• © 2005 by American Society of Clinical Oncology

Will Future Progress in Non–Small-Cell Lung Cancer Be Step by Step … or by Leaps and Bounds?

“We cannot sit and wait for the promise of tomorrow so long as stepwise progress can be made with the tools at hand today.”¹

In terms of lung cancer therapy, these words of Dr Howard Skipper from 1979 seem to apply just as much today as they did a quarter of a century ago. Before 1980, the standard of care for locally advanced/unresectable non–small-cell lung cancer (LA-NSCLC) was definitive radiation therapy (RT) alone. However, the 5-year survival rate was disappointingly low, at only 5%.² The treatment paradigm began to shift after a landmark trial by the Cancer and Leukemia Group B (CALGB 84-33) demonstrated that induction chemotherapy (with two cycles of cisplatin and vinblastine) followed by RT significantly improved survival compared with RT alone.³

Yet, the results with sequential chemotherapy followed by RT in CALGB 84-33 seemed too good to be true. There was almost a “triple jump” improvement in the 5-year survival from 6% with RT alone to 17% with sequential therapies.³ Indeed, two subsequent randomized trials utilizing the same regimen as CALGB 84-33 (with substantially more patients) demonstrated more modest long-term survival rates of 8% to 10%.^4,5

The next incremental gain in LA-NSCLC came from the use of concurrent rather than sequential chemotherapy and RT. In 1999, Furuse et al reported in the Journal of Clinical Oncology the results of a randomized trial with 320 patients, comparing sequential chemotherapy (mitomycin, vindesine, and cisplatin) followed by RT (56 Gy) versus concurrent chemoradiotherapy.⁶ They found a statistically significant improvement in 5-year survival of 16% for concurrent chemoradiotherapy compared with 9% for sequential therapies (P = .01). In another randomized trial with more than 600 patients, Radiation Therapy Oncology Group (RTOG) trial 94-10, Curran et al reported remarkably similar superior results of once-a-day RT (63 Gy) concurrent with chemotherapy over the same therapies delivered sequentially.⁷ Zatloukal et al also performed a randomized phase II trial favoring concurrent therapy.⁸ Patterns of failure data suggest that the benefit of concurrent chemoradiotherapy is due to a synergistic effect that leads to an improvement in local, rather than distant, control.⁹

In this issue of the Journal, Fournel et al add yet another study corroborating the clinical benefit of concurrent over sequential chemoradiotherapy.¹⁰ In this phase III trial, 212 patients were randomly assigned between induction chemotherapy with cisplatin and vinorelbine followed by RT (66 Gy), compared with the same RT concurrent with cisplatin and etoposide followed by consolidation therapy with cisplatin and vinorelbine. While not statistically significant, they report potentially clinically meaningful differences in the median and 4-year survival rates, favoring the concurrent arm (16.3 v 14.5 months and 21% v 14%, respectively). The key reason for the borderline result was the lack of adequate statistical power in this study to reject the null hypothesis. According to their study design, with a calculated sample size of 210 patients, Fournel et al only had the ability to detect an improvement in 2-year survival from 20% in the sequential arm to 35% in the concurrent arm. This relative improvement of 75% was simply too high a bar, too great a leap, and did not follow the general rule of “stepwise” improvements that one would expect in this setting. Moreover, unlike RTOG 94-10, this trial was confounded by the addition of consolidation chemotherapy in the experimental arm. Fortunately, this was not the first trial on this subject, or some may have incorrectly concluded that concurrent therapy is no better than sequential treatments.

What is the next step? With the goal of improving systemic control, it is logical to test the role of either induction or consolidation chemotherapy delivered before or after the concurrent chemoradiotherapy. The preliminary results from CALGB 39801, however, do not support the use of induction chemotherapy before definitive chemoradiotherapy.¹¹ Similarly, upfront chemoradiotherapy proved to be the optimal arm in a randomized phase II (Locally Advanced Multimodality Protocol [LAMP]) trial evaluating paclitaxel and carboplatin in conjunction with RT.¹² In this regard, LA-NSCLC is similar to its “sister” disease of limited-stage small-cell lung cancer, in which the benefit of integrating RT earlier, rather than later, with chemotherapy has been demonstrated.¹³ Regarding consolidation chemotherapy, the jury is still out. In Southwest Oncology Group (SWOG) trial 9504, all patients received upfront chemoradiotherapy (with cisplatin and etoposide) followed by consolidation with docetaxel. This study reported a promising 3-year survival rate of 37%.¹⁴ Currently, there are ongoing randomized trials further testing related strategies.

What about the radiation therapy itself? Should we continue to use “conventional” doses of approximately 60 Gy? Or, in the interest of improving local control, are there tools that will enable us to safely escalate the dose of RT? Socinski et al reported a phase I/II trial of 62 patients with stage III NSCLC treated with two cycles of induction chemotherapy followed by concurrent chemoradiotherapy using conformal RT with doses escalated from 60 to 74 Gy.¹⁵ They found a promising 3-year survival of 40%. Only one patient had grade 2 RT pneumonitis, and less than 10% had grade 3 to 4 esophagitis. Interestingly, in a recent secondary analysis of RTOG lung cancer trials activated during the 1990s (stratified by chemotherapy use and accrual year), the biologically equivalent dose (BED) of RT was a key factor on multivariate analysis that significantly predicted for both improved local control and survival. For example, an increase in 10 Gy BED was associated with a reduction in mortality of 14% (RTOG, personal communication, May 2005). While this hypothesis needs to be tested prospectively, if chemotherapy and radiation are going to “dance together,” should we not optimally choreograph these therapies such that they “dance in step”?

During the last quarter of a century, important strides have been made in LA-NSCLC. From the 1980s to the 1990s to the present, the 5-year survival rates for patients with LA-NSCLC has gradually improved, from approximately 5% with RT alone, to 10% with sequential chemotherapy followed by RT, to approximately 15% with concomitant chemoradiotherapy. Randomized trials along the way have demonstrated that this 200% improvement in long-term survival (5% to 15%) is more than just stage migration. However, at this rate (approximate 5% increase in survival per decade), most of us will be retired before the long-term survival rate even reaches 40%! How do we step up the pace of improvement so that it moves forward by leaps and bounds?

Currently, we have entered an exciting era of molecular biologic therapies. In a randomized, placebo-controlled trial (National Cancer Institute of Canada BR.21) with more than 700 patients, erlotinib, an epidermal growth factor (EGFR) tyrosine-kinase (TK) inhibitor, was shown to prolong survival in advanced NSCLC after first- or second-line chemotherapy (P = .001).¹⁶ Such new agents are now being tested in patients with LA-NSCLC, both sequentially and concurrently with chemoradiotherapy. Yet, I believe the real promise is not simply the advent of new biologic therapies, but rather the novel way that researchers are critically thinking about these molecular strategies. Recently, mutations in the TK domain of EGFR were found to be strongly associated with sensitivity to erlotinib or gefitinib in patients with NSCLC.¹⁷ This breakthrough suggests that mutational profiling will allow these agents to be applied to the patients most likely to benefit. In all likelihood, additional mutations will be discovered that will refine our capacity to select appropriate therapies. This is a clear paradigm shift from the “shotgun” approach applied to lung cancer therapies in the past. Lung cancers accumulate a wide variety of mutations in oncogenes and/or tumor suppressor genes.¹⁸ If future therapies are tailored to the individual's molecular fingerprints, the potential for success is boundless. Is this strategy a leap of faith? I think not … but in the meantime, we must continue to build on the consistent, stepwise gains to further improve outcome for our patients with NSCLC.