Foundation Models
The AI engine for the future of precision medicine
Tempus combines AI expertise with real-world, multimodal data to transform how we understand and treat disease, starting with cancer.
Our proprietary foundation models enable rapid AI application development and meaningful insights across the end-to-end drug development lifecycle, from discovery to clinical development and commercialization.
Models require appropriate validation and market authorization, as required by applicable law, prior to clinical deployment.
Our foundation models
A comprehensive suite of foundation models for oncology, powered by one of the largest, de-identified multimodal datasets in the world.
oFM — Oncology Foundation Model
Unifies pathology, genomics, clinical notes and patient trajectories to create holistic, patient-level disease representation
How it’s used
Prediction of payload sensitivity, treatment benefit and progression risk
Biological interrogation to identify and quantify the factors driving patient response or resistance
Phenotype identification associated with treatment benefit
Clinical trial cohort enrichment and retrospective trial analysis
Digital biomarker development from multimodal longitudinal patient trajectories
Patient-level Digital Twins can be used to understand how patients may respond on alternative treatment regimens based on individual biology and clinical history
Training data
pFM — Pathology Foundation Models
A portfolio of proprietary foundation models designed to capture how disease biology is represented in tissue.
How it’s used
Virchow Family
Creates fine-grained representations of tissue morphology that are highly generalizable for applications based on pathology images
Diagnostic tasks such as cancer detection, classification, and subtyping across both common and rare tissue types
Phenotype-genotype mapping enabling biomarker detection, expression and regulation
Novel biomarker discovery to predict payload resistance and overall therapeutic response
PRISM Family
Combines pathology images and clinical reports for enhanced contextual reasoning
Diagnostic tasks, including pan-tissue cancer detection and complex histological subtyping
Predict genomic biomarkers, IHC expressions, and morphology-associated phenotypes, even when working with limited training data
Advanced prognostic and predictive tasks such as predicting patient survival or treatment benefit
Training data
How it works
From vast multimodal data to task-specific precision
We train our foundation models on massive, diverse datasets to capture deep disease context and complex biological associations. Instead of building a new AI model from scratch for every specific research question, we use these pre-trained, context-rich representations as a powerful starting point. This allows us to rapidly adapt and fine-tune models for your exact use case, requiring significantly less additional data and compute.
Case studies
Our foundation models can be tailored to support a wide range of critical tasks across the drug development lifecycle.
Deployed Applications
Explore the AI-enabled applications powered by our foundation models today.
Genomic biomarker prediction using H&E alone
*For research use only
Paige Predict* is an AI application that predicts the status of ~1,600 biomarkers across >500 genes from a single H&E slide, enabling accelerated genomic insights across cancer types. Predictions can help researchers link morphological phenotypes with genotypes, prioritize use of tissue and pre-screen samples for biomarkers of interest at scale.
AI application for pan-cancer detection
*For research use only
Paige PanCancer Detect* is a groundbreaking application capable of detecting suspicious tissue across more than 40 cancer types to support research with pre-screening of all biopsy and resected tissue. Developed on Virchow, this application demonstrates the generalizable power of our foundation models.
Partnership Framework
We offer flexible frameworks to align our AI capabilities with your needs
Partner with us on a customized project tailored to your unique needs, from an initial proof-of-concept through deploying and commercialization an AI application
01 R&D
- Rapidly assess feasibility and data scale effects, across modalities
- Mature successful AI models under design controls
02 Validation
- Use real-world data for robust validation and benchmarking
- Generate clinical evidence necessary for market authorization and adoption
03 Regulatory
- Navigate global regulatory pathways relevant to AI-enabled software
- De-risk a streamlined path to market
04 Deploy & Commercialize
- Launch through Tempus’ extensive network to drive initial access and adoption
- Leverage decentralized deployment channels for global scale
Foundation Model Licensing
Access foundation model embeddings for further development in your own environment
Control over training data and downstream AI development
Leverage your own compute resources
Combine access across multiple teams
Pre-built AI applications
Leverage off-the-shelf AI applications for rapid insights, across our data or yours
Tools for rapid insights at the scale you need
Pan-cancer digital biomarker screening
TME characterization and spatial biology
Our scale
~1M
tests run annually on our labs
~5K+
connected healthcare institutions
~65%
of academic medical systems in the US are connected to Tempus
~50%
of oncologists in the U.S
4
regions globally with AI-enabled footprint
50+
operational countries across North, Central and South America, Europe, the Middle East and Asia
Featured Resources
Hear from our experts
Article
Webinar
News
Press Release
Research
Manuscript
Abstract
This is AI-enabled precision medicine
This is the future of healthcare.