Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Hemodialysis I: Introduction

Hemodialysis I: Introduction

Hemodialysis (HD) is a medical treatment that artificially removes waste products, excess fluids, and toxins from the blood when the kidneys are no longer able to perform these functions effectively. In this process, blood is filtered through a semipermeable membrane, allowing for the selective removal of waste while preserving necessary components like blood cells and proteins. Hemodialysis is typically performed in patients with end-stage renal disease (ESRD) or severe kidney...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Renal R2* as a noninvasive predictor of hepatorenal syndrome-acute kidney disease in cirrhotic patients with ascites: a retrospective cohort study.

Quantitative imaging in medicine and surgery·2026

Same author

A Clinical-Imaging Score Reclassifies Indeterminate Diastolic Dysfunction and Predicts Posttransplant Cardiac Events in Cirrhosis.

Clinical transplantation·2026

Same author

Bone marrow R2* correlates with liver iron load and is associated with decreased survival in patients with cirrhosis without liver iron overload.

European radiology·2026

Same author

Low-Dose Neostigmine at 0.7 Train-of-Four Ratio Safely Accelerates Neuromuscular Recovery: A Prospective Observational Study.

Drug design, development and therapy·2026

Same author

Catheter-Directed Sclerotherapy Versus Ovarian Cystectomy for Unilateral Ovarian Endometrioma: A Pilot Randomized Controlled Trial.

AJR. American journal of roentgenology·2026

Same author

Uterine Artery Embolization for Pure Adenomyosis: Predictive Factors Affecting Outcomes.

Journal of vascular and interventional radiology : JVIR·2026

Search research articles

Related Experiment Video

Updated: Aug 8, 2025

A Novel Murine Model of Arteriovenous Fistula Failure: The Surgical Procedure in Detail

A Novel Murine Model of Arteriovenous Fistula Failure: The Surgical Procedure in Detail

Published on: February 3, 2016

A deep learning algorithm to quantify AVF stenosis and predict 6-month primary patency: a pilot study.

Jae Hyon Park¹, Jongjin Yoon¹, Insun Park²

¹Department of Radiology, Yonsei University College of Medicine, Seoul, Republic of Korea.

Clinical Kidney Journal

|March 3, 2023

Summary

This summary is machine-generated.

A novel deep convolutional neural network (DCNN) model effectively predicts arteriovenous fistula (AVF) stenosis and 6-month patency using shunt sounds. This AI approach surpasses traditional machine learning models for improved AVF management.

Keywords:

angioplasty arteriovenous fistula auscultation convolutional neural network primary patency

More Related Videos

Murine Model of Central Venous Stenosis using Aortocaval Fistula with an Outflow Stenosis

Murine Model of Central Venous Stenosis using Aortocaval Fistula with an Outflow Stenosis

Published on: July 11, 2019

Author Spotlight: Establishment and Confirmation of a Postnatal Right Ventricular Volume Overload Mouse Model

Author Spotlight: Establishment and Confirmation of a Postnatal Right Ventricular Volume Overload Mouse Model

Published on: June 9, 2023

Related Experiment Videos

Last Updated: Aug 8, 2025

A Novel Murine Model of Arteriovenous Fistula Failure: The Surgical Procedure in Detail

A Novel Murine Model of Arteriovenous Fistula Failure: The Surgical Procedure in Detail

Published on: February 3, 2016

Murine Model of Central Venous Stenosis using Aortocaval Fistula with an Outflow Stenosis

Murine Model of Central Venous Stenosis using Aortocaval Fistula with an Outflow Stenosis

Published on: July 11, 2019

Author Spotlight: Establishment and Confirmation of a Postnatal Right Ventricular Volume Overload Mouse Model

Author Spotlight: Establishment and Confirmation of a Postnatal Right Ventricular Volume Overload Mouse Model

Published on: June 9, 2023

Area of Science:

Medical Imaging and Artificial Intelligence
Cardiovascular Surgery
Nephrology

Background:

Arteriovenous fistula (AVF) dysfunction is a common complication in hemodialysis patients.
Accurate prediction of AVF stenosis and primary patency (PP) is crucial for effective patient management.
Current methods for assessing AVF status can be invasive or lack precision.

Purpose of the Study:

To develop and evaluate a deep convolutional neural network (DCNN) model for predicting AVF stenosis and 6-month primary patency (PP).
To assess the model's performance using AVF shunt sounds converted to melspectrograms.
To compare the DCNN model's diagnostic performance against various machine learning (ML) models utilizing clinical data.

Main Methods:

Prospective recruitment of 40 dysfunctional AVF patients.
Recording AVF shunt sounds pre- and post-percutaneous transluminal angioplasty using a wireless stethoscope.
Converting audio recordings to melspectrograms for analysis by a ResNet50-based DCNN model.
Comparing the DCNN model with logistic regression, decision tree, and support vector machine models trained on clinical data.

Main Results:

Melspectrograms qualitatively correlated with AVF stenosis severity, showing higher amplitude at mid-to-high frequencies during systole for more severe stenosis.
The melspectrogram-based DCNN model accurately predicted the degree of AVF stenosis.
The DCNN model achieved a superior area under the receiver operating characteristic curve (≥0.870) for predicting 6-month PP compared to ML models using clinical data (LR: 0.783, DT: 0.766, SVM: 0.733) and a spiral-matrix DCNN model (0.828).

Conclusions:

The developed melspectrogram-based DCNN model demonstrates high accuracy in predicting AVF stenosis.
This AI-driven approach significantly outperforms traditional ML models based on clinical data for predicting 6-month AVF primary patency.
The findings suggest a promising non-invasive method for monitoring AVF health and improving patient outcomes.