NEW YORK – Bolstered by recent deals to acquire lab operations and license digital pathology technologies from PathAI, Quest Diagnostics plans to accelerate its transition to digital pathology across its US pathology labs and expand its use of artificial intelligence to help diagnose and treat cancer.
The firm has previously tested the waters for digital pathology and AI-based pathology tools through three labs in New Jersey, Colorado, and Florida. Kristie Dolan, senior VP and general manager of oncology and pathology for Quest Diagnostics, said that the company had set a deliberate pace for its adoption of digital pathology because slide digitization alone has limited value, but the firm has recently seen greater potential benefits as AI technologies have advanced and scanner prices have fallen.
Through the combined use of those technologies, Quest anticipates that digital pathology can help to improve the efficiency of cancer diagnosis and quickly connect patients with follow-up testing, Dolan said. She also said that Quest can use those tools to help address shortages of pathologists and laboratory technicians in the US by offering slide preparation, digitization, and analysis.
The company aims to fully digitize its US network of about two dozen pathology labs and equip them with PathAI's digital pathology software and AI-based tools from other firms that have received FDA approval or clearance, Dolan said. She estimates that the adoption of digital pathology across those facilities will take three to five years.
Quest said last week that it completed its acquisition of PathAI's anatomic and digital pathology laboratory services business and laboratory facilities in Memphis, Tennessee, which PathAI had acquired three years ago from Poplar Healthcare Management. The lab will become Quest's national AI and digital R&D hub as well as a support center for the firm's AmeriPath and Dermpath Diagnostics businesses. That facility has about 200 employees, including a dozen pathologists.
As part of the deal, PathAI will continue operating a research laboratory at the Memphis site to support its biopharmaceutical clients. Dolan noted that the PathAI laboratory in Memphis is also a few miles from a FedEx hub, which will help broaden the reach of its lab services.
Quest also struck a deal this spring to license PathAI's AISight digital pathology management system and a menu of algorithms to help its pathology labs transition from reviewing slides on a microscope to doing so on a screen. Dolan said that the AISight system will help accelerate Quest's transition to digital pathology by offering seamless slide preparation and interpretation operations.
"We believe that PathAI has designed an image management system that meets and, in some ways, anticipates the needs of customer pathologists," Dolan said.
Nick Brown, chief growth officer for PathAI, said the AISight software integrates with laboratory information systems and slide scanners to organize cases across laboratories and provide access to those images across a network along with features for slide annotation, collaboration, and consultation. The software includes a suite of algorithms that are used to prioritize the analysis of slides with suspected tumor cells, aid the identification of biomarkers such as PD-L1 and HER2, and alert pathologists about blurred areas on slides that need to be rescanned to avoid delays.
"The thesis of our partnership is really around not just going digital but how do we really improve the accuracy and quality of diagnostic testing [and] really also drive efficiency within that lab workflow," he said.
The companies have also agreed to establish Quest as a preferred provider of PathAI's biopharma and clinical lab services, and Quest will provide PathAI access to digitized slides and pathology reports that PathAI will use to continue developing AI-based algorithms in collaboration with pharma companies and other partners. The companies plan to collaborate on the development and design of AI-based algorithms for research and clinical use.
Quest expects that the various developers of AI-based algorithms will excel at the analysis of certain tissue types and intends to partner with the companies that develop the most clinically useful products. Dolan noted that Quest has partnerships and research collaborations with about a dozen more companies that are developing AI-based software, and she anticipates that the firm will be able to announce some of those partnerships with research results in the coming weeks.
Quest also has an existing collaboration with Paige and uses the de novo authorized Paige Prostate software in its digital pathology labs to aid the diagnosis of prostate cancer.
PathAI also has existing algorithm development partnerships with the likes of Roche, Agilent, Cleveland Clinic, and Laboratory Corporation of America. Labcorp invested in PathAI and has a partnership with Mount Sinai Health System to establish a digital pathology and AI center of excellence.
Brown said that PathAI sold only its laboratory business to Quest and remains an independent company that is "supporting labs all over the world" in their transition to digital pathology using AISight.
By selling its clinical diagnostic laboratory to Quest, PathAI has refocused on its core business of supporting pathology lab customers and pharma partners, he said.
Meanwhile, Dolan said that Quest sees value in immediately connecting a cancer diagnosis with its strong molecular testing menu. She said that rapid access to such tests as Quest's 523-gene next-generation sequencing-based Solid Tumor Expanded Panel can be critical for therapy selection for, say, a late-stage cancer patient.
"Without that genomic profile at the day-one visit, potentially that late-stage cancer patient will start with chemo and only later in their journey will they learn that they could have been on a targeted therapy for their specific mutation," she said.
Dolan noted that there is a great deal of research being conducted into how much information can be discerned from an H&E slide or even an IHC or immunohistochemistry slide, she said.
"The potential to infer a genomic profile at diagnosis is incredibly powerful," she said.
As those researchers refine their digital pathology algorithms, the tools that they develop could be used to identify tumor types or infer that a patient is at higher risk of certain types of cancers that could be detected using NGS, Dolan said. If an AI-based algorithm could aid the identification of, for example, an NTRK mutation, that information could be used to guide patients to more effective therapies.
"The therapy is highly effective, but it's not cost-effective today to sequence every single tumor for NTRK nor is it to do an IHC on every single tumor," she said. "But, if that AI could infer it at the diagnosis, then you could enrich and only send a subset of those tissues for confirmatory NGS."