Every few months I see a patient who reminds me of the fallability of mutation testing.

One of the brightest rays of hope in the management of lung cancer and much of cancer care in the last decade has been the transition toward treatment guided by the presence or absence of particular molecular markers like EGFR mutations, ALK or ROS1 rearrangements, and potentially others being studied now.  However, I'd like to sound a note of caution about focusing too much on the results of a test. Specifically, I have seen a few patients, including one in the last couple of days, who highlight how tragic it would be for us to limit the use of the oral EGFR tyrosine kinase inhibitor (TKI)  Tarceva (erlotinib) to patients who are tested and have an EGFR mutation.

To backtrack, over the last few years it has become clear that most of the people who are major beneficiaries of an EGFR TKI are those with an activating EGFR mutation. We've also seen several studies that show that in the minority of advanced NSCLC patients whose tumors have an activating EGFR mutation, an EGFR TKI is extremely likely to be more effective than chemo as first line treatment.  Tarceva had already been approved by the FDA as a second or third line treatment for a broad range of molecularly unselected patients (no EGFR mutation requirement) based on a randomized trial that showed a 2 month survival benefit overall, so the question was whether when you take out the profound benefit seen in a minority, there's still enough benefit in the "EGFR wild type" patients (no activating mutation) to have it be worth treating such patients with Tarceva.  

Though there are many who minimize the value of Tarceva in patients with EGFR wild type, the evidence indicates that such patients still receive an approximately 1.5-2 month improvement in median overall survival with it, so I routinely recommend Tarceva as a second or third line therapy after initial chemo in patients who test negative for an EGFR mutation. Most of the benefit is in the form of prolonged stable disease or at most a minor response, which can still translate to living a few months longer. In the back of my mind, as well, is the possibility that a small minority of such patients will respond extremely well, reminiscent of what we'd expect to see in someone with an EGFR mutation -- because we know that the response rate to Tarceva in tested, EGFR wild type patients is low, but it's not zero. 

This week I saw a 77 year-old Burmese never-smoking woman in follow up who illustrates the danger of putting our complete faith in molecular oncology.  When I first met her 10 months ago, I tested her metastatic adenocarcinoma for a driver mutation and found she had none. She had declined significantly before her diagnosis and treatment, had become weak and lost a lot of weight, but I was able to start her on carboplatin/Alimta (pemetrexed) and see how she would do. Each cycle for the first few months she improved steadily, gaining weight and energy, and her scans demonstrated a good partial response. After 4 cycles of the doublet, we switched to single agent Alimta, which she continued on for another 3 months, by which time her scans had worsened enough that I favored changing treatments. Specifically, because she had tested negative for an EGFR mutation and had responded well to chemo, I favored either Taxotere (docetaxel) or a clinical trial with Taxotere vs. an investigational agent.

At that time, she said that rather than switch treatments, she felt well enough that she wanted to take a break from all treatment. I followed her closely, and though her scans showed more progression six weeks later, she still wanted to wait on treatment. Unfortunately, things came apart just a few weeks later, when she suddenly lost her appetite and energy again, really experiencing a sudden tailspin. Repeat scans now showed extensive disease in her liver and throughout her abdomen, where fluid was now collecting (called ascites) and she was now having increasing pain. 

But by the point that she was convinced to pursue chemo, in November, I was no longer confident that she could tolerate it safely. I considered the limited options, perhaps very low dose chemo, perhaps supportive care only, but I thought that Tarceva was potentially tolerable, could confer a realistic chance of a modest survival benefit, and would offer a small chance of a more dramatic response. 

In fact, she improved within days, just as you'd hope and have a significant probability of seeing in someone with an activating mutation. Her appetite, energy, and pain improved quickly and dramatically, and I suspected by my first follow-up visit with her that her scans would corroborate her symptomatic improvement. In fact, that's what we saw this week.

It's disturbing for me to think that many oncologists, including some lung cancer specialists, would discount the potential value of Tarceva in patients with EGFR wild type so much that they would say it's not worth trying. We don't know where the disconnect is. What are some of the possible explanations:

1) Perhaps the lab made a mistake and she's actually got an activating mutation, possibly a rare one that the specific lab assay didn't detect.

2) Perhaps she has a mutation in some/most of her tumor, but the specific tissue sampled at the biopsy didn't happen to have it. We know that there can be molecular heterogeneity, meaning that the molecular profile of one area/nodule/metastatic lesion of cancer is different from that of another.

3) Perhaps her cancer acquired the mutation since the initial diagnosis.

4) Perhaps a small percentage of people who truly don't have an EGFR mutation just happen to have a great response to EGFR TKIs.

I can't say which of these is the right answer, though I'll be working with my pathologist to clarify whether her initial biopsy is truly negative or just missed an EGFR mutation.  But I can say that she's not the first patient I've seen who follows this pattern, and while I don't know how long her response will last, the others who followed this pattern have shown a duration of response in the range of years, just as you'd hope to see from a patient with an activating EGFR mutation.  I'd say I see at least 1-2 per year who fit in this category, despite working with a very reputable lab that has plenty of experience with mutation testing.  Most have little or no smoking history and an adenocarcinoma, but there is just enough variability among them to lead me to avoid any strong rules.

Given this possibility, I always want to give my patients with advanced NSCLC an opportunity with Tarceva. I definitely don't favor it over chemo as first line treatment in someone who tests as EGFR wild type, in whom chemo has been shown to be a significantly superior option here. But since it's approved for a broad population, regardless of the presence or absence of an EGFR mutation, and since there's a real, albeit small chance that the person will have a wonderful and long-lasting response despite the molecular testing predicting they shouldn't, Tarceva is a treatment I want to always offer to my patients at some point. In this case, it's great to be wrong about what is supposed to happen. 

As always, I welcome your comments and questions.