Digital Pathology AI

Panakeia makes the most technically ambitious claim in this lane. Every other pathology vendor here reads morphology from an image and reports what a pathologist could in principle see. Panakeia predicts molecular status from the same image, inferring microsatellite instability and mismatch repair deficiency directly from routine haematoxylin and eosin stained slides, without sequencing, immunohistochemistry or any additional laboratory test. The company describes itself as the world's first in silico multi omics company. The clinical rationale is straightforward and the value is in the time. Microsatellite instability and mismatch repair deficiency status determines immunotherapy eligibility in colorectal cancer and identifies Lynch syndrome, a hereditary cancer predisposition with implications for the patient's relatives as well as the patient. That status normally arrives weeks after diagnosis through separate molecular testing. PANProfiler Colorectal returns it in minutes from an image the laboratory has already produced, with no change to existing workflow and no additional tissue consumed. The validation is the strongest in this lane and it was not conducted by the company. A blinded multi site clinical validation led by the University of Leeds and Leeds Teaching Hospitals NHS Trust was published in npj Digital Medicine in January 2026 and is described as the largest real world blinded multi site validation of artificial intelligence enabled molecular profiling from routine diagnostic images. Earlier blinded multi site results were presented at the ASCO gastrointestinal cancers symposium in January 2025, and proof of concept across more than 30 cancer indications was published in Communications Medicine in 2024. The company also submits to comparison it does not control. Its breast product was included in an independent multi vendor study assessing agreement on HER2 expression, alongside nine other models and three pathologists. The published result is uncomfortable for the whole field rather than for Panakeia specifically: the ten models agreed with each other 65 percent of the time, artificial intelligence and pathologists agreed 65 percent of the time, and the three pathologists agreed with each other 70 percent of the time, with low HER2 expression difficult for both. Participating in work that produces that finding is a form of transparency almost nobody in this index practises. The company also participates in the Friends of Cancer Research digital pathology agreement project and PANProfiler Colorectal has been selected for the UK regulator's AI Airlock programme for real world evaluation of artificial intelligence medical devices. Regulatory status is split deliberately. Two clinical products for colorectal and breast cancer are UKCA marked and clinically deployed in the United Kingdom, while the wider PANProfiler platform covering more than 30 cancer types is research use only, sold to pharmaceutical companies for trial screening, biomarker discovery and patient selection. Headquartered in Cambridge, United Kingdom, founded and led by chief executive Pahini Pandya.

Qritive is the closest thing in this index to Deep Bio's counterpart, and the pair are worth reading together. Both are Asian computational pathology companies anchored on prostate. Deep Bio builds algorithms only and distributes them through other companies' platforms. Qritive builds both the algorithms and the platform they run in, and then also distributes the algorithms into competitors' platforms, which is an unusual position to hold on both sides of the same market. Pantheon is the platform: a vendor agnostic, cloud ready image management system supporting all major whole slide image formats, covering case management, slide viewing, artificial intelligence supported analysis and structured reporting, with telepathology and remote consultation. Vendor agnosticism is stated as a design commitment and appears to be genuine, since the company's own modules also run inside competitors' systems. The QAi module family is the intelligence and it spans more cancers than most specialists attempt: prostate, colon, breast, lymph node and gastric, plus immunohistochemistry marker quantification and lymph node metastasis detection. QAi Prostate analyses whole slide images of core needle biopsies, identifies prostatic adenocarcinoma, segments and classifies benign and malignant areas, and reports tumour size and percentage per slide or region of interest. In its most recent deployment the prostate module is described as detecting malignant glands, identifying tumour architecture, grading according to International Society of Urological Pathology criteria, characterising Gleason patterns, quantifying tumour burden and flagging suspicious regions, with final clinical decisions remaining with pathologists. The regulatory position is real and modest in class. Pantheon is registered with Singapore's Health Sciences Authority as a class A medical device and is CE marked, described in one account as CE marked for in vitro diagnostic use, and the company holds ISO 13485 certification for design, development, manufacture, distribution and installation of software medical devices. Deployment is the strongest evidence and it is geographically unusual. Three Indian institutions adopted Pantheon and the artificial intelligence modules, including Metropolis Healthcare, the Rajiv Gandhi Cancer Institute and CORE Diagnostics. In June 2026 M42's National Reference Laboratory integrated the prostate module into its workflow at Cleveland Clinic Abu Dhabi, introducing artificial intelligence prostate diagnostics in the United Arab Emirates. The prostate grading module also runs on Roche's digital pathology platform, and the modules integrate with Corista's DP3. Founded in 2017 in Singapore by Dr Aneesh Sathe, a mechanobiology doctorate who built computer vision systems for identifying cancer cells, and Dr Kaveh Taghipour, whose doctorate is in natural language processing, both from the National University of Singapore. Operations extend to India and the United States, and the company is led by chief executive Bruno Occhipinti.

Deep Bio is a narrow specialist in a lane full of platforms, and the narrowness is the point. The company builds deep learning algorithms for prostate cancer histopathology and has taken them through a regulator rather than to a marketplace. DeepDx Prostate analyses whole slide images of haematoxylin and eosin stained prostate core needle biopsies, detecting and localising malignancy on Gleason patterns and quantifying tumour to tissue ratios, in a stated 30 seconds per core. Reported performance is 99 percent sensitivity and 97 percent specificity. Output is delivered as colour coded heatmap overlays highlighting suspect lesions alongside quantitative measures, and the company redesigned its reports specifically because pathologists in early deployments were not noticing the heatmaps, which is an unusually candid account of an adoption problem. DeepDx Prostate Pro extends this from detection to grading. It classifies histological severity automatically and generates Gleason grades and scores, returning a no grade result where tissue does not fall within the system rather than forcing a classification. In evaluation studies it showed 98.7 percent concordance in grade group classification and 96.9 percent in no grade classification against a reference standard created by three pathologists. The regulatory record is the strongest element and the firsts are real rather than promotional. DeepDx Prostate was the first artificial intelligence pathology tool in Korea to receive class 3 in vitro diagnostic device approval from the Ministry of Food and Drug Safety. DeepDx Prostate Pro then received what the company describes as the world's first regulatory clearance for an artificial intelligence device classifying prostate cancer severity by Gleason grade, again as a class 3 device, and was later designated an innovative product by Korean regulators and by the public procurement service. The products are CE marked. The evidence base is genuinely independent in part. An external validation study was published in Modern Pathology in October 2022 with named academic authors, and the company reports participating in the United States and Canadian Academy of Pathology meeting annually since 2018 with papers each year. Training data reportedly included thousands of slides at approval with more than 500,000 core images from the United States added subsequently for quality control. Distribution is through other companies' platforms rather than a sales force. The prostate algorithm is integrated into Roche Diagnostics' Navify Digital Pathology, Visiopharm's App Center, and Techcyte's platform, the latter two both separately indexed here. Five South Korean hospitals ran trial deployments ahead of full workflow integration. Based in Seoul, with work extending to breast cancer lesion differentiation and frozen section analysis.

Lumea is built on a claim the rest of this lane does not make: that digital pathology fails upstream of the software, and no viewer however sophisticated can compensate for fragmented, disoriented or compromised tissue. The company therefore starts at the biopsy rather than at the scanner, and the record covers four layers rather than one. The first is physical. BxChip is a patented sectionable matrix holding up to six biopsy cores through laboratory processing, with a stated average increase of 14.5 percent in histologic tissue surface area, and the matrix is stated not to interfere with immunohistochemistry, in situ hybridisation or sequencing. BxBoard and BxFrame preserve spatial orientation of specimens from the operating room into the paraffin block and onto the slide. Radio frequency identification tags link tissue to a patient encounter and provide chain of custody tracking from collection onward. Six United States patents are cited on the tissue technology alone. The second is grossing. BxCamera applies models at the grossing station to standardise measurement and description and document specimens automatically, addressing manual transcription between the bench and the laboratory system. The third is software and it carries the strongest credential on the record. Viewer+ received FDA 510(k) clearance for primary clinical diagnosis in March 2025 under number K242244. That is the distinction that separates this from most of the lane: a United States pathologist can sign out primary diagnoses on these images, where several neighbouring records are research use only or signal their status with an asterisk. BxLink is the combined viewer and laboratory information system for laboratories replacing an existing system, and the company states plainly on every page that BxLink itself is not FDA approved for digital primary diagnosis, so the clearance attaches to the viewer rather than to the whole catalogue. The fourth is the marketplace. Rather than building diagnostic models, Lumea integrates them, with a partner marketplace covering artificial intelligence, molecular tests and scanners, and announced partnerships with Paige and with Valar Labs for bladder cancer treatment decisions, both separately indexed here. Molecular test ordering sits one click from the diagnostic workspace. Founded in Lehi, Utah by Dr Matthew Leavitt, a dermatologist, with co founder Dr Jared Szymanski, a board certified pathologist, both among the earliest practitioners to attempt a full clinical digital transition. Leavitt moved to a board role in 2026 and the company announced a chief executive transition the same year. Specialty focus is urology, gastroenterology and dermatology, with prostate the anchor. One thing a reader should weigh. The company is candid that specimen quality drives diagnostic accuracy and sells the products that improve it, so its clinical argument and its commercial interest point the same way, which is not a criticism and is worth naming.

Gestalt occupies the layer between the scanner and the algorithm, and it is the only record in this lane that hosts other vendors' models rather than only selling its own. PathFlow is an image management system and digital pathology platform that centralises whole slide images, patient and case information, and artificial intelligence outputs in one workspace, and the positioning is vendor neutral by design: it supports multiple artificial intelligence tools simultaneously so an organisation can evaluate, adopt and change its model strategy without changing platforms. That is a genuine architectural stance rather than marketing. Partnerships have brought outside algorithms onto the platform, including a French vendor's tumour biomarker models for skin tissue and an Indian vendor's quality control models, and the company has an established relationship with one of the largest computational pathology vendors, which is separately indexed here. A platform whose commercial interest lies in hosting competitors' models is a different proposition from the analysis vendors that surround it in this category. The most distinctive product on the record follows from that position. The PathFlow AI Algorithm Evaluator, launched in 2024 and described as patent pending, lets pathologists and researchers assess and compare the performance of artificial intelligence algorithms inside the platform, securely and against their own material. In an index whose recurring finding is that vendors publish no performance data, a tool built so customers can generate their own is worth noting. The company's own models sit mostly upstream of diagnosis. AIRE, the artificial intelligence requisition engine, reads pathology requisition forms including handwritten ones and, in the company's description, insurance cards photographed upside down, validates patient information against the existing database, flags discrepancies for correction and auto accessions cases above a customer set confidence threshold of up to 95 percent, with accessioning time reduced by up to 80 percent. The company states it trains against each new form format and improves over time. Around the platform sit PathCloud for web based image sharing without heavy infrastructure, an education and proficiency testing module used for credentialing, and an information technology services business. Founded in 2017 in Spokane, Washington, with a Series A of $7.5M announced in 2025. Leadership includes president and chief strategy officer Lisa-Jean Clifford, who sits on the governing council of the pathology informatics professional association, and a chief medical officer. The company states eight granted United States patents, two CE marked in vitro diagnostic certifications, compliance to the United States health privacy statute and service organisation control 2, and integrations with Epic, Cerner and other laboratory and record systems. Customers named publicly are mid sized rather than marquee, including a private laboratory in Oklahoma and a veterinary diagnostic laboratory at Purdue University. One thing a reader should weigh. The company describes its artificial intelligence team winning first place in two global competitions, and competition performance is a real but narrow signal that does not establish clinical performance in a laboratory.

Pramana sits one step upstream of every other record in this lane. PathAI, Paige, Ibex, Aignostics and the rest analyse whole slide images. Pramana makes them. It builds autonomous whole slide imaging scanners and applies models during acquisition rather than after it, which makes this the only record in the pathology category where the artificial intelligence runs before a pathologist ever sees anything. The technical claim is hardware and software co designed together. Scanners perform fast focus sampling and dynamically calculate the z plane of the tissue, replicating the fine focus action of an optical microscope, and make inline decisions based on tissue type, thickness and slide artifacts such as annotations and debris. Edge models execute feature detection, quantification and classification in real time during the scan, so results are available the moment scanning completes rather than in a later processing pass. Quality assurance runs inline: each slide is annotated in real time for focus, stitching, bubbles, folds and other detectable errors, abstracted into tile based feature overlays and optionally a numeric quality metric, with autonomous rescanning when the scanner judges its own output inadequate. The company holds granted United States patents on the approach. Two product lines exist. A high throughput configuration handles routine and large caseloads, with a single scanning cluster stated at more than 1,000 slides a day. A desktop scanner takes four manually loaded slides and is positioned for labs entering digital pathology, satellite sites, teaching institutions and tumour boards. Sample coverage spans anatomic pathology, cytopathology, hematopathology, clinical pathology and microbiology, in brightfield. Alongside the hardware the company sells Digital Pathology as a Service, in which an institution sends glass slides and receives quality assured images without buying scanners or hiring staff to run them. That model produced the strongest evidence on this record. Mayo Clinic engaged the company in 2021 to evaluate throughput and quality, then awarded a multi million slide archival digitisation contract through a competitive process. The published result: 23,916 slides digitised over 30 days by a single human operator, roughly 800 a day, drawn from a tissue archive spanning the 1950s to the present, with no cleaning or preparatory steps, each slide annotated in real time by on scanner quality models. Founded in 2021 in Cambridge, Massachusetts by nference, a health data analytics company, with Matrix Capital Management named among investors. The company has since been acquired by Evident, the life science imaging business formerly part of Olympus, and the brand continues to be sold under its own name as the Pramana HT and Pramana M scanners, which is why it is enrolled rather than treated as absorbed. One regulatory fact belongs at the top rather than buried. The scanners are CE marked under the European in vitro diagnostic regulation and licensed by Health Canada, and in the United States they are research use only with a 510(k) submission pending. A United States laboratory cannot use these images for primary diagnosis today, and the company states that plainly on its product pages rather than obscuring it.

Veracyte builds genomic classifiers, each one a machine learned model over RNA whole transcriptome data, and it is the direct competitor to the artificial intelligence prostate tests indexed alongside it. Founded in 2008, based in South San Francisco with laboratories there and in San Diego, and listed on Nasdaq, it reported testing revenue of 135.8 million dollars in the fourth quarter of 2025, up 21 percent. The flagship is Decipher Prostate, a 22 gene classifier developed with machine learning that estimates the risk of metastasis and informs how intensively to treat. More than 300,000 patients have been tested since launch, over 100,000 of them in 2025 alone, with roughly 27,200 tests in the fourth quarter and a fifteenth consecutive quarter of volume growth above 20 percent. Market penetration is stated at about 33 percent. Its standing in the guidelines is the strongest of any record in this index. Decipher Prostate is the only gene expression test to reach Level I evidence and inclusion in the risk stratification table of the National Comprehensive Cancer Network guidelines for prostate cancer, and those guidelines uniquely recommend using its score to guide whether to add hormone therapy after prostatectomy, a recommendation resting on a Phase 3 randomised trial with a median of twelve years of follow up. Performance has been examined in more than 85 studies covering over 200,000 patients, and results have been linked to real world outcomes through the National Cancer Institute's SEER database. The classifier is currently under investigation in seven Phase 3 randomised trials sponsored by that institute. The wider portfolio uses the same method in other cancers: Decipher Bladder, a 219 gene classifier sorting tumours into five molecular subtypes; Afirma for thyroid nodules; Prosigna for breast cancer; Percepta for lung, including a nasal swab test whose trial has completed enrolment; and Envisia for interstitial lung disease, with lymphoma and renal tests in development alongside a tumour informed test for minimal residual disease. The company holds exclusive global access to the nCounter analysis system, which lets laboratories run its tests locally rather than shipping samples. HalioDx is among the businesses it has absorbed.

Deciphex is a Dublin company founded in 2017 by Donal O'Shea, its chief executive, and Mark Gregson, addressing the shortage of pathologists rather than the accuracy of any single diagnosis. Its chief medical officer is Runjan Chetty. It has raised roughly 56 million dollars in total, including a 31 million euro Series C in January 2025 led by Molten Ventures with ACT Venture Capital, Seroba, Charles River Laboratories, IRRUS Investments, the HBAN Medtech Syndicate and Nextsteps Capital. The business has two halves and a buyer should understand which one they are purchasing. Patholytix is software: a non clinical workflow platform for toxicologic and preclinical pathology, used by pharmaceutical and biotechnology companies during drug safety assessment. In April 2024 Charles River Laboratories, the largest preclinical research organisation in the sector and also an investor here, launched Patholytix Foresight jointly with Deciphex, a decision support tool built on Patholytix 4.0 that pairs artificial intelligence classifiers with whole slide images to speed primary evaluation and peer review. The two extended that into an exclusive image management arrangement in February 2025. Diagnexia is the other half and it is a service rather than a product: a network of more than 250 subspecialty pathologists who report cases digitally for healthcare providers, positioned explicitly against the traditional locum model, with a research variant called Diagnexia Analytix serving drug development. This record is scoped principally to the software, since Patholytix is separately licensable while Diagnexia is a staffed diagnostic service, and both are described here because the artificial intelligence and the human network are sold together. The company states its platforms let pathologists work up to 40 percent faster while maintaining accuracy, is expanding across the United States, United Kingdom, European Union, Canada and Japan, holds a partnership with Novartis on artificial intelligence for drug discovery pathology, and has stated it will use its image repository to build pathology foundation models.

Aignostics builds foundation models for computational pathology and sells their output to biopharmaceutical companies for drug discovery, translational research, clinical trials and companion diagnostic development. It was established in 2018 inside Charité Universitätsmedizin Berlin and the Berlin Institute of Health, alongside TU Berlin and Fraunhofer HHI, and spun out in 2020. Founders include Frederick Klauschen of Charité, Viktor Matyas and Maximilian Alber. It is based in Berlin with a New York presence and more than 75 staff. The models are the product and they are documented in public preprints rather than in marketing copy. RudolfV, the first, was built on a deliberately heterogeneous dataset drawn from more than 15 laboratories covering 58 tissue types and 129 histochemical and immunohistochemical staining modalities, with pathologist knowledge built into the curation rather than applied afterwards. Atlas followed, developed with Mayo Clinic and Charité, a vision transformer of roughly 632 million parameters trained on 1.2 million whole slide images from more than 490,000 cases, sampled into about 520 million tiles at four magnifications, and evaluated against 21 public benchmarks alongside named rival models. Atlas 2, announced January 2026 with Mayo Clinic, LMU Munich and Charité, is around 2 billion parameters trained on more than 5 million slide images and reports the highest average performance across 80 public benchmarks, with distilled smaller versions released for compute constrained settings. The most recent product, Atlas H&E-TME, is a self service application profiling the tumour microenvironment at single cell resolution from routine stained images. The company states that Atlas 2 ships with clinical grade regulatory documentation intended to support integration into medical devices built by others. More than 55 million dollars has been raised including a 34 million dollar Series B in October 2024, with ATHOS, Wellington Partners and the Boehringer Ingelheim Venture Fund among investors. Development partnerships are named with Bayer and Mayo Clinic.

Taiwan based medical imaging AI company spanning digital pathology infrastructure and diagnostic algorithms, whose aetherSlide image management platform received both FDA clearance and IVDR certification in 2026, positioning it for global expansion beyond Asia. Its algorithm portfolio includes a CE marked lymph node metastasis detector for gastric cancer that classifies and quantifies positive and negative nodes, and aetherAI Hema, described as the first bone marrow differential AI system, trained on a curated dataset of more than one million cells and reporting 15 subtype differential counts. Deployed across Chang Gung Memorial Hospital branches and developed in collaboration with National Taiwan University Hospital and Chi Mei Medical Center. Backed by Quanta Computer and Cathay Venture, and has applied to list on the Taiwan Innovation Board.

Formerly Owkin Dx, the CE-IVD marked diagnostics arm of Owkin, since spun out as an independent company named Waiv with $33M in financing. This record keeps its original address for continuity. It remains indexed separately from Owkin itself under the index rule of indexing by AI product rather than by company, and Owkin holds a separate record covering its research and pharmaceutical business. Two products established the position, both category firsts in Europe. RlapsRisk BC predicts risk of distant relapse in ER positive HER2 negative early invasive breast cancer from haematoxylin and eosin stained whole slide images combined with clinical variables including age, node involvement and tumour size, described as the first CE-IVD approved digital pathology AI diagnostic predicting relapse risk. MSIntuit CRC pre-screens colorectal tumours from routine histology to rule out microsatellite stable phenotypes, reducing the volume of confirmatory MSI testing required. Both are positioned as cheaper and more accessible alternatives to molecular and gene expression testing, using slides laboratories already produce. Following the spin out the portfolio is presented under an AI native digital pathology platform named Destra. Buyers should note that certifications and compliance disclosures published by Owkin cannot be assumed to transfer to the spun out entity, and should be confirmed with Waiv directly.

Oslo based company whose Histotype Px Colorectal predicts patient outcome in stage II and III colorectal adenocarcinoma from standard H and E stained slides, stratifying patients into low, intermediate, and high risk groups to inform whether adjuvant chemotherapy should follow surgical resection. Reported as the first CE marked product using AI to predict patient outcome from image analysis. Built on research from the Institute for Cancer Genetics and Informatics at Oslo University Hospital and trained on close to 100 million image tiles. Addresses a specific overtreatment problem the company states plainly: most stage II and III patients are cured by surgery alone, so a large majority receiving adjuvant chemotherapy gain no benefit while incurring its harms.

Paris based pathology AI company whose Cleo tools detect and quantify cancer biomarkers in whole slide images across two specialties. Cleo Breast is CE-IVDR marked as a primary diagnostic solution covering breast tissue biomarkers including mitosis counting and lymph node metastasis detection, while Cleo Skin is described by the company as the first AI tool for dermatopathology to receive CE-IVDR certification. Distinctive for a deliberately multi platform distribution strategy, with Cleo integrated into PathAI, Proscia, PathPresenter, and Gestalt environments rather than requiring its own workspace, and for running a dermatopathology clinical validation study measuring pathologist performance with and without the tool.

Stockholm based company whose Stratipath Breast is described in peer reviewed literature as the first CE-IVD marked AI image analysis tool for primary breast cancer risk stratification available for routine clinical use. Rather than detecting or quantifying a biomarker, it stratifies patients into low and high risk groups from routine hematoxylin and eosin stained slides already produced in standard care, positioning it as a prognostic alternative to molecular multigene assays that carry longer lead times and higher cost. Validated retrospectively across 2,719 primary breast cancer patients from two Swedish hospitals, and since 2025 also distributed through PathAI's AISight platform.

Palo Alto precision oncology company whose Vesta platform predicts treatment response and prognosticates outcomes from routine hematoxylin and eosin stained slides, built on a Computational Histology AI foundation trained on more than 500,000 pathologist annotations of tumors and their microenvironment. Reaches the clinic through a different regulatory route than most pathology AI indexed here: tests are offered as laboratory developed tests from the company's own CLIA certified and CAP accredited laboratory rather than as cleared devices, spanning bladder, prostate, and pancreatic cancer. Vesta Bladder Risk Stratify Dx received FDA Breakthrough Device Designation in May 2026, reported as the first AI powered digital pathology prognostic test in bladder cancer to do so.

Hamburg based computational pathology company focused on biomarker quantification from H and E, immunohistochemistry, and multiplex immunofluorescence tissue images. Holds a reported eleven CE-IVD marked algorithms spanning breast cancer markers HER2, ER, PR, and Ki-67 plus lung PD-L1, with research use only assays for prostate and other tumor types. Its BreastIHC product was described at launch as the first deep learning solution certified to distinguish tumorous from non tumorous structures at cellular level for primary diagnosis. Since 2025 the portfolio is also distributed through PathAI's AISight platform, which is separately indexed here, so a buyer may encounter these algorithms either directly or embedded in another vendor's workspace.

Danish precision pathology software company whose diagnostic products are packaged as APPs, self contained algorithms for specific biomarker scoring tasks, covering HER2, Ki67, PD-L1, lymph node metastasis detection, and invasive tumor detection across breast, lung, prostate, and colorectal cancer. Distinguished by regulatory volume under Europe's IVDR, with the PD-L1 application described as its ninth IVDR clearance, and by a commercial alliance with Agilent under which the APPs are validated for Agilent assays and the Omnis platform and sold as the Visiopharm Diagnostic Package. The APP model reflects a structural reality of pathology AI: clearances are narrow, tied to a specific tumor type, biomarker, antibody clone, and assay vendor, so breadth is achieved by accumulating many separately certified algorithms rather than one general model.

New Mexico based digital pathology company whose HALO AP Dx is an enterprise platform FDA cleared for primary diagnosis of surgical pathology slides, with clearances tied to specific scanner hardware: K232833 with the Hamamatsu NanoZoomer S360MD and K252762 adding the Leica Biosystems Aperio GT 450 DX. Operates as both a diagnostic workspace and an algorithm host, with third party AI available through an associated store, positioning it alongside Proscia and PathAI as platform infrastructure rather than a single algorithm vendor. The company maintains a deliberate two product split in the United States: HALO AP Dx for clinical primary diagnosis, and HALO AP as research use only, explicitly not FDA cleared for diagnostic use.

Biomedical AI company building generative and agentic tools for pathology, whose PathChat co-pilot combines pathology foundation models pretrained on histology image and image text datasets with a custom trained multimodal large language model, enabling conversational analysis of high resolution slides and clinical data. The copilot is presented as usable from a microscope, a slide viewer or a smartphone, which removes the usual prerequisite of a completed digital pathology programme. A second product, Judith, is an agent for automating AI model development for biomedical image analysis. Extends research published in Nature from an academic lab at a major Boston hospital system; the company's whole slide foundation model TITAN is described in the literature as pretrained on 335,645 whole slide images and fine tuned partly on synthetic captions generated by PathChat itself. Received FDA Breakthrough Device Designation for its diagnostic version, which the company states plainly does not imply clearance or approval. In January 2026 the company announced its acquisition by AstraZeneca to advance AI driven oncology research and development at global scale; this record grades the products rather than the parent, and the ownership relationship is treated on the governance axis. Earlier research collaborations were announced with Techcyte, also indexed here, and with illumiSonics.

Publicly traded Finnish deep learning pathology company offering cloud based image analysis across both clinical diagnostics and preclinical research. Holds IVDR certification and describes itself as Europe's leading provider of CE-IVD marked digital pathology AI, with clinical suites spanning breast, lung, prostate, colorectal, and gastric cancer plus lymph node metastasis detection. Recent models are built on a foundation engine designed to perform across real world variation in sample quality, staining, and scanning.

Digital pathology company whose AISight image management system holds FDA 510(k) clearance for primary diagnosis, paired with a large portfolio of AIM interpretation modules covering commercially significant oncology biomarkers. Its AIM-MASH tool became the first AI powered pathology Drug Development Tool to receive FDA and EMA qualification, allowing pharmaceutical sponsors to use it for endpoint assessment in registrational trials. Sold its diagnostics laboratory business to a national reference lab in 2024 while licensing the platform back, and entered a definitive merger agreement with Roche in May 2026.

Cancer diagnostics company describing itself as the first and most widely deployed AI platform in pathology, with prostate, breast, and gastric solutions used in routine clinical practice worldwide. Its prostate product received FDA 510(k) clearance as an in vitro diagnostic that generates heatmaps flagging small and rare cancers missed on initial assessment, functioning as a second read safety net rather than a primary reader. Notable in this index for holding an unusually complete set of published security and quality certifications.

Computational pathology company that produced the first FDA authorized AI based digital pathology product, Paige Prostate, cleared via the De Novo pathway for cancer detection support. Subsequently pursued foundation model scale work in oncology imaging and received FDA Breakthrough Device designation for a pan cancer detection tool spanning common and rare variants across multiple tissue types. Acquired by Tempus in August 2025 for a reported 81.25 million dollars, making it now part of an indexed vendor rather than an independent competitor.

Digital pathology software company whose Concentriq platform manages and analyzes whole slide pathology data for clinical laboratories and life sciences organizations. Its diagnostic edition is FDA 510(k) cleared and CE-IVDR marked for primary diagnosis, while the research edition remains labeled Research Use Only, a distinction the company draws explicitly. Embedded AI recommends stain panels and ancillary tests from tissue appearance and scores image analysis biomarkers including PD-L1, HER2, and Ki-67 in the viewer.

Spatial biology company applying computer vision and machine learning to tissue imaging, integrating high plex spatial proteomics, histopathology, and clinical data to identify predictive spatial biomarkers. Positioned primarily for pharmaceutical R&D, supporting patient stratification and trial enrichment for antibody drug conjugates, bispecifics, and immunotherapies, with an emerging diagnostics application. Reported as the first spatial AI tool used by pathologists for clinical trial patient selection tied directly to a drug development program.

Note on naming: this is Artera of Los Altos, California (artera.ai), developer of multimodal AI cancer tests. It is unrelated to the separately indexed Artera of Santa Barbara (artera.io), which makes patient communication agents. The two companies share a name and nothing else. The ArteraAI Prostate Test is a multimodal artificial intelligence model combining digitized histopathology images from an existing biopsy with structured clinical data including age, PSA, and tumor stage, producing a continuous risk score from 0 to 1 with pre established cut points corresponding to roughly 3 percent and 10 percent estimated ten year risk of distant metastasis. It requires no additional procedure because it reads tissue already taken. Beyond prognosis it is predictive: for NCCN intermediate risk patients it estimates whether adding short term androgen deprivation therapy to radiation will reduce risk, which the literature describes as the first predictive biomarker for that decision. The algorithm is pre established and locked at version 1.2, developed from large datasets and multiple phase 3 randomized trials with up to 15 years of follow up. Regulatory and guideline standing is unusually complete: FDA De Novo marketing authorization in August 2025, inclusion in the NCCN Clinical Practice Guidelines for Prostate Cancer as the first and only AI risk stratification tool, and CMS coverage with a payment rate effective January 2024. In June 2026 the company introduced a digital pathology based test providing individual risk estimates in metastatic hormone sensitive prostate cancer. Co-founded by Felix Y. Feng of UCSF; CEO Andre Esteva.

AI digital pathology platform spanning human, veterinary, and environmental laboratories, developing in this field since 2013. Techcyte Fusion is positioned as the first digital workflow integrating anatomic and clinical pathology in one platform, with deep learning models identifying parasitic cells, bacteria, and cancer cells, plus digitization enabling remote case review. Product areas include parasitology, bacteriology, hematology with a peripheral blood smear classifier, cervical cytology, and bladder cancer surveillance. Regulatory status is the decisive buyer fact and the company states it plainly: the anatomic and clinical pathology platform and its human products are Research Use Only in the United States, while Fusion carries CE-IVDD marking in Europe. The architecture is deliberately open rather than proprietary, connecting to numerous slide scanners, laboratory information systems, and third party AI developers through a Fusion Partner Program; one such partner, Modella AI, brought its PathChat research copilot to the platform and was subsequently acquired by AstraZeneca in January 2026. A long running Mayo Clinic collaboration provides access to the Safe Harbour dataset of more than 17 million de-identified slides and pathology reports for model development. The company reports its veterinary segment is already profitable, with human and environmental segments targeted for profitability by 2027, and raised $15 million to expand the platform. Headquartered in Orem, Utah; CEO Ben Cahoon.