Occasionally at meetings, oncologists are confronted with a marketing study done by the pharmaceutical industry that reveals that something like half of patients diagnosed with lung cancer never receive any treatment. In fact, epidemiologic studies using the massive Surveillance, Epidemiology, and End-Results (SEER) database suggest that only a minority of patients receive chemotherapy for advanced lung cancer (see here and here for examples). Most oncologists find it hard to believe that so many people fail to receive a treatment that has a consistently demonstrated survival benefit. Yet these results are out there.

Among the potential explanations for a gulf between what we perceive and what the data tell us are that at least the SEER database is linked to Medicare data and therefore represents older patients who may be less likely to receive treatment than younger patients, and also that there is a lag of many years between the data included in SEER and its subsequent reporting, so that the data may be obsolete compared to more current practice.

In the current issue of the Journal of Clinical Oncology, a group of researchers at the University of Texas-Southwestern in Dallas reported their more contemporary data on what proportion of patients of any age received chemotherapy. And while better than the SEER data would suggest, to me they still highlight how many patients with advanced NSCLC go untreated, for one reason or another.

The authors reviewed the experience of a total of 718 patients diagnosed with stage IV NSCLC at UT-Southwestern over the 8 year period from the beginning of 2000 to the end of 2007, and they had insurance information, a relevant variable, on 97% of them. In contrast with the broader experience in the US, where the median age of a new diagnosis of lung cancer is 71, their patient population had a younger median age of 60. Their stage IV population was 58% male and 45% Caucasian.

The key finding was that 49% of the patients received chemotherapy at their academic center in the modern era, and over the interval from 2000 to 2007, there wasn't a clear trend of greater probability of receiving chemotherapy later in the time period. Not surprisingly, younger patinets (<65) were more likely to receive chemotherapy than older ones (53% vs. 41%; p = 0.003), and patients with insurance were more likely to receive chemo than those without it, varying from a high of 60% for those with private insurance to 40% for patients with no insurance. Though the financial aspects presumably are a factor, this is confounded by the fact that uninsured patients also have more extensive medical issues.

Not surprisingly, survival was markedly superior for the patients who received chemotherapy:

(click on image to enlarge)

There are likely to be several important factors operating here. The benefit conferred by chemotherapy for advanced NSCLC is well established, but the patients not given chemotherapy likely had a poorer performance status and more medical problems, which also would have contributed to their worse survival overall.

Though a message from the paper itself is that the approximately 50% chemo administration rate exceeds that seen in the SEER-based studies, I was still surprised to see that it was that low. The authors also point out that rates of chemo administration are higher in other cancers, and that systemic therapy rates are in the 80% range for breast and prostate cancer, where less toxic hormone therapy is often used before or instead of chemotherapy.

At this point, it's hard to presume that these results are an aberration and that in truth most patients must certainly receive chemotherapy but are somehow missed or that UT-Southwestern must be unusually unwilling to offer chemotherapy to patients with lung cancer. Though the authors include a caveat that these results are from a single institution, it's a large academic center in a major city, so I strongly suspect that the numbers would be lower if we looked at a VA hospital or community practice in a mid-size town. (And perhaps on the up side, we can rid ourselves of the notion that oncologists are invariably hell-bent on giving unhelpful chemo to everyone they encounter, whether it's beneficial or not.)

Rather, I believe that these results are real but that many lung cancer patients are mentally discounted and become invisible to use when we mentally track our clinic populations and estimate that the vast majority are getting chemo. Certainly, some patients are never referred to an oncologist, either because the doctor doing the workup sees no value in it or because the patient wants nothing to do with treatment. And while some patients may lose a potential benefit from not receiving chemotherapy, I suspect that many of these patients do, in fact, have a marginal performance status and significant competing medical problems.

But I also suspect that oncologists, myself included, don't quite appreciate how many patients they see and follow for whom they don't recommend chemotherapy because they are too ill for it or because a patient refuses it (I hear a fair amount of, "my sister got chemo and was miserable -- I don't want any of it"). We sometimes never see these patients after an initial consultation, so between our attention being occupied more by the patients receiving ongoing treatment and the fact that we see these patients more often, we probably overestimate the proportion of our patients who receive active treatment.

Regardless of how and why, the evidence shows that a huge fraction of patients don't get chemotherapy, in large part because there are so many patients for whom our standard treatments don't apply bcause they aren't the same patients who were enrolled on the clinical trials that defined these standards. My next post will cover some interesting evidence on the gap beween the patients on our trials and the patients we see in real clinical practice.