Oral Potentially Malignant Disorders Clinical and Image Data for AI Research

Supporting data for a doctoral thesis on AI-driven evaluation of oral potentially malignant disorders. The package contains two datasets: longitudinal clinical records for patients with oral leukoplakia and lichenoid mucositis, and high-resolution clinical photographs of oral epithelial dysplasia and leukoplakia. Author Jingwen Li published the data on figshare in April 2026 under a CC-BY-NC-4.0 license.

Use Cases

• Developing predictive models for malignant transformation based on longitudinal clinical data.
• Training image classification models for oral epithelial dysplasia based on high-resolution clinical photographs.
• Evaluating AI integration for the visual assessment of dysplastic features in oral leukoplakia.

Strengths

• Data is ethically approved and anonymized, with IRB approval from the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Reference: UW23-094).
• Comprehensive documentation for variables, coding systems, and metadata is reported to be included within the spreadsheets.

Limitations

• Column-level documentation is absent from the provided metadata; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

Hospital Authority Clinical Management System (HA-CMS) and clinical image curation.

Collection Method

Retrieved from clinical systems and curated for research.

Time Range

null

Freshness

Last updated 2026-04 23 07:50:03; freshness should be verified.

Geography

Likely Hong Kong, based on the approving institution.