Iressa was approved by the US FDA in May of 2003 as a third-line therapy, and for the next 18 months it was the only EGFR inhibitor on the market. At the ASCO conference a year later, Frances Shepherd, from the Princess Margaret Hospital in Toronto and chair of the Lung Cancer Committee for the NCI-Canada, first presented the positive results of the BR.21 trial randomizing patients in a 2:1 fashion to either Tarceva or a placebo as second- or third-line treatment for NSCLC; this was subsequently published in the New England Journal of Medicine (abstract here). I’ve summarized the trial and the demonstrated survival benefit in a prior post. So we came out of that meeting knowing that at least one EGFR inhibitor had a proven significant benefit, and while Iressa was the only EGFR TKI on the market, it continued to be prescribed and generally presumed to have similar benefit, given that the two drugs had the same mechanism of action, very similar side effect profile, and shared characteristics for patient populations who seemed to demonstrate the most impressive responses, such as Asians and patients with adenocarcinomas, especially BAC. I also gave an oral presentation confirming responses, some dramatic and long-lasting, for Iressa in BAC, at that ASCO meeting in 2004 (subsequently published, abstract here).

For a little more than six months, we expected the approval of Tarceva based on the proven benefit, and we saw similar characteristics with Iressa. Most of us in the field considered the two drugs to be like Coke and Pepsi, with some mild distinctions, but more similar than different. We were waiting on a trial of Iressa against placebo, but until November of 2004, Iressa was the only commercially available option, so there wasn’t much of a question of which drug to prescribe. People began to switch with Tarceva’s approval, but the big shocker came in December of 2004, when the first announcement of the ISEL (for Iressa Survival Evalaution in Lung Cancer ) trial was released, indicating that there was no significant survival benefit of Iressa compared to a placebo (abstract here).

(click to enlarge)

Shortly after that, the FDA no longer allows new prescriptions of Iressa, but patients who aren't progressing can stay on it, and it's also in other clinical trials. It continues to be widely used outside of the US, especially in Asia, and we’ll talk about why.

One of the reasons pretty much all of the lung cancer experts were surprised to see the ISEL trial emerge as negative was that it was so big, with 1692 patients, and larger trials have a greater chance of having even small improvements in absolute terms reach statistical significance. So even if Iressa gave only a 1.2 month improvement in survival vs. placebo compared with the 2 month survival benefit with Tarceva, it would have likely been statistically significant. This would have led to a catfight about whether there were really meaningful differences between the two drugs. But the ISEL trial showed just a 0.5 month improvement in median overall survival, and less than a month improvement even looking just at the patients with adenocarcinomas, who were expected to show more benefit.

The trial did include a planned analysis also of never-smokers, and of Asian patients as a distinct subgroup, with both of these showing significant improvements in median survival (2.8 months for never-smokers, 4 months for Asian participants who received Iressa), despite the negative overall results. Here are the curves split by smoking status:

But overall, it was too little benefit, no benefit in the overall trial, and since then Tarceva is now the EGFR inhibitor used in the vast majority of lung cancer management in the US.

So what happened? There were some subtle differences in the populations of the two trials, and the ISEL trial participants had to have shown recent progression within 2 months, so they may have been a bit more resistant as a group than the people on the Tarceva trial. However, another potential explanation, and the one that most experts favor, has to do with the effective dose between the two drugs at the “standard dose” for each drug. Remember that the strategy with Tarceva was to use the highest dose that could be reliably managed, while the strategy with Iressa was to use the lowest dose that could be helpful. The actual blood levels of active drug for Tarceva at 150 mg per day are actually in the ballpark of a daily dose of Iressa at more than 500 mg, actually in the ballpark of 700 mg per day:

So many people think the difference in outcomes for these two similar drugs has to do with the drug levels achieved at the doses used.

Interestingly, the ISEL trial did also show that the response rate for Iressa was 8%, comparable to the results of Tarceva on the BR.21 trial and really about the same as taxotere or alimta in previously treated patients, for that matter. We have also seen a small proportion of patients achieve great results with Iressa. My explanation of the overall results from this work is that the patients with EGFR mutations, which were first identified in April of 2004, are very sensitive to EGFR inhibitors and may benefit from doses as much as 10-fold less than less sensitive tumor cells without the mutation. Therefore, either Iressa or Tarceva may work equally well in the folks with EGFR mutations, and this may be why Asian populations have been encouraged by ongoing favorable results with Iressa and continue to use it (although the V-15-32 trial results may lead to questioning of that view). But the results on BR.21 demonstrate survival benefits in a broader population of patients who experienced stable disease, and I suspect this benefit may depend on higher drug levels of anti-EGFR activity required to help the much larger group of patients who don’t have an EGFR mutation. If Iressa at 250 mg daily provides lower drug levels than Tarceva 150 mg daily, then the vast majority of patients, the ones who didn’t have an EGFR mutation, may not have received enough active drug to show a survival benefit with stable disease/non-progression.

I’ll qualify my conclusions by saying that the explanation of why Tarceva worked and Iressa didn’t in the large trials against placebo are based on my best interpretation, not ironclad proof. But the fact of the matter is that Tarceva at the current dose showed a benefit, and Iressa at its chosen dose didn’t. The question is whether the results with Iressa are really negative, or maybe just not positive enough. I think the trials with Iressa can still provide some useful insights to generalize for using Tarceva, but there are enough important differences that I don’t consider the two to be interchangeable anymore. I’d be much less inclined to substitute Iressa for Tarceva at this point than I would to accept a Diet Pepsi instead of the Diet Coke I prefer.