Hunt and peck

Last week, Clinical Data announced the sale of its in vitro diagnostics (IVD) division to focus its efforts in pharmacogenomics (PGx), making it just the latest in a string of companies moving away from IVD. Interestingly, Clinical Data is making this move at a time when all market indications would suggest that companies should be expanding their IVD portfolios, not shrinking them.

According to a 2006 report by BCC Research, the market for point-of-care testing systems (e.g., glucose monitors) was expected to grow at an annual average rate of 13.8% to reach almost $10 billion by 2010. Frost & Sullivan similarly reported that the cancer IVD market would grow by 2012 to $2.3 billion in the United States alone. And more recently, Kalorama Information suggested the global market for all molecular diagnostics (including pharmacogenetics) should surpass $92 billion by 2016.

I quite understand the appeal of PGx—and genomics more generally. While never wishing to imply that genomics is easy, it offers the perquisite of dealing with a relatively static component of the metabolic machinery: the genome. This means that correlations between gene sequence and drug response or disease state make themselves known pretty quickly. Who, for example, can deny the inherent beauty of the Her2/Herceptin pairing?

But what will happen when all the low-hanging fruit are gone?

At the physiological level, the distance between genes and metabolites—where all the real action is—is vast. Too vast, I fear, to be left to genomic technologies alone, which tend to provide black-and-white answers to annoyingly gray questions. You can't have 37% of a genetic marker, but you can easily have 37% enzyme activity. Or 59%. Or 72%.

Human disease and drug response are invariably multifactorial, even in situations that are predominantly linked to a single gene. If it were otherwise, appropriate patient dosing wouldn't be as difficult as it appears to be, and it wouldn't matter if mother likes to take her pills with cranberry juice.

Thus, I don't see genomics technologies as a solution in and of themselves, but rather as parts of a vastly more complex solution comprised of genes, proteins, lipids, and other metabolites.

Just as basic research has been forced to alter its focus from isolated biomolecules to a systems approach, I think that next-generation diagnostics will have to move from a single-biomarker paradigm to a multifactorial approach. It may be the only way for us to move from hunt-and-peck trial-by-error analysis to truly informed treatment decisions.

In the meantime, I appreciate the challenges facing companies like Clinical Data that have to balance near-term returns and long-term progress. I wish them well and hope they make the most of their choice.