Statistical modeling of atherosclerotic plaque in carotid B mode images--a feasibility study

A Study on the Effects of Depth-Dependent Power Loss on Speckle Statistics Estimation

Characterization of the interference patterns observed in B-mode images (i.e., speckle statistics) is a valuable tool in tissue characterization. However, changes in echo amplitudes unrelated to speckle, including power loss due to attenuation and diffraction, can bias these metrics, undermining their utility. Tissue with high attenuation such as the uterine cervix are especially affected. The purpose of this study was to demonstrate and quantify the effects of attenuation and diffraction on speckle statistics and to propose methods of compensation. Analysis was performed on simulated diffuse scattering phantoms of varying attenuation with simulated transducers at 9 and 5 MHz center frequency. Application in the in vivo macaque cervix using a clinical scanner is also presented. Nakagami and homodyned K distribution parameters were calculated in parameter estimation regions (PERs) of varying size within simulations and experiments. Changes in speckle statistics parameters with respect to PER size and depth were compared with and without two different compensation schemes. It has been shown that compensation for attenuation and diffraction is necessary to produce speckle statistics estimates that do not depend on medium attenuation or PER size. Reducing the dependence on these factors connects speckle statistics estimates more closely with the microstructure of the probed medium.

Review of Envelope Statistics Models for Quantitative Ultrasound Imaging and Tissue Characterization

The homodyned K-distribution and the K-distribution, viewed as a special case, as well as the Rayleigh and the Rice distributions, viewed as limit cases, are discussed in the context of quantitative ultrasound (QUS) imaging. The Nakagami distribution is presented as an approximation of the homodyned K-distribution. The main assumptions made are (1) the absence of log-compression or application of nonlinear filtering on the echo envelope of the radiofrequency signal and (2) the randomness and independence of the diffuse scatterers. We explain why other available models are less amenable to a physical interpretation of their parameters. We also present the main methods for the estimation of the statistical parameters of these distributions. We explain why we advocate the methods based on the X-statistics for the Rice and the Nakagami distributions and the K-distribution. The limitations of the proposed models are presented. Several new results are included in the discussion sections, with proofs in the appendix.

Carotid artery ultrasound image analysis: A review of the literature

Stroke is one of the prominent causes of death in the recent days. The existence of susceptible plaque in the carotid artery can be used in ascertaining the possibilities of cardiovascular diseases and long-term disabilities. The imaging modality used for early screening of the disease is B-mode ultrasound image of the person in the artery area. The objective of this article is to give a widespread review of the imaging modes and methods used for studying the carotid artery for identifying stroke, atherosclerosis and related cardiovascular diseases. We encompass the review in methods used for artery wall tracking, intima-media, and lumen segmentation which will help in finding the extent of the disease. Due to the characteristics of the imaging modality used, the images have speckle noise which worsens the image quality. Adaptive homomorphic filtering with wavelet and contourlet transforms, Levy Shrink, gamma distribution were used for image denoising. Learning-based neural network approaches for denoising give better edge preservation. Domain knowledge-based segmentation approaches have proved to provide more accurate intima-media thickness measurements. There is a requirement of useful fully automatic segmentation approaches, 3D, 4D systems, and plaque motion analysis. Taking into consideration the image priors like geometry, imaging physics, intensity and temporal data, image analysis has to be performed. Encouragingly more research has focused on content-specific segmentation and classification techniques. With the evaluation of machine learning algorithms, classifying the image as with or without a fat deposit has gained better accuracy and sensitivity. Machine learning-based approaches like self-organizing map, k-nearest neighborhood and support vector machine achieve promising accuracy and sensitivity in classification. The literature reveals that there is more scope in identifying a patient-specific model in a fully automatic manner.

Assessment of Carotid Artery Plaque Components With Machine Learning Classification Using Homodyned-K Parametric Maps and Elastograms

Quantitative ultrasound (QUS) imaging methods, including elastography, echogenicity analysis, and speckle statistical modeling, are available from a single ultrasound (US) radio-frequency data acquisition. Since these US imaging methods provide complementary quantitative tissue information, characterization of carotid artery plaques may gain from their combination. Sixty-six patients with symptomatic ( n = 26 ) and asymptomatic ( n = 40 ) carotid atherosclerotic plaques were included in the study. Of these, 31 underwent magnetic resonance imaging (MRI) to characterize plaque vulnerability and quantify plaque components. US radio-frequency data sequence acquisitions were performed on all patients and were used to compute noninvasive vascular US elastography and other QUS features. Additional QUS features were computed from three types of images: homodyned-K (HK) parametric maps, Nakagami parametric maps, and log-compressed B-mode images. The following six classification tasks were performed: detection of 1) a small area of lipid; 2) a large area of lipid; 3) a large area of calcification; 4) the presence of a ruptured fibrous cap; 5) differentiation of MRI-based classification of nonvulnerable carotid plaques from neovascularized or vulnerable ones; and 6) confirmation of symptomatic versus asymptomatic patients. Feature selection was first applied to reduce the number of QUS parameters to a maximum of three per classification task. A random forest machine learning algorithm was then used to perform classifications. Areas under receiver-operating curves (AUCs) were computed with a bootstrap method. For all tasks, statistically significant higher AUCs were achieved with features based on elastography, HK parametric maps, and B-mode gray levels, when compared to elastography alone or other QUS alone ( ). For detection of a large area of lipid, the combination yielding the highest AUC (0.90, 95% CI 0.80-0.92, ) was based on elastography, HK, and B-mode gray-level features. To detect a large area of calcification, the highest AUC (0.95, 95% CI 0.94-0.96, ) was based on HK and B-mode gray level features. For other tasks, AUCs varied between 0.79 and 0.97. None of the best combinations contained Nakagami features. This study shows the added value of combining different features computed from a single US acquisition with machine learning to characterize carotid artery plaques.

Echolucency-based phenotype in carotid atherosclerosis disease for risk stratification of diabetes patients

AIM

The study investigated the association of carotid ultrasound echolucent plaque-based biomarker with HbA1c, measured as age-adjusted grayscale median (AAGSM) as a function of chronological age, total plaque area, and conventional grayscale median (GSM_conv).

METHODS

Two stages were developed: (a) automated measurement of carotid parameters such as total plaque area (TPA); (b) computing the AAGSM as a function of GSM_conv, age, and TPA. Intra-operator (novice and experienced) analysis was conducted.

RESULTS

IRB approved, 204 patients' left/right CCA (408 images) ultrasound scans were collected: mean age: 69 ± 11 years; mean HbA1c: 6.12 ± 1.47%. A moderate inverse correlation was observed between AAGSM and HbA1c (CC of -0.13, P = 0.01), compared to GSM (CC of -0.06, P = 0.24). The RCCA and LCCA showed CC of -0.18, P < 0.01 and -0.08; P < 0.24. Female and males showed CC of -0.29, P < 0.01 and -0.10, P = 0.09. Using the threshold for AAGSM and HbA1c as: low-risk (AAGSM > 100; HbA1c < 5.7%), moderate-risk (40 < AAGSM < 100; 5.7% < HbA1c < 6.5%) and high-risk (AAGSM < 40; HbA1c > 6.5%), the area under the curve showed a better performance of AAGSM over GSM_conv. A paired t-test between operators and expert (P < 0.0001); inter-operator CC of 0.85 (P < 0.0001).

CONCLUSIONS

Echolucent plaque in patients with diabetes can be more accurately characterized for risk stratification using AAGSM compared to GSM_conv.

Automated localization and segmentation techniques for B-mode ultrasound images: A review

B-mode ultrasound imaging is used extensively in medicine. Hence, there is a need to have efficient segmentation tools to aid in computer-aided diagnosis, image-guided interventions, and therapy. This paper presents a comprehensive review on automated localization and segmentation techniques for B-mode ultrasound images. The paper first describes the general characteristics of B-mode ultrasound images. Then insight on the localization and segmentation of tissues is provided, both in the case in which the organ/tissue localization provides the final segmentation and in the case in which a two-step segmentation process is needed, due to the desired boundaries being too fine to locate from within the entire ultrasound frame. Subsequenly, examples of some main techniques found in literature are shown, including but not limited to shape priors, superpixel and classification, local pixel statistics, active contours, edge-tracking, dynamic programming, and data mining. Ten selected applications (abdomen/kidney, breast, cardiology, thyroid, liver, vascular, musculoskeletal, obstetrics, gynecology, prostate) are then investigated in depth, and the performances of a few specific applications are compared. In conclusion, future perspectives for B-mode based segmentation, such as the integration of RF information, the employment of higher frequency probes when possible, the focus on completely automatic algorithms, and the increase in available data are discussed.

Evaluation of carotid plaque echogenicity based on the integral of the cumulative probability distribution using gray-scale ultrasound images

OBJECTIVE

Carotid plaque echogenicity is associated with the risk of cardiovascular events. Gray-scale median (GSM) of the ultrasound image of carotid plaques has been widely used as an objective method for evaluation of plaque echogenicity in patients with atherosclerosis. We proposed a computer-aided method to evaluate plaque echogenicity and compared its efficiency with GSM.

METHODS

One hundred and twenty-five carotid plaques (43 echo-rich, 35 intermediate, 47 echolucent) were collected from 72 patients in this study. The cumulative probability distribution curves were obtained based on statistics of the pixels in the gray-level images of plaques. The area under the cumulative probability distribution curve (AUCPDC) was calculated as its integral value to evaluate plaque echogenicity.

RESULTS

The classification accuracy for three types of plaques is 78.4% (kappa value, κ = 0.673), when the AUCPDC is used for classifier training, whereas GSM is 64.8% (κ = 0.460). The receiver operating characteristic curves were produced to test the effectiveness of AUCPDC and GSM for the identification of echolucent plaques. The area under the curve (AUC) was 0.817 when AUCPDC was used for training the classifier, which is higher than that achieved using GSM (AUC = 0.746). Compared with GSM, the AUCPDC showed a borderline association with coronary heart disease (Spearman r = 0.234, p = 0.050).

CONCLUSIONS

Our experimental results suggest that AUCPDC analysis is a promising method for evaluation of plaque echogenicity and predicting cardiovascular events in patients with plaques.

Identification of Ultrasonic Echolucent Carotid Plaques Using Discrete Fréchet Distance Between Bimodal Gamma Distributions

OBJECTIVE

Echolucent carotid plaques are associated with acute cardiovascular and cerebrovascular events (ACCEs) in atherosclerotic patients. The aim of this study was to develop a computer-aided method for identifying echolucent plaques.

METHODS

A total of 315 ultrasound images of carotid plaques (105 echo-rich, 105 intermediate, and 105 echolucent) collected from 153 patients were included in this study. A bimodal gamma distribution was proposed to model the pixel statistics in the gray scale images of plaques. The discrete Fréchet distance features (DFDFs) of each plaque were extracted based on the statistical model. The most discriminative features (MDFs) were obtained from DFDFs by the linear discriminant analysis, and a k-nearest-neighbor classifier was implemented for classification of different types of plaques.

RESULTS

The classification accuracy of the three types of plaques using MDFs can reach 77.46%. When a receiver operating characteristics curve was produced to identify echolucent plaques, the area under the curve was 0.831.

CONCLUSION

Our results indicate potential feasibility of the method for identifying echolucent plaques based on DFDFs.

SIGNIFICANCE

Our method may potentially improve the ability of noninvasive ultrasonic examination in risk prediction of ACCEs for patients with plaques.

Impact of element pitch on synthetic aperture ultrasound imaging

PURPOSE

Synthetic aperture imaging was introduced in medical ultrasound to obtain high-quality images. In synthetic aperture ultrasound imaging, spherical transmit waves illuminate a target region from different positions, resulting in low-resolution images for each transmission. By coherent compounding of the resulting low-resolution images, a high-resolution image is obtained. Multiple steered receiving beams need to be created to obtain each low-resolution image and, thus, grating lobes should influence the image quality. In the present study, an array ultrasonic probe with a small element pitch was introduced to reduce the influences of grating lobes, and the effect of element pitch on image quality was examined in detail.

METHOD

A linear array ultrasonic probe at a nominal center frequency of 7.5 MHz with an element pitch of 0.1 mm has been introduced. This probe does not produce grating lobes within the imaging region in theory because the element pitch of this probe is half of the ultrasonic wavelength. The contrast of an ultrasonic image was evaluated using a cyst phantom.

RESULTS

The contrasts obtained by synthetic aperture imaging with element pitches of 0.1 and 0.2 mm were 4.88 and 4.69 dB, respectively, which were similar to the 4.67 dB obtained by conventional beamforming with focused transmit beams, when the number of transmissions was 121. The contrast obtained with an element pitch of 0.1 mm was similar (4.34 dB) even when the number of transmissions was decreased to 61. However, the contrast obtained with an element pitch of 0.2 mm showed a larger degradation (3.77 dB) at 31 transmissions.

DISCUSSION AND CONCLUSION

Even with larger element pitches, good image contrast could be obtained when the number of transmissions was large. This is because echoes from grating lobes are incoherent among transmissions, and they are suppressed by compounding low-resolution images obtained by individual transmissions. On the other hand, an array probe with smaller element pitches achieves good image contrast even with a smaller number of transmissions and, thus, it would be preferable to realize a higher frame rate.

Statistics of boundaries in ultrasonic B-scan images

The existence of edges and boundaries in regions of interest (ROIs) in B-scan images alters the statistics of the backscattered echo from the ROI. Boundaries are the result of at least two different types of scattering scenarios in tissue, and the Nakagami model, which is being used extensively in ultrasound, is unlikely to fit the statistics of the backscattered echo under these conditions. Furthermore, there are very few other statistical models exist that describe the statistics of the backscattered echo from regions containing boundaries. In this work, the gamma mixture density and the recently proposed McKay density are explored as two viable models to fill this void. Justifications of these models are presented along with methods for estimating their parameters. Random number simulations and studies on tissue-mimicking phantoms indicate that the McKay and gamma mixture densities are the best for the modeling of the backscattered echo intensity when boundaries are present in the regions of interest.

Gamma mixture classifier for plaque detection in intravascular ultrasonic images

Carotid and coronary vascular incidents are mostly caused by vulnerable plaques. Detection and characterization of vulnerable plaques are important for early disease diagnosis and treatment. For this purpose, the echomorphology and composition have been studied. Several distributions have been used to describe ultrasonic data depending on tissues, acquisition conditions, and equipment. Among them, the Rayleigh distribution is a one-parameter model used to describe the raw envelope RF ultrasound signal for its simplicity, whereas the Nakagami distribution (a generalization of the Rayleigh distribution) is the two-parameter model which is commonly accepted. However, it fails to describe B-mode images or Cartesian interpolated or subsampled RF images because linear filtering changes the statistics of the signal. In this work, a gamma mixture model (GMM) is proposed to describe the subsampled/interpolated RF images and it is shown that the parameters and coefficients of the mixture are useful descriptors of speckle pattern for different types of plaque tissues. This new model outperforms recently proposed probabilistic and textural methods with respect to plaque description and characterization of echogenic contents. Classification results provide an overall accuracy of 86.56% for four classes and 95.16% for three classes. These results evidence the classifier usefulness for plaque characterization. Additionally, the classifier provides probability maps according to each tissue type, which can be displayed for inspecting local tissue composition, or used for automatic filtering and segmentation.

Fully automatic plaque segmentation in 3-D carotid ultrasound images

Automatic segmentation of the carotid plaques from ultrasound images has been shown to be an important task for monitoring progression and regression of carotid atherosclerosis. Considering the complex structure and heterogeneity of plaques, a fully automatic segmentation method based on media-adventitia and lumen-intima boundary priors is proposed. This method combines image intensity with structure information in both initialization and a level-set evolution process. Algorithm accuracy was examined on the common carotid artery part of 26 3-D carotid ultrasound images (34 plaques ranging in volume from 2.5 to 456 mm(3)) by comparing the results of our algorithm with manual segmentations of two experts. Evaluation results indicated that the algorithm yielded total plaque volume (TPV) differences of -5.3 ± 12.7 and -8.5 ± 13.8 mm(3) and absolute TPV differences of 9.9 ± 9.5 and 11.8 ± 11.1 mm(3). Moreover, high correlation coefficients in generating TPV (0.993 and 0.992) between algorithm results and both sets of manual results were obtained. The automatic method provides a reliable way to segment carotid plaque in 3-D ultrasound images and can be used in clinical practice to estimate plaque measurements for management of carotid atherosclerosis.

Statistical distributions of potential interest in ultrasound speckle analysis

Compound statistical modelling of the uncompressed envelope of the backscattered signal has received much interest recently. In this note, a comprehensive collection of models is derived for the uncompressed envelope of the backscattered signal by compounding the Nakagami distribution with 13 flexible families. The corresponding estimation procedures are derived by the method of moments and the method of maximum likelihood. The sensitivity of the models to their various parameters is examined. It is expected that this work could serve as a useful reference and lead to improved modelling of the uncompressed envelope of the backscattered signal.