The marker known as an anaplastic lymphoma kinase (ALK) translocation has been all over the lung cancer news in recent weeks, most notably in the setting of being the marker in about 4% of patients with non-small cell lung cancer (NSCLC) that is correlated with a high probability of response to the ALK inhibitor crizotinib, which was just approved by the FDA for patients whose tumors demonstrate this marker on a test in a central reference laboratory. But this marker may also be correlated with responsiveness to Alimta (pemetrexed), providing a readily available treatment option before or after these patients start crizotinib.

If this sounds familiar, it's because I actually described such an association of particular benefit in ALK-positive NSCLC with Alimta back in February, when our friend, Dr. Ross Camidge, from the University of Colorado, published a retrospective review of patient outcomes with Alimta at the University of Colorado. They evaluated the outcomes of 89 patients tested for an EGFR mutation, KRAS mutation, or ALK translocation and also found that the median progression-free survival of their 19 ALK positive patients was significantly higher, at 9.0 months, than that seen in 37 who were wild type (WT, or "no mutation) for all three markers (4.0 months), and that neither patients with EGFR nor KRAS mutations demonstrated a significant difference compared with triple negative patients.

Though Dr. Camidge and I both noted that these results can only be taken so far, as a retrospective analysis from a single institution, these conclusions were corroborated by a remarkably similar retrospective review that comes out of Korea by Lee and colleagues, who evaluated 95 patients with advanced NSCLC and who had received Alimta as a second line or later therapy between March, 2007 and April, 2010. Specifically, they assessed outcomes as a function of whether these patients had an EGFR mutation (N = 43, or 45%), an ALK translocation (N = 15, or 16%), or WT for both (N = 37, or 39%), finding that patients with an ALK translocation had a more favorable outcome with Alimta across several endpoints. These included a significantly higher objective response rate compared to those with an EGFR mutation or WT for both (47% vs. 5% and 16%, respectively; p = 0.001), as well as disease control rate (87% vs. 26% and 57%, respectively; p < 0.001) and median time to progression (TTP) (9.2 vs. 1.5 and 2.8 months, respectively; p = 0.001).

(click on image to enlarge)

ALK positivity was a significant predictor of response (hazard ratio (HR) = 0.07, p = 0.001), as well as TTP (HR = 0.44, p = 0.007) and remained so regardless of whether Alimta was administered as a second line or later therapy. As a potential explanation for this effect, the authors reported lower levels of thymidylate synthase mRNA, a possible but still not prospectively validated correlate to sensitivity to Alimta, in ALK positive tumor cells compared with control cells.

Taken together, these results don't provide the definitive weight of a prospective trial, but we need to recognize that the number of prospective trials that can be done on a subgroup that comprises 4-5% of the entire NSCLC population is very limited and need to be prioritized. Still, they demonstrate to me the impact of well conducted retrospective studies that, I would contend, converge to make compelling point that leads me to favor treatment with Alimta in ALK-positive patients, before or after crizotinib. Beyond this single helpful observation, it suggests that the move toward more molecularly-defined oncology will continue to help us better predict which specific subsets of patients will be best served by one treatment over another, including some of our already established treatments.

The promise of moving well beyond mediocre results in lung cancer will be achieved by our recognizing the diversity of many small subgroups, many still to be recognized. But as we focus on smaller and smaller subgroups, from the 10% with an EGFR mutation to the 4% ALK translocation to the 2-3% with a BRAF mutation, we will need to adapt by learning new ways to ask clinical questions in order to glean the most information from such limited patient populations. I think it will be an important development to see more of these types of trials, as we to recognize that large prospective randomized clinical trials are far less feasible in a new model of molecular oncology that divides NSCLC into small subsets of the larger whole.