In a recent issue of the Journal of Clinical Oncology, Dr. Vince Miller and colleagues published the results of an important trial of the EGFR inhibitor tarceva (erlotinib) in the unusual NSCLC subtype bronchioloalveolar carcinoma, or BAC (abstract here). This work was predicated on the observation, also by Dr. Miller and his colleagues at Memorial Sloan Kettering Cancer Center in NYC, that EGFR inhibitors, and specifically gefitinib in the earliest clinical experience, appeared to be most associated with great responses in patients with adenocarcinoma and particularly BAC, and also in never-smokers (abstract here). Over the past several years, much of our greatest interest in drugs like iressa and tarceva has focused on BAC, including a trial with iressa that I led with the Southwest Oncology Group (SWOG) and subsequently published (abstract here), and also the trial with tarceva that Dr. Miller ran with collaborators at Vanderbilt-Ingram Cancer Center in Nashville and MD Anderson Cancer Center in Houston. Over several years, these investigators enrolled 101 patients, most of whom had adeno/BAC and not the pure form of BAC (see prior post for discussion of this spectrum), which is representative of the eligibility of most or all of our current trials that we say are for "BAC": the majority, and sometimes the vast majority, are BAC mixed with some invasive adenocarcinoma. In truth, different pathologists, even great ones, differ in their definitions of what is BAC, what is adenocarcinoma with BAC features, and what is BAC with focal invasive adenocarcinoma.

Regardless, these 101 patients received single agent tarceva and continued on it until the time of progression or very problematic side effects. Three quarters received tarceva as their first treatment, while the remaining quarter had received prior chemo. The investigators reported a response rate (RR) of 22%, median progression-free survival (PFS) of 4 months, and a median overall survival (OS) of 17 months. These numbers are comparable and, in fact, a little better than what we had seen with iressa in BAC in the slightly larger SWOG study, which was done very similarly. I wasn't surprised to see slightly better results with tarceva, given that it overall appears to be a modestly better version of a drug remarkably similar to iressa, at least at the doses that were commonly used in clinical trials (iressa at 250-500 mg daily, tarceva at 150 mg daily). But the authors also attempted to determine which patients did particularly better or worse with tarceva, looking at both clinical variables like patient sex and prior chemo, as well as molecular variables like EGFR mutations, KRAS mutations, and other molecular markers that remain controversial but potentially important. Looking first at clinical variables, they found that tarceva produced very similar results between those patients who had not received prior chemo and those who did. There was also no particular difference in activity for those patients with a good performance status compared with those patients who were more marginal. The investigators also saw the same general response rates in patients with adeno/BAC compared with those who had "pure BAC", and when they looked at survival, the patients with adeno/BAC had a far better PFS (5 vs. 1 months) and OS (19 vs 4 months), although neither was quite statistically significant. Women, who outnumbered men nearly 2 to 1 (very unlike general lung cancer numbers, but typical for BAC trials), had a modestly higher response rate (27% vs. 15%), median PFS (5 vs. 3 months), and median OS (21 vs. 16 months); none of these sex-based differences was statistically significant. As we've seen in other trials, smoking status was associated with some clear differences. The response rate among patients who had never-smoked or had smoked less than 15 pack-years (the product of number of years smoking and average number of packs smoked per day) was 45%, compared with just 5% for those patients with a more significant smoking history. To me, this suggests that much of what we think of as an impressive response to EGFR inhibitors in BAC is seen in minimal or never-smokers. However, despite this very high response rate in never-smokers or light former smokers, median PFS and OS were quite similar -- so it seemed like many patients still had pretty indolent disease and did just as well whether they had a transient but impressive radiologic response or not. Both the minimal smokers and the significant smokers had exactly the same median survival, 17 months. The opposite trend was seen when looking at the outcomes of patients who did or did not receive prior surgery. You can imagine that most of the patients who had undergone prior surgery now had metastatic disease because of a relatively small burden of BAC that was now visible, recurrent, a few months or a few years after surgery. That's very different from patients who present with metastatic disease throughout their lungs and who were never candidates for surgery. In fact, the response rates are very similar for those patients who did or did not have prior surgery (26% vs. 22%), but the survival was far, far longer in patients who had undergone prior surgery (median OS 35 vs. 15 months). Here, it's not that surgery will give you a better survival, but the patients who had presented with localized cancer that made surgery a good initial option still had a minimal cancer burden after progressing on tarceva. And patients with a small amount of cancer do better than patients with a large amount of cancer in their lungs. The final factor they described was weight loss of 6% of prior baseline or more (not sure why they picked 6% and not 5% or 10%, but that's what they did), which was seen in 16% of their patients. The patients with weight loss had a strong trend toward a lower RR (7% vs. 26%), and a statistically significantly worse median PFS (2 vs. 5 months) and OS (7 vs. 21 months) than patients who did not have significant weight loss. On clinical trials in lung cancer, we almost always see that patients with significant weight loss do worse than counterparts who haven't lost much weight. So there are some interesting signals of which patients received greater or lesser benefit with tarceva, but it's interesting also to see the dissociation between response rate and survival: never-smokers had a strikingly better RR but no significant improvement in OS, while the opposite was seen in the patients who had received prior surgery. This trial also included tissue collection to do molecular analyses of EGFR and KRAS to see which molecular variables were associated with benefit or no benefit on tarceva for BAC. I'll cover that fascinating analysis next.