Paige Digital Pathology (Prostate, Breast, PanCancer)

Paige Digital Pathology, available through Tempus, is a comprehensive suite of AI-powered image analysis applications designed to assist pathologists and researchers in detecting, grading, and gaining deeper insights from whole-slide images across a broad range of cancer types. The platform is built for clinical laboratories, academic and research centers, and life sciences organizations seeking to increase diagnostic confidence, improve workflow efficiency, and optimize tissue stewardship across oncology indications.

The suite encompasses three primary product families — Paige Prostate, Paige Breast, and Paige PanCancer — alongside tissue optimization and biomarker prediction capabilities under the Paige Predict umbrella. Applications span regulatory statuses from FDA-authorized and CE-IVDR cleared tools for clinical use to Research Use Only (RUO) applications supporting translational and discovery research.

Paige Prostate: Detection and Grading of Prostate Cancer

• Paige Prostate Detect (CE-IVDR, FDA-Authorized) — The first AI-based software application in pathology to receive FDA authorization to aid in the primary diagnosis of prostate cancer. It assists pathologists in identifying even minute foci suspicious for cancer on H&E-stained whole-slide images of prostate needle biopsies, offering high patient-level sensitivity and negative predictive value while reducing pathologist reading times by 20%. The application provides coordinates for the single location on the image with the highest likelihood of cancer for further pathologist review and is intended as an adjunctive tool, not a replacement for full pathologist evaluation.
• Paige Prostate Grade & Quantify (CE-IVDR, RUO in the US) — Categorizes areas suspicious for cancer by Gleason pattern, providing predictions for primary and secondary Gleason grades, as well as percentage and linear measurements of tumor burden.
• Paige Prostate Perineural Invasion (PNI) (CE-IVDR, RUO in the US) — Aids in the detection of suspicious foci around nerve fibers to identify the presence of perineural invasion, helping to streamline the final diagnostic process.

Paige Breast: Efficiency in Breast Cancer Detection and Staging

• Paige Breast Detect & Neoplasm (RUO) — Highlights areas most likely to harbor invasive cancer and in situ neoplasms, directing pathologist review to the most critical regions of the slide.
• Paige Breast Mitosis (RUO) — Identifies mitotic hotspots and provides a count and density of mitotic instances, enabling easy review of individual mitotic figures to support tumor grading.
• Paige Breast Lymph Node (RUO) — Detects foci suspicious for breast cancer in lymph node specimens. In a study of anatomic pathologists, use of this application improved pathologist sensitivity while cutting slide read times by more than half.

Paige PanCancer: Broad Cancer Detection Across 40+ Types

• Paige PanCancer Detect (RUO) — Developed using Paige's foundation model for computational pathology, this application identifies foci suspicious for cancer across more than 40 cancer types. It supports case triage through prediagnostic screening of biopsy and resected tissue, as well as postdiagnostic quality control. The application achieved a 0.95 specimen-level AUC across nine common and seven rare cancer types.

Paige Predict: Tissue Optimization and Biomarker Prediction

• Paige Predict: Tissue Optimization (RUO) — Forecasts total nucleic acid (TNA) yield from an H&E image to identify samples at risk of NGS failure, helping researchers make informed decisions prior to molecular profiling and conserve scarce tissue.
• Paige Predict: Biomarker Prediction (RUO) — Predicts the status of approximately 1,600 biomarkers across 505 genes from a single H&E slide, delivering results in approximately 5 minutes. This accelerates sample selection for NGS, hotspot testing, IHC, and special stains, and enables pre-screening of cohorts at scale to support research and clinical study enrollment. Predictions are particularly valuable in cancer types with limited tissue availability, such as NSCLC, or less well-tested indications such as melanoma and endometrial cancer.

Paige Predict Deployed Within Tempus Clinical Workflows

• Tempus has integrated Paige Predict's tissue optimization capabilities into its clinical laboratory NGS workflow with no additional ordering steps required, reducing tissue utilization by 18% and expected to lower the overall quantity not sufficient (QNS) rate by 15%.
• Paige Predict's biomarker predictions for 123 clinically relevant features across 16 cancer cohorts have been validated for clinical use as a Tempus laboratory developed test and are performed for eligible xT QNS cases, helping to inform confirmatory testing strategies.

Applications in Limited Feedback Release

• Paige Breast Detect & Subtype (RUO)
• Paige GI Detect & Subtype (RUO)
• Paige GYN Detect & Subtype (RUO)
• Paige Bladder Detect & Subtype (RUO)

Paige Digital Pathology applications support flexible deployment models, allowing laboratories to integrate Paige applications within their existing digital pathology ecosystem — including their preferred scanner, image management system (IMS), and viewer — or to leverage Tempus's own image management infrastructure. Compatible scanners include the Philips Ultra Fast Scanner, Leica Aperio AT2, Leica GT450, and Hamamatsu S360, with visualization supported through the Paige FullFocus Image Viewer. All Paige Predict results are probabilistic predictions and are not intended to determine treatment eligibility or serve as evidence of actual biomarker status.