While the differences in anticipated clinical benefits from EGFR tyrosine kinase inhibitors like tarceva and iressa are well known (summarized in prior post), less well appreciated is the potentially significant differences in results with garden variety standard chemotherapy. These differences are likely to be very relevant as the US has a harder time getting large trials completed, especially placebo-controlled ones, as we look more to trials done in other parts of the world to prove what agents are beneficial and well tolerated. For instance, the original trial of alimta in mesothelioma (abstract here) was a worldwide effort, and the BR.10 trial of tarceva vs. placebo (abstract here) was led by NCI Canada but really conducted throughout the world, with a very minimal contribution of patients from the US. So we really need to know whether conclusions can be generalized from one race to another, one continent to another, or whether we might get very different results in the US than they see in Japan, even if both places conduct their clinical research well.

To test whether there really are meaningful differences in outcomes when different racial/genetic/geographic populations receive the same chemotherapy, Dr. David Gandara and the SWOG Lung Cancer committee collaborated with leaders of lung cancer trials to ensure that both North American and Japanese groups would develop trials using the same chemo regimen (carbo/taxol at identical dose and every three week schedule) in trials done in the US or Japan for patients with the exact same eligibility. This would then allow for comparison of the results across different trials that were run essentially concurrently (many years of difference in timeline could have led to differences over time in supportive care or subsequent treatments available), producing a "common arm analysis". The carbo/taxol arms from three different trials have been compared, two from Japan (abstracts here and here) and one from the US (abstract here). As you'd hope for trials with the same eligibility, the enrolled populations were quite comparable to each other:

(Click on image to enlarge)

While you'd expect this, it's not a foregone conclusion. Social patterns, or differences in the characteristics of NSCLC on one continent vs. another could lead to significant differences in factors like age, balance by patient sex, and definitely histology. For instance, European trials have historically had more patients with squamous NSCLC than adenocarcinoma, unlike the US and definitely Asian trials; they also typically have 70-90% men, which may be from lung cancer still being a very disproportionately male disease in Europe, and perhaps from men being more likely to be recommended for or agree to participate on European clinical trials. Regardless, as shown above, the common arms in the Japanese and US trials included similar patients by demographic and basic cancer features.

They differed greatly, however, in how patients fared, both in terms of survival and side effect profile. Here you can see the summary table of basic efficacy endpoints, which demonstrated no differences in response rate to the regimen but a significantly higher survival among Japanese patients (on both Japanese trials) who received the same regimen as US recipients:

Interestingly, the more favorable results among Japanese patients didn't occur because they tolerated the carbo/taxol regimen better. In fact, the toxicity information demonstrated that Japanese patients actually had a significantly higher likelihood of a very low white blood cell count (neutropenia) and a fever in that setting (febrile neutropenia):

In fact, I look at those numbers and think the differences are quite remarkable, considering that we're talking about patients receiving the exact same drugs and doses inthese studies.

So we see that patients in Japan getting carbo/taxol regimen "by the book" have a much lower drop in their white blood cell counts, on average, than US-based patients. While you'd predict that this would lead to a higher risk of infection and overall worse outcomes, in fact Japanese patients had a quite superior survival. We can't say why this is the case. Could supportive care and access to care be so much better in Japan than in the US? Could patients in Japan be overall healthier despite having the same disease as in the US? Is advanced NSCLC a more indolent cancer in Japan? Perhaps more patients in Japan received profound benefits from EGFR inhibitors after first line chemo than American patients...it's not possible to draw conclusions from this analysis.

I'll delve into this issue a little further in my next post, covering differences between Japanese and US-based work in SCLC, and some biological underpinnings of differences in how chemotherapy is processed in different individuals.