Back
For best output, select "Paper Size" as "A4" and "Margin" as "0" or "None".
To save or print to PDF, please select Print Destination > Save as PDF, enable Background Graphics under "More Settings", then click "Save".
During the congress, E-Posters will be accessible to all participants on the congress website 24/7, as well as in the E-poster stations in the congress center.
Preparing your E-Poster
Please review the E-Poster format requirements carefully when preparing your E-Poster. Should your E-Poster not meet the mentioned requirements, it may not be displayed as described above.
E-Poster Submission Deadline
Please prepare and upload your E-Poster no later than March 14, 2026 11.59PM CET. After this date, you will no longer be able to prepare and upload your E-poster and it will not be displayed and accessible on the congress website.
Please follow the instructions below to input your abstract title.
Abstract titles should be brief and reflect the content of the abstract.
To date, several studies have been reported on shunt sounds that correlate with brachial blood flow volume (FV) measured using vascular ultrasound. However, only a few have identified acoustic signatures, which are combinations of acoustic features that are highly correlated with the FV. In this study, we used machine learning to explore the acoustic signatures that are useful for distinguishing FV. Furthermore, we examined whether integrating patient background factors improves the accuracy of FV prediction.
This retrospective study was conducted at a single institution. We targeted 51 audio recordings obtained from 33 patients who underwent arteriovenous fistula surgery at our hospital, followed by a vascular ultrasound examination and shunt sound recordings. Shunt sounds directly above the anastomosis were recorded using an electronic stethoscope (Littmann CORE Digital Stethoscope) for 10 seconds and saved as WAV files. The saved audio recordings were converted into acoustic features using Python libraries to extract 48 acoustic features, including physical and perceptual features. These features were grouped into three categories: physical, perceptual, and combined. Using machine learning (random forest and recursive feature elimination with cross-validation), the five acoustic features that contributed the most to classifying the FV (threshold of 400 mL/min) were identified for each group, and the performance of the models was comparatively evaluated. Next, we constructed an "acoustic-only model" and an "integrated model" that included clinical information such as age and body weight, and compared the performance of both models using group-based cross-validation on a per-patient basis. Performance was evaluated using the Area Under the Curve (AUC).
Among the three groups, the perceptual group showed the highest performance (AUC 0.942), with Mel-Frequency Cepstral Coefficients and spectral contrast identified as important features for FV discrimination. Although the integrated model, which included clinical information, demonstrated a higher AUC than the acoustic-only model (0.942 vs. 0.970), no statistically significant difference was observed between the two models (p>0.05).
Shunt acoustic signature analysis is a promising noninvasive method for assessing shunt flow. Although the integration of clinical information has shown the potential for improving accuracy, validation with larger datasets is necessary to demonstrate its statistical superiority.
