Treating Small-Cell Lung Cancer: More Consensus Than Controversy

I am a little surprised that Ganti et al question or caution about issues for which a fair amount of level I evidence is available. Having just worked with the late John Murren and George Simon on the American College of Chest Physicians' evidence-based guideline, and having written with Nevin Murray a concise review of first-line treatment for small-cell lung cancer (SCLC) in the new Journal of Thoracic Oncology,[1] I find a lot more consensus than controversy on most of these issues.

A review needs to identify the mainstream of information important to today's therapy, including the controversial areas, but needs to point a way out of the thicket of conflicting trials with rational guiding principles. Authors have the duty to provide clear prose and balanced information, but must distinguish ideas that represent their point of view from that which is historically true or established fact. Unless they have some prevailing currency, old issues and old trials need to be allowed to settle to the bottom of the stream, and should not be dredged up to muddy the waters.

Background, Incidence, and Staging

As our review[1] pointed out, the incidence of SCLC as a clinical entity seems to be decreasing, from about 20%-25% in the 1970s to as low as 13.8% now. In 1980, there were about 50 American Society of Clinical Oncology (ASCO) lung cancer abstracts for each cell type, ie, small-cell and non-small-cell lung cancer (NSCLC). In 2004, there were still 50 SCLC abstracts, but the upsurge in interest in NSCLC resulted in 350 abstracts—a sevenfold increase. Industry realized that lung cancer offered a big market for selling many products, but SCLC is too small a niche for them, and clinical research in this setting has been retarded and suppressed. Due to the lack of modern trials, older trials are "immortalized," never being eclipsed by newer trials, so the meta-analyses go back 20 to 30 years to find "evidence," often neglecting that staging tools have changed as well as the basics of standard therapy.

Writing about what constitutes limited disease as opposed to extensive disease can be a contentious business. With a financial grant from Eli Lilly, the International Association for the Study of Lung Cancer (IASLC) has invested resources to redefine stages for SCLC and NSCLC. That report is due out any day now. Pleural effusions, as seen on chest x-ray, are clearly extensive disease and have been dealt with in this fashion for 20 years (the era in which the role of thoracic radiotherapy has been ensconced). The argument about pleural effusion and stage periodically arises and stirs debates that have no clinical relevance. When older trials looked at pleural fluid as a prognostic marker for survival in chemotherapy trials, it did not seem to make a difference. Once we accept the evidence that thoracic radiotherapy improves local control and survival—and that evidence is supported by two meta-analyses, as well as guidelines from the National Comprehensive Cancer Network (NCCN) and American College of Chest Physicians (ACCP)—one cannot treat the entire hemithorax with radiotherapy; it is not a "reasonably sized" target.

Another piece of evidence that did not fit into the

New England Journal of Medicine

report of the Intergroup trial was that although pleural effusions were specifically excluded, a few patients enrolled in the study had pleural fluid. The survival for the pleural effusion subset was significantly worse. We see cases otherwise limited with bulky central disease, and computed tomography (CT) scans are much more capable of detecting pleural fluid. On the other hand, they can also detect pleural studding and small nodules. Whether we call these cases limited or extensive may have importance for clinical trials, but these factors may influence other aspects of clinical management as well.

Management of Limited Disease

The Intergroup study set the high-water mark for survival and brought to light some rather interesting factors that I feel the need to reemphasize. The study compared two pilot studies that were roundly accepted as demonstrating the best outcomes that people had seen up until that time. The exception from the original pilot studies was chemotherapy duration and the agents used. We used only four cycles of cisplatin/etoposide. We did not use the "alternating" tactic with cyclophosphamide, doxorubicin [Adriamycin], and vincristine (the components of the widely used CAV regimen). Interestingly, none of the meta-analyses on the role of thoracic radiotherapy used cisplatin/etoposide as upfront treatment. When one looks back at this time, the term "alternating therapy" was prevalent; however, it has now fallen completely away.

When the study was designed, no one used doses of more than 50 Gy. Nonetheless, the disease melted, and as Ganti et al noted in their paper, it was considered a step forward to prove that 37 Gy was better than 25 Gy. No one considered using a dose greater than 50 Gy in the era without CT-based treatment planning, because esophagitis was very debilitating and commonly led to stricture. Rapid early response and excessive toxicity was attributed to the thoracic radiotherapy, and there was controversy about whether there was any role for it in a disease that disseminated early. When CAV or the alkylator-based treatments were used, esophageal strictures were common, and concurrent therapy was damned by "intolerable" toxicity.

The current theory that the Intergroup trial used the wrong control or comparative arm, because one arm was "biologically more active" than the other, is either a misunderstanding of history or a revisionist fantasy. When one calculates the difference using alpha/beta ratios and attempts at biologically equivalent doses, one is hard pressed even in retrospect to show clear or compelling differences. The equipoise that allowed the trial was that each treatment had performed rather well in pilots conducted by the Eastern Cooperative Oncology Group (ECOG) for the twice-daily regimen, or the Southwest Oncology Group (SWOG) for the once-daily regimen. Many forecast that this twice-daily idea was daft and would fail in a large national trial. That did not prove to be the case.

Early reports were remarkable because of how similar outcomes were with a median follow-up of 24 months. Treatment planning was in its infancy, and CT-based treatment planning had not yet dawned. Escalating doses beyond 50 Gy started after the Intergroup trial was completed and in follow-up. While doses of 60 to 70 Gy can now be administered, they have never been compared in prospective trials; the assumption has been that a higher dose must be better, but there are no compelling data to support that belief. Since protraction of radiotherapy dose takes 6 to 7 weeks or longer, it also overlaps with more cycles of chemotherapy, and may lead to more dose reductions and treatment breaks—both of which can be the cause of inferior local control and poorer survival.

Local Failure

My 1998 ASCO submission was the first report after a sufficient number of events had been recorded to discern a survival effect of cisplatin/etoposide plus concurrent radiotherapy in this setting. Although we showed a minimum 5-year follow-up and a significant survival difference,[2] because of the two prior reports (one called "final" in the title), ASCO leadership chose not to make it a plenary presentation. (The urge to report early results before sufficient events have accumulated is not isolated to this trial.)

Moreover, this abstract[2] also reported that local failure was counted for all patients with partial responses, which were scored in about one-third of both groups. The 5-year survival in the twice-daily arm was nearly identical (23%) to that of the overall group, but only 8% of the once-daily patients survived for 5 years. This means that local failure was exaggerated, and some patients deemed to have experienced local failure on the twice-daily arm were actually cured, surviving 5 years or longer. I've estimated the local failure to be 10% to 15%, not the number reported errantly in the statistical report and New England Journal paper. In addition, the cases were not planned using three-dimensional techniques, and there were major protocol failures in treatment delivery, which were significantly worse on the twice-daily arm. As I have remarked, it is indeed better to hit the target once each day than it is to miss it twice.

Local failure is an interesting endpoint but can be rather difficult to discern; the finding depends on what studies are done and when they are declared as controlled, failed, or stable. The Schild update of the North Central Cancer Treatment Group (NCCTG) study used two daily fractions but planned a 2-week treatment interruption, and this is in no way similar to the accelerated scheme also employing two daily treatments. The NCCTG total treatment time of the once-daily and twice-daily treatments was quite similar,[3] so there was no accelerated treatment arm. As done, this study actually provides evidence that by decreasing the intensity of the treatment, the desired effect (improved survival) is lost. Protracting treatment time (either by planned breaks as in this study or unplanned breaks in practice) and increasing dose by a small amount produced no significant results and might be a cause of resistance and local failure.

Timing of Radiotherapy

The argument about early vs late use of thoracic radiotherapy arises in studies that continue to use cyclophosphamide and/or an anthracycline for chemotherapy. In many ways, the framing of the question about radiotherapy timing is more about the chemotherapy than the radiotherapy. While this is a throwback to the days when this issue was debated, the evidence points to four cycles of cisplatin/etoposide as standard therapy, given that this combination can be used in full doses with full doses of thoracic radiotherapy. The same cannot be said of cyclophosphamide and anthracycline chemotherapy—it is their toxicity and inability to be used with thoracic radiotherapy that raises this red herring.

The recent British paper by Spiro et al[4] recreated an earlier Canadian trial that had found an advantage for earlier treatment. Since this trial went back to older chemotherapy regimens, and all of the current guidelines advocate four cycles of cisplatin/etoposide, one has to question the relevance of such a trial. The theoretic debate concerns proliferation of tumor cells that are resistant to either chemotherapy, radiotherapy, or both. The advocacy for delay centers on the ability to deliver full doses of chemotherapy, which might be impaired or cause dose reductions or delays when early radiotherapy is used.

This concept was addressed by the Cancer and Luekemia Group B (CALGB) in their trial of "early" vs "late"—a trial that remains the banner effort in the crusade for delaying treatment with radiotherapy. The doses of chemotherapy in the "early" arm were markedly reduced. Survival in this trial is remarkable for its low numbers in all arms, probably due to patient selection and the fact that the chemotherapy was less effective—ie, no cisplatin/etoposide. When one eliminates the alkylator/anthracycline-based trials, the argument and controversy dissolve. Three meta-analyses[5-7] and an editorial[8] have focused on the subject of timing. While conclusions may be controversial because of the old trials, the evidence points to early being better when you stick to cisplatin/etoposide as the only chemotherapy. It is when you try to use more chemotherapy cycles[9] and revisionist CAV chemotherapy that early vs late radiotherapy can again become controversial.

The practice of delay in order to "get all the chemotherapy in" may be based on theory or the economics of wanting to deliver four to six cycles without the bother of involving radiotherapy. The institution of a planned break in the radiotherapy schedule also delays treatment delivery. DeRuysscher and colleagues introduced the concept of "start to end of radiotherapy" (SER) to analyze the empiric data using meta-analytic techniques.[10] According to their assessment, starting chemotherapy or any treatment stimulates the proliferation of surviving tumor cells, especially those that resisted the first cycle. Survival is superior when radiotherapy treatment is completed in 30 days—a little more than 4 weeks. Obviously, if this assessment is correct, it would preclude delay of initiation of chemotherapy until after completion of four to six cycles of chemotherapy. I find these data most compelling.

Prophylactic Cranial Irradiation

There is no question about the benefit of prophylactic cranial irradiation (PCI). Its role is established by the meta-analysis published by Auperin et al,[11] who reported a reduction in brain relapse and a 5% improvement in survival. Due to the lack of modern trials addressing the subject, all of the studies in the meta-analysis are dated, with the overall survival blunted by inadequate staging, cyclophosphamide/anthracycline-based initial chemotherapy, and variable assessment of systemic response to therapy and baseline brain imaging (usually CT, not magnetic resonance imaging, if even these)—actual long-term survivors from these studies are few.

Current research does not center on the use of PCI, but rather, on what is the best dose to minimize brain relapse and neurocognitive deficits. Neurocognitive deficits were more commonplace in the past, with protracted chemotherapy rather than the current limit of four cycles. The problem of decrease in executive function and activities of daily living has been inadequately addressed by the older studies. The best example is the French work—the largest trial in the Auperin meta-analysis—which followed patients for 24 months with regular "mini-mental" assessments. The frequency of deficits was less than 10% regardless of whether PCI was used.

The methodologies of these trials load the dice against PCI. The magnitude of the difference diminishes by not continuing to assess cognitive mental function in the "no-treatment" control group who suffer CNS relapse. Neurocognitive function is clearly worse for those suffering brain failures, either from not having PCI or not having it in a sufficient dose to prevent brain relapse. The mythic, anecdotal data indicting PCI as the cause of deficits needs to be addressed by careful trials. One such study is currently being conducted on each side of the Atlantic, testing high-dose (36 Gy) vs low-dose (25 Gy) PCI.

As a final consideration, no PCI results in brain relapse rates of 50% to 60%, and the neurocognitive functional impairments are not even recorded for these patients. Again, measurable deficits occur in less than 10% of patients whether PCI is used or not. It seems injudicious to raise concerns about a safe and effective therapy when withholding such therapy is so risky. On the other hand, efforts to optimize the timing and dose are on target, and recording late events seems a very prudent course.

—Andrew T. Turrisi, MD

Disclosures:

The author has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.

References:

1. Murray N, Turrisi AT: A review of the first-line treatment for small-cell lung cancer . J Thorac Oncol 1:270-278, 2006.

2. Turrisi AT, Kim K, Sause W, et al: Observations after 5 year follow-up of Intergroup Trial 0096: 4 cycles of cisplatin (P) etoposide (E) and concurrent 45 Gy thoracic radiotherapy (TRT) given in daily (QD) or twice-daily (BID) fractions followed by 25 Gy PCI. Survival differences and patterns of failure (abstract 1757). Proc Am Soc Clin Oncol 17:457a, 1998.

3. Schild SE, Bonner JA, Shanahan TG, et al: Long-term results of a phase III trial comparing once-daily radiotherapy to twice-daily radiotherapy in limited small cell lung cancer. Int J Radiat Oncol Biol Phys 59:943-951, 2004.

4. Spiro SG, James LE, Rudd RM, et al: Early compared with late radiotherapy in combined modality treatment for limited disease small-cell lung cancer. A London Lung Cancer Group multicenter randomized clinical trial and meta-analysis. J Clin Oncol 20:3823-3830, 2006.

5. Fried DB, Morris DE, Poole C, et al: Systematic review evaluating the timing of thoracic radiation therapy in combined modality therapy for limited-stage small-cell lung cancer. J Clin Oncol 22:4785-4793, 2004.

6. Hunracheck M, McGarry R: A meta-analysis of timing of chest irradiation in the combined modality treatment of limited stage small cell lung cancer. Oncologist 9:6665-6672, 2004.

7. DeRuysscher D, Pijl-Johannesma M, Vansteenkiste J, et al: Systematic review and meta-analysis of randomized, controlled trials of the timing of chest radiotherapy in patients with limited-stage, small-cell lung cancer. Ann Oncol 17:543-552, 2006.

8. Jeremic B: Timing of concurrent radiotherapy and chemotherapy in limited-disease small-cell lung cancer: "Meta-analysis of meta-analyses." Int J Radiat Oncol Biol Phys 64:981-982, 2006.

9. Work E, Nielsen OS, Bentzen SM, et al: Randomized study of initial versus late chest irradiation combined with chemotherapy in limited-stage small-cell lung cancer. Aarhus Lung Cancer Group. J Clin Oncol 15:3030-3037, 1999.

10. DeRuysscher D, Pijls-Johannesma M, Bentzen S, et al: Time between the first day of chemotherapy and the last day of radiation is the most important predictor of survival in limited-disease small-cell lung cancer. J Clin Oncol 24:1057-1062, 2006.

11. Auperin A, Arrigada R, Pignon JP, et al: Prophylactic cranial irradiation for patients with small cell lung cancer in complete remission. Prophylactic Cranial Irradiation Overview Collaborative Group. N Engl J Med 341:476-484, 1999.