Artificial Intelligence-Based Assessment of Indices of Right Ventricular Function

Area of Science:

• Cardiology
• Echocardiography
• Cardiac Imaging

Background:

• Right ventricular (RV) function assessment in echocardiography traditionally relies on subjective visual estimation.
• Objective myocardial performance assessment using regional strain analysis is often limited by vendor-specific software.
• A need exists for a standardized, vendor-neutral approach to RV strain analysis.

Purpose of the Study:

• To investigate the statistical correlation between RV region-specific strain and conventional echocardiographic parameters of RV function.
• To evaluate the utility of a vendor-neutral RV-specific strain assessment program.

Main Methods:

• Retrospective study involving 107 patients undergoing cardiac surgery.
• Comprehensive intraoperative echocardiography of RV function.
• Off-line analysis of global, longitudinal, and septal RV strain using vendor-neutral software.

Main Results:

• Strong statistical significance was found between RV strain values and echocardiographic parameters.
• The strongest correlations were observed between RV fractional area change (FAC) and RV longitudinal strain (r² = 0.83, p < 0.001).
• Significant correlations were also found between tricuspid annular plane systolic excursion and lateral S' velocity (r² = 0.80, p < 0.001).

Conclusions:

• RV function can be objectively assessed using strain analysis across different platforms with vendor-neutral software.
• Statistically significant correlations exist between RV strain values and conventional 2D echocardiographic parameters.
• This approach offers a reliable, objective method for RV function evaluation.