By Tom MIller
GreyBird Ventures was founded with an exclusive investment focus in medical diagnostics. Specifically, we invest in precision diagnostics in which the sensitivity and specificity of our portfolio technologies would enable treatment or prognostic uncertainty to be minimized.
At the time of our founding, investing in diagnostics of any kind (in-vitro, in-vivo, imaging, IT, etc.) was at a nadir as we quickly learned from the many fundraising rejections that we received. It seemed that almost all investment funds were chasing new pharmaceuticals and therapeutic medical devices. While laboratory diagnostics influence 85% of healthcare decisions, they only account for 3% of healthcare spending. This was fully reflected in how capital was being allocated.
COVID-19 abruptly shifted public attention and, concurrently, the focus of medical investing. For two years, diagnostic testing has dominated the front page of every newspaper and became a part of presidential addresses to the public. New testing technologies and the performance of available diagnostic tests that allow us to safely reopen our societies have become the topic of dinner conversations, political discourse, and of course, investment opportunities. While GreyBird was a pioneer in its original investing focus, we are no longer alone.
We wish that it did not take a global catastrophe to underscore our original thesis; that cost-effective medicine requires innovation in diagnostics. However, now that we appear to have brilliantly forecasted our present, we must articulate why we decided to invest entirely in precision diagnostics eight years ago, as the fundamentals have not changed. More importantly, we want to outline why this remains a tricky endeavor as scientific advances simultaneously help and hinder the field.
The cost of healthcare as a percent of GDP continues to rise in the United States and almost all developed economies. The US healthcare system has proven extremely difficult to fix, with policies and practices largely driven by the conflicting incentives of patients, payors, regulators, providers, publicly funded researchers, and privately funded industries. These problems have been written about extensively and are, unfortunately, not unique to the USA.
What has received little attention are the dramatic improvements in outcomes coupled with reduced costs offered by just getting an early and highly specific diagnosis correct in the first place. After all, there is nothing more wasteful to both health and money, than treating someone for a disease they don’t have, or with a treatment that will not work, or to detect an illness well after the point when the cost to treat, and the likelihood
of cure, is well past. For every patient that is helped by the ten highest grossing medications, between 3 and 24 patients receive no benefit (Nature 520, 609-611; 2015). Eliminating this waste of both money and health would likely be of greater benefit to the world’s healthcare systems than any policy change under consideration – and this is precisely the investment focus of GreyBird Ventures. Additionally, despite the many billions invested in new treatments for cancer, the advent of the diagnostic screening technologies of the mammogram, the Pap smear, and colonoscopy have demonstrably saved many more years of life.
Recent scientific developments are both helpful and harmful to this challenge. On the positive side, the benefits of developments in such areas as low-cost genetic testing, new imaging and detection technologies, and advances in machine learning all contribute to our ability to gather more information about a patient’s biology prior to treatment. However, to understand the scientific trend that makes this task more difficult, we need to provide some foundation.
The performance of any diagnostic test can often be reduced to two numbers: sensitivity and specificity. Perfect sensitivity means that the test never fails to identify the disease when present. Perfect specificity means that the test never identifies a person as having a disease that they don’t have. While this is well-known and very simple to understand, there is a brutal math at work that creates the need to understand the medical context of how the test is being used to determine whether a test is good or, better said, “appropriate”.
Let’s take a hypothetical diagnostic test that has 100% sensitivity and 95% specificity for a specific disease. Now let’s apply this test to a hypothetical population of one million people in which the natural incidence of the disease is one person in every thousand (somewhat similar to the average incidence of reported COVID-19 infections over the course of the pandemic).
In this case, our brilliant test would identify every individual having the disease. However, it would lump them in with the 50,000 who also received a positive result. Assuming that one would treat every positive case, this would mean that the number needed to treat (NNT) is 50. If the treatment is an aspirin, no problem. If major surgery, a risky biopsy, or an expensive course of medication, then the test performance is poor or needs to be augmented by a subsequent test with improved specificity.
Now, the scientific trend that works against the field is simply the growing list of defined diseases. As our understanding of the biology of a medical condition (defined by a set of symptoms) improves, what was once a single disease becomes many. For example, 100 years ago a patient that presents as tired and anemic might have a “disease of the blood”. Twenty years later this might be diagnosed as leukemia or lymphoma. Today, this could be diagnosed as any one of 80 separate diseases of the blood each requiring a different management, treatment, and probable outcome. The International Classification of Disease demonstrates this trend clearly with ICD 11 containing more than ten times the number of diagnostic codes as thought to exist when the author received his graduate degree in the 1980’s.
It is easy to understand what this trend does to the performance requirements of diagnostic tests. If the test performance does not improve and the number of defined diseases increase, the incidence of any given disease goes down. All else remaining the same, this means that we will administer more ineffective treatments (actually, this was happening already, we were just unaware of it).
This trend has been paralleled by an influx of new therapeutics. While pharmaceutical companies are more likely to focus on medications for diseases with large markets of potential patients, the above progress in medical science has led to a greater need of treatments for more narrowly defined conditions with more limited numbers of patients (assuming they are correctly identified with a precision diagnostic). To maintain revenues and to justify the investment expense, costs per treatment skyrocket. To avoid this unsustainable trend, diagnostics must become equally specific and such diagnostic testing must be mandated, and appropriately reimbursed, prior to treatment.
These trends create an increasingly critical requirement that we look for in our investments. The entrepreneur seeking funds to develop and commercialize a new diagnostic technology needs to be informed, in detail, of the clinical course and epidemiological context that will define the required test performance. This does not mean that we seek perfection in new diagnostic tests as this does not exist. But it does mean that the entrepreneur should have a thorough understanding of the above trends for the condition to be diagnosed to best attract capital.
Two of our portfolio companies can serve to illustrate the attributes we are seeking.
Ceres Nanosciences has commercialized a technology that solves a fundamental problem in diagnostics: At early stages of disease, the biomarkers most specific to that disease are low in biological abundance. The Ceres Nanotraps™ capture, concentrate, and preserve desired biomarkers to increase abundance in samples to be analyzed, effectively increasing specificity by increasing analytical sensitivity. One dramatic use case is that the Ceres Nanotraps™ are now the gold standard in capturing COVID-19 virus existing in very low abundance in wastewater. Additionally, early data shows that this technology will be equally powerful in liquid biopsy applications in which circulating DNA from tumors are also in very low abundance in a typical blood sample.
Genetika+ is pioneering precision psychiatry. Their NeuroKaire™ test combines genetic analysis with a test of medications for depression in a unique “brain-in-a-dish” (cortical neurons grown from a patient’s own blood cells). Early clinical results indicate improved medication selection by a factor of four or better. While not a perfect test (as none are), in the context of today’s medical practice, this is a true breakthrough.
This brings us to the heart of our investment thesis: Precision Medicine requires Precision Diagnostics. We can no longer justify overspending on useless or harmful treatments. The goal of diagnostics should ultimately be to limit the misuse of therapeutics by requiring more strict diagnostic criteria before prescribing treatment thus achieving the promise of precision medicine – increasing beneficial outcomes while simultaneously reducing costs. The time to invest in precision diagnostic technologies to reduce costs, prevent harm, and save not only lives but our medical economy, has clearly arrived.
After earning a graduate degree from the Harvard/MIT Health Sciences and Technology program, Tom Miller joined Siemens where he ran the global MRI business. He has also served as the CEO of the global medical operations of Carl Zeiss, the CEO of Analogic Corporation, and Chairman and CEO of LightLab Imaging, a start-up he helped to establish and sell. Tom re-joined Siemens in 2002 serving as a member of the Global Operating Board and Division CEO of Siemens Healthcare with 26,000 employees in over 130 countries. In 2013, Tom started GreyBird Ventures, an investment firm focused on technologies for precision medicine diagnosis. Tom currently serves on the boards of eight medical technology companies.
