Scarcity of Samples Threatens Personalized Medicine
Plenty of factors can conspire to slow down progress in molecular diagnostics and personalized medicine. There’s never enough money to run all the experiments scientists can imagine, it’s hard to gather data that squares with traditional medical evidence standards, and society is willing to pay a lot more money for drugs than diagnostics.
But there’s another factor increasingly limiting diagnostics entrepreneurs and the investors who back them—a scarcity of high-quality tissue samples.
Samples—everything from tumor biopsies to blood and urine to feces and saliva—are the crucial raw material that goes into the whiz-bang science of seeking diagnostic answers. Microscopes, real-time PCR machines, next-generation gene sequencers—they all depend on high-quality samples. In this case, “high quality” means it’s been stored in a central place, preserved the right way, shipped and handled properly, and ideally, matched with a patient’s long-term medical records. Sometimes, those records might even say something about how that patient’s case ended up so researchers can connect the dots between genes that are out of whack and a state of disease.
I’ve been hearing lots of anecdotal comments over the past year from entrepreneurs and investors about how tricky it can be to get ahold of enough high-quality samples to advance their visions for genetic-based personalized medicine. Sue Siegel, the former managing director at Mohr Davidow Ventures now with GE, called biological samples “liquid gold” at a genomics event I organized last October. Brook Byers, the partner at Kleiner Perkins Caufield & Byers, said at the Personalized Medicine Conference last November in Boston that his firm has gone so far as to connect its portfolio companies with “sample hunters” to help the companies get access to precious samples.
There are many reasons why samples are scarce, and valuable. There’s a surge in understanding of basic biology, which is creating demand for more samples to run through experimental instruments. When patients go in to see the doctor, nobody wants the patient to get poked, prodded or asked to come to the clinic for repeat sample donations any more than necessary. It also takes money, manpower, and a long-term organizational focus to really manage a biobank so that it’s truly a great resource for physicians and researchers.
Like many other things in life, this is really about supply and demand. The supply of good samples is short, and the demand is increasing.
“Cancer screening is getting more and more aggressive,” says Brad Gray, the CEO of Seattle-based NanoString Technologies, which is developing a digital gene expression test to gauge breast cancer risk. Physicians today “catch tumors early, and we catch them when they are small. That’s good. But it means you’re starting with a small amount of tissue, and the amount of things you want to know keep going up as our understanding of the molecular biology of cancer is going up.” And, obviously, once a small tumor biopsy gets used in an experiment, it can’t be used again. “This is a non-renewable resource,” Gray says.
Lung cancer is a great example of the sample bottleneck. Since the lungs are a sensitive internal organ, getting samples of tumor tissue there isn’t easy, and is generally done by a physician with a tool to scrape out as little tissue as he or she needs for analysis. There are traditional pathology tests that need some of that sample to classify the tumor, but beyond that, there are now single-purpose tests to look for EGFR status, or a mutation to the ALK gene, to see whether the patient is a good candidate for a certain targeted therapy. Each of those diagnostics consumes a little bit of the precious sample.
Hospitals and physicians know how rich in information their samples are, but the first priority is going to be using the supply to best manage the patient’s care. Research, especially the really exploratory kind, doesn’t always strike them as the best use of their limited resource, Gray says. It’s easier to make a case for using samples as part of an application to the FDA for approval of a new diagnostic test, which presumably should have a positive impact on patients.
Bonnie Anderson, the CEO of South San Francisco-based Veracyte, says the scarcity of samples is an issue that every diagnostic entrepreneur needs to think about early in a company’s life. Her company, which works in partnership with Cambridge, MA-based Genzyme, developed a test that looks at how 142 different genes are expressed in thyroid tissue, to help doctors determine when a lump is truly benign and doesn’t need to be surgically removed. Veracyte went after this market because there’s currently a lot of confusion about what to do with ambiguous conventional diagnostic results, and a lot of money wasted on unnecessary thyroid surgeries.
Importantly, Veracyte’s test was designed from the get-go to require very small samples of less than 50 nanograms. By extracting some valuable information from such a small sample, Veracyte only required doctors to make a very small change to their sample-gathering procedure in clinical trials, which left all the rest of the thyroid sample they needed for conventional analysis, Anderson says.
Essentially, getting the maximum amount of information out of the smallest amount of sample is the name of the game. The pressure to do that only grows when you have four or five different companies competing in a diagnostic category, all beating a path to the same few biobanks with the right kind of samples they need, Anderson says.
The good news is that some of the most advanced biological technologies out there today like high-speed/low-cost gene sequencing instruments are capable of answering a lot of questions based on a single run of a tiny amount of sample.
Still, powerful as those instruments are, they won’t reach their potential unless people and institutions to cooperate. Many samples are obtained today in clinical trials that have pretty limited use for the modern tools of biology, because they were given in the context of a researcher that only wanted to ask one specific question, says Rowan Chapman of Mohr Davidow Ventures, a leading investor in personalized medicine companies.
To get more informational bang out of the tissues, patients need to give informed consent for future use of their samples in a wide array of experiments. There are obviously privacy considerations to think about here—nobody wants to be identified to an insurer as possibly having an increased risk of lung cancer that may or may not be real—so the care and handling of samples requires a lot of thought.
But this is a conversation that patients and doctors and innovators and bioethicists all ought to be paying a lot of attention to. Access to biological samples doesn’t need to be, and shouldn’t be, a major barrier to progress in the genomic age.