Deep Learning-Based Method of Diagnosing Hyperlipidemia and Providing Diagnostic Markers Automatically.

Area of Science:

• Artificial Intelligence in Medicine
• Deep Learning for Healthcare
• Medical Diagnostics

Background:

• Auxiliary diagnosis systems face challenges in interpretability and credibility.
• Traditional diagnostic marker research is time-consuming and costly.
• A model for automatic diagnosis and basis provision is needed.

Purpose of the Study:

• To develop an attention-based deep learning tool for hyperlipidemia diagnosis.
• To automatically predict diagnostic markers using hematological parameters.
• To enhance the interpretability and reliability of AI-driven medical diagnosis.

Main Methods:

• An attention deep learning algorithm was employed.
• Human physiological parameters were used as model input.
• The model's attention layer provided a diagnostic basis.

Main Results:

• The model achieved 94% accuracy, 97.48% AUC, 96% sensitivity, and 92% specificity.
• Predicted diagnostic markers for hyperlipidemia were in full agreement with gold standards.
• The system demonstrated robust performance on clinical data.

Conclusions:

• The developed system provides accurate preliminary diagnoses for hyperlipidemia.
• It shows potential for discovering novel diagnostic markers.
• The tool can improve clinical diagnosis efficiency and shorten research timelines.