Tarceva Life Cycle Leader
Name: Don MacLean
Position: Life Cycle Leader (Project Leader) for Tarceva, Roche
Background: : US Head of Regulatory Affairs, Roche
Education: PhD, Pharmaceutical Chemistry, University of London
As reported in last week's Pharmacogenomics Reporter, a clinical trial by researchers at the Princess Margaret Hospital in Toronto found Tarceva to be efficacious in most subgroups under scrutiny. However, a New England Journal of Medicine study concerning the trial noted that gene-copy number by FISH testing and positive EGFR protein expression by immunohistochemistry seemed capable of identifying patients who would respond to the non-small cell lung cancer drug. But the number of tumor samples used for that part of the Tarceva research was too small to generate significant results.
That left a few questions about the intentions of the three partner companies overseeing the drug Genentech, Roche Pharmaceuticals, and OSI Pharmaceuticals. So late last week, we spoke to Don MacLean, Roche Pharmaceuticals project leader for Tarceva, in the hope that he would shed some light on Tarceva's current situation.
The drug is currently a second- and third-line therapy, and a move to a first-line indication would mean it could be prescribed to a responding subset, at the very least, of the estimated 138,000 Americans diagnosed with NSCLC annually.
Can you tell me if Roche is interested in pursuing a diagnostic to identify Tarceva responders?
I guess there's been a lot of discussion on this, and I think in the area in general, all the work that's gone on has all been retrospective subset analysis of a limited number of patients from all trials. That's why the data are basically underpowered, that no real conclusions can be drawn.
So, in our future trials, which we're setting up now, that we're doing prospectively, we're going to look at potential markers whether it be FISH, EGFR [immunohistochemistry], mutations in the EGFR pathway or other pathways and try and correlate that with outcome. And for that, we need to get tissue, obviously paraffin tissue, which we're having as an entry criterion.
So, we're trying to answer the question moving forward in a robust, scientific way, but at the moment it's basically up in the air. We don't believe that at the moment there's any need to test, because our pivotal trial demonstrates benefit in a broad range of patients, and when you start slicing and dicing subsets, you just lose power, and all the chance creeps in. However, we do understand that there is a potential that something may be a marker, and we're keen to find that out.
Assuming you establish that people without a marker don't gain survival, is that a drawback to running studies on markers?
At the moment it's not a drawback, because the question is open, so you can ethically run trials, which we're doing. Now, if we find out in future randomized trials that you can clearly select or deselect patients based on a marker, and it may be something that we don't know about today but we'll do some broad testing in our trials then we would obviously amend our labels accordingly. But at the moment, there's no requirement for testing in our labels that are approved, and that's our position moving forward until we get more data.
Is there interest within Roche to get Tarceva approved as a first-line treatment?
Yes, we're moving ahead with our first-line trials with our partners, Genentech and OSI. We basically agree what to do, divide up the work, and get on with it.
In order to get first-line approval, will it be necessary to identify responders?
In these first-line trials, these are the trials where we'll try and determine if a marker exists and if it does predict response. But there's a mandatory requirement to have paraffin tissue before you can get into the trial. So, we'll be able to answer that question definitively.
It's hard to get tissue in BL21, after phenomenal effort, we got paraffin tissue from 45 percent of the patients, and of that 45 percent, a certain percent was used for either IHC or FISH or mutation [testing] it varied a bit depending on the test. But if you want to do a proper trial, you need to get 100 percent of patients with tissue if you can, and plan it prospectively.
Where did tissue come from?
It was archived tissue these are primary diagnostic tissue samples. So again, there are similar limitations on that, because there needs to be a sufficient amount available, but our protocol was pretty specific in what we need.
An analyst suggested last week that, if a test was required for first-line use, doctors might be hesitant to prescribe Tarceva for second- and third-line in the absence of a test.
That's a fair assumption, I think. At the moment, there's no real strong data to support a marker either selecting or deselecting patients for whatever line of therapy. But if our first-line trials show that, then we would take appropriate action to amend our prescribing information. But you need to have the data to do that. I think the FDA understood that if you look at the FDA label for Tarceva, which is pretty long and convoluted like most labels they put the data in there, but they also understand that it's an unanswered question, and we have commitment trials to answer them.
When do the first-line therapy trials begin?
They're due to start they're basically starting this quarter. We should be enrolling patients this quarter. They'll have a big Phase III trial, so they'll take probably 18 months to enroll, and you've got follow up, so it's a three-year, two-year report, basically.
Do any of the tests look better than the others?
Well, it's speculation. You've seen the data, so you could say maybe FISH seems a reasonable sort of guess, but there are other things that may be better. Maybe a combination [of IHC, FISH, and EGFR-mutation testing]. But we're going to look at a number of markers in our future trials. You have to be pretty clear to have a marker that, if you deselect a patient from treatment, it has to be pretty clear that they're not going to get any benefit.
We've looked at the data pretty deeply, and the more you slice and dice it, the more you get into uncertainty. And [in light of] all the literature you see from ASCO and [in the] literature I guess the final conclusion is that this warrants randomized prospective trials.
How large is the trial?
It's going to be about 850 randomized patients. One sample of tissue per patient. We are due to enroll in the third quarter.