Is there gold in that thar genomic medicine?

I've written a lot about a doctor named Stanislaw Burzynski who claims to have much better outcomes in treating deadly brainstem tumors than conventional oncology does. Although the way he claims to do it is through the use of substances he calls "antineoplastons," which he claimed to have isolate from the urine of patients. Over 35 years after having formed his own clinic and "research institute" to use these compounds to treat cancer and after having had over 60 phase I and phase II clinical trials registered with ClinicalTrials.gov, with none of these completed trials having been published, Burzynski continues on, charging large sums of money in "case management fees" and in general acting as though the same rules that apply to universities, pharmaceutical companies, and biotech companies don't apply to him. Meanwhile, Burzynski's propagandist Eric Merola makes yet another conspiracy mongering paean to his hero while melting down on Twitter whenever anyone criticizes him or his hero.

One of the interesting things (to me, at least) about the whole Burzynski saga is not so much his antineoplaston therapy. Basically, Burzynski needs to put up or shut up (not to mention stop treating patients with them) and, whatever happens, admit that antineoplastons are chemotherapy with potential side effects and adverse reactions as bad as some chemotherapies. No, what frequently irks me even more is how Stanislaw Burzynski tries to coopt the concept of "personalized medicine," except that he calls it "personalized, gene-targeted cancer therapy." I call it "personalized, gene-targeted therapy for dummies." Even more irritating, Burzynski is arrogant enough to claim that he is, in essence, the originator of the concept of personalized, gene-targeted medicine, and that big cancer centers like M.D. Anderson, Memorial Sloan-Kettering, and the like are only now following the trail that he blazed.

As much as I do try to get away from Burzynski from time to time, I couldn't help but think of that smirking, arrogant face pontificating about his genius and bragging about his incompetently administered personalized medicine as I read yesterday's New York Times article about what real cancer centers are doing to try to develop personalized cancer therapy. Contrast this with Burzynski's rinky-dink operation:

Electric fans growl like airplanes taking off and banks of green lights wink in a basement at Mount Sinai’s medical school, where a new $3 million supercomputer makes quick work of huge amounts of genetic and other biological information.

Just a couple of miles away, a competitor, Weill Cornell Medical College and NewYork-Presbyterian Hospital/Weill Cornell hospital are building a $650 million research tower. Across the street is a newly completed $550 million tower housing labs for another competitor, Memorial Sloan-Kettering Cancer Center.

Major academic medical centers in New York and around the country are spending and recruiting heavily in what has become an arms race within the war on cancer. The investments are based on the belief that the medical establishment is moving toward the routine sequencing of every patient’s genome in the quest for “precision medicine,” a course for prevention and treatment based on the special, even unique characteristics of the patient’s genes.

Contrast this sort of expenditure and scientific effort to what Stanislaw Burzynski does, which is to use a commercially available test that looks at various immunohistochemical markers and cDNA microarray processing. True, more recently the company offers next-generation sequencing, but does anyone think that Burzynski has the necessary skill to analyze the results. He thinks he does, but an effort to really do personalized medicine requires vast expenditures that Burzynski hasn't done, has shown no interest in making, and probably can't do. Worse, the issues in bringing personalized medicine to fruition so that it can be used in every patient and so that it results in improved outcomes are highly complex. They are issues that this article actually discusses surprisingly well for an article designed for the lay press.

Here's the idea, and it's an unproven idea: That sequencing every cancer patient's genome will lead to our identifying better treatments that lead to better outcomes, even cures. Another part of the idea is that sequencing patients' genomes will allow doctors to predict who will and will not ultimately develop cancer and, if so, which one. Given that it now costs between $5,000 and $10,000 to sequence a complete genome, leaving out the costs of analyzing and interpreting the information. In the search for cures, we're creating veritable torrents of data, petabytes upon petabytes of sequence and expression data from various next generation sequencing techniques. We still don't know what to do with this information, nor do we know yet how to interpret the information in a way that guides therapy:

Even optimists warn that medicine is a long way from deriving useful information from routine sequencing, raising questions about the social worth of all this investment at a time of intense fiscal pressure on the health care system.

“What’s the real health benefit?” said Dr. Robert C. Green, a Harvard professor and a medical geneticist at Brigham and Women’s Hospital in Boston. “If you’re a little bit cynical, you say, well, none, it’s foolish.”

Dr. Green is part of a federally sponsored research project that is looking at the economic and medical impact of whole genome sequencing. “One of the most prominent downsides is you start chasing risks for a whole lot of disease you’ll never have, and generate a lot of cost for little benefits,” he said.

Of course, sequencing every genome we can sequence is, in part, in essence screening an asymptomatic population for potentially thousands of conditions that could be found by analyzing gene sequences. Every time anyone's genome is sequenced, it's virtually inevitable that mutations or sequence variants associated with diseases will be found. the problem is that, in most cases, the risk of diseases associated with those mutations is not that well characterized; so it becomes problematic what to tell patients. Genetics is not destiny. On the other hand, if the information is there, it's hard not to treat it as such. We've already seen this sort of problem with various direct-to-consumer tests marketed as gene tests, such as 23 and Me.

None of this is to say that personalized therapy, which is now increasingly being called "precision medicine," much the same way that "complementary and alternative medicine" has become "integrative medicine." (Yeah, I'm being a bit sarcastic here because the name change to me is pretty pointless.) However, none of this is to say that I don't think there will be a role for "personalized medicine" or precision medicine or whatever you want to call it. Genomics is here to stay, and I have little doubt that in the relatively near future (as in five to ten years) we'll be sequencing patient genomes for all manner of diseases and conditions and that many more hospitals than just NCI-designated cancer centers with huge endowments and tons of money to build research institutes costing hundreds of millions of dollars will be sequencing cancer genomes to guide therapy, and it'll be done right, not the way someone like Stanislaw Burzynski does it.

The problem, as I've described before, is that cancer genomes are messed up. Really messed up. I mean, so incredibly messed up that it's rarely just one set of mutations driving cancer growth. Add to that evolution constantly selecting for more aggressive clones, and cancer might actually be a less fruitful area for personalized medicine than everyone understands. After all, if you find mutations for which there are no targeted therapies, it doesn't help much, if at all. Even so, contrary to the claims of certain people trying to co-opt personalized therapy, it actually is possible to do clinical trials testing personalized therapy. for instance:

The promise of whole genome sequencing can be seen in trials like one for bladder cancer at Memorial, where the effects of a drug normally used for breast cancer were disappointing in all but one of about 40 patients, whose tumor went away, Dr. Baselga said. Investigators sequenced the patient’s whole genome. “The patient had a mutation in one gene that was right on the same pathway as the therapy,” Dr. Baselga said. “And that explained why this worked.”

As I've pointed out before, it will take some rethinking of our clinical trial design paradigms to incorporate new genomic technologies routinely into clinical care. We need to know when it's useful to sequence a patient's genome and when it's not, what we can tell from it, and what we can't. However, the biggest task is going to be to figure out the clinical significance of hundreds, if not thousands, of mutations and sequence variants. In cancer, the task will be even more difficult because no two cancer genomes are alike, and evolution guarantees that different areas of the tumor can be very different from each other and have different drug sensitivities and resistances.

So what does this mean? Heck if I know right now. I'm actually fairly optimistic, but I also believe that personalized medicine tends to be oversold. There are two things that I think I know right now. First, the NYT article was right to characterize it as a "gold rush," because that's what universities are behaving like. The question is whether there is actually any gold at the end of that rush; there might be a lot less than scientists think there are. Finally, I know that the the benefits and scientific advances in personalized therapy will come from real scientists delving into the mysteries of the genome with rigorous science. It won't come from personalized medicine wannabes like Burzynski.