Although EGFR tyrosine kinase inhibitors and chemotherapy agents have been tested in previously treated patients with advanced NSCLC, and tarceva, alimta, and taxotere are all approved by the US FDA in this setting, we haven't had studies directly comparing chemo to targeted therapy. However, we're starting to get the first glimpses of information, including a randomized Phase III trial out of Japan that gave previously treated advanced NSCLC patients either iressa or taxotere. In addition to the results being put out in the public domain on the internet last week (here, but in Japanese only), one of the leading lung cancer experts in Japan, Dr. Nagahiro Saijo from the National Cancer Center Hospital, described the results (in English).

The trial (known as V-15-32), which started back in 2003, was done in several countries but enrolled largely from Japan, and the design is shown here:

(click to enlarge)

It accrued just under 500 patients who had metastatic disease or had a recurrence after surgery and had been treated with 1-2 lines of prior chemotherapy (>80% in both groups had received just one prior chemo regimen). The two arms were well balanced, with 62% men, 45% 65 or older, and about 45% having demonstrated a response to prior chemo on both arms. Never-smokers accounted for 29% on the iressa arm and 35% on the taxotere arm (at least twice the proportion we typically see in the US). Just over 75% on both arms had adenocarcinomas: Japanese populations tend to have a higher proportion of adenocarcinomas than Europe or the US (European lung cancer populations tend to have more squamous than adeno, while the US has more adenos, but not as disproportionately adenocarcinoma-heavy as in Asia). We are also seeing higher response rates and survival in Japanese patients with chemo, and have definitely seen much higher EGFR mutation rates and higher response rates to EGFR inhibitors in Asian populations vs. the US or Europe. Survival also tends to be higher overall in many of the Asian, at least Japanese trials, compared to the US, and there tends to be different toxicity profiles even with the exact same regimens and eligibility criteria, so there appear to be true genetic differences operating here. Because of the high response rates to iressa in Asia, this agent is still very widely used as an early therapy there, even though a large randomized trial (Iressa Survival Evaluation in Lung cancer, or ISEL) that compared iressa 250 mg by mouth daily to placebo in previously treated patients with advanced NSCLC showed no overall survival benefit (Thatcher ISEL abstract here), as shown:

The ISEL trial actually showed that there was a survival benefit in never-smokers as well as Asian patients, even though the overall trial didn't show a survival benefit with iressa.

The preliminary results from this direct comparison of iressa to taxotere as 2nd or 3rd line therapy was quite interesting. Although the response rate was nearly twice as high for the iressa arm (22.5% vs. 12.8%), the rate of "disease control" that includes stable disease was equal at about 33-34% for both arms. More improtantly, the one-year survival was actually higher in the taxotere arm, at 54% vs. 48%, although both groups had a much higher survival than we see in second/third line trials in the US, (usually 20-30% at one-year). Interestingly, just as we have seen survival benefits with a rather low response rate, here we see one-year survival better in the arm with the lower response rate. Treatment discontinuation was about the same in both arms, 13.5% for iressa and 17.6% for taxotere. The results are summarized in this table:

The objective of the overall trial was to show statistical "non-inferiority" for iressa, or basically that it was definitely as good as taxotere. The trial didn't quite show that, so there remains a statistical possibility that iressa was not as effective a treatment as taxotere IV every three weeks.

What can we say about what this means for general practice? While it certainly raises questions about whether EGFR inhibitor therapy is as effective as chemo overall in the second/third line setting, there are problems with making broad conclusions. First, tarceva demonstrated a significant survival benefit compared to placebo overall in the BR.21 trial (abstract here), but iressa did not, so it's not clear whether the results would really be the same with tarceva. On the other hand, iressa and the EGFR inhibitors in general have looked their best in Asia, so this largely Japanese trial not looking better for iressa in a part of the world where you'd expect iressa to do its best are a potential problem. There are other trials coming out in the next year that directly compare EGFR inhibitor therapy to chemo, and those will help us clarify the situation better. We also don't know how iressa or tarceva would do in a selected population of never-smokers or patients with EGFR mutations. But in an unselected population, these are the findings we've got, at least for now.