Cognitive Machine-Learning Algorithm for Cardiac Imaging: A Pilot Study for Differentiating Constrictive Pericarditis...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Imaging

Background:

• Clinical judgment relies on associating patient profiles with prototypical cases from experience.
• Cognitive computing tools may learn and recall multidimensional attributes from medical data.
• Speckle tracking echocardiography (STE) generates complex data for cardiac diagnoses.

Purpose of the Study:

• To test a machine-learning algorithm for differentiating constrictive pericarditis (CP) from restrictive cardiomyopathy (RCM) using STE data.
• To evaluate the performance of an associative memory classifier (AMC) in cardiac diagnosis.
• To compare AMC performance with traditional echocardiographic parameters.

Main Methods:

• Developed an AMC using STE data from 50 CP patients and 44 RCM patients, normalized against 47 controls.
• Evaluated the diagnostic area under the receiver operating characteristic curve (AUC) for differentiating CP from RCM.
• Compared AMC performance with early diastolic mitral annular velocity and left ventricular longitudinal strain.

Main Results:

• AMC achieved an AUC of 89.2% using only STE variables, improving to 96.2% with additional echocardiographic data.
• AMC outperformed traditional parameters like early diastolic mitral annular velocity (AUC 82.1%) and left ventricular longitudinal strain (AUC 63.7%).
• AMC demonstrated superior accuracy and faster learning curves compared to other machine-learning methods.

Conclusions:

• A cognitive machine-learning approach is feasible for analyzing echocardiographic patterns.
• Integrating machine-learning algorithms can standardize cardiac imaging assessments.
• These tools can enhance interpretation quality, particularly for novice readers.