doses in combination regimens resulted in lower benefit in terms of response rates (3 agents vs 2 agents and 2 agents vs single agents) and survival (2 agents vs single agents). This observation is of interest and merits further evaluation.

Chemotherapy provides a meaningful survival benefit to patients with advanced, incurable NSCLC. Two cytotoxic drugs are better than 1; however, the addition of a third cytotoxic agent to a 2-drug regimen does not appear to be beneficial in terms of improving survival in advanced NSCLC. It seems clear that alterations in the components or schedules of standard cytotoxic regimens are unlikely to result in major improvements in outcome in NSCLC. Instead, ability to ultimately win the war on lung cancer is likely to stem from a deeper understanding of cancer biology with the development of drugs aimed at targeting pathways specific for cancer cells, such as the epidermal growth factor receptor pathway inhibitors and the angiogenesis inhibitors. These targeted therapies have already shown promise in lung cancer (as is the case of the epidermal growth factor receptor inhibitor, gefitinib, in recurrent NSCLC)¹² or other malignancies (as is the case of the vascular endothelial growth factor monoclonal antibody, bevacizumab, in colorectal carcinoma)¹³ and are the subject of intense ongoing research. It is anticipated that continued development of these and other novel targeted agents will have a major impact in the treatment of cancer, including NSCLC, in the future.