Computer-assisted image processing 12 lead ECG model to diagnose hyperkalemia.

Area of Science:

• Cardiology
• Medical Diagnostics
• Biomedical Engineering

Background:

• Hyperkalemia, a condition of elevated serum potassium, poses significant health risks.
• Traditional electrocardiogram (ECG) analysis has limitations in accurately diagnosing hyperkalemia.
• There is a need for improved diagnostic models for hyperkalemia.

Purpose of the Study:

• To develop and validate an improved 12-lead ECG model for diagnosing hyperkalemia.
• To incorporate novel ECG parameters alongside traditional ones for enhanced diagnostic accuracy.
• To assess the sensitivity and specificity of the developed ECG model.

Main Methods:

• Retrospective analysis of ECGs from hyperkalemic and normokalemic patients.
• Development of a predictive model using general linear mixed models and standardized variables.
• Utilized T-wave width, descending T-wave slope, and QRS prolongation as key parameters.
• Optimized sensitivity and specificity through ROC-AUC analysis.

Main Results:

• The developed ECG model demonstrated a ROC-AUC of 0.78 in the validation set.
• The model achieved a maximum specificity of 84% for detecting potassium levels above 5.91 mEq/L.
• The model incorporated descending T-wave slope, T-wave width, and QRS prolongation.

Conclusions:

• The novel ECG model significantly improves hyperkalemia diagnosis compared to traditional methods.
• Incorporating T-wave width, descending T-wave slope, and QRS prolongation enhances diagnostic capabilities.
• This model offers a more accurate non-invasive tool for identifying hyperkalemia.