One of the general rules in oncologist is that we typically use our most effective treatments first, and often early, though there are certainly exceptions. Women with metastatic breast cancer may have a higher response rate by receiving combination chemotherapy than single agent chemo, but when a gentle single drug chemotherapy or hormone therapy option will do very well and provide fewer side effects, that’s usually the approach we recommend until bigger guns are needed. Similarly, patients with slow-growing cancers like chronic lymphocytic leukemia (CLL) often do well for many years and sometimes forever without needing treatment, so we often hold on initiating any treatment unless there’s a good reason to (and I think the same argument can be made for some lung cancers, most commonly indolent bronchoioloalveolar carcinoma, or BAC). So we don’t want to treat patients with less therapy than they need, but we’d also prefer not to treat them with more.

When you have multiple potential treatments to pursue, one question that becomes relevant is whether the sequnce matters. If you have choices A, B, and C, does it matter if you give AB followed by C, or C followed by AB or even A followed by B and then C? First, we need to take a step back and recognize that this is a good dilemma to have, because 10 years ago we didn't have multiple effective treatments to try to position. For advanced NSCLC, the value of giving chemo for metastatic disease was debated until about 10-12 years ago, when the survival benefit of platinum-based doublet chemo regimens became a clear standard. In 2000 or so, second line taxotere emerged as a second line treatment option that improved survival and was approved by the FDA in that setting. A few years later, additional treatment options for previously treated patients with advanced NSCLC were approved, ushering in agents like alimta and EGFR inhibitors (initially iressa and then tarceva, with iressa removed from the US marketplace. These all became viable options, and over the last few years all of these agents have also been studied as first line treatments. Taxotere and alimta are now both FDA approved as first line combination therapy with platinum, and the EGFR inhibitors have also been studied as first line treatment and have appeared particularly attractive in certain subsets of patients, such as those with an EGFR mutation or patients with advanced BAC.

So just getting to the point, does the order of treatment matter? Most of the time we ask this specifically with regard to the question of giving an EGFR inhibitor before or after chemotherapy. Chemotherapy is a more time tested and standard approach, but don’t see 70% response rates or durations of treatment responses lasting a year or more with standard chemo: we only see these results by giving a targeted therapy like an oral EGFR inhibitor (tarceva, iressa), to patients whose cancers are heavily driven by that target. You might make the argument that the patients who are destined to be the great responders to an EGFR inhibitor should get that immediately, since it’s possible that they could benefit less if they decline significantly before getting that treatment. On the other hard, the patients not destined to do well with an EGFR inhibitor, who would be better served by chemo, could decline significantly before starting chemo if we were to give an EGFR inhibitor first. In fact, a trial of patients with marginal performance status, who were not selected for having EGFR mutations or other features predictive of a good result with EGFR inhibitor therapy, did significantly better if they started with chemotherapy than if they started an EGFR inhibitor, presumably because many of these patients “missed their chance” by declining too much to benefit from subsequent chemotherapy.

Last year, I was involved in a debate at a national lung cancer meeting in which I was arguing against routine testing for EGFR mutation testing. Among the arguments against it was my review of some limited data that it didn’t seem to matter whether patients received an EGFR inhibitor first line or second line or later, so if you were planning to give tarceva as a second or third line therapy, there wasn’t a clear incentive to rush it into your first line strategy. I noted that in the early work in iressa, the evidence showed that the response rate was every bit as good for patients who had received many lines of prior therapy as for the patients who had received fewer lines of prior therapy:

(Click to enlarge)

I also reviewed the data from a couple of trials in advanced BAC, a trial of iressa for BAC that I led (SWOG 0126), and another led by investigators at Memorial Sloan Kettering that administered tarceva. Both studies showed no real suggestion of meaningful differences in outcome whether a patient received a tyrosine kinase inhibitor as first line therapy or as a later intervention:

Finally, I noted that even in some work with patients who were identified as having an EGFR mutation, the survival benefit of treatment with tarceva vs. chemo in a Spanish population of patients was every bit as good whether these mutation positive patients received tarceva first line or later.

Even recently, I discussed the results of the very provocative IPASS trial that compared iressa to carbo/taxol as first line treatment among Asian never-smokers or light former smokers. The results of this trial demonstrated that the subset patients with a tumor harboring an EGFR mutation had a remarkably better progression-free survival with first line iressa, while the opposite was true for the patients who did not have an EGFR mutation.

One additional thing that is striking is the very rapid drop on the PFS curve for recipients of iressa who don’t have an EGFR mutation (highlighted by the red oval in the figure on the right, above). If this leads to a clinical decline that precludes some of these patients from getting chemo that would have been beneficial for them, the order of treatment definitely matters.

The IPASS trial also reported preliminary overall survival results, which weren’t different for the trial population as a whole, but when looked at broken down by EGFR mutation status are worthy of further discussion:

Though the differences don’t reach statistical significance, the same trend is present as was seen with progression-free survival: patients with the mutation show a strong trend toward living longer if they received an EGFR inhibitor to begin with, while the patients without an EGFR mutation show a strong trend toward a superior survival with initial chemotherapy. And these results represent just the third of patients who had tissue available, so we might have seen statistically significant differences if more patients had been included in the mutation analysis.

My prior arguments against the importance of treatment sequence are still there, but I consider the IPASS mutation breakdown to represent the best test of whether sequence of treatment matters. I think that these findings are at least suggestive if not conclusive that it may be important to choose the best treatment as possible as an early treatment for a patient. What we want to avoid is a swing and a miss that allows a patient to decline before they get the opportunity with the best drug or regimen for them. It's also an argument for more molecular variable- and histology-based treatment, because there's likely to be a real cost to flying blind, which is essentially what we've done up until now.

Postscript

For those of you who have followed my writings on these subjects, I'll acknowledge that this is a revision of my prior view. Though I don’t like to be inconsistent, I do think it’s important to change your perspective when new information becomes available. For me, having the chance to see and study the survival curves of the IPASS trial has led me to at least step back and strongly consider that for patients with a significant possibility of carrying an EGFR mutation (patients with an adenocarcinoma or BAC, never or remote prior smokers), the optimal approach could be to try to clarify as well as possible whether EGFR inhibitor therapy is likely to be a good early option. The limited availability of tissue and the delay of several weeks before obtaining mutation results are real barriers, which leads me to consider the appeal of a blood-based test for favorable outcomes with an EGFR inhibitor, but I would say that I’m becoming a convert on the value of molecular testing for at least a subset of patients. I think the sequence of treatment may truly matter now.