Physiologically-guided signal processing

Digital Health Tools for non-invasive diagnosis of lung disease 

Pulmonary hypertension (PH) is a strong prognostic factor in many cardiovascular diseases, such as heart failure, valvular disease and congenital heart disease. Driven by the need for reliable non-invasive techniques, engineering tools must be developed to monitor, detect and predict PH using algorithms which integrate standard echocardiographic images of tricuspid regurgitant and pulmonary flow. Artificial intelligence methods can automate such image analyses, improving robustness. The three specific aims are (1) to develop simple algorithm to automate the assessment of tricuspid regurgitation signals, (2) to establish an artificial intelligence-based framework for automated detection of signal quality, and (3) to investigate physiological features for standardized assessment of right heart function. If proven effective, these algorithms will reduce the need to perform serial risky invasive procedures, improve the early detection and treatment of PH, and increase patient survival.

Publications

Frida C. Nordgren Hermansson, B.E. Celestin, S.P. Bagherzadeh, E. J. Santana, N. Mantri, Roham T. Zamanian, A. L. Marsden, M.Salerno, S.A. Dual, F. Haddad. Cardiac Contour Analysis: A Novel Method for Right Ventricular Assessment in Healthy Controls and Patients with Pulmonary Arterial Hypertension. Manuscript submitted

P. Sattar, C. Verdonk, F.Hermansson, X.Tang, A. Marsden, F. Haddad, S.A. Dual. Enhanced Quality Assessment of Echocardiographic Images for Pulmonary Hypertension Using Convolutional Neural Networks. 2024 Computing in Cardiology (CinC). pp.1‒4  https://cinc.org/archives/2024/pdf/CinC2024-430.pdf

S.A. Dual, C. Verdonk, M. Amsallem, J. Pham, C. Obasohan, P. Nataf, D.B. McElHinney, T. Kuznetsova, R. Zamanian, J. Feinstein, A. Marsden, F. Haddad, Improving Right Ventricular Systolic Pressure Estimation using a novel Doppler Signal Interpolation Method. Pulmonary Circulation. 12:e12125, 2022.