CEO SUMMARY: Pathologists were surprised this summer when a company developing image analysis software announced the acquisition of one of the nation’s largest independent anatomic pathology (AP) laboratories. The technology company PathAI is now the owner of Poplar Healthcare Management, in Memphis. It may turn out that this provocative combination of a pathology image analysis company and a sizeable AP laboratory accelerates the use of AI in pathology image analysis.
ONE OF THIS YEAR’S MORE INTRIGUING DEVELOPMENTS in the anatomic pathology (AP) profession came in July when PathAI—a technology company in Boston that develops image analysis software—acquired Poplar Healthcare Management, of Memphis, an anatomic pathology (AP) group with a large regional and national base of clients.
This acquisition creates a unique player in the anatomic pathology profession. It combines a company developing artificial intelligence-powered digital imaging tools with a regional and national provider of anatomic pathology services. Poplar Health’s lab facilities, management team and 350 employees now comprise PathAI’s diagnostics division. In 1995, Patrick J. Dean, MD, founded Poplar Healthcare, which today has 25 pathologists.
Moreover, with this bold action, PathAI may have accelerated its ability to achieve three goals. First, it gains access to Poplar Healthcare’s sizeable archive of about two million or more glass slides and diagnoses reaching back several years. These materials have significant value for PathAI as it tunes its machine-learning algorithms and its artificial analysis software to analyze whole-slide images.
Second, ownership of Poplar Healthcare allows PathAI to pursue contracts for pharmaceutical research studies and clinical trials. Not only is this business potentially profitable on its own, but it expands the company’s access to more images and case data to improve its machine-learning software tools.
Preparing for FDA Review
Third, as PathAI prepares to submit its image analysis products to the federal Food and Drug Administration (FDA), Poplar Healthcare can be a trial site, cutting the time PathAI will need to gather the clinical data to support its application to the agency for pre-market review.
A fourth possible benefit is that Poplar Healthcare’s staff and client pathologists will become the first to use PathAI’s analytical tools and artificial intelligence algorithms in daily clinical care. That development may help Poplar Healthcare gain additional market share and revenue.
In addition, PathAI’s digital analysis tools are likely to contribute to improved productivity and accuracy among Poplar Healthcare’s pathologists, cutting costs and increasing the laboratory’s profitability as part of the diagnostics division of PathAI.
Artificial Intelligence Tools
Founded in 2016, PathAI uses artificial intelligence and machine learning software tools to expedite the examination of human tissue and to facilitate the work that AP groups do every day. It already has pharmaceutical companies and contract research organizations (CROs) as customers, and it expects that the addition of Poplar Healthcare will help it expand its presence in both markets.
“Pharma companies, CROs, clinical laboratories, and other healthcare companies are developing companion diagnostics for patients with cancer and other chronic conditions,” said Andrew H. Beck, MD, PhD, PathAI’s co-founder and CEO. “Each of these organizations can use the artificial intelligence tools that PathAI has developed to make that work more efficient.”
PathAI plans to deploy its AI image-analysis system in a variety of clinical situations that previously have been closed to anatomic pathologists, he noted. “We’ve built a platform that allows our technology to be deployed in different settings, such as translational research, prospective clinical trials, and in clinical settings.”
Over the past five years, PathAI has been building that capability while also raising $255 million in venture capital to support its efforts, Beck noted. In the spring, PathAI closed on $165 million in funding through the healthcare provider organization Kaiser Permanente and D1 Capital Partners. Earlier, PathAI received funding from General Catalyst, a venture-capital company in Cambridge, Mass., and from the pharmaceutical company Bristol Myers Squibb, according to published reports.
Some of that recent funding was used to acquire Poplar Healthcare. PathAI also has used that capital to develop image-analysis software to identify complex patterns in patient specimens to detect the presence of cancer and other diseases.
One immediate consequence from the acquisition of Poplar Healthcare is that the AP lab now has full access to PathAI’s image analysis software and its algorithms to increase the efficiency and accuracy of its diagnostic work, Beck noted.
“Poplar is a top-tier laboratory with a dedicated team known for their accuracy of diagnosis and turnaround time,” Beck said, when the company announced the deal on July 26. “PathAI’s investments in digital pathology and artificial intelligence will further enhance Poplar Healthcare’s value proposition to providers across the United States.”
Improving AP Workflow
In an interview with The Dark Report, Beck explained the plan that PathAI will follow to improve the workflow for digital pathology.
“We’ve invested heavily in building robust generalizable algorithms to address the most challenging areas of pathology that we think could have the biggest impact on patient outcomes,” he said. “To do that, we have a large technology-focused team here in Boston of more than 225 employees who will build the platform to enable us to learn what we can from Poplar’s millions of slides.
“In addition, we have a large network of more than 400 board-certified pathologists who help us develop the annotation data we need to build and validate some of the models we’re developing for AP groups to use in their work,” Beck added.
“With the acquisition of Poplar Healthcare Management, the PathAI diagnostics division is now centered in Memphis,” he continued. “Our goals are to improve workflow efficiency, quality, and reproducibility across the entire spectrum of work that happens to each specimen in anatomic pathology.
Manual Processes in AP
“To date, our AI tools have largely been deployed in research settings,” Beck commented. “But they also can be used in clinical settings to help anatomical pathologists work faster and be more accurate and to work with colleagues more efficiently.
“With the acquisition of Poplar Healthcare Management, we intend to use our AI software tools to optimize workflow for increased speed, reproducibility, and accuracy,” he added. “Those three factors are the core processes in any AP group, particularly reproducibility and accuracy.
“For anatomic pathologists, increased speed is a by-product of the way that computational tools work, primarily because AI allows pathologists to accomplish many steps in parallel,” Beck noted. “But the real mission is to get the right diagnosis every time. That’s our main focus.
“Increased efficiency also can have an indirect effect on workflow because boosting throughput may mean a pathologist can see more cases,” he commented. “Greater throughput may enable the pathologist to provide value to more patients and to more referring physicians each day.
Accuracy and Reproducibility
“While increased efficiency is important, our top priorities are focused on accuracy and reproducibility,” he added. “That way we can ensure that all patients—whether in research settings or in clinical practice settings—are getting the right diagnosis.
“We are not tied to any specific biomarker or any specific tissue type,” Beck noted. “Our AI tools are used across many major cancers and ultimately we’re aiming to be involved in helping pathologists and researchers working on all cancers. By that I mean, PathAI aims to work across any type of tissue that can be viewed on a whole-slide image. That can include certain other types of diseases.
“One such example is a type of liver disease called NASH or non-alcoholic steatohepatitis,” Beck explained. “We currently work with researchers to develop tools for NASH. There are other major non-oncology diseases for which we are developing analytical tools. Our technology enables us to develop AI-powered tools for any number of diseases that require pathology slides for diagnosis.”
AI Software Can Help Improve Workflow
“CURRENTLY, MUCH OF THE WORKFLOW OF PATHOLOGISTS is done manually as they review image after image from referring physicians,” said Andrew H. Beck, MD, PhD, CEO and co-founder of PathAI. “Our artificial intelligence software is designed to help pathologists analyze stained images.
“If you think about all the steps that occur within a pathology laboratory for both research and clinical practice, the opportunity exists to optimize those steps, many of which are manual,” Beck explained. “We are developing our AI tools to provide the most efficient workflow for pathologists. These tools can have a positive effect on a lot of those steps—including the steps that involve the interpretation of images.
“In addition, AI can improve other workflow considerations, such as data entry, image management, triaging specimens, and the reporting of results,” he added. “When you think about the full range of what happens daily in every AP group, there are a lot of elements that, at their core, involve the transfer of information. By leveraging artificial intelligence into computing and informatics systems, we can help pathologists make the right diagnosis as efficiently as possible.”
Contact Andrew H. Beck, MD, PhD, at 617-500-8457 or Andy.Beck@PathAI.com.