New Predictive Diagnostic Method for Cardiac Dynamics Based on Probability Distributions

Area of Science:

• Cardiology
• Biomedical Engineering
• Data Science

Background:

• Holter monitor evaluation is complex.
• Diagnostic support tools leveraging probability theory and dynamic systems can simplify this process.
• Existing methods may lack accuracy in differentiating various cardiac dynamics.

Purpose of the Study:

• To develop and validate a diagnostic support tool for cardiac dynamics analysis.
• To differentiate between normal, chronic, acute, and pacemaker cardiac conditions using frequency ranges and probability theory.
• To assess the tool's diagnostic performance in terms of sensitivity, specificity, and kappa coefficient.

Main Methods:

• A study involving 80 Holter tests (over 21 hours each) from adult patients.
• Development of four prototypes based on normal, chronic, acute, and pacemaker diagnoses.
• Application of probability theory to frequency repetition ranges (1000-2000 and 2001-3000 Hz) within a defined heart rate probability space.
• Blinded analysis of remaining Holter tests using the developed methodology.

Main Results:

• The diagnostic tool demonstrated high accuracy in differentiating cardiac dynamics.
• Sensitivity, specificity, and kappa coefficient were calculated to be 1.
• Probability analysis across different frequency ranges effectively distinguished normal from abnormal (chronic, acute, pacemaker) cardiac states.

Conclusions:

• A novel diagnostic support tool for cardiac dynamics has been successfully developed.
• The tool utilizes frequency range appearance and probability theory for clinical applications.
• The developed tool enables accurate differentiation of normal, chronic, acute, and pacemaker cardiac dynamics.