Towards Clinical-Grade Pathology AI: From Whole-Slide Diagnosis to Clinically Validated Foundation Models

PhD Thesis Proposal Defence


Title: "Towards Clinical-Grade Pathology AI: From Whole-Slide Diagnosis to Clinically
Validated Foundation Models"

by

Mr. Zhengrui GUO


Abstract:

The field of computational pathology has advanced rapidly with the 
digitization of whole-slide images and the emergence of pathology foundation 
models, but clinical-grade pathology AI requires more than high performance 
on retrospective benchmarks. A deployable system must reason over gigapixel 
tissue context, learn from limited annotations, communicate findings in 
clinically meaningful language, and be validated under realistic diagnostic 
workflows. This proposal focuses on developing a staged research progression 
towards clinical-grade pathology AI: context-aware whole-slide image 
diagnosis through query-aware long-context modeling, label-efficient evidence 
discovery for few-shot whole-slide image classification, visual-to-language 
translation through histopathology report generation, and a lung-specific 
pathology foundation model with retrospective, external, prospective, and 
pathologist-AI interaction evaluation. Together, these studies support the 
central argument that pathology AI becomes clinically meaningful only when 
representation learning, data efficiency, semantic output, and clinical 
validation are considered together.


Date:                   Monday, 18 May 2026

Time:                   10:00am - 12:00noon

Venue:                  Room 2128A
                        Lift 19

Committee Members:      Dr. Hao Chen (Supervisor)
                        Dr. Dan Xu (Chairperson)
                        Dr. Terence Wong (CBE)