The FLEX trial, a European study of cisplatin/Navelbine (vinorelbine) with or without the monoclonal antibody against EGFR Erbitux (cetuximab), was a technically positive study that was initially reported at ASCO 2008. However, showing an improvement in median survival of just 1.2 months, most oncologists came away feeling that the trial illustrated the difference between a statistically vs. clinically significant result. Currently, the use of Erbitux in clinical practice is just a very small percentage of potentially eligible patients, though I think this would change if we could identify patients who are significantly more likely to benefit substantially and spare the others who actually don't benefit from the cost, weekly infusions, and side effects of it. A recent report suggests that there may be such a method.

One of the most prominent presentations from the World Conference on Lung Cancer last month was a retrospective analysis of the results from the FLEX trial that divided patients based on the extent to which their tumor expressed EGFR protein on cancer cells as measured by the technique of immunohistochemistry (IHC). The investigators developed an "EGFR IHC score" that was a product of the percentage of cancer cells positive for EGFR protein on the cell surface multiplied by the overall intensity of staining (ranging from 0 to 3+), producing a number from 0 to 300.

(click on image to enlarge)

The investigators then defined two groups, one of low EGFR IHC score of less than 200, with 69% of the patients on the study, and another with high EGFR IHC scores of 200 or higher, who comprised 31% of their patient population. Interestingly, there weren't any significant differences in clinical variables of the low vs. high EGFR IHC groups except that those with a squamous NSCLC tumor were a little more likely to have high EGFR IHC scores (approximately 50% adenocarcinoma, 30% squamous (and 20% "other") in the low EGFR IHC group, vs. approximately 40% adenocarcinoma and 40% squamous in the high EGFR IHC group).

With this breakdown, it appeared that the benefit was confined to the patients in the high EGFR IHC subgroup, with their overall survival (OS) curves shown on the right, while the patients with low EGFR IHC showed no difference in OS:

For those with a high EGFR IHC score, this translated to a 27% improvement in survival, with a difference in median OS of 2.4 months (12.0 vs. 9.6 months), one year survival 50% vs. 37%, and two year survival of 24% vs. 15%.

Importantly, the investigators also looked at outcomes in the high EGFR IHC score group for patients with both adenocarcinoma and squamous cell NSCLC, with the results showing that both subtypes showed a similarly favorable trend of benefit. Though patients with adenocarcinoma tended to do better overall, the squamous cell patients demonstrated what appeared to be a little more of a relative benefit with cetuximab compared with chemotherapy alone.

So what can we say overall? Though these results are provocative, they don't lead the entire lung cancer community to wholesale adoption of Erbitux in advanced NSCLC, even for the subset of patients with a high EGFR IHC score. The fact that it took over three years from initial presentation of the FLEX trial to a retrospective analysis that produced a subset of patients who appeared to do very well leads to suspicion that the investigators might have done many, many different analyses, potentially with different scoring systems and different cutoff points, before finding one that worked to separate the beneficiaries from those who appeared to not benefit from addition of Erbitux. As the saying goes, "if you torture the data enough, it will tell you anything you want it to": doing many analyses until you get the result you want to see isn't the pinnacle of good science. Ideally, we'll see validation of EGFR IHC score as a predictive marker of benefit with Erbitux.

In the meantime, it's fair to recognize that the overall trial was positive and that the National Comprehensive Cancer Network (NCCN) includes mention of Erbitux as an appropriate consideration for first line treatment of advanced NSCLC. How it fits in for patients who are appropriate candidates for Avastin (bevacizumab) is a particularly open question -- we can't just assume that adding it to Avastin-containing regimens will provide additive benefits. But knowing these results will make me far more inclined to add Erbitux for patients who aren't receiving Avastin and who have a high EGFR IHC score, as these results now substantially raise the value proposition for patients.

And this leaves us with one other key question: will insurers pay for it??