Evaluation of an automatic intraluminal edge detection technique for intravascular ultrasound images

Automatic Lumen Border Detection in IVUS Images Using Deep Learning Model and Handcrafted Features

In the clinical analysis of Intravascular ultrasound (IVUS) images, the lumen size is an important indicator of coronary atherosclerosis, and is also the premise of coronary artery disease diagnosis and interventional treatment. In this study, a fully automatic method based on deep learning model and handcrafted features is presented for the detection of the lumen borders in IVUS images. First, 193 handcrafted features are extracted from the IVUS images. Then hybrid feature vectors are constructed by combining handcrafted features with 64 high-level features extracted from U-Net. In order to obtain the feature subsets with larger contribution, we employ the extended binary cuckoo search for feature selection. Finally, the selected 36-dimensional hybrid feature subset is used to classify the test images using dictionary learning based on kernel sparse coding. The proposed algorithm is tested on the publicly available dataset and evaluated using three indicators. Through ablation experiments, mean value of the experimental results (Jaccard: 0.88, Hausdorff distance: 0.36, Percentage of the area difference: 0.06) prove to be effective improving lumen border detection. Furthermore, compared with the recent methods used on the same dataset, the proposed method shows good performance and high accuracy.

Automated three-dimensional assessment of coronary artery anatomy with intravascular ultrasound scanning

BACKGROUND

Angiography allows the definition of advanced, severe stages of coronary artery disease, but early atherosclerotic lesions, which do not lead to luminal stenosis, are not identified reliably. In contrast, intravascular ultrasound scanning allows the precise characterization and quantification of a wide range of atherosclerotic lesions, independent of the severity of luminal stenosis.

METHODS

Three-dimensional (3-D) reconstruction of entire coronary segments is possible with the integration of sequential 2-dimensional tomographic images and allows volumetric analysis of coronary arteries.

RESULTS

Automated systems able to recognize lumen and vessel borders and to display 3-D images are becoming available.

CONCLUSION

These systems have the potential for on-line 3-D image reconstruction for clinical decision-making and fast routine volumetric analysis in research studies. This review describes 3-D intravascular ultrasound scanning acquisition, analysis, and processing, and the associated technical challenges.

Validation of an automated system for luminal and medial-adventitial border detection in three-dimensional intravascular ultrasound

The precise tomographic assessment of coronary artery disease by intravascular ultrasound (IVUS) is useful in quantitative studies. Such studies require identification of luminal and medial-adventitial (MA) borders in a sequence of IVUS images. We have developed a three-dimensional (3D) active-surface system for border detection that facilitates the analysis of many images with minimal user interaction. To assess the validity of the technique, luminal and MA borders in 529 end-diastolic images from nine coronary arterial segments (58.8 +/- 14.2 images per patient) were traced manually by four experienced observers. The computer-detected borders were compared with borders determined by the four observers using a modified Williams' index (WI), the ratio of inter-observer variability to computer-observer variability. While manual tracing required 49.2 +/- 12.1 min for analysis, the analysis system identified luminal (R2 = 0.92) and MA borders (R2 = 0.97) in 13.8 +/- 4.0 min, a decrease of 35.4 min (p < 0.000001). The computer minus observer differences in lumen area and MA area were -0.88 +/- 0.90 and -0.07 +/- 0.63 mm2. Therefore, the computer system underestimated both lumen and MA area, but this effect was very small in MA area. The WI values and 95% confidence intervals were 0.98 (0.89,1.06) for luminal border detection and 0.99 (0.95,1.04) for MA border detection. Plaque volume measurements, a common endpoint of clinical trials, also verified the accuracy of the technique (R2 = 0.98). The proposed 3D active-surface border detection system provides a faster and less-tedious alternative to manual tracing for assessment of coronary artery anatomy in vivo.

Mechanism of lumen enlargement with direct stenting versus predilatation stenting: influence of remodelling and plaque characteristics assessed by volumetric intracoronary ultrasound

OBJECTIVE

To compare the effects of arterial remodelling and plaque characteristics on the mechanisms of direct stenting and predilatation stenting. Direct stenting has become routine in some laboratories and differs technically from predilatation stenting.

METHODS

Pre- and post-interventional volumetric intravascular ultrasound (IVUS) was undertaken in 30 patients with direct stenting and in 30 with predilatation stenting of non-calcified native coronary lesions, using the same stent design and stent length. Lumen, vessel (external elastic membrane (EEM)), and plaque (plaque + media) volumes were calculated. Remodelling was determined by comparing the EEM area at the centre of the lesion with the EEM areas at proximal and distal reference sites. Plaque eccentricity was defined as the thinnest plaque diameter to the thickest plaque diameter ratio. Plaque composition was characterised as soft, mixed, or dense.

RESULTS

All volumetric IVUS changes were similar in the two groups. Pre-intervention remodelling remained uninfluenced after direct stenting, but was neutralised after predilatation stenting. Eccentric lesions responded to intervention by a greater luminal gain owing to greater vessel expansion in direct stenting. Plaque composition influenced luminal gain in direct stenting, the gain being greatest in the softest plaques; in predilatation stenting, luminal gain was equivalent but vessel expansion was greater for "dense" plaque and plaque reduction greater for "soft" plaque.

CONCLUSIONS

In non-calcified lesions, the mechanisms of lumen enlargement after direct or predilatation stenting are significantly influenced by atherosclerotic remodelling, plaque eccentricity, and plaque composition.