FLEX was a European trial that enrolled 1125 previously untreated patients with advanced NSCLC who were all screened and found to have expression of the EGFR protein on their tumors (because that’s what the monoclonal antibody binds to). There are various levels of stringency for the amount of protein expression, but this trial was as lenient as you could get, with evidence of the protein by immunohistochemistry (IHC) on just one single cell being considered enough to get waved into the trial. A total of 15% of patients who were screened didn’t have any EGFR expression by IHC, which may be a reason why this trial was positive for a survival benefit but the results from the BMS-099 trial have been less clearly favorable (a rumor last month of it being positive for survival was wrong and based on miscommunication — we don’t have any survival data yet for that).

Regardless, the FLEX trial randomized patients to either cisplatin/navelbine alone or the same chemo with erbitux weekly, and patients who didn’t progress after 6 cycles would receive erbitux weekly as a maintenance therapy until progression or serious problematic toxicity. It was reported in the plenary session because it was only the second trial that has shown a significant survival benefit from adding a targeted agent to chemo (the first being avastin), and this is the first that applies to a much broader patient population, since it included patients with a marginal performance status and didn’t exclude patients with squamous tumors, on blood thinners, unlike the trial of avastin (although the FLEX trial also excluded patients with brain metastases). But the overall difference in median survival was only 1.2 months, or 5 weeks — 10.1 months in the chemo alone arm, compared with 11.3 months with erbitux. There was only a 5% difference in one-year survival between the two groups, although both groups did better than we’ve seen in older studies (42% vs. 47%). The trial just barely met the pre-defined criteria for what would be considered a statistically significant improvement in survival, and many in the audience, both general oncologists and also the lung cancer specialists, were largely left struggling with the question of whether the difference was really large enough to be considered clinically meaningful.

There were some very interesting differences between Caucasian and Asian patients. For starters, Asian patients did startlingly better than Causasian ones overall (regardless of assignment to chemo or chemo/erbitux). How much better?? how about a median of 19.5 months for the Asian population vs. 9.6 months for Causasian patients? Asian patients were far more likely to be never-smokers — 52% (!!) vs. 17% — and they were very likely to get oral EGFR tyrosine kinase inhibitors such as tarceva (erlotinib) or iressa (gefitinib) than Caucasian patients (61% vs. 17% overall, pooled across chemo and chemo/erbitux). Asian patients didn’t receive a significant benefit from erbitux, perhaps because the patients who were assigned chemo alone were much more likely to get a drug like tarceva later (72% vs. 44%). But this point is worth considering: if the survival in Asian patients, including more than 50% never-smokers, was markedly improved because of second or third line oral EGFR inhibitors, they seemed to work after erbitux (median survival 17.6 months for prior erbitux recipients, who as noted above were far less likely to be given tarceva or iressa later than the chemo alone patients, who had a median survival of 20.4 months).  If we start using erbitux commonly as a first line therapy, we’re going to need to figure out whether and when drugs like tarceva would still have a role, since both agents target the same molecule, although from two different angles.

The survival benefit in Caucasians was about 2 months, so that is arguably in the same range that we’ve seen and accept as worthy of becoming a new standard of care, in the same ballpark as that with avastin.  But some of the lung cancer experts are bothered by the fact that there was no improvement at all in the progression-free survival, which is almost always seen when there is a survival benefit from a new drug (it’s hard to understand why patients will live longer if they progress on the new treatment just as quickly as with the old treatment).  There was a small increase in the response rate with erbitux, from 29% to 36%.  On the downside, erbitux was associated with a rash very similar to tarceva but probably even more of a challenge, and a high rate of fevers with low blood counts (15% with chemo alone, 22% with chemo/erbitux).

Overall, there isn’t a clear consensus yet that chemo with erbitux is a standard, and I think many oncologists will think that the survival difference of 1.2 months combined with the added side effects and cost, in the context of another trial that hasn’t been declared positive, isn’t enough to change practice yet.  I suspect most oncologists will continue to use chemo and avastin for avastin-eligible patients, but chemo with erbitux will be a strong consideration and will likely be the approach I use for avastin-ineligible patients, such as those with squamous tumors, and probably those with brain metastases as well.  That won’t apply to my Asian patients, based on the lack of benefit in that group.  And some but probably most oncologists won’t be giving erbitux with cisplatin/navelbine, despite the fact that we haven’t seen positive results with the trial that gave carbo/taxol or carbo/taxotere with erbitux.

We’re going to continue debate what this trial means for a while.  I’ll have a few concluding thoughts on it in my next post.

    Some financial analysts have already described the findings and potential financial implications (here), but I’ll give you an oncologist’s perspective, and a sense of what my colleagues feel about the strength or weakness of the results and what they’ll mean to the changing standards for lung cancer treatment, after the session Sunday.

In the next few weeks, I’ll cover each of the topics I’m covering in some detail, but I thought I’d just describe some of the research I’ll be covering in this whirlwind tour.  There are two studies on using molecular markers to guide post-operative treatment decisions.  One uses a “gene signature”, a pattern of activity of a collection of genes that can add a better sense of whether a patient will be a long-term survivor after surgery.  The other trial, out of Spain, looks at expression of a gene called BRCA1, a gene first identified as being associated with breast cancer (hence the name BRCA1), but this marker also may correlate with doing well on some chemo drugs but not others.  The study I’m discussing is a small trial just looking at 83 patients who had their post-operative treatment guided by the degree of their tumor’s BRCA1 expression.

Next are several studies that look at pathologic variables as prognostic tools for patients with NSCLC.    Several trials are looking at how well the number of lymph nodes removed at surgery actually predicts how well a person will do (more is better), as well as how helpful it is to know the actual number of nodes that are positive, rather than just staging based on whether any nodes are positive in a certain location.

Another trial looks at a large US database of patient information (anonymous) to see whether patients with node-positive bronchioloalveolar carcinoma (BAC) have the same or a better prognosis than patients with non-BAC node-positive tumors.

Finally, a study out of Japan looks at the clinical factor of age in a previously reported study of adjuvant chemotherapy to see whether in this trial the benefit of post-op chemo was restricted to a certain age group.

Although I’m going to storm through all of these topics in just a few minutes tomorrow, I’ll cover all of them in some detail in the next couple of weeks.

But now I’ve got to get to sleep.  Big talk tomorrow.

6)  Molecular signatures to predict which patients benefit from adjuvant chemo (abstract here):  Using tissue from 133 patients enrolled on the JBR.10 trial of adjuvant cisplatin/navelbine vs. observation (New England Journal of Medicine abstract here), investigators from NCI-Canada were able to identify a 15-gene signature (pattern) that can discriminate between high risk patients, who received a very striking benefit from adjuvant chemo, and low risk patients, who received a significant detrimental effect from chemo.  This is very encouraging, but whether this will end up being more useful and widely adopted than any of several other gene signature concepts or even just ERCC1 (described in a prior post) is not clear.

7)  Randomized trial of carbo/alimta vs. carbo/etoposide in SCLC (abstract not yet available): we haven’t seen much evidence of activity for alimta in SCLC thus far, but this trial will likely clarify whether it offers a benefit compared to our current standard of platinum/etoposide.   It would be nice to have something new: now that it appears irinotecan isn’t a real improvement in the US (abstract here), unlike the more favorable experience in Japan, we’re left with a standard that goes back several decades without our ability to improve on it.   A phase II trial with cisplatin or carboplatin combined with alimta (abstract here) looked OK, if not especially encouraging, compared to our typical results with a platinum and etoposide. (more…)

1) FLEX trial (abstract withheld): The only one that we don’t have an abstract for, because it’s been selected for the plenary session and designated as among the most important at the meeting.  This phase III randomized trial of cisplatin/navelbine alone vs. same chemo with the EGFR inhibitor erbitux (cetuximab)  as first line treatment for advanced NSCLC has already been declared positive for a survival benefit in a press release (see post here) from all the way back in September, 2007, with no additional information since then).  With that news, it already becomes a player.    Where it fits in relative to current standards really depends on the actual numbers.  The lung cancer world will be watching…

2) SWOG 0124 (abstract here): This trial is probably the most important one being presented on the topic of SCLC this year.  As a background, we’ve had years of back and forth about whether the cisplatin/irinotecan combination that has appeared so beneficial in Japan is actually better than our current US standard of platinum/etoposide in a North American population (see discussion of this saga in a prior post here).  The SWOG 0124 trial randomized 671 patients with previously untreated ED-SCLC randomizes patients to cis/irinotecan vs. cis/etoposide to provide that answer.   Sadly, it wasn’t worth the wait: both looked very comparable, with more blood count-related side effects with etoposide, more diarrhea with irinotecan, but overall no meaningful differences in activity.  The hope is that we can use genetic profiles of patients to predict which drugs will work better in which patients. (more…)

   But another article was published (abstract here) on the combination of celebrex (at a pretty high dose of 400 mg twice daily) and iressa (at a typical dose of 250 mg daily) as a first line therapy for advanced NSCLC, in an unselected population (not chosen by smoking status, EGFR mutations, COX-2 protein expression on the tumor, etc.).   This study was conducted by the Hoosier Oncology Group (affectionately known as HOG) in Indiana, who enrolled 31 patients.  They found that there was certainly activity, but the response rate of 16% and median survival of 7 months was on the low side for what you’d expect in a group of patients with advanced NSCLC who started with the standard approach of platinum-based doublet chemo (with or without avastin).   Another concerning issue was that 2 of these 31 patients developed fatal interstitial pneumonitis, and inflammation of the lungs, a complication that can rarely occur with EGFR inhibitors, but I’ve never had a life-threatening or fatal case occur in the well over 200 patients I’ve treated with iressa or tarceva in the past 5-6 years, so having this occur in 2 of 31 patients, or 6%, is pretty concerning for the possibility that there is a danger to this combination.

    It’s possible that the results would have been better if they only enrolled patients who never smoked, or who had EGFR mutations or high expression of the COX-2 protein on their tumors.  We can’t say that this combination wouldn’t be good for anyone, but this trial was a clear disappointment to the investigators and corroborates my belief that there is definitely no reason to deliberately combine these drugs in hopes of improving the impact as a lung cancer treatment outside of a clinical trial.

   The best evidence we have to address this issue is a trial by the Eastern Cooperative Oncology Group, ECOG 3503 (abstract here), in which 137 patients (118 eligible) with previously untreated advanced NSCLC were treated with first line tarceva, starting at the typical dose of 150 mg daily.  But the dose was then escalated by 25 mg every two weeks until patients developed either grade 2 rash (scattered bumps or spots or general skin redness with itching or other symptoms) or significant other side effects that precluded dosing up to a maximum of 250 mg per day.  

    Only half of the patients (60 of 118) ended up pursuing the dose escalation, presumably because the other half already had enough of a rash or other side effects that increasing tarceva dose wasn’t feasible.  Only 15 (13%) were escalated up to 250 mg daily.   The overall response rate was only 7% (8/112, with one complete response), and the overall survival wasn’t any better than you’d expect. 

    This trial was actually presented at ASCO last year, but it’s one that we heard almost nothing about.  It hasn’t been published as a full manuscript (that’s not unusual, as it can easily take many months or a year or more to get a full paper drafted and accepted in a journal), but I suspect that this information hasn’t gotten out there into the world because the results were pretty disappointing.

   But they do make an important point.  So to recap, this trial showed that with aggressive dose escalation of tarceva, the response rate was less than was seen in the larger trial with 150 mg and then dose reduction as needed (abstract here); survival was also no better than you’d expect.  So this really suggests that there isn’t any incremental benefit to escalating dose.  Patients who don’t benefit on tarceva don’t appear to be underdosed.  The standard dose of 150 mg per day seems to be adequate, and there doesn’t seem to be an incentive to increasing side effects.   This isn’t especially surprising for me, since I see that some of the patients who do very well on tarceva also do it on a reduced ongoing dose of 100 mg or sometimes lower when they’re experiencing toxicity.  You don’t need to experience pain to receive the gain.     

    There were a few limitations to that work.  First, some prior, generally smaller studies didn’t clearly support the conclusion that maintenance or early second line chemo is definitely superior.  Because of that, most experts felt that it would be helpful to get another study that supported maintenance chemo before we declared it a standard of care.  Second, the prior trial waited a full three months before doing a repeat scan that would trigger a start of chemo in the delayed chemo arm — and about 1/3 of the patients on that arm were too sick to get chemo by the time they were found to have progression.  That’s too long, in my opinion, to wait before checking for progression, which is often found radiographically before a patient gets too sick for chemo.   With so many people in the delayed chemo arm not getting it, the trial was in some ways a study of everyone getting immediate chemo vs. 2/3 getting delayed chemo — not fair.

    But yesterday there was a press conference sponsored by ASCO to highlight the results of a trial sponsored by Eli Lilly and being presented at the oral presentation on advanced lung cancer at ASCO in two weeks (abstract here), and this result added to the prior study will likely change the standard of care, in my opinion.   The new trial, called JMEN by Lilly (every common has their own cryptic coding for trial names, and I don’t know if ANYONE really knows what JMEN refers to — it’s not an acronym), asked a very similar quesiton to the one from last year — does maintenance chemo (or early second line chemo, depending on your point of view) improve progression-free survival, the time before someone shows cancer progression and needs to change treatment plans?  (more…)

   ASCO changed its policy this year to give everyone equal access to these abstracts online.  They were just released and are available here.   The only ones that aren’t available yet are the highest stakes ones that are in the plenary session, including the “FLEX” trial of chemo with or without the EGFR monoclonal antibody Erbitux (cetuximab), which is reportedly positive for a signficant survival benefit according to a September, 2007 press release (with the 8 months since then a ridiculously long time to have no further information offered). 

   There are hundreds of abstracts just on lung cancer, and they are broken down into early stage/adjuvant, locally advanced (including LD-SCLC), and metastatic lung cancer.   We’re going to have LOTS to cover over the next few months.  So look around, and I’ll try to provide background and commentary on as much as possible over time.

The groups were similar, remember this was not a randomized prospective study; the median age was 65 for those receiving Tarceva and 74 for those on Iressa. About 26% of the whole group were never/former smokers. They all were pretty physically fit (ECOG PS of 0 or 1), and most of them (55-58%) had adenocarcinoma.

There were no complete responses and the rate of partial response was 6% and stable disease of 26% for the people on Tarceva and 13% partial response and 29% stable disease for those on Iressa. These figures, although low, are entirely in keeping with previous study results for both drugs given to unselected patient populations.

What was interesting was that for the people who had a partial response the median survival was 9.7 months, but it was 9.1 months for those who had stable disease, and only 3.7 months for those who progressed on treatment. This was a trend noted in other studies and one I certainly see in my clinic, but it was reassuring to see it reproduced. And hopefully reassuring to anyone who might have “only” stable disease. Dare I say “size isn’t everything”?