Experimental Sensitivity Analysis for LVAD Obstruction Detection Using Acoustic Sensing
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
This research explores the sensitivity of sensor selection, flow rate, and phantom tissue thickness associated with a simplified left ventricular assist device (LVAD) system. Previous studies have applied obstruction detection methods for LVADs in mock circulatory loops, however, insight into the design considerations of the experimental setup related to variation in operating conditions is missing. The goal of this research is to assess the sensitivity of obstruction detection in a simplified flow setting to inform the experimental design of LVAD studies and further the development of an AI-enabled digital stethoscope. Acoustic recordings were collected using two digital stethoscopes for several operating conditions, including three occlusion severities (0\%, 25\%, and 50\% by diameter), three phantom tissue thicknesses (6.35, 12.7, and 19.05 mm), and three flow rates (2, 5, and 10 LPM). Recordings from each sensor were recorded in separate instances to assess variability and repeatability. The dataset was used to characterize the influence of severity, flow rate, and attenuation on the relevant condition monitoring signals. Frequency domain features were extracted using an FFT and used to train support vector machine (SVM) and random forest (RF) models. The second part of the study considers pulsed flow, which is closer to the reality of LVAD operating conditions, where the flow through the LVAD is dependent on the pressure drop across the pump, which is highly influenced by native heart function. Here, the continuous wavelet transform (CWT) was applied to the acoustic measurements to analyze the time-frequency domain at specific regions of interest. Spectral features were extracted from the CWT regions and were used to train SVM and RF models. Detection of obstructions is sensitive to sensor selection, and the operating condition boundaries of each sensor are apparent. The commercially available sensor performs well at higher flow rates and with thinner phantom tissue. Conversely, the custom sensor is effective at low-to-medium flow rates and maintains effectiveness across a range of tissue thicknesses. This work will directly inform experimental design decisions for a custom mock circulatory loop LVAD study and provide the basis for the development of an AI-enabled stethoscope specifically for LVADs.
How to Cite
##plugins.themes.bootstrap3.article.details##
LVAD, AI-enabled Stethoscope, Sensitivity, Condition Monitoring
Baronti, L., Castellani, M., Hefft, D., & Alberini, F. (2022, March). Neural network identification of water pipe blockage from smart embedded passive acoustic measurements. Can J Chem Eng, 100(3), 521–539.
Bhatnagar, R., Fonarow, G. C., Heidenreich, P. A., & Ziaeian, B. (2022). Expenditure on Heart Failure in the United States. JACC: Heart Failure, 10(8), 571–580.
Bordoloi, D. J., & Tiwari, R. (2017, Aug). Identification of suction flow blockages and casing cavitations in centrifugal pumps by optimal support vector machine techniques. Journal of the Brazilian Society of Mechanical Sciences and Engineering, 39(8), 2957–2968.
Bozkurt, B., Ahmad, T., Alexander, K., Baker, W. L., Bosak, K., Breathett, K., . . . Ziaeian, B. (2025). HF STATS 2024: Heart Failure Epidemiology and Outcomes Statistics An Updated 2024 Report from the Heart Failure Society of America. J Card Fail, 31(1), 66–116. (Publisher: Elsevier)
Cao, S., Hu, Z., Luo, X., & Wang, H. (2021, March). Research on fault diagnosis technology of centrifugal pump blade crack based on PCA and GMM. Measurement, 173, 108558.
Cortes, C., & Vapnik, V. (1995, September). Support-vector networks. Machine Learning, 20(3), 273–297.
D’Souza, G. A., Rinaldi, J. E., Meki, M., Crusan, A., Richardson, E., Shinnar, M., & Herbertson, L. H. (2023, November). Using a Mock Circulatory Loop as a Regulatory Science Tool to Simulate Different Heart Failure Conditions. Journal of Biomechanical Engineering, 146(011004).
Fonarow, G. C., Ahmad, F. S., Ahmad, T., Albert, N. M., Alexander, K. M., Baker, W. L., . . . Ziaeian, B. (2026). HF STATS 2025: Heart Failure Epidemiology and Outcomes Statistics An Updated 2025 Report from the Heart Failure Society of America. Journal of Cardiac Failure, 32(2), 439–498.
Gwak. (2015). Model-Referenced Cardiovascular Circulatory Simulator Construction and Control. Artif. Ograns.
Kim, H. C., Park, S. K., Hindrichs, H. L., Bearnson, G. B., Khanwilkar, P. S., & Olsen, D. B. (1995, sep). Noninvasive Diagnosis of Mechanical Failure of the Implanted Total Artificial Heart Using Neural Network Analysis of Acoustic Signals. ASAIO Journal, 41(3), M271.
Kumar, A. (2023, March). Support Vector Machine (SVM) Python Example.
Lee, S., Fukamachi, K., Golding, L., Moazami, N., & Starling, R. C. (2013, May). Left Ventricular Assist Devices: From the Bench to the Clinic. Cardiology, 125(1), 1–12. (Num Pages: 12 Place: Basel, Switzerland Publisher: S. Karger AG)
Misumi, Y., Miyagawa, S., Yoshioka, D., Kainuma, S., Kawamura, T., Kawamura, A., . . . Sawa, Y. (2021, June). Prediction of aortic valve regurgitation after continuous-flow left ventricular assist device implantation using artificial intelligence trained on acoustic spectra. J Artif Organs, 24(2), 164–172.
Muralidharan, V., Sugumaran, V., & Indira, V. (2014, September). Fault diagnosis of monoblock centrifugal pump using SVM. Engineering Science and Technology, an International Journal, 17(3), 152–157.
Panda, A. K., Rapur, J. S., & Tiwari, R. (2018, December). Prediction of flow blockages and impending cavitation in centrifugal pumps using Support Vector Machine (SVM) algorithms based on vibration measurements. Measurement, 130, 44–56.
Parke, W. C. (2020). Biophysics: A Student’s Guide to the Physics of the Life Sciences and Medicine. Cham: Springer International Publishing.
Prechtl, I. R., Day, S. W., & Kolodziej, J. R. (2018). An Acoustic Time-frequency Approach to Condition Monitoring of Ventricular Assistive Devices. Proc. Annu. Conf. Progn. Health Manag. Soc. PHM.
Rapur, J. S., & Tiwari, R. (2019, December). Experimental fault diagnosis for known and unseen operating conditions of centrifugal pumps using MSVM and WPT based analyses. Measurement, 147, 106809.
Reuter, S. P. (2019). LVAD Occlusion Condition Monitoring Using Boost Classification Trees (Unpublished master’s thesis). Rochester Institute of Technology, Rochester New York. (ISBN: 9781392340943)
Semiz, B., Hersek, S., Pouyan, M. B., Partida, C., Blazquez- Arroyo, L., Selby, V., . . . Inan, O. T. (2020, July). Detecting Suspected Pump Thrombosis in Left Ventricular Assist Devices via Acoustic Analysis. IEEE Journal of Biomedical and Health Informatics, 24(7), 1899–1906. (Conference Name: IEEE Journal of Biomedical and Health Informatics)
Vaidya, Y., Riaz, S., & Dhamoon, A. S. (2025). Left Ventricular Assist Devices. In StatPearls. Treasure Island (FL): StatPearls Publishing.
Vapnik, V. (1999, September). An overview of statistical learning theory. IEEE Transactions on Neural Networks, 10(5), 988–999.
Wood, J., & Barry, D. (1995, March). Time-frequency analysis of the first heart sound. IEEE Engineering in Medicine and Biology Magazine, 14(2), 144–151. (Conference Name: IEEE Engineering in Medicine and Biology Magazine)
Yost, G. L., Royston, T. J., Bhat, G., & Tatooles, A. J. (2016). Acoustic Characterization of Axial Flow Left Ventricular Assist Device Operation In Vitro and In Vivo. ASAIO journal (1992), 62(1), 46. (Place: United States)
Zhao, Q., Geng, S., Wang, B., Sun, Y., Nie, W., Bai, B., . . . Hong, S. (2024, January). Deep Learning in Heart Sound Analysis: From Techniques to Clinical Applications. Health Data Sci, 4, 0182.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.