I would consider the recently published IPASS trial that compared Iressa (gefitinib) to standard chemo of carbo/taxol (paclitaxel) to be an extremely influential trial in lung cancer that has essentially ushered in a new era of molecularly-defined guidance of our treatment for many patients with advanced NSCLC, and we can expect that this is how we’ll be approaching a much broader population of lung cancer patients. But there was actually another very similar trial done in Korea that was presented in the Plenary Session of the World Conference on Lung Cancer a couple of months ago in San Francisco that provides another opportunity to directly compare chemo to an EGFR inhibitor.

The First-SIGNAL trial (First-line Single Agent Iressa versus Gemcitabine and cisplatin Trial in Never-smokers with Adenocarcinoma of the Lung) randomized 313 Korean never-smokers with previously untreated stage IIIB or IV lung adenocarcinoma to either Iressa at 250 mg by mouth daily to standard chemo with cisplatin/gemcitabine as initial therapy. It asked whether first line Iressa would be associated with a significant improvement in overall survival (OS) and also looked at progression-free survival (PFS), response rate (RR), side effect profile, and quality of life between the two treatment arms. This trial had a rather ambitious goal of showing a 40% improvement in survival and enrolled only about ¼ the number enrolled on the IPASS trial, so the statistics reported here need to be taken with a grain of salt because the study is pretty underpowered to show many meaningful statistical differences.

As in the IPASS trial, the patient population was startlingly different from the broader population of lung cancer that we typically see in North America. Specifically, the median age was 57, compared with a median age just over 70 for patients newly diagnosed with lung cancer in the US. And while lung cancer remains disproportionately seen in men vs. women, 89% of the patients on the First SIGNAL trial were women. Over 90% had a good or excellent performance status.

The response rate was numerically (not significantly) higher (53% vs. 45%), but there were no real suggestions of differences in OS (21.3 vs. 23.3 months for Iressa and chemo, respectively) or PFS (6.1 vs. 6.6 months) in the overall trial populations:

(click on image to enlarge)

As you can see on the left, the PFS curves again showed the crossover pattern visible in the IPASS trial results and suggestive of their being two distinct populations in there:

Based on the results from the IPASS trial, we can guess that these two subpopulations on the First-SIGNAL trial reflect EGFR mutation status (positive vs. negative). As on the IPASS trial, only about a third (31%) of patients had tissue available for mutation testing, and in this Korean never-smoker population, the EGFR mutation incidence was 44%, and pretty comparable across patient sex, performance status, and other variables. The results largely recapitulated the findings of the IPASS trial:

1)RR was 85% vs. 26% for Iressa recipients with vs. without an EGFR mutation (26% RR being much higher than the 1% RR for EGFR mutation negative patients on the IPASS trial). For EGFR mutation positive patients who received first line chemo, the RR was 38%, which was actually lower than the 52% RR for chemo among the EGFR mutation negative patients.

2)Median PFS was better with chemo than with Iressa (6.4 vs 2.1 months) for EGFR mutation negative patients. It trended toward longer with Iressa for EGFR mutation positive patients (8.5 vs. 6.7 months).

3)OS results washed out completely, probably because of cross over. We know that 80% of the patients assigned to initial chemo later received an EGFR inhibitor, but we don’t know what proportion of patients assigned to initial Iressa received chemo after that.

As we’d expect, the trial showed that the two approaches had very different side effects. Chemo was associated with significantly more hair loss, fatigue, and decreased kidney function, while Iressa was associated with a significantly higher incidence of rash, itching, diarrhea, and liver function test (LFT) abnormalities (blood test out of normal range). Overall, patients who received Iressa rated their overall functional status and quality of life more favorably on several indicators than the patients assigned to initial chemo.

My overall impression of this trial is that it reinforces the same general conclusions that come out of the IPASS trial. Molecular variables and particularly EGFR mutation status are more powerful predictors of benefit with chemo vs. targeted therapies like Iressa or Tarceva (erlotinib) than clinical factors like smoking status, patient sex, adenocarcinoma, and other factors. Nevertheless, the First-SIGNAL study shows that the correlation of EGFR mutation status and outcome may not be as striking as suggested in the IPASS trial, where the RR was only 1% to Iressa among EGFR mutation negative patients. The Korean population reported here showed a 26% RR to Iressa, even among EGFR mutation negative patients. And unlike the IPASS experience, EGFR mutation positive patients in this Korean study did not have a markedly higher RR to chemotherapy than patients without an EGFR mutation (in fact, it was lower in the EGFR mutation positive patients than in the EGFR mutation negative).

At the same time, the median OS for the patients on this trial was closer to two years than the typical one year we see in North American trial, and it was still over 18 months even among EGFR mutation negative Korean never-smokers, so there are still some unknown factors in play here that lead many, never-smokes, perhaps particularly Asian never-smokers, to do so much better than a broader NSCLC population, even when an EGFR mutation doesn’t explain it.

Along with the relatively modest size of the trial in general, the subset who had tissue available was less than a third of the entire trial population, limiting the accuracy of the results in these molecular subsets because they now amount to just a few dozen patients in each. This highlights the biggest bottleneck we are now seeing in our enthusiasm for using molecular markers in lung cancer: availability of tissue. Along with that, there is also the difficulty that many of our pathology departments are having in figuring out how to get these tests performed, especially in less than 2-3 weeks, as patients and their oncologists become increasingly anxious about waiting on tests to finalize treatment plans. We are currently experiencing the growing pains of practice patterns that are evolving in real time.