An acquaintance recently sent me a copy of an article from the December 7, 2015 edition of The New Yorker magazine, describing efforts of a neurosurgeon to use an unconventional approach to treating terminal brain cancer. Follow the link and read the article for the whole story, but the physician in question was acting on anecdotes of people whose brain tumors had improved dramatically after serious post-operative wound infections. (Surgery is the front-line, but generally inadequate, treatment for these tumors.) The physician purposely infected the wound of one of his patients with Enterobacter aeroenes, a bacterium normally found in the bowel and in human feces.
The first patient had a dramatic improvement and lived for a few years after—a highly unusual, positive result. The second one did not do so well. There was a small number of people treated in this way, and results were mixed at best.
From the article, it seems clear that this was a good doctor trying to help his patients—one who had done a practice-changing clinical trial and who understood that some treatments may “work”—that is, have a biologic effect as expected—but not help the patient—that is, the effect does not translate into any actual clinical benefit against the disease. This is a common problem in the development of new treatments, especially for cancer, where clinical trials often do no better than showing “a correlation with a correlation with something you care about,” as a former senior official of the National Cancer Institute put it to me (not about this neurosurgeon’s approach, but about something else) some years ago.
“Innovative therapy” is not the same as “clinical research” or “human subjects research.” The former is performed by a licensed physician using his clinical judgment to try to apply available means for the best care of his patient. The latter tries to collect generalizable knowledge and, one hopes, help the study subjects in the process. Often in clinical trials, especially the earliest trials of the newest approaches to therapy, the latter (“direct benefit” to subjects) is not attainable, but the former (the “indirect benefit” that may translate into future direct benefit for other people, or for society at large) is. In early cancer trials, overoptimism, implicit or explicit, is often a concern, leading to something called a “therapeutic misconception.”
In this case, the doctor took pains NOT to create a therapeutic misconception. He didn’t oversell the idea to his patients. With the first case, the local IRB—the ethics board—did not think its oversight was needed, calling the approach “innovative therapy.” As the number of patients receiving the attempted treatment increased, the IRB started to think it should get more involved.
To make the story short, enough was done without formal IRB review or FDA regulatory review that it was decided, in retrospect, that several people should be disciplined for, in essence, getting out over the end of their skis with the idea. The author of the article contrasts this case with the Duke University studies into using a modified poliovirus to treat brain cancers, something that was the subject of a widely-noted report on 60 Minutes last spring (subscription required to view online). That approach was studied in the laboratory for years before being tested in patients.
A big challenge with “alternative” or unorthodox treatment approaches for desperate diseases is that they are typically blind shots. Now, if you know you don’t have long to live, you might be willing to let a blindfolded marksman shoot a wild animal off your head, knowing he might hit your head instead. And, frankly, many new treatments are still “semi-blind” at least. But (and leaving aside for now the whole question of how much of what kind of regulation is best) there is something of an ethical imperative for doctors exploring new treatments to perform the most disciplined scientific research they reasonably can to try to understand what is going on with their new idea, how to improve their chances of success, and how to limit undue risk to their patients. It’s part of integrating medical research into good patient care, and it’s called “translational,” trying to translate a set of observations or laboratory results into a defined treatment with some reasonable estimates of the risks and benefits.