In addition to developing new agents or refining which patients should or shouldn't be given to particular patients, another way to potentially gain ground in fighting cancer is to improve the tolerability of a standard treatment in a way that either makes it more effective or reduces side effects, thereby making it possible to give higher doses. The drug Abraxane (nanoparticle albumin bound paclitaxel) uses the approach of coating taxol with albumin to reduce the side effects associated with standard taxol (paclitaxel), making it possible to give it without steroids (which can be a rather bothersome issue for many patients, causing problems from severe insomnia to very high blood sugars and more) and also reduce some other taxol-associated side effects like joint and muscle aches (see prior post on Abraxane for more background information). It's also possible to give doses that may be high enough to work better than standard taxol. Abraxane has been approved in breast cancer since 2005, and it's used here and there in lung cancer, but there is relatively little data about it in lung cancer. However, earlier this week, it was reported that a large trial of carboplatin/Abraxane vs. carboplatin/taxol for initial treatment of advanced NSCLC was "positive", at least in terms of showing a significant improvement in response rate. But what does this mean for the field of lung cancer?

The short answer is that we don't know because we need much more information, to be presented at the large ASCO meeting in early June, to really be able to assess the impact that this trial should have. For starters, response rate isn't the most important endpoint in lung cancer, and I would consider it to be of relatively little importance compared with overall survival, tolerability/quality of life, and progression-free survival. Response rates represent the proportion of patients with tumor shrinkage beyond a strict set of criteria, but there are many reasons that the impact of a treatment may be dissociated from response rate. You'd definitely prefer to see major tumor shrinkage in more patients rather than fewer, but some treatments like the EGFR inhibitor Tarceva (erlotinib) or the investigational ALK inhibitor crizotinib (PF-02341066) may have a relatively or extremely low response rate in a broad population but be hugely effective in that group in which the cancer shrinks.

Still, it's good to see more people with major tumor shrinkage, but a key issue is who the comparator group is (the control arm). In this trial, patients on the control arm received carbo/taxol, a perfectly acceptable and plain vanilla combination. It's a very acceptable comparator because it's essentially completely average. However, in North America today, many patients who are eligible for the anti-angiogenic agent Avastin (bevacizumab) receive that with carbo/taxol, because the three drug combination clearly beat carbo/taxol alone in terms of response rate, but also progression-free survival, and overall survival. That's a combination of improvements that we'd hope to see. Importantly, the response rate in the trial of carbo/taxol with or without Avastin, the chemo alone arm had a response rate of only 15%. In contrast, carbo/taxol/Avastin had a response rate of 35% in that trial. Meanwhile, in a worldwide trial of cisplatin/Gemzar (gemcitabine) vs. cisplatin/Alimta (pemetrexed), the objective response rates were 28.2% and 30.6% respectively (the latter percentage including squamous cell NSCLC patients who are now known to not benefit from Alimta). My point is that carbo/taxol is completely fine, but beating it on response rate is not especially impressive. Meanwhile, we have many other treatments available now that can deliver on that (admittedly, comparing across trials in many cases), some associated with significant improvements in survival as well.

One other key advantage of Abraxane that is cited is that it doesn't need to be given with steroids and spares patients some taxol-related side effects. That's good, but we can already do that with Gemzar, Navelbine (vinorelbine), and some other readily available treatments.

Essentially, I'd say that this news is welcome, that it's always good to get some positive trial results in lung cancer, but we really need to learn much more before we try to interpret the clinical significance of these results. The FLEX trial of chemo with or without the antibody against EGFR Erbitux (cetuximab) showed a rather marginal survival benefit along with an improvement in response rate, but it hasn't been FDA approved and isn't widely used or even felt to be especially appealing among many (probably most) lung cancer experts because of the lack of perceived significant benefit compared with the added side effects, requirement for weekly treatment, and very high cost. If Abraxane delivers only an improvement in response rate without even a significant improvement in progression-free survival or overall survival, I can't imagine that the oncology world will be impressed. On the other hand, if it delivers a 2-3 month improvement in median survival, then it's a legitimate contender as an attractive treatment for a broad range of patients. We'll see when the actual data are presented.

As a few final points, I'll mention that it's relevant where the trial was conducted and how the patients do in absolute terms. One of the limitations of the maintenance therapy trials that have shown a survival benefit in advanced NSCLC is that only about 20% of patients on the placebo arms ever received the opportunity for the same active drug later, so access to what we'd consider optimal treatment was very questionable. I think that's a problem with running trials in places with very disadvantaged health care systems just so that you can get the trials done quickly. If you run a trial primarily in central or eastern European countries where you're lucky to get anti-nausea medication with your chemo and there's no opportunity for second line therapy, are results generalizable to a health care system in which effective second and third line options are routinely available? If the Abraxane trial shows an improvement in median survival from 9.5 months with carbo/taxol to 12 months with carbo/Abraxane, that's truly impressive. If it shows an improvement from 7.2 months with carbo/taxol to 9.0 months with carbo/Abraxane, that just means that carbo/Abraxane was average and beat very unimpressive results with carbo/taxol.

And then there's cost. Abraxane is very, very expensive, and that's perhaps justifiable if it offers major value. Avastin, Alimta, and Tarceva are also painfully expensive, and several other non-generic chemo agents aren't exactly a bargain either. Taxol is available as a generic, so it's cheap as an off-patent comparably effective chemo, and Abraxane is about an order of magnitude more expensive.

A key question is whether a plain vanilla bowl of generic store brand ice cream (taxol) costs $1, and you add a few sprinkles, how much better is that? Even if it's better than plain vanilla without sprinkles, is it worth $10? What if you can get Ben & Jerry's Chunky Monkey for the same price or less? Carbo/Abraxane may beat the plain vanilla without sprinkles, but is it better than cisplatin/alimta, cisplatin/gemzar, carbo/taxol/avastin, or carbo/alimta/avastin? We don't know whether carbo/Abraxane will be offer an improvement that is quite small (sprinkles) or pretty major (cookie dough and brownies) vs. plain vanilla, but it's priced in the super-premium shelf of the ice cream section.

The press release doesn't really provide us any meaningful information, and it provides yet another example of what seems like a marketing-driven hijacking of medical information that jacks up the price of the stock. We'll see if actual evidence delivers, but right now I can tell you that a difference in response rate over carbo/taxol alone is not especially impressive on its own.