One of the major questions in the field of EGFR mutation-positive advanced NSCLC is whether we should continue patients on EGFR tyrosine kinase inhibitor (TKI) therapy as we transition to new treatment options because of acquired resistance after an initial good response.  I've written several posts about this always evolving question about how best to manage acquired resistance, as summarized in this FAQ, but in this post specifically about the question of whether to continue the EGFR TKI when deciding it's time to start chemotherapy, we have been left primarily to our best judgment far more than data.  On the one hand, there is no evidence that giving an EGFR TKI concurrent with chemotherapy improves outcomes in any setting relative to chemotherapy alone, and it may just give the side effects of both chemo and the EGFR TKI, even after we now have evidence that the cancer is becoming resistant to the targeted therapy. On the other hand, we sometimes see a "flare reaction" from discontinuing a targeted therapy, even after patients are experiencing acquired resistance, and we might presume that only a subset of the cancer cells are resistant to the EGFR TKI, especially if progression is limited and somewhat slow. In that setting, perhaps it makes sense to treat the newly EGFR TKI-resistant cancer cells with chemotherapy, while continuing the EGFR TKI for the residual subset of cancer cells that seem to remain suppressed and are presumably still EGFR TKI-sensitive.

Up until now, we've had very little direct evidence to speak to this. A single randomized phase II trial in this setting showed no suggestion of benefit from continuing Tarceva (erlotinib) beyond progression, but that trial closed early due to weak accrual, with only 39 patients, and they were just required to have clinical benefit, not necessarily a proven EGFR mutation and clear response to first line EGFR TKI.

But what we really needed was a larger, randomized phase III trial of patients with a proven EGFR mutation who were all going to be randomized to the same chemo regimen with or without ongoing EGFR TKI after acquired resistance emerged.  And we just got the first results of such a trial.

The trial in question, known as IMPRESS, been looming on the horizon for years; specifically, 265 patients with an activating EGFR mutation and having developed acquired resistance to Iressa (gefitinib) , still widely used outside of the US and clearly very active for EGFR mutation-positive NSCLC, are randomized between cisplatin/Alimta (pemetrexed) and ongoing Iressa or the same chemo with placebo.

One of the principal investigators for the study, Dr. Tony Mok from Hong Kong, just presented the results of this trial at the European Society for Medical Oncology (ESMO) conference in Madrid, Spain, revealing the clear and practice-changing result that there is no benefit to continuing Iressa, with even a trend toward a detrimental effect.  The primary endpoint of the trial, progression-free survival, was the same, with a median of 5.4 months, on both of the two arms.  More notable was the difference in median overall survival, which was 14.8 vs. 17.2 months, in favor of the chemo alone arm (P = 0.029); the data were noted as relatively immature still, with many patients still alive, so the survival results may well change over time, but Dr. Mok noted that the findings were important enough that they merited being presented and discussed.   

And so they are.  Over the past few years, I and many other experts in lung cancer with extensive experience in managing patients with acquired resistance have come to favor continuing many and even most of our patients on EGFR TKI therapy in anything short of very striking progression.  There are some modest imbalances favoring the placebo arm -- more patients with a complete response in the placebo group than ongoing Iressa (76% vs. 68%, both still very high numbers), fewer with brain metastases on the placebo arm (23% vs 33%), and more patients on the placebo arm going on to receive additional chemotherapy (12.9% vs. 3.8%) and/or another EGFR TKI (33.3% vs. 26.6%).  Frankly, those percentages for follow-up treatment are surprisingly low to me, since many patients with an EGFR mutation continue to have a very good performance status, and many may benefit from 3rd line chemotherapy, +/- another EGFR TKI (I suspect that many of these patients were enrolled in China or parts of Europe where off-protocol therapy beyond first line tends to be rather "minimalist").  But it's hard to argue today that the data are more compelling today in favor of  ongoing TKI.

Though these data are with gefitinib, we have always presumed, I think safely, that the same would hold up with erlotinib. Other trials with a similar design are being done in the US with erlotinib, though the completion of the IMPRESS trial may undermine that effort.  I'd be interested in confirming that the same conclusions hold up if we look at EGFR exon 19 deletion and exon 21 L858R substitution patients separately, as they may possibly show different patters. I'd also like to confirm that results aren't different depending on the country patients were treated in, since post-progression treatments may have been far more readily offered and administered in some countries than others.

But the biggest change likely to obviate this question, at least for a large subset of EGFR mutation-positive patients,  is the potential availability of novel 3rd generation EGFR TKIs with activity in the 60% of patients with acquired resistance mediated by the new appearance of the  T790M mutation, such as CO-1686 (newly named rociletinib) and AZD9291, as discussed in this post on new options for T790M-positive patients.  We're finally breaking the impasse and getting better data for patients with EGFR mutations and acquired resistance, though our options for T790M-negative patients are still limited enough that the results of the IMPRESS trial will likely change my treatment recommendations moving forward.