Retrospective Pathology Foundation Models

Development and Clinical Application of Deep Learning-Based Retrospective Pathology Foundation Models

By integrating retrospective multimodal data such as pathology and imaging, AI technologies offer novel solutions for disease classification, tumor grading, histological and molecular subtyping, selection of chemotherapy regimens, risk stratification, and treatment-response prediction. This research direction not only deepens our understanding of tumor biological characteristics but also provides essential support for precision medicine and individualized therapy. It holds significant theoretical and practical value and has important implications for mitigating strained medical resources and improving the accuracy of therapeutic decision-making, representing a cutting-edge application with substantial translational potential.

Study Overview

• Status: Enrolling by invitation
• Conditions: Pancancer

• Study Type: Observational
• Enrollment (Estimated): 2000

Contacts and Locations

Study Locations

• China
  • Guangdong
    • Guangzhou, Guangdong, China, 510515
      • Nanfang Hospital, Southern Medical University
  • Shandong
    • Jinan, Shandong, China, 250014
      • Qianfoshan Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

Inclusion criteria comprised patients aged 18-75 years with definitive pathological diagnoses. All cases were retrospectively collected from Nanfang Hospital, Southern Medical University (NFHSMU) .

Description

Inclusion Criteria:

1. Aged 18-75 years old.
2. Patients with complete pathological slides and clinical information.

Exclusion Criteria:

1.Patients with missing data or specimens not meeting quality control requirements for analysis.

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort
QFSH external validation dataset; 1000 slides from 1000 eligible individuals were obtained in the Qianfoshan Hospital (QFSH, Jinan, China) between January 2020 and July 2025, which was used to validate the pathology foundation models.
NFH dataset; We conducted a validation study to compare the diagnostic performance among pathologists, our pathology foundation model, and pathologist-with-AI-assisted diagnosis. This study was initiated at Nanfang Hospital, Southern Medical University (NFHSMU), with patient enrollment from January 1, 2011 to July 31, 2024.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Area under ROC curve (AUC)	Area under the curve	Diagnostic evaluation will be performed within 1 week when the WSIs are obtained

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Specificity	The true negative rate (TNR) of the diagnostic platform, which is the ratio between the number of negative individuals correctly categorized by platform and the total number of actual negative individuals (%).	Diagnostic evaluation will be performed within 1 week when the WSIs are obtained
Sensitivity	The true positive rate (TPR) of the diagnostic platform, which is the ratio between the number of positive individuals correctly categorized by platform and the total number of actual positive individuals (%).	Diagnostic evaluation will be performed within 1 week when the WSIs are obtained

Collaborators and Investigators

• Sponsor: Nanfang Hospital, Southern Medical University
• Investigators: Study Director: Li Liang, Nanfang Hospital, Southern Medical University

Study record dates

Study Major Dates

• Study Start (Actual): August 1, 2025
• Primary Completion (Estimated): July 1, 2027
• Study Completion (Estimated): July 1, 2028

Study Registration Dates

• First Submitted: November 16, 2025
• First Submitted That Met QC Criteria: November 16, 2025
• First Posted (Actual): November 20, 2025

Study Record Updates

• Last Update Posted (Actual): November 20, 2025
• Last Update Submitted That Met QC Criteria: November 16, 2025
• Last Verified: November 1, 2025

More Information

Terms related to this study

• Other Study ID Numbers: NFEC-2025-419

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: Requests for the data collected and analyzed in this study will be considered if the application is in line with public benefits and the applicant is willing to sign a data access agreement. Contact can be through the corresponding author.