ISPAD: An Explainable AI Framework for Pernicious Anemia Diagnosis

A 305.3 KB PDF authored by Nora Boumela, last updated on 2026-04-28. It describes the ISPAD (Intelligent System for Pernicious Anemia Diagnosis) framework, an explainable AI-based probabilistic system designed to integrate heterogeneous clinical information for diagnosing pernicious anemia. The document presents a proof-of-concept approach tested on published diagnostically challenging cases.

Use Cases

• Developing explainable AI models for medical diagnosis based on the described probabilistic framework.
• Formalizing complex diagnostic reasoning for autoimmune gastritis and pernicious anemia based on the described methodology.
• Integrating heterogeneous clinical, histological, immunological, and biochemical information for diagnostic support as outlined in the document.
• Testing AI systems on challenging diagnostic cases like antibody assay interference or seronegative presentations as mentioned in the results.

Strengths

• The dataset is a structured, peer-reviewed research document (305.3 KB PDF) describing a formal diagnostic framework.
• The methodology addresses a specific, complex clinical problem (pernicious anemia diagnosis) with heterogeneous data sources.
• The work is licensed under CC-BY-4.0, allowing for open sharing and adaptation.

Limitations

• The primary content is a descriptive PDF (305.3 KB), not a traditional dataset of records; row and column counts are unknown.
• Column-level documentation is absent; the document describes a framework rather than a structured data table.
• The data scale is very limited, consisting of a single document file.

Provenance

Source

figshare, authored by Nora Boumela.

Collection Method

Describes the development and testing of the ISPAD AI framework on published case studies.

Freshness

Last updated 2026-04-28 10:46:03; freshness should be verified.