Clinical Diagnostics: Moving Beyond Diagnostics to Prognostics and Then Predictive

These are the seven trends that will shape the field of clinical diagnostics in the coming decade.

Ajit Singh, PhD

Currently, there is no hotter area in global healthcare than diagnostics. Collectively, this is a set of services and tools that accounts for only 3% of direct global healthcare costs but impacts 70% of all healthcare costs that occur afterward, and 100% of clinical outcomes.

With such an outsized impact on all points of the healthcare value chain, it's no wonder that diagnostics is undergoing a wave of disruption and innovation from startups not typically seen in the medical field. Importantly, this innovation is happening around the world and with a scientific rigor that will ensure lasting, beneficial change for patients and medical professionals. It will also disrupt many existing paradigms for healthcare delivery--at hospitals, at clinics, at doctors' offices, and in many new settings--be it work, play, or home.

Improving diagnostics is a priority for healthcare systems around the world and is equally important in oncology, cardiology, infectious disease, neurology, and overall wellness (screening/monitoring). 

When it comes to healthcare, the world is truly flat. Whether it is cancer or heart disease, chronic illnesses or acute, the disease burden is becoming more and more uniform across the globe. Infectious diseases that used to be concentrated only in some geographies are quickly becoming global. A virus or bacteria only takes one carrier introducing it into a new population for it to spread there. So, not only is the issue a global one, but the startups we're seeing are also spread around the world. This is good for innovation as it brings fresh approaches and ideas from varying backgrounds.

The following seven trends will underscore the developments in clinical diagnostics over the next decade: consumerization, a move from centralized to distributed service models, access to remote expertise, disintermediation, genomic and molecular technologies, use of artificial intelligence and deep learning, and the integration of radiology and pathology.

1) Consumerization: Several factors will accelerate consumerization. In general, as the cost of healthcare increases globally, the consumer--the one person who should be most vested in their health--will have to be more integrated into the delivery chain. Second, molecular technologies are becoming more accurate, less expensive, and much easier to use. Recent diagnostic examples in the areas of HIV, various STDs, respiratory infections, and more highlight the move to a near-consumer setting. Net-net, more informed patients will make for more cost-efficient healthcare.

For instance, with an infectious disease diagnosis in today's delivery model, one has to get an appointment with a physician or go to urgent care, and then wait for the test results. This often means it takes several days before antibiotics or an antiviral can be prescribed, if necessary. Frequently, it is too late at that stage: either the illness gets worse, or unnecessary antibiotics are administered. A test that educates the patient in a few minutes at home (or at a pharmacy) about whether they need to see a specialist will eliminate all of pain points highlighted above. Add to it the mitigation of anxiety and privacy concerns, for instance in the area of STDs, and the benefits multiply.

Similarly, the benefit of routine monitoring of biomarkers at home for chronic illnesses is obvious. For instance, Click Diagnostics, one of my portfolio companies, has developed a technology for rapid detection of hundreds of infections in a small one-time use device. Think home pregnancy test but for nearly anything. Click technology has been tested with real clinical samples from patients and will seek to enter the regulatory phase shortly.

2) Decentralization: Nearly three decades ago, the advent of reference labs was a natural corollary of the fact that testing was complex, and centralization allowed for standardization, better quality control, and economies of scale (even though they added additional logistics of sample shipping and movement). But over time, as any diagnostic technology matures, it becomes less complex, less expensive, and more automated with respect to quality control. Why continue with a decentralized model? Why not bring the lab to the patient or to the point of care? Decentralization simplifies sample movement logistics and can be of tremendous benefit in emerging economies where the logistics infrastructure is not well developed. The recent acquisition of Cepheid by Danaher for $4 billion marks a key trend in this direction. Cepheid has been one of the pioneers in point-of-care testing, an approach that directly impacts decentralization.

3) Access to global expertise: One of the advantages of central reference labs was the concentration of the required clinical expertise in any specialty at the lab; the economies of scale allowed for that. This would be impractical in a decentralized setting. Each decentralized location couldn't possibly afford a super-specialist in every area of expertise; it would not be cost-effective. Fortunately, remote consultation technologies--digital imaging, cloud based computing, inexpensive network communications--allows leveraging of clinical expertise wherever it is available. This trend is already in the making: CORE Diagnostics, one of our portfolio companies based in India, for instance, gets second opinions on complex oncology cases from expert pathologists in the United States. The trend will only accelerate over time.

4) Disintermediation: Disintermediation is a natural corollary of decentralization and remote consultation. In the past, it was important that a specialist like a radiologist be where the technology and/or patient is. As digital technologies proliferate, all of the needed resources do not need to be in the same place at the same time. That creates efficiency. For instance, multiple opinions on a complex case can be obtained simultaneously. This eliminates or reduces the need for intermediaries, and makes the overall value chain of healthcare delivery more efficient. SecondOpinions.com is an up-and-coming startup based on the premise of disintermediation in the area of radiology diagnostics.

5) Genomic and Molecular Technologies: As the cost of genomic sequencing continues to fall and the informatics platforms that leverage genomic data become more mature, the clinical adoption of these technologies will accelerate. In the areas of noninvasive prenatal testing, genetic testing for rare diseases, and cancer diagnostics, genomic technologies are well past the early adoption stage and will continue to become more prevalent. This helps improve specificity of diagnoses, aids in determining prognosis, and allows for more accurate therapy selection. This will have a tremendous impact on quality of life and on downstream cost of treatment and follow up. This space is very quickly becoming extremely crowded--especially in the areas of sequencing hardware, informatics, and applications. Companies like Illumina, Life Technologies (now part of ThermoFisher Scientific), Roche, and Foundation Medicine have been pioneers in this domain, with several hundred others covering specific applications. For instance, Verinata Health (for noninvasive prenatal testing), now owned by Illumina, and CellMax Life (for cancer screening, diagnosis, and follow up), a portfolio company of Artiman, are two good examples of startups that are scaling up well.

6) Artificial Intelligence (AI) and Deep Learning: These technologies have matured significantly over the past decades, fueled primarily by the dramatic decrease in the cost of computing and widespread availability of cloud-based distributed, massively parallel computing. AI and deep learning have found applications in the discovery phase, as well as in clinical deployment. Similar to the introduction of other leading-edge technologies, I expect to see additional, unforeseen applications as new use cases are developed. For example, Cellworks, an Artiman portfolio company, uses computational modeling, simulation, and annotation of tumor genomics for clinical therapeutics predictions. Specifically, the system can (a) identify patients who will respond or not respond to a specific treatment, (b) design patient specific cocktails of existing approved agents for unmet treatment needs, and (c) identify predictor biomarker signatures (companion diagnostics) for a drug agent to target clinical trials to the right patients. At a more macro level, it can also identify novel target indications for a drug--for drug rescue and repositioning.

7) Integration: The key phrase is morphology + molecules. One cannot replace the other. As a consequence, integration of diagnostic information from radiology and pathology, early on in the clinical cycle, will be a key trend. We are beginning to see startups in two broad areas--software systems and services. BioImagene (now owned by Roche), for instance, built a comprehensive system to bring the entire diagnostic profile of a cancer patient--with information from routing pathology, radiology, and molecular diagnostics, on to one single platform. Similarly, services like CORE Diagnostics are providing an integrated patient report--combining relevant information from radiological imaging and pathology, including any second opinions. There is sufficient anecdotal evidence that this improves patient care.

Taken together, whether you're a physician, a patient, a startup entrepreneur, or an investor, this is an ideal time to be involved in diagnostics. We'll see more innovation here in the next decade than we've seen in the past 30 years. Most importantly, patients around the world will be better off. I'm excited about what's to come.

Ajit Singh, PhD is a Silicon Valley-based managing director at Artiman Ventures, an early-stage venture fund investing in white space companies creating or disrupting multi-billion dollar markets. He is also a consulting professor in the School of Medicine at Stanford University and holds a doctorate in computer science from Columbia University.

[Image courtesy of COOLDESIGN/FREEDIGITALPHOTOS.NET]