For years it has been generally accepted that the choices for the second drug in a platinum doublet for treating metastatic non-small cell lung cancer (NSCLC) were pretty much interchangeable. The question of whether cisplatin is better than carboplatin is a separate question, one which GRACE’s own Dr. Sanborn recently reviewed quite nicely. For the second drug, as long as the choice was a “newer generation” drug, oncologists would mix and match cisplatin or carboplatin with Taxol (paclitaxel), Taxotere (docetaxel), Gemzar (gemcitabine), or Navelbine (vinorelbine) based mostly upon which drug they liked the best or had the most experience with.

This perception that the specific second drug was irrelevant was not just pulled out of a hat; there has been reason to think they are all equally good. The most often quoted study to make this point has to be the ECOG 1594 trial (Comparison of Four Chemotherapy Regimens for Advanced Non–small-cell lung cancer), which randomized NSCLC patients to one of 4 arms: cisplatin/Taxol, carboplatin/Taxol, cisplatin/Gemzar, and cisplatin/Taxotere. The survival curves were essentially identical (see below), as were the response rates of the tumors.

Well, now a few studies have been published that are challenging the notion that it doesn’t matter which of these agents you combine with platinum. For example, a meta-analysis (a type of study which combines the results of multiple smaller studies to magnify effects too small to see in the smaller trials) published in 2005 suggested that combinations with Gemzar were superior to non-gem regimens.

In this month’s issue of The Oncologist, a group led by Dr. Francesco Grossi published a meta-analysis of 45 trials with 11,867 patients, comparing platinum doublets with all of the primary new generation drugs (Gem, Navelbine, Taxotere, and Taxol). They then compared not overall survival, which tends to be diluted out over time by patients getting effective second and third-line chemotherapy, but the percentage of patients who progressed at the earliest time points.

This is a fascinating endpoint, one which is definitely unorthodox but makes sense. The response rates for the trials and the drugs were all about the same, which we could have guessed. But not everyone who benefits from chemo responds! Some only have stable disease, but lack of progression (synonymous with stable disease) is every bit as good as response to the patient who is not dying of his/her cancer. In this trial, the authors looked at the number of patients who had progressive disease at the earliest evaluation point, usually at about 6-8 weeks after starting chemo, and compared this rate among the drugs.

What they found was very interesting. They confirmed the benefit of Gem-containing regimens previously shown, which had a 14% lower risk of immediate progression than non-Gem-containing regimens. But what was most intriguing about this study was the result with the Taxol-containing regimens.

Carboplatin and Taxol has been the most commonly used combination in the United States for many years. It is the backbone of the ECOG 4599 regimen (carbo/taxol/Avastin), and is the comparison arm for many phase III trials that test new combinations. You would assume that something so important was definitely at least as good as other combinations, right?

Well, maybe we should think again. In the Grossi study, regimens that contained Taxol again had similar response rates to other combinations, but also had a statistically significant 22% higher chance of the patient progressing at the earliest time point compared to the other combinations, which indicates that significantly fewer patients were getting disease stabilization from the Taxol combinations than from other drugs.

In truth, Taxol has been falling out of favor for many oncologists for a while now. It has to be given very slowly and has a high rate of anaphylactic reactions, and the regimen causes complete hair loss which is a deal-breaker for many when regimens like carbo/Gem do not. There are also newer versions of paclitaxel on the market, such as Abraxane, which do not carry the same risk of allergic reactions and may deliver more drug than traditional Taxol.

What does this mean for oncologists? Should we stop using Taxol now and just use Gemzar instead? I think the jury is still out. For one thing, the significance of this new endpoint is unproven. How important is a lower rate of early progression? I have no idea, but it is certainly troubling to think that so many of my patients getting Taxol are having the “It isn’t working” talk so very early. I would be curious to hear what our other oncologists have to say on this topic.