Non-Invasive Pressure-Only Aortic Wave Intensity Evaluation Using Hybrid Fourier Decomposition-Machine Learning...

Area of Science:

• Cardiovascular Physiology
• Biomedical Engineering
• Machine Learning

Background:

• Wave intensity (WI) analysis is clinically significant for diagnosing cardiovascular and cerebrovascular diseases.
• Current WI methods require simultaneous pressure and flow waveform measurements, limiting clinical translation.
• A Fourier-based machine learning (F-ML) approach was developed to overcome these limitations.

Purpose of the Study:

• To develop and validate a novel F-ML approach for estimating WI using only pressure waveform data.
• To assess the accuracy and reliability of the pressure-only F-ML model compared to traditional methods.

Main Methods:

• Utilized tonometry recordings of carotid pressure and ultrasound measurements of aortic flow from the Framingham Heart Study (2640 participants).
• Developed and blind-tested a Fourier-based machine learning model using pressure waveform data.
• Compared F-ML estimates with method-derived WI parameters.

Main Results:

• The F-ML model showed significant correlations for forward wave peak amplitudes (Wf1, r=0.88; Wf2, r=0.84) and peak times (Wf1, r=0.80; Wf2, r=0.97).
• Strong correlations were observed for backward wave components (Wb1 amplitude, r=0.71; peak time, r=0.60).
• The pressure-only F-ML model outperformed the analytical reservoir model, with negligible bias shown by Bland-Altman analysis.

Conclusions:

• The proposed pressure-only F-ML approach accurately estimates WI parameters.
• This method expands the clinical utility of WI analysis to non-invasive and cost-effective settings, including telemedicine.