Improved automated lumen contour detection by novel multifrequency processing algorithm with current intravascular ultrasound system.

OBJECTIVES: The aim of this study was to evaluate a new fully automated lumen border tracing system based on a novel multifrequency processing algorithm. BACKGROUND: We developed the multifrequency processing method to enhance arterial lumen detection by exploiting the differential scattering characteristics of blood and arterial tissue. The implementation of the method can be integrated into current intravascular ultrasound (IVUS) hardware. METHODS: This study was performed in vivo with conventional 40-MHz IVUS catheters (Atlantis SR Pro™, Boston Scientific Corp, Natick, MA) in 43 clinical patients with coronary artery disease. A total of 522 frames were randomly selected, and lumen areas were measured after automatically tracing lumen borders with the new tracing system and a commercially available tracing system (TraceAssist™) referred to as the "conventional tracing system." The data assessed by the two automated systems were compared with the results of manual tracings by experienced IVUS analysts. RESULTS: New automated lumen measurements showed better agreement with manual lumen area tracings compared with those of the conventional tracing system (correlation coefficient: 0.819 vs. 0.509). When compared against manual tracings, the new algorithm also demonstrated improved systematic error (mean difference: 0.13 vs. -1.02 mm(2) ) and random variability (standard deviation of difference: 2.21 vs. 4.02 mm(2) ) compared with the conventional tracing system. CONCLUSIONS: This preliminary study showed that the novel fully automated tracing system based on the multifrequency processing algorithm can provide more accurate lumen border detection than current automated tracing systems and thus, offer a more reliable quantitative evaluation of lumen geometry.

Challenges in atherosclerotic plaque characterization with intravascular ultrasound (IVUS): from data collection to classification.

In vivo plaque characterization is an important research field in interventional cardiology. We will study the realistic challenges to this goal by deploying 40 MHz single-element, mechanically rotating transducers. The intrinsic variability among the transducers' spectral parameters as well as tissue signals will be demonstrated. Subsequently, we will show that global data normalization is not suited for data calibration, due to the aforementioned variations as well as the stringent characteristics of spectral features. We will describe the sensitivity of an existing feature extraction algorithm based on eight spectral signatures (integrated backscatter coefficient, slope, midband-fit (MBF), intercept, and maximum and minimum powers and their relative frequencies) to a number of factors, such as the window size and order of the autoregressive (AR) model. It will be further demonstrated that the variations in the transducer's spectral parameters (i.e., center frequency and bandwidth) cause inconsistencies among extracted features. In this paper, two fundamental questions are addressed: 1) what is the best reliable way to extract the most informative features? and 2) which classification algorithm is the most appropriate for this problem? We will present a full-spectrum analysis as an alternative to the eight-feature approach. For the first time, different classification algorithms, such as k-nearest neighbors (k-NN) and linear Fisher, will be employed and their performances quantified. Finally, we will explore the reliability of the training dataset and the complexity of the recognition algorithm and illustrate that these two aspects can highly impact the accuracy of the end result, which has not been considered until now.

Characterisation of atherosclerotic plaque by spectral similarity of radiofrequency intravascular ultrasound signals.

AIMS: The aim of this paper is to report a method of atherosclerotic plaque tissue characterisation based on pattern recognition and assess its accuracy under conditions of potential clinical relevance. METHODS AND RESULTS: Excised saline infused human arteries were imaged using IVUS with RF acquisition. 40% of the vessels were re-imaged with human blood infusion. A database of approximately 12000 image regions-of-interest (ROIs) of histologically established types was used to design a pattern recognition algorithm to predict the tissue type of a given ROI by comparing its RF-spectrum against the database, and also to estimate the confidence of prediction. Ex vivo validation demonstrated accuracies at the highest level of confidence as: 97%, 98%, 95%, and 98% for necrotic, lipidic, fibrotic and calcified regions respectively. Good agreement with histology was shown in an in vivo swine animal model. CONCLUSIONS: Ex vivo validation demonstrated the ability to characterise plaque tissue using an IVUS+RF system and a method incorporating (1) full spectral information (2) spectral similarity (3) estimating confidence of characterisation and, (4) ability to characterise plaque imaged through blood. Promising results were demonstrated in a live animal model. This approach may have potential for accurate and reproducible plaque characterisation in vivo.