Machine learning-based model for identifying liver injury in patients with thyroid-associated ophthalmopathy

Area of Science:

• Ophthalmology
• Hepatology
• Medical Informatics

Background:

• Thyroid-associated ophthalmopathy (TAO) is an autoimmune condition affecting the eyes.
• Liver injury can be a complication in patients with TAO, necessitating early detection.
• Developing effective diagnostic tools for liver injury in TAO patients is crucial.

Purpose of the Study:

• To investigate the association between TAO and liver injury.
• To develop and validate a machine learning model for identifying liver injury in TAO patients.
• To provide an effective diagnostic aid for liver damage in this population.

Main Methods:

• Retrospective analysis of clinical data from 318 TAO patients (2016-2022).
• Patients categorized into liver injury (104) and normal liver function (214) groups.
• Machine learning models (logistic regression, random forest, SVM, decision tree) were trained and tested for liver injury prediction.

Main Results:

• Random forest identified key risk factors for liver injury.
• The random forest (RF) model achieved high accuracy (0.937) and AUC (0.977) in cross-validation.
• Support vector machine (SVM) demonstrated the highest AUC (0.986) in the test set, with SVM and logistic regression (LR) achieving 0.914 accuracy.

Conclusions:

• Machine learning classification models can accurately identify the risk of liver damage in TAO patients.
• These models offer a promising approach for early diagnosis and management of liver injury in TAO.
• The study highlights the potential of AI in improving diagnostic capabilities for TAO-related complications.