The fast-growing arena of personalized genetic medicine provides an intoxicating brew of hype and hope. The technology will purportedly provide revolutionary benefits in medical care. This promise has spawned new laboratories and research firms, eager to study strands of DNA to identify who is at risk for contracting a disease, to guide more exacting selected application of existing treatment, and to provide blueprints for formulation of novel therapies. President Obama recently called on Congress to authorize spending $215 million in 2016 on personalized medicine, calling it, “One of the greatest opportunities for new medical breakthroughs”. Federal funds would create a genetic databank of millions of donors accessible to research groups, just a part of the $6 billion industry that government and private investors are eager to finance.
This hoopla is predicated on a much distorted perception of the cause-effect relationship of genes and disease. One thing the genome project has definitely revealed is how much “junk” is in our DNA, or how small the percentage of an individual’s DNA material actually imparts individuality, and the similarities among species and the pervasiveness and frequency of genetic mutations. In an age enthralled by “big data”, the illusion persists that “mining” large DNA banks can provide exact answers to specific questions such as disease incidence, response to medication, and prognosis. At one extreme, we know that combined immunodeficiency (“bubble babies”) is caused by specific adenosine deaminase enzyme deficiency. Insert the genetic snippet that codes to synthesize this enzyme and you cure the disease. On an intermediary level, we know that the same genetic mutations are common in some malignancies from entirely different organs and provide rationale for formulating treatment strategies that are mutation-specific, not organ origination-specific. In other words, there is commonality of biological behavior when sorted by mutation, not site. However, this genetic commonality is identified in the vast minority of cancers.
The least useful “personalized” genetic testing is for the most common chronic diseases, like atherosclerosis, hypertension and diabetes mellitus. These have a multifactorial pathophysiology mixing family history with environmental exposure, diet, nutrition, and infection. Supercomputers can sort the genetics into correlations, but drilling it down further to provide reliable causation and therapeutics unique to a particular patient is an illusory. This is the medical equivalent of Google and Facebook selling your online history to a marketing firm so that when you log on to a particular website the pop-up ads most likely to pique your fancy will appear. Ergo, the propensity for you to respond to targeting advertising is more likely than responding to a uniform marketing pitch. But this is a statistical difference; the fact is that the number of times you buy versus the number of targeted ads you see is infinitesimal. Is that the methodology you want to determine what antihypertensive to take? Given the enormous “promise” of genetically-driven outcomes, the potential for fraud by the testing laboratories in both billing and results is enormous. Some things are just too good to be true, and as yet, personalized genetic medicine is one of them.
By Norman Silverman, MD, with Ryan McKennon, DO and Ren Carlton