This past week, I saw a new patient who had just moved from another part of the country and needed long-term management of her high risk lung cancer. A never-smoking Asian woman, she was found to have a stage IIIA lung cancer with “N2″ mediastinal lymph nodes involving cancer in her mid-chest. As is typically done, she received chemotherapy with a cisplatin/Alimta (pemetrexed), a very strong treatment option for her lung adenocarcinoma, and she then proceeded to surgery. There, she was found to have a small amount of residual cancer in her mediastinal lymph nodes, which suggests a high risk of recurrence. Her oncologist recommended post-operative radiation to treat this area.
Though radiation to the mid-chest is commonly and I think appropriately favored for patients with residual N2 nodal disease after surgery, especially if they hadn’t received it before surgery, this woman declined the radiation. She was far more receptive to more systemic therapy, especially in the form of targeted therapy once it was discovered that her tumor has a ROS1 rearrangement. This is rare (only about 1-1.5% of lung cancers, disproportionately in younger never-smokers with an adenocarcinoma) and has been found to very often respond well to XALKORI (crizotinib), the same agent used for the 4-5% of patients in the US (more in Asia) who have an ALK rearrangement, for whom XALKORI is approved. XALKORI isn’t technically FDA approved for people with a ROS-1 rearrangement, but it’s often covered by insurers when they are made to understand the profound value of this agent for these rare patients.
But when we talk about XALKORI for patients with an ALK or ROS1 rearrangement, or an EGFR tyrosine kinase inhibitor (TKI) like Tarceva (erlotinib), Iressa (gefitinib), or Gilotrif (afatinib) for those with an EGFR mutation, we’re focusing on the overwhelmingly favorable data obtained in patients with metastatic cancer. It’s an entirely different situation with curable and potentially cured lung cancer. When someone has undergone surgery or chemo/radiation or some other combination for stage I-III NSCLC, we know that some of these people are already going to be cured without adding any other treatment. When we give targeted therapies for lung cancer who may already be cured, we often presume we can only be helping them. However, that wasn’t true when adding Iressa after chemo/radiation for patients with stage III NSCLC — in fact, Iressa was significantly harmful and shortened survival. We decided, with the benefit of hindsight, that this was because the patients on that trial were not molecularly selected as having an EGFR mutation; they didn’t check for or require EGFR mutations, so we presume that 90% didn’t have an EGFR mutation, and we know such patients typically benefit modestly at best from EGFR TKIs. However, when we saw the results of a trial of post-operative (adjuvant) Iressa in early stage NSCLC patients, we saw that not only was there a trend toward worse outcomes for the recipients of Iressa in the overall, molecularly unselected population, but the trend of worse outcomes was especially pronounced in patients with an EGFR mutation. If there was ever an observation that was humbling in highlighting how wrong we could be in our presumptions, it was that one.
That hasn’t clearly dampened the enthusiasm of some believers, who note that there are promising early results from the SELECT trial, a single arm study of patients with an EGFR mutation who underwent surgery for early stage lung cancer and then were placed on Tarceva as adjuvant therapy. It’s true that they demonstrate an impressive recurrence-free survival, at least for a couple of years after surgery, with results arguably better than you’d expect to see without this treatment, but there isn’t any evidence that patients actually live longer, and a suggestion that patients may be at high risk to relapse as soon as you stop the therapy.
Here’s my fear: We have no evidence that you cure anyone who wasn’t already cured by adding a limited duration of a targeted therapy, even in people with no evidence of disease. The idea of adjuvant chemotherapy is that you can give 3-4 cycles of chemotherapy to kill a few stray microscopic cancer cells that may still be present after surgery, but after that time-limited treatment, you’re done. If adjuvant targeted therapy suppresses but doesn’t kill enough cancer cells, even after surgery, that you don’t need it indefinitely (an average of 5 months on the Canadian BR19 trial led to an apparent rebound progression after it was discontinued, so 2 months longer than the maximal duration of adjuvant chemotherapy didn’t help in any way), adjuvant targeted therapy is only postponing recurrence that was going to happen anyway, for as long as it takes before acquired resistance to develop. Then you’ve exhausted your most effective therapy by giving it perhaps 2-3 years before there might be visible evidence of cancer to treat, while a patient had no symptoms, but you’ve given them side effects of treatments and incurred the costs of years of treatment. And if half of the people with a stage I cancer positive for an EGFR mutation, for instance, would have never recurred because they’d have been cured by the surgery, so they’re just taking Tarceva for years or indefinitely and experiencing the known and unknown side effects of years of treatment that isn’t helping them at all.
In other words, adjuvant treatment with targeted therapy seems to be very good at doing exactly what we expect it should be able to do: postpone evidence of recurrence in the short term. We’d expect this because many patients are already cured, and those who would have progressed at 15 months, for instance, might have their recurrence postponed until 2.5 or 3 years out from surgery, when their cancer develops acquired resistance — and then they’ve exhausted their best treatment at a time when they had no evidence of visible cancer, experiencing side effects during that time. They may end up with a shorter survival than the patients who just had the cancer recur at 15 months, then go on Tarceva at recurrence and respond very well for the same time they would have had by taking it earlier. But if we only give it to the patients with recurrence, we avoid giving people who are already going to be cured the targeted therapy for years or indefinitely, with all of the costs and side effects that come with it.
As I see it, the counterargument is that perhaps acquired resistance can be slowed by giving targeted therapy when there’s minimal cancer that can develop resistance. However, we’ve never seen evidence that time to development of acquired resistance is predicted by or correlated with the “tumor burden”, the volume of cancer in the body, when a person starts targeted therapy.
This is a question that begs for a proper, randomized trial to clarify whether adjuvant targeted therapy helps, harms, or has no effect on overall survival. It will clearly increase costs, and expose some patients to completely unneeded treatments. If it helps cure patients, it’s worth it, but we can’t presume that’s going to be the case.
What do you think? Would everyone here take a targeted therapy indefinitely after surgery if they were found to have an EGFR mutation, ALK rearrangement, or ROS1 rearrangement, and had a 50% chance of being cured without any further treatment? What if, because it hasn’t been proven to be helpful, it wasn’t covered or had a significant co-pay and required hundreds or thousands of dollars out of your pocket every month?