Randomized Phase III Study of Gemcitabine-Cisplatin Versus Etoposide-Cisplatin in the Treatment of Locally Advanced or Metastatic Non-Small-Cell Lung Cancer
Reviewers: John Han-Chih Chang, MD
Source: Journal of Clinical Oncology 1999; Volume 17: pages 12 ? 18
Chemotherapy (ChT) for advanced non-small-cell lung cancer (NSCLCa) has been a fairly controversial topic for many years. There have been randomized trials from multiple institutions (France, Spain, the European Organization of Research and Treatment of Cancer, MD Anderson and the Cancer and Leukemia Group B) that have supported the use of ChT as an adjunct to surgery and/or radiotherapy (RT). Still other trials have been equivocal about the benefits. Recent meta-analyses have demonstrated a benefit for survival when adding ChT.
Since we have established that ChT may be beneficial in NSCLCa, we need to find out what regimens are the most beneficial. Despite all our attempts to improve the treatment for this disease, we have very few long-term survivors of NSCLCa, unless they present in the early stages. We truly have a long way to go, and this paper may help us along that journey. This article details a multi-institution phase III randomized trial performed in Spain of a relatively novel ChT agent, Gemcitabine, in combination with cisplatin (CDDP). The most frequently utilized regimen, historically, included CDDP and etoposide (VP-16), the combination of which serves as the control arm of this trial for patients with locally advanced and/or metastatic disease.
Materials and Methods
Patients eligible for this trial had stage IIIB (locally invasive into mediastinum or vascular structures or lymph node involvement in the opposite mediastinum, hilum or supraclavicular region) or IV (metastatic disease). They were without prior ChT, but were allowed to have had RT. The patients must have had at least one lesion that could be measured in 2 dimensions on scans. Patients with brain metastases were not eligible.
The two treatment regimens were: arm 1 ? CDDP + VP-16 versus arm 2 ? CDDP + gemcitabine. The dose of gemcitabine was 1250 mg/m2 on day 1 and 8 of each 21-day cycle, while the VP-16 dose was 100 mg/m2 on days 1, 2 and 3. CDDP was given at 100 mg/m2 on day 1. Patients continued for a maximum of 6 cycles if they remained without progressive disease. Multiple patients (14 in arm 1 and 21 in arm 2) received RT during or after their ChT. The RT was mostly palliative for arm 1, while about half were for curative intent in the gemcitabine arm.
One hundred and thirty-five patients were enrolled for this trial between July 1995 and June 1996. Two were deemed ineligible. Table 1 in the paper demonstrates the equivalency of the two treatment arms in patient characteristics.
Minimum follow-up was 16 months on the survivors. Compliance with treatment administration was better on arm 2 where 30 of 69 (43%) received all 6 cycles, while only 26% (17 out of 66) received the total 6 cycles on the VP-16 arm. The median number of cycles for arm 1 was 4, while the gemcitabine arm had a median of 5 cycles completed. The mean dose per infusion was over 90% of the planned dose (not many needed a significant dose reduction)
Responses to treatment were assessed and seem to heavily favor the gemcitabine arm (response rate of 40.6% or 28 of 69). There were 14 responders (21.9%) out of 64 in the VP-16 arm. Significant prognostic factors that predicted for benefit in response (of gemcitabine/CDDP over VP-16/CDDP) were age 65 or less, female gender (small numbers ? question complete validity) and stage IV disease.
Toxicity data were summarized in tables 3 and 4 of the article. The most common hematological toxicities were myelosuppression and thrombocytopenia. Febrile neutropenia (fevers in the setting of a lowered white blood cell count) were more common in arm 1, while non-symptomatic severe (grade 3 ? 4) thrombocytopenia was much more common in the gemcitabine arm. While the grade 3 ? 4 thrombocytopenia did not result in any more bleeding episodes in arm 2 than 1, the increased febrile neutropenia did cause greater time of hospitalization for the VP-16 patients. One toxic death did occur on the gemcitabine arm. The most common non-hematological toxicity was nausea and vomiting. It was much more common in the gemcitabine arm, but was not statistically significant. Alopecia in the VP-16 arm was statistically much more common than in arm 2.
Quality of life was assessed with questionnaires given with each cycle of ChT. Over 85% compliance in filling out these surveys were observed. Based on analysis of those questionnaires, no difference in the quality of life was noted between the two treatment arms.
Survival was assessed, and a benefit was seen in those that had received gemcitabine along with CDDP, though not statistically significant. The figures 1 and 2 depict this data. Below are some of the survival data:
Arm 1: VP-16/CDDP
Arm 2: gemcitabine/CDDP
Median time to progression
6.9 months (p = 0.01)
8.7 months (p = non signif)
32% (p = 0.19)
This randomized trial has brought to light the fact that more responses can be obtained with gemcitabine than with VP-16 when combined with CDDP. Gemcitabine did have a benefit in increasing the time to progression, though the progression free survival curves for both treatment arms become very similar after 10 months (figure 1). Survival was not improved, though the trial was not really designed or empowered for that end point. Toxicities were fairly similar for both treatment arms.
Despite demonstrating a well-balanced trial in table 1, biases in the treatment arms may have been alluded to when they stated that 3 of the patients in arm 1 received RT for curative intent versus 10 in the gemcitabine arm. If a greater percentage are being radiated for curative intent during or after ChT, it seems to suggest a better patient population in arm 2. Another area that needs clarification was the difference in the number of cycles completed in the two arms (favoring arm 2). Did more patients in arm 1 not get full treatment because of progressive disease ? which would support the premise of the paper or is it related to patient or physician non-compliance or treatment-related toxicity? Being that the mean dose per infusion was close to full dose, progressive disease or non-compliance were the most likely etiologies for the decreased number of cycles, however, this was not well explained. This problem could bias the results to favor gemcitabine, since you cannot get a response to a drug if the patient does not receive it.
Despite some points of contention, this article places a promising outlook on the future of gemcitabine for NSCLCa. The problem is that we need to translate a response benefit to a survival benefit. Perhaps future studies will address gemcitabine before or after RT for locally advanced NSCLCa. There continues to be many avenues to pursue as the treatment outcomes for NSCLCa remain very humbling for oncologists.