Share this Post:
October 30, 2018 – The true barrier for pathology AI is the business model, not the technology. To get pathology AI into the clinical laboratories, payers need to provide a value-based model that creates a viable business case.
We estimate the U.S. anatomic pathology market for tissue image analysis, based on the current reimbursement model (e.g. computer assistance = CPT 88361 – CPT 88360 = $7-8), to be about $550 million with about $7-8 per test, even though it is unlikely that CMS (Centers for Medicare & Medicaid Services) is going to add $550 million to their reimbursements.
The problem is that the anatomic pathology market is segmented into sub-specialties, which correspond to different tissue types (e.g. breast), each of which has a list of different tests (e.g. H&E diagnosis, IHC Her2, ER, PR), which typically correspond to different stains. This creates a myriad of “tissue – stain – clinical outcome”-specific tests, each with its own little market segment that we estimate to be on average about $11 million (e.g. breast – IHC HER2 – score), that is shared by multiple manufacturers.
Pathology AI is dependent on the adoption of digital pathology, which by itself does not have a tangible business case (unlike radiology). Depending on whether a laboratory has a scanner, any additional reimbursement for computer assistance may need to fund the purchase of the digital pathology equipment as well. Ultimately pathology AI will drive the adoption of digital pathology, providing a return on investment (ROI)!
When you consider the costs associated with building and commercializing a pathology AI system as a medical device, this business case becomes a challenge. See our previous article Pathology AI as a Medical Device for a discussion on what it takes to commercialize a pathology AI system as a medical device.
What is the killer app?
Applications that provide the same results as pathologists using a microscope, just providing better consistency or saving time, make almost no difference in the market.
We have seen this very clearly with the tissue image analysis IHC HER2 test for breast cancer, the poster child for these kinds of applications. The adoption started very strong between 1998 and 2002 when the additional reimbursement was very high, about an additional $170 per test. By 2002, about 450 ACIS systems (the first commercial product) were placed. The reimbursement decreased to $60 in 2003, and by 2007, only 250 ACIS systems were still in the market. Today, the additional reimbursement is under $10 per test.
Interestingly, several additional tissue image analysis IHC HER2, ER, PR, etc. medical devices have been developed over the years, all by digital pathology manufacturers (who each have a completely different business case in mind) to market their digital pathology equipment for the clinical market. Given that there was a predicate device, this allowed for an easy route.
Rare event detections, including pathogens, like Acid Fast Bacillus (AFB), or cellular patterns, like mitotic figures, which could save pathologists a lot of time, didn’t’ even get that far. Why would the diagnosis of cancer, the latest application that everybody is talking about, be any different? After all, the gold standard is a pathologist using a microscope, why change to potentially lose money?
The adoption of pathology AI, under the current reimbursement model, will only be driven by “microscope impossible” tests that require a pathologist to use pathology AI. Today, immuno-oncology (IO) is the killer app with a massive business case behind it. There is an extensive need for tissue context data that other modalities, like next generation sequencing, cannot provide. The required analysis of the tissue is far too complex for a pathologist, if we just give him a microscope.
What is the right business model? Where is the value?
Pharmacogenomics allow us to identify the patients who are more likely to respond to particular therapies or who require dose modifications. Stratification of clinical trials, even retrospective, boosts efficacy and eliminates toxicity. How much is that worth?
Prescribed cancer treatments are effective in only about 25 percent of cancer patients, making them inefficient, expensive and detrimental to patient health. Adverse drug reactions annually in the U.S. alone, account for 100,000 patient deaths, $100 billion healthcare costs, and is the 4th leading cause of mortality. Between 1997-2004, 19 drugs were removed from the market (based on 2008 data).
Continuing with our example of HER2, the drug Herceptin acts on the HER2/neu (erbB2) receptor. In normal cells, HER2 controls aspects of cell growth and division. When activated in cancer cells, HER2 accelerates tumor formation. About 20-30 percent of breast cancers over express HER2, meaning those patients may be candidates for the drug, which costs about $70,000 for a full course of treatment. One of the significant complications of Herceptin is its effect on the heart. It is associated with cardiac dysfunction in 2-7 percent of cases. An IHC HER2 test that identifies the 20-30 percent of patients who would benefit from a $70,000 treatment and eliminates the risk of a cardiac dysfunction in the 2-7 percent of the 70-80 percent of patients that would not benefit from the treatment would save the healthcare system about $50,000 per test; not to mention the saved lives.
The true opportunity for pathology AI is personalized medicine with big data. This means that we could run a single test in a clinical laboratory (for any given tissue type), a standardized panel with multiple markers that provides rich information data for tissue. Treatment decisions that include the full spectrum of all available and future drugs could be based on that single test. New diagnostics (Dx), prognostics (Px) and companion diagnostics (CDx) could be created by clinicians in the field, correlating existing or emerging health conditions with this clinical “live” database. Drug developments could be done faster and cheaper with a smarter patient selection that could be achieved through better characterization of a patient population using the rich information data from that test. Diagnostics (Dx), prognostics (Px) and companion diagnostics (CDx) that were based on that same test could significantly simplify the regulatory pathways. See our previous article, Healthcare Big Data for Pathology for a more detailed discussion on our vision of Healthcare Big Data for Pathology and what it takes to get there.
Pharma companies who want to bring their drugs to market and payers who want to improve patient care and lower healthcare costs need to provide a viable business model that is based on the value that is provided by pharmacogenomics and big data.
Do you think a computer-aided-diagnostics (CADx) system for the diagnosis of cancer has a viable business case? How much are the real costs for a pathology AI system (it is not just $7-8 per test)? What does it take to establish a value-based reimbursement model for pathology AI?
Flagship Biosciences have been developing our own pathology AI system over the last 8 years which has been working in our clinical laboratory and meeting the needs of the market. The technology is there! With the right business opportunity, we are ready to commercialize our pathology AI system as diagnostics (Dx), prognostics (Px) or companion diagnostics (CDx).
If you want to learn more about the technology behind our pathology AI system, check out our previous LinkedIn articles, where we presented a short lecture series about the different aspects about pathology AI intended for a broad non-technical audience.
Holger Lange, PhD
Chief Technology Officer