ASN's Mission

ASN leads the fight to prevent, treat, and cure kidney diseases throughout the world by educating health professionals and scientists, advancing research and innovation, communicating new knowledge, and advocating for the highest quality care for patients.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on Twitter

Kidney Week

Abstract: PO1022

Prediction of Stenosis in Arteriovenous Fistula Using Video Image Analysis

Session Information

Category: Dialysis

  • 703 Dialysis: Vascular Access

Authors

  • Zhu, Fansan, Renal Research Institute, New York, New York, United States
  • Wang, Lin-Chun, Renal Research Institute, New York, New York, United States
  • Cherif, Alhaji, Renal Research Institute, New York, New York, United States
  • Thwin, Ohnmar, Renal Research Institute, New York, New York, United States
  • Tisdale, Lela, Renal Research Institute, New York, New York, United States
  • Tao, Xia, Renal Research Institute, New York, New York, United States
  • Paneque Galuzio, Paulo, Renal Research Institute, New York, New York, United States
  • Shtaynberg, Norbert, Azura Vascular Care, New York, New York, United States
  • Preddie, Dean C., Azura Vascular Care, New York, New York, United States
  • Kotanko, Peter, Renal Research Institute, New York, New York, United States
Background

We developed a video image processing (VIP) technique with frequency domain analysis to predict stenosis in AVF. The study aimed to evaluate whether the degree of stenosis can be assessed using parameters from frequency domain signal analysis.

Methods

We employed VIP in 100 hemodialysis patients (age 63.3 ±14.1 years, 47 females) prior to endovascular arterio-venous fistula (AVF) interventions. A 1-minute video of the AVF area was recorded using a smartphone (Fig 1). We constructed time series based on pixel changes between two consecutive video frames (Fig 2-A) and used Fast Fourier Transform (FFT) to transform the time domain signals into the frequency domain (Fig 2-B). Parameters in the frequency domain included maximal (Max) and minimum (Min) amplitude, and frequency (FMax and FMin). ΔF was defined as FMin - FMax. M2 was calculated by the squared ratio of the Max-to-median magnitude. The degree of AVF stenosis (%ST) was determined by angiography, the access flow (AF) by thermodilution (HVT100; Transonic Systems Inc., Ithaca, NY, USA).

Results

Data from 98 patients were analyzed. %ST was categorized into three groups: 60% stenosis (n=8), 70-80% (n=76), and 90% (n=14). AF correlated with %ST. Max, FMax, FMin and ΔF were associated with %ST (Table 1). An algorithm was developed to predict degree of %ST based on patient characteristics and parameters of frequency domain analysis. In the respective three %ST groups the sensitivities to detect AVF stenoses were 88%, 86% and 100%, and the specificities 99%, 82% and 98% (Table 1).

Conclusion

VIP applied to videos taken with a smartphone may provide a contact-free method to estimate the degree of AVF stenosis. Validation studies in independent cohorts are needed to further assess the diagnostic capability of the proposed method.