Paige
The first FDA-authorized AI in pathology — built on 7M+ annotated slides to help pathologists detect and grade cancer faster.
Overview
Paige is a computational pathology company founded in 2017 as a spin-off from Memorial Sloan Kettering Cancer Center. It achieved a historic first in September 2021 when its Paige Prostate Detect product became the very first AI-based software to receive FDA De Novo authorization for use in primary cancer diagnosis. Built on a dataset of nearly 7 million clinically annotated whole-slide images spanning 45 countries, Paige's platform covers cancer detection, grading, molecular biomarker prediction, and research acceleration across prostate, breast, gastrointestinal, and 40+ other cancer types. In August 2025, Paige was acquired by Tempus AI for $81.25 million to anchor the largest oncology foundation model project in the industry.
Under Tempus ownership, the Paige product portfolio has expanded significantly. The open-sourced Virchow and PRISM foundation models (released August 2024 on Hugging Face under Apache 2.0 / CC-BY-NC-ND 4.0) allow the research community to build on Paige's pathology AI. Paige Predict, launched January 2026, predicts 123 biomarkers across 16 cancer types directly from standard H&E-stained slides, removing barriers when tissue is insufficient for full molecular profiling. The platform integrates with eight major digital pathology vendors (including Roche, Sectra, Indica Labs, and PathAI) via open APIs, and is available on Microsoft Azure Marketplace for enterprise deployment.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 4 Sep 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Virchow (Apache 2.0)
- PRISM (CC-BY-NC-ND)
- Non-clinical research use only
- Paige Prostate and diagnostic suite
- 30-day trial on request
- Also available via Azure Marketplace
Use cases
What you can produce with Paige
- FDA-authorized AI flag identifying suspicious prostate cancer regions on whole-slide images
- Gleason grade group and tumor quantification outputs for prostate biopsy slides
- HER2-complete expression score and breast cancer detection report from H&E slides
- Paige Predict report predicting likelihood of oncogenic biomarkers across multiple cancer types
- Open-source Virchow and PRISM foundation model weights for custom pathology AI development
- Structured case-level diagnostic summaries via the Alba research-use copilot integrating EHR + LIS + image data
ASEAN Perspective
Paige in Southeast Asia
Paige holds FDA, CE-IVD, and UKCA authorizations for its prostate cancer AI suite, but no announced regulatory approvals in ASEAN or broader APAC jurisdictions as of mid-2026. Its training dataset spans 45 countries, providing some demographic diversity relevant to Asian populations, though slide representation from Southeast Asia specifically is unconfirmed. The Microsoft Azure cloud infrastructure supporting the platform enables global deployment in principle, which is relevant to Singapore, Australia, and Japan — but commercial rollout in ASEAN markets depends on local health-technology assessment and reimbursement frameworks that have not yet been publicly addressed by Tempus/Paige. ASEAN pathology labs with digital pathology infrastructure and an interest in AI-assisted cancer detection should engage Tempus directly to assess regional availability and compliance pathways.
Paige is the standard-bearer for clinical AI in digital pathology — the only product holding FDA De Novo authorization for primary cancer diagnosis in pathology, now backed by Tempus AI's oncology data engine. Its clinical validation is unusually rigorous: 97.7% sensitivity and 99.3% specificity across 1,876 prostate biopsy slides, with pathologist-assisted reads showing a 70% drop in false negatives. The open-sourced Virchow and PRISM foundation models give outside researchers a real head start, and the January 2026 Paige Predict suite — 123 biomarkers across 16 cancer types from a single H&E slide — is a technically ambitious step forward.
The main caveats are enterprise complexity and cost opacity. There's no self-serve tier for the clinical products; deployment requires integration with a compatible scanner or LIS platform and a formal onboarding process, and pricing for Paige Predict and the diagnostic suites is unpublished. The Tempus acquisition brings broader oncology data synergies but also roadmap uncertainty as the two brands consolidate. No dedicated APAC infrastructure has been announced, and clinical use in individual ASEAN markets would need separate national approvals beyond the existing FDA/CE-IVD/UKCA marks.
What people say
97.7% sensitivity, 99.3% specificity — the published numbers behind Paige Prostate Detect, the first AI software to win FDA De Novo authorization for primary cancer diagnosis, back in September 2021. Nothing else in computational pathology carries quite that regulatory pedigree, and per Paige the company now holds 13 CE-IVD marks and 13 UKCA marks alongside it.
The August 2025 acquisition by Tempus AI, worth $81.25 million in stock, folded Paige's roughly 7-million-slide annotated dataset (from 800-plus labs across 45 countries) into Tempus's oncology foundation-model push. The first visible product of that combination, Paige Predict, launched in January 2026 and reads 123 biomarkers across 16 cancer types straight off a standard H&E slide — useful when there isn't enough tissue left for full molecular profiling.
What holds it back for smaller buyers is straightforward: enterprise-only, quote-based pricing and an integration process that assumes a compatible scanner and LIS setup already in place. There's no confirmed APAC regulatory pathway or regional support desk as of mid-2026, so hospitals outside the US, EU, and UK are largely on their own for local clearance. Crowdsourced reviews are essentially nonexistent for a product like this — the evidence base here is peer-reviewed clinical data and regulatory filings, not star ratings.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- Paige's FDA De Novo authorization in September 2021 was the very first for any AI product in the entire history of digital pathology.
- The Virchow foundation model was trained on 1.5 million MSK slides — the largest pathology training dataset publicly disclosed at the time of its release in 2024.
- Paige was acquired by Tempus for $81.25 million in August 2025 — a valuation representing less than 40% of its total $241M raised, reflecting a tough funding environment for health-tech.
- The company's name is a tribute to James Paget, the 19th-century British surgeon who described several conditions still taught in pathology today.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Paige's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Paige unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Paige directly →
Spotted something out of date? Suggest an update →
Alternatives to Paige
More in Healthcare
