Automated cardiac auscultation for detection of pathologic heart murmurs

Area of Science:

• Cardiology
• Biomedical Engineering
• Medical Diagnostics

Background:

• Distinguishing pathologic heart murmurs from innocent ones can be challenging, especially for less experienced clinicians.
• Certain murmur characteristics, like harsh, pansystolic murmurs at the left upper sternal border (LUSB) with intensity grade ≥3, are often indicative of underlying pathology.

Purpose of the Study:

• To develop and evaluate an automated system for detecting systolic heart murmurs associated with various cardiac conditions.
• To assess the correlation between the relative intensity of murmurs and the likelihood of cardiac pathology.

Main Methods:

• Recorded and digitized cardiac auscultatory examinations from 194 children and young adults.
• Utilized automated spectral analysis with continuous wavelet transforms for murmur detection.
• Compared findings between a control group (no heart disease, n=95) and a patient group with diagnosed cardiac conditions and pathologic murmurs (n=99).

Main Results:

• The automated system achieved 96% sensitivity and specificity for detecting pathologic murmurs with intensity grade ≥2.
• For murmurs with intensity grade ≥3, the system demonstrated 100% sensitivity and specificity.
• Significant correlation found between murmur intensity and the likelihood of pathology.

Conclusions:

• Automated cardiac auscultation shows high accuracy in identifying pathologic systolic murmurs.
• This technology can serve as a valuable diagnostic aid, supporting clinical decision-making in cardiology.
• Potential to improve the accuracy and efficiency of heart murmur assessment.