Though I didn't make it to the American Association for Cancer Research (AACR) meeting in mid-April, I did catch discussion that followed presentation of some preliminary data from the Biomarker-integrated Approaches of Targeted Therapy for Lung Cancer Elimination (BATTLE) trial, conducted at MD Anderson Cancer Center over the last several years. My friend and colleague, Dr. Ed Kim, presented the provocative early results. The study enrolled 255 patients with previously treated advanced NSCLC who had previously received a median of two prior lines of therapy. They had fresh biopsies done there, which were analyzed for 11 biomarkers (EGFR, KRAS, and BRAF mutations, EGFR and Cyclin D1 copy number by fluorescence in situ hybridization (FISH); VEGF, VEGFR, 3 RXR receptors, and Cyclin D1 by immunohistochemistry (IHC)). The patients were then randomized to four treatment groups: Tarceva (erlotinib), Zactima (vandetanib), Tarceva plus Targretin (bexarotene)(from a class of drugs called rexanoids, and not successful in lung cancer thus far), and Nexavar (sorafenib). Initially, the randomization was equal across the four treatments, but the trial employed a novel statistical technique called adaptive randomization, in which arms on which patients were doing better became more heavily weighted for subsequent randomization, at the expense of the arms on which patients were not doing as well.

The primary endpoint of the trial was disease control (stable disease or tumor shrinkage) after eight weeks, which is a rather low bar as endpoints go, but probably reasonable for heavily pretreated patients. The disease control rate for the trial overall was 46%, which was claimed to be clearly bettter than historical controls, but I would be skeptical about comparing MD Anderson patients, who came from all over and have surpassed many barriers to get there, to a wider experience in the real world. The one year survival overall was 38%, with a median overall survival of 9 months.

There was some fanfare around the concept that patients who received treatments particularly suited to their molecular defects did well. Not surprisingly, 100% of patients with an EGFR mutation and FISH amplification for EGFR had disease control. That looks favorable when compared with the 0% disease control rate patients with similar molecular characteristics showed to Nexavar, but I don't think this tells us anything new about the value of EGFR inhibitors for patients with an EGFR mutation.

One quite encouraging, albeit preliminary, insight was that the patients with KRAS mutations seemed to do well with Nexavar (disease control rate 61%), higher than other groups and clearly more than the 22% disease control rate seen with Tarceva. Thus far, we haven't seen especially good signals of benefit with Nexavar in a broader population with NSCLC, but perhaps it will emerge as a valuable treatment if we hone the focus on the 20% of NSCLC patients with a KRAS mutation, a group for whom we haven't been able to identify any particularly effective interventions. And while this is just a first step, I am doubly encouraged when viewing these results alongside the just published brief report of favorable results among 10 patients with a KRAS mutation who received Nexavar.

We'll get more details soon, but in the meantime, I think it's worth taking a step back to see what we can glean from the BATTLE trial. One key element was proving feasibility of doing mandated tissue collection in a large cohort of patients, and this was feasible (at least at MD Anderson) and not associated with prohibitive risk (11% experiencing a pneumothorax (lung collapse, to some degree), and no life-threatening or fatal complications). However, there was some discussion of this approach possibly representing a "new standard of care", and I think we need to be very cautious about declaring anything a new standard of care until it's shown to be feasible and beneficial in a much broader setting than the largest cancer center in the entire world, which attracts a very unique patient population. There's no question that the people who find their way there have a different level of fitness, motivation, support, and other resources (on average) than a broader community oncology population. Importantly as well, this small study didn't really compare a biomarker driven strategy to that of giving patients established treatments.

On a cautionary note, we shouldn't forget that figitumumab, the IGF-1R inhibitor, had a >50% response rate with chemo in a phase II trial from MD Anderson, leading to the development of two phase III trials that not only were negative but were closed early, in the face of toxicity issues and possibly worse overall outcomes. It would be a mistake to be led down the primrose path and make too much out of any single institution experience.

Still, the BATTLE trial represents an achievement in moving forward the concept of molecular oncology, and research that collects tissue from the majority or full cohort of patients is the way we're going to move forward. We're gaining ground at a faster pace than ever before, and much of that is from identifying subgroups who benefit from treatments that we might overlook based on far less impressive results in an unselected population.