for so long as to become a sad cliché. But changing a patient’s treatment based on a genetic test and especially initiating a treatment that would otherwise not have even been considered—that is a watershed. An idea like Foundation’s, in which you scan the genome of an individual patient for variations in more than 200 genes, is a medical reality today that was barely even conceivable five years ago.
Furthermore, Foundation hardly depends on its test revenues at the moment. The bulk of its revenues (something like 85 percent, I’ve heard) still come from partnerships with pharmaceutical companies. Its investors, both private and public, may well grant Foundation the time it will need to achieve reimbursement and make a compelling case to enough physicians to drive test adoption and growth.
Critics have correctly observed that there is little evidence for the utility of most of the genes on Foundation’s first panel, FoundationOne. Something like two hundred genes are assayed when barely 20 are known to be drivers of cancer. As I understand it, this is where Foundation’s entrepreneurial strategy comes into play. By aggregating data on the next 180 genes rather than focusing just on the 20 genes of known relevance to cancer patients, Foundation hopes to bring a much greater degree of clarity and utility to cancer therapy, which has traditionally been based on a brutal process of trial-and-error. Many patients (and their physicians) don’t have enough time or scientific insight to go through a series of single-gene diagnostic tests to find out which drug might be best for them. Even if patients demanded this one-at-a-time approach, it is not at all part of current medical practice. For the sake of cancer patients, I hope Foundation Medicine succeeds with its broader approach.
Critics have also observed that Foundation’s business model is predicated on the company being paid $5,000 or more for a test (according to Xconomy, recently out-of-pocket payment by patients or one-off payments from insurers have been running more at the $3,800 level). But the cost of sequencing is very low! Can’t the test be less expensive? Where does all that money go? The answer, to me, is clear: the money goes to research. The model reminds me of crowdsourcing, a funding mechanism that has just become a viable mechanism for funding biotech companies. In Foundation’s case, it is a way to raise money from people who have a real need (cancer patients), provide them with sufficient value (sequences of genes with known implications for cancer therapy) and then increase the incremental value of the test for the next round of patients.
To succeed, this approach has to scale. That is, insights obtained from the first 3,000 patients have to become more valuable for the next 30,000 and so on. There have to be increasing returns or else there will be a backlash at the level of pricing and adoption. In the absence of reimbursement, the only way to make this work is to raise a lot of money (through IPOs, secondary share offerings, pharmaceutical industry partnerships, self-pay from patients, international adoption or whatever) and pour it back into the company. The field of genomics spent several years wandering in the wilderness of “genome-wide association studies” (GWAS) which were supposed to identify canonical mutations that affected large numbers of individuals. That barely turned out to be the case. Now mutation hunters have come to the opposite conclusion: it is individual mutations, perhaps even those with an “N” of just one person, that will matter the most in improving cancer therapy. The company or entity that builds the largest database of these mutations—and applies them in cases where there is an “N” of two or 20—will become a go-to source for insight into specific patients’ cancers.
There are three dangers here: first, that scaling cannot be achieved quickly enough to justify reimbursement. The tests Foundation is doing are by their very nature outside of the parameters such as sensitivity and specificity that are traditional metrics for payers. So their results have to be so overwhelmingly good that payers change the rules in order to reimburse for the tests. That is likely to happen slowly if it happens at all.
Second, unless great insights arise from the additional genes, Foundation—with no real intellectual property on the content of its assay—will fall prey to commodity entrants offering tests at much lower price points. That is reminiscent of the dynamic I see playing out in non-invasive prenatal testing (NIPT).
Third, what if Foundation succeeds and gains insights from its database (paid for by patients) that lead to a true competitive advantage? Won’t there be a clamor for public release of Foundation’s data, similar to what happened when Myriad Genetics lost its Supreme Court case and no longer had patent coverage over its BRCA1/2 test? It will be interesting to see this play out.
In my view, Foundation’s IPO is a turning point that will only boost the many efforts to make the genome a powerful ally in the fight against cancer. Given the massive drop in sequencing costs and today’s vote of confidence, it will not take too long for similar insights into other diseases to follow.
