Extraction of the best frames in coronary angiograms for diagnosis and analysis

Abstract

X-ray coronary angiography has been a gold standard in the clinical diagnosis and interventional treatment of coronary arterial diseases for decades. In angiography, a sequence of images is obtained, a few of which are suitable for physician inspection. This paper proposes an automatic algorithm for the extraction of one or more frames from an angiogram sequence, which is most suitable for diagnosis and analysis by experts or processors. The algorithm consists of two stages: In the first stage, the background and illumination in the angiogram sequence are omitted. By analyzing the histogram of the sequence, a feature is attributed to each frame. These features, determining the visibility of the vessel tree, are clustered by a fuzzy c-means method. In the second stage, the cardiac phase for each frame is specified. Using the results of both stages, the best frames in an angiogram sequence are obtained. To evaluate the proposed method, it has been tested on angiogram sequences from several patients. The results demonstrate the accuracy of the method. The performance and speed of our method indicate its usefulness in clinical applications.

TCT-231 An artificial intelligence based solution for fully automated cardiac phase and end-diastolic frame detection on coronary angiographies

“…Currently, the selection of the end-diastolic frame is performed either manually or automatically based on the simultaneously acquired ECG signal [5]. The ECG signal may not always be available, and the ECGbased cardiac phase detection has several drawbacks: the signal-to-noise ratio may be too low to accurately detect end-diastole or the signal may present artefacts [11], [12]. Herein we report, to the best of our knowledge, the first deep learning based workflow for purely image-based cardiac phase classification of angiographic frames validated on a large, real-world dataset.…”

“…Currently, the selection of the EDF and the identification of the cardiac phase are performed either manually or automatically based on simultaneously acquired ECG signals [5]. This has a number of drawbacks: ECG signals may not always be available, and cardiac phase detection based on ECG can be challenging if the signal-to-noise ratio is too low to accurately detect end-diastole or the signal presents artefacts [11], [12]. Methods for automated cardiac phase detection on medical images, without the need for processing ECG signals, have been described in the past for use cases involving cardiac echocardiographic images [13][14][15][16][17], cardiac angiographic images [18], cardiac MRI [19][20], and intravascular ultrasound [21].…”

“…Methods for automated cardiac phase detection on medical images, without the need for processing ECG signals, have been described in the past for use cases involving cardiac echocardiographic images [13][14][15][16][17], cardiac angiographic images [18], cardiac MRI [19][20], and intravascular ultrasound [21]. Furthermore, a proof-of-concept study focusing on image based cardiac phase detection on invasive coronary angiography was previously proposed and evaluated on a dataset consisting of 15 patients [12]. Deep learning (DL), and specifically convolutional neural networks (CNNs), represent the state-of-the-art machine learning technique that has been shown to provide unprecedented performance at learning patterns in medical images, providing great promise for helping physicians during clinical decision making processes [22].…”

Retraction: Extraction of the best frames in coronary angiograms for diagnosis and analysis

“…The original article entitled “Extraction of the Best Frames in Coronary Angiograms for Diagnosis and Analysis, published in Journal of Medical Signal and Sensors, on pages 150-157, Issue 3, Volume 6, 2016,[ 1 ] has a number of unattributed sections of content with high rate of similarity, with an article titled “Local feature fitting active contour for segmenting vessels in angiograms”, published in IET Computer Vision, on pages 161-170, Issue 3, Volume 8, 2014. [ 2 ] Plagiarism, unethical publication or redundant publication violates the editorial policy of Journal of Medical Signal and Sensors, which follows best practice guidelines given by the International Committee of Medical Journal Editors (ICMJE) and Committee on Publication Ethics (COPE) mentioned on the Information for Authors and as codified in the signed statements made by the authors regarding the copyright of their work.…”