Concurrent Chemoradiotherapy in Stage III Non–Small-Cell Lung Cancer: What Is the Best Regimen?

The accepted standard treatment for inoperable stage III non– small-cell lung cancer (NSCLC) today typically consists of concurrent chemoradiotherapy that is based on so-called platinum-doublet chemotherapy. The majority of treatment centers administer cumulative radiation doses between 60 and 66 Gy in conventionally fractionated schedules. There is currently an ongoing discussion regarding the optimal radiation dose and schedule; data from North America provides at least a note of caution concerning radiotherapy doses higher than 60 Gy, whereas other groups around the world have demonstrated 63 to 66 Gy to be safe and effective. Nevertheless, the optimal dose and fractionation of radiotherapy in this setting will remain a matter of continuing discussion. If we sum up the published evidence from randomized trials, a meta-analysis that was based on individual patient data, and further published experience from several large phase II trials, there is currently no consensus regarding which chemotherapy is best to combine with radiotherapy in this curative setting. At the end of the day, the number of randomized trials in stage III NSCLC with convincing outcomes has unfortunately been far too small. The broadest evidence concerning this issue comes from trials that have included cisplatinbased doublets—particularly, cisplatin and etoposide or cisplatin and vinca alkaloid. Despite this accepted evidence, there has been a strong trend in North America and even in parts of Europe to prefer outpatient administration of weekly low-dose carboplatin and paclitaxel combination schedules that are simultaneously administered with outpatient radiotherapy, on the basis of the assumption that this is more convenient and possibly just as effective as cisplatin-based doublets. But is this decision evidence based? In the article that accompanies this editorial, Santana-Davila et al report their investigation of data derived from the large Veterans Health Administration (VHA) database. Over a 10-year period (2001 to 2011), they combined patient groups receiving either cisplatin and etoposide, or a combination of carboplatin and paclitaxel, administered concurrently with curative doses of radiation. Several baseline parameters could be extracted from the VHA database, including several prognostic but also some therapeutic parameters. For statistical methods, they used Cox models that were adjusted to the propensity scores and instrumental variables methods. The aim of their study was to compare the two chemotherapy protocols with respect to the survival outcome of the treated patients. From their findings, they concluded that carboplatin and paclitaxel, when administered concurrently with radiotherapy, resulted in survival outcomes that were comparable to cisplatin and etoposide in a comparable clinical setting. Given that the statistical methods used by the authors are often used and are now generally accepted, probably nobody would have challenged their interesting findings. However, a critical evaluation of this data set raises the possibility of an underlying patient selection bias in their patient population, which should be carefully taken into account before we draw definitive conclusions from the work. First, using the VHA database almost naturally resulted in a patient group that consisted primarily of men (over 97% were men). Does this reflect the current typical population with stage III NSCLC referred to our thoracic centers? Frankly, no. Given that, in recent years, the number of women has steadily increased, we would currently expect 35% to 40% of the patients presenting with and requesting curative-intent treatment for stage III NSCLC to be women. In addition, the ratio of approximately 22% adenocarcinomas and between 45% and 50% squamous cell carcinomas would also not represent the typical patient subset that is currently being diagnosed with stage III disease at our institutions. Both in North America and in Europe, and even more so in Asia, the rate of adenocarcinoma histopathologies would today generally exceed 50%, and even be as high as 60% in selected regions. In this regard, it is not clear that squamous cell carcinoma in a man is the most prevalent problem facing oncologists these days. Rather, we do see considerably more women and also some men with adenocarcinomas presenting for first-line curative therapy. So can we still learn something from the presented retrospective outcome analysis? Yes, we can, given that in men with predominantly squamous cell carcinoma, the combination of carboplatin and paclitaxel did emerge as a regimen that may still have clinical relevance. Although much smaller in number, the men with an adenocarcinoma also seemed to have comparably good outcomes on the basis of the statistical analysis in this study. Finally, there is an additional note of caution when generalizing the results of this retrospective analysis. Both comparisons of patients groups treated later in time (2005 to 2007 and 2008 to 2010), when analyzed against the group treated from 2001 to 2004, seemed to support the value of the cisplatin-based doublet as a partner to radiation. This finding could theoretically be related to the newer antiemetics now in general clinical practice, namely the 5HT3 receptor JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 33 NUMBER 6 FEBRUARY 2