2015 Journal Publications

73.   Automated IVUS-based Coronary Calcium Volumes have Mild correlations with Coronary Lumen Volume, Coronary Vessel Volume and Coronary Atheroma Volume in Japanese Diabetic Cohort

    Tadashi Araki, Sumit K. Banchhor, Narendra D. Londhe, Nobutaka Ikeda, Petia Radeva, Luca Saba, Ayman El-Baz, Andrew Nicolaides, Shoaib Shafique, John R. Laird, Jasjit S. Suri.

    International Angiology.2015. (Under Review). [Impact Factor:1.462]

    Abstract: Two automated IVUS-based coronary calcium volume (aCCV) measurement methods (threshold and k-means-based) and its correlation with three expert coronary parameters namely: coronary vessel volume (eCVV), coronary lumen volume (eCLV), and coronary atheroma volume (eCAV) was determined in a Japanese diabetic cohort. Diabetic patients with stable angina pectoris who underwent percutaneous coronary interventions were eligible for this study. Nineteen patients (17 M/2 F, age: 66±12 years), IRB approved, Toho University, Japan, with written informed consent, consisting of 38,760 coronary intravascular ultrasound (IVUS) frames were collected using 40 MHz IVUS catheter (Atlantis® SR Pro, Boston Scientific®, pullback speed of 0.5 mm/sec). The methodology consisted of automatic compution of the coronary calcium volume using k-means based pixel classification technique. This technique was compared against previously published threshold-based method. The coefficient of correlation (CC) was then established between these two automated coronary calcium volumes and three experts reading namely: eCVV, eCLV and eCAV computed using iMAP software Boston Scientific®. Calcium mean volume for k-means and threshold-based method were 27.79±10.94 mm3, and 35.92±16.44 mm3 respectively. A mild CC was observed between threshold-based and three manual readings (eCVV, eCLV and eCAV): 0.18, 0.17 and 0.16. A moderately higher CC was observed between k-means and three manual readings: 0.24, 0.29 and 0.19. The CC between (i) age and hemoglobin (HbA1c) was 0.31 and between (ii) HbA1c and eCLV was 0.14 respectively. No correlation was observed between HbA1c and eCVV and eCAV. In general, automated coronary volume measurement using k-means method was more strongly correlated with the expert’s readings as compared to previous published threshold-based method in this diabetic cohort.

74.   Accurate Lumen Diameter Measurement in Curved Vessels based on Iterative Spatial Transformation and Scale Space

    Krishna Kumar P, Tadashi Araki, Jeny Rajan, Luca Saba, Francesco Lavra, Nobutaka Ikeda, Aditya M. Sharma, Shoaib Shafique, Andrew Nicolaides, John L Laird, Ajay Gupta, Jasjit S. Suri.

    Medical & Biological Engineering & Computing.2015. (Under Review). [Impact Factor:1.726]

    Abstract: Carotid degree of stenosis (cDOS) using carotid ultrasound is used for monitoring of carotid and cerebrovascular diseases as they are considered as an important predictor of stroke risk. Pursuing lumen size measurement in ultrasound via manual quantification of lumen diameter is subjective, slow, and erroneous. There has been a clear need to automatically measure lumen diameter. Further, due to the curvature associated with the vessels along with non-uniformity in the plaque growth poses challenge to the automated lumen detection and quantification systems. The present study uses a novel and generalized approach for LD measurement based on a combination of spatial transformation and scale-space. It is an iterative process where the scale-space is first used to get the lumen axis which is then used with spatial image transformation paradigm to get transformed image. The scale-space is then again applied to finally retrieve the lumen region and boundary in the transformed framework and later inverse transformed to bring back to the original framework. The LD is measured using polyline distance method. The automated LD measurements were compared and validated against the manual LD borders. 202 patients’ left and right CCA artery (404 carotid images) B-mode ultrasound images retrospectively analyzed (ethics approval with IRB was granted by Toho University, Japan) with mean age 69±15.9 years. Mean HbA1c, LDL, HDL and Cholesterol of patients were 6.28±1.1 mg/dl, 101.27±31.6 mg/dl, 50.26±14.8 mg/dl and 175.04±38 mg/dl, respectively. The validation of our automated system is performed against the two Manual expert tracers. The mean LD for the automated system was 6.58±2.15 mm, while for the two Manual tracings were 6.43±2.10 mm and 6.49±2.14 mm, respectively. The CC between the two Manual tracings for LD were 0.98 (p<0.0001) and 0.99 (p<0.0001), respectively. The precision of merit (PoM) between the manual expert tracings and the automated system were 97.7% and 98.7%, respectively. Image transformation based approach for LD measurement was more accurate compared to conventional approach. The system is fully automated.

75.   Automated Carotid Lumen Diameter: A Biomarker and Link for Identification of High Risk Coronary Artery Disease Patients

    Tadashi Araki, Krishna Kumar P, Nobutaka Ikeda, Ajay Gupta, Luca Saba, Jeny Rajan, Francesco Lavra, Aditya M. Sharma, Shoaib Shafique, Andrew Nicolaides, John L Laird, Jasjit S. Suri.

    Ultrasound in Medicine and Biology.2015. (Under Review). [Impact Factor:2.214]

    Abstract: Carotid lumen diameter (LD) and inter-adventitial diameter (IAD) can be non-invasively measured using B-mode ultrasound, but their relationship to coronary artery disease (CAD) is uncertain. This study has two novel objectives: (i) to automatically measure LD/IAD in light-medium plaque to moderate stenosis in the carotid wall, and compare against two manual tracings; (ii) to measure the correlation coefficient (CC) between carotid LD/IAD and intravascular ultrasound (IVUS) based coronary vessel area (measured using iMAP software – Boston Scientific®) and use carotid LD as a biomarker and link for detection of high risk CAD patients for the risk of myocardial infarction.One hundred and sixty six patients (mean age 69±15.9 years) left and right carotid artery (332 US images) were analyzed retrospectively (IRB approved, Toho University, Japan). Mean HbA1c, LDL, HDL and cholesterol was 5.8±1.0 mg/dl, 99.9±31.3 mg/dl, 50.6±15.5 mg/dl and 174.2±36.6 mg/dl, respectively. In objective one, we showed mean LD & IAD was 6.24±0.87 mm & 7.60±0.92 mm, respectively. The LD/IAD errors against two manual readings were: (a) 0.28±0.26 mm and 0.24±0.26 mm for LD, & (b) 0.23±0.23 mm and 0.24±0.24 mm for IAD, respectively. The intra-observer CC between the two manual tracers for carotid LD and IAD were 0.94 (p<0.0001) and 0.98 (p<0.0001), respectively. In objective two, using an optimization procedure (i.e., coronary VA > 15 mm2 & carotid LD < 4.2 mm), 45 patients for high risk CAD were identified. Best high risk CAD detection accuracy was 83% (with coronary vessel area (VA) > 17 mm2 & carotid LD < 4.2 mm). The CC between stenotic carotid LD/IAD & coronary VA showed moderate-to-high values of -0.37 (p < 0.0134) and -0.35 (p < 0.0198), respectively. The automated carotid LD/IAD system was reliable, accurate, and reproducible. The carotid-coronary relationships are consistent with previous studies.

76.   Longitudinal assessment of carotid atherosclerosis after Radiation Therapy using Computed Tomography: A case control Study.

    Anzidei M, Suri J, Saba L, Sanfilippo R, Laddeo G, Montisci R, Piga M, Argiolas GM, Raz E.

    European radiology.2015 Sep 25. [Epub ahead of print] [Impact Factor: 4.014]

    Abstract: To study the carotid artery plaque composition and its volume changes in a group of patients at baseline and 2 years after head and neck radiation therapy treatment (HNXRT).In this retrospective study, 62 patients (41 males; mean age 63 years; range 52-81) who underwent HNXRT and 40 patients (24 males; mean age 65) who underwent surgical resection of neoplasm and did not undergo HNXRT were assessed, with 2-year follow-up. The carotid artery plaque volumes, as well as the volume of the sub-components (fatty-mixed-calcified), were semiautomatically quantified. Mann-Whitney and Wilcoxon tests were used to test the hypothesis. In the HNXRT group, there was a statistically significant increase in the total volume of the carotid artery plaques (from 533 to 746 mm3; p = 0.001), in the fatty plaques (103 vs. 202 mm3; p = 0.001) and mixed plaque component volume (328 vs. 419 mm3; p = 0.034). A statistically significant variation (from 21.8 % to 27.6 %) in the percentage of the fatty tissue was found. Results of this preliminary study suggest that HNXRT promotes increased carotid artery plaque volume, particularly the fatty plaque component. HNXRT increases carotid plaque volume. • Plaque volume increase is mainly due to increase.in fatty plaque component • Patients who undergo HNXRT have a progression of carotid artery disease.

77.   A Review on Carotid Ultrasound Atherosclerotic Tissue Characterization and Stroke Risk Stratification in Machine Learning Framework

    Aditya M. Sharma, Ajay Gupta, P. Krishna Kumar, Jeny Rajan, Luca Saba, Ikeda Nobutaka, John R. Laird, Andrew Nicolades, Jasjit S. Suri.

    Current Atherosclerosis Reports. 2015 Sep;17(9):529. doi: 10.1007/s11883-015-0529-2.  [Impact Factor: 3.417]

    Abstract: Cardiovascular diseases (including stroke and heart attack) are identified as the leading cause of death in today’s world. However, very little is understood about the arterial mechanics of plaque buildup, arterial fibrous cap rupture, and the role of abnormalities of the vasa vasorum. Recently, ultrasonic echogenicity characteristics and morphological characterization of carotid plaque types have been shown to have clinical utility in classification of stroke risks. Furthermore, this characterization supports aggressive and intensive medical therapy as well as procedures, including endarterectomy and stenting. This is the first state-of-the-art review to provide a comprehensive understanding of the field of ultrasonic vascular morphology tissue characterization. This paper presents fundamental and advanced ultrasonic tissue characterization and feature extraction methods for analyzing plaque. Additionally, the paper shows how the risk stratification is achieved using machine learning paradigms. More advanced methods need to be developed which can segment the carotid artery walls into multiple regions such as the bulb region and areas both proximal and distal to the bulb. Furthermore, multimodality imaging is needed for validation of such advanced methods for stroke and cardiovascular risk stratification.

78.   Association between carotid artery plaque inflammation and brain MRI: Preliminary results

    Luca Saba, Elisabetta Tamponi, Letizia Lai, Pierleone Lucatelli, Roberto Sanfilippo, Roberto Montisci, Jasjit S. Suri, Mario Piga, Gavino Faa.

    AJNR Am J Neuroradiol. 2015. (Under Review). [Impact Factor: 3.167]

    Abstract: To explore the association between presence of inflammatory cells in the carotid plaques surgically treated and brain MRI findings. 40 consecutive patients were prospectively analyzed. Brain MRI was performed using a 1.5 Tesla system and lesions pertinence of the anterior circulation was recorded. All patients underwent carotid endarterectomy “en bloc”; carotid plaques histological sections were prepared and immuno-cytochemical analysis was used to characterize and quantify the presence of inflammatory cells. ROC curve analysis, Pearson Rho correlation and Mann-Whitney test were applied.The immuno-cytochemical analysis demonstrated that ruptured cap plaques of symptomatic patients (stroke\TIA; n=25) had many inflammatory cells, principally macrophages (CD68) compared with those caps observed in ruptured plaques from patients without symptoms (Mann-Whitney = p < 0.001, ROC curve area = 0.901). Correlation analysis showed a statistically significant association between the number of brain non-lacunar infarct and the entity of macrophages (p < 0.001); whereas no association with lacunar infarct (p = 0.1934) was found.Results of this preliminary study suggest that presence of inflammatory cells within carotid artery plaque is associated with cerebrovascular events and with the number of MRI brain detectable infarct.

79.   Inter-Adventitial Diameter is better related to Plaque Score than Lumen Diameter in Diabetic Patients: An Automated Carotid Ultrasound Study

    Luca Saba, Tadashi Araki, Krishna Kumar P, Jeny Rajan, Francesco Lavra, Nobutaka Ikeda, Aditya M. Sharma, Shoaib Shafique, Andrew Nicolaides, John L Laird, Ajay Gupta, Jasjit S. Suri.

    Stroke. 2015. (Under Review) [Impact Factor: 6.158]

    Abstract: Carotid lumen diameter (LD) and inter-adventitial diameter(IAD) can be reliably determined using B-mode ultrasound. This study compares the strength of correlation between automatically measured carotid LD and IAD with Plaque Score (PS). Diabetic patients who underwent bilateral common carotid B-mode ultrasound were eligible for this study. LD and IAD were computed automatically using an advanced method of edge detection. The correlation coefficients (CC) between (a)LD and PS and (b) between IAD and PS were calculated and compared.Our study included 202 patients (155 males, 47 females) with mean age 69±15.9 years. Mean HbA1c, LDL, HDL and Cholesterol of patients were 6.28±1.1 mg/dl, 101.27±31.6 mg/dl, 50.26±14.8 mg/dl and 175.04±38 mg/dl, respectively. The CC between (i) LD and PS was 0.19 (p<0.007), and (ii) IAD and PS was 0.25 (p<0.0006). On excluding the outliers, the corresponding CC between (i) LD and PS was 0.38 (p<0.0001); (ii) IAD and PS was 0.25 (p<0.0001), respectively. We validated our automated LD/IAD measurements against two Manual expert tracers. The corresponding errors for LD were: 0.27±0.26 mm and 0.25±0.24 mm, while for IAD, the errors were: 0.23±0.23 mm and 0.24±0.24 mm (sensitivity:100%, specificity:100%). The intra-observer CC between the two Manual tracings for LD and between the two Manual tracings for IAD were 0.95 (p<0.0001) and 0.98 (p<0.0001), respectively.

80.   Automated calcium burden measurement in internal carotid artery plaque with ct: a hierarchical adaptive approach.

    Saba L, Bhavsar AV, Gupta A, Mtui EE, Giambrone AE, Baradaran H, Lavra F, Laird JR, Nicolaides A, Suri JS.uca Saba, Eytan Raz, Michele di Martino, Roberto Montisci, Jasjit Suri, Mario Piga.

    International Angiology, .2015 Mar 31. [Epub ahead of print] [Impact Factor: 1.462]

    Abstract: Calcium burden measurement in internal carotid artery (ICA) plaque could play an important role in assessing stroke risk and stenosis quantification in the ICA. We propose an automatic method for labelling calcified plaques in ICA in CT images. Our approach builds upon the mean shift paradigm via an adaptive thresholding strategy. The data consists of single CT slices from 75 patients, with variety of plaque sizes and number of calcium regions. The manual measurements were carried out by a neuroradiologist for benchmarking. The calcium burden was measured as the area of the labelled plaque. Various metrics were employed to compare manual and automated measurements including correlation coefficient (CC), dice similarity (DS), Jacard Index (JI), polyline distance metric (PDM) and precision of merit (PoM). We found that our automated method of calcium area characterization performed accurately compared to manual measurements with CC = 0.978, and PoM = 0.915. The PDM, DS, and JI, also indicate a good performance with a mean DS = 0.85 (SD = 0.085), a mean JI = 0.747 (SD = 0.12), and a mean PDM = 0.195 (SD = 0.177). The proposed approach for calcium burden measurement, yields reasonably accurate labelling of calcified plaque when benchmarked against manual measurements. The approach is independent of the number and size of calcium regions, and the prototype design shows encouraging results to be adaptable to clinical practice.

81.   Carotid Inter-Adventitial Diameter is better related to Plaque Score than Lumen Diameter in Diabetic Patients: An Ultrasound Study

    Luca Saba, Tadashi Araki, Krishna Kumar P , Jeny Rajan, Francesco Lavra, Nobutaka Ikeda, Aditya M. Sharma, Shoaib Shafique, Andrew Nicolaides, John L Laird, Ajay Gupta, Jasjit S. Suri.

    Stroke. 2015. (Under Review) [Impact Factor: 6.158]

    Abstract: Duplex ultrasound only provides a wide range of possible level of stenosis. Automated measurement of the carotid lumen and inter-adventitial diameters via B- mode ultrasound provides an alternative way for risk quantification. Plaque score (PS) has been used for the quantification of plaques in the carotid arteries and can be associated with stroke risk and CAD. Our purpose was to compare the strength of correlation between carotid Lumen Diameter (LD) and Inter-Adventitial Diameter (IAD) with Plaque Score (PS) to improve risk stratification. Diabetic patients who underwent bilateral common carotid B-mode ultrasound were eligible for this study. The following imaging features were analyzed. (1) LD/IAD Measurement: A higher order Gaussian derivative filter followed by spectral peak detection method is applied on B-mode carotid US images to find the pair of peaks corresponding to adventitial boundary. Lumen region is isolated using the pixel classification paradigm. (2) Plaque Score Measurement: The PS was computed by adding the maximal thickness in millimeters of plaques in each of the four segments from the ICA and CCA. We compared the coefficient of correlation between plaque score and both IAD and LD. 202 patients (males: 155) who underwent bilateral common carotid artery ultrasound. Our study included 202 patients (155 males) with mean age 69±15.9 years. Mean HbA1c, LDL, HDL and Cholesterol of patients were 6.28±1.1 mg/dl, 101.27±31.6 mg/dl, 50.26±14.8 mg/dl and 175.04±38 mg/dl, respectively. The coefficient of correlation (CC) between PS and IAD was 0.29 and between PS and LD was 0.19. If we remove the 6-10 outliers, it will become 0.38 and 0.25. Plaque score is strongly correlated to carotid IAD than carotid LD.

82.   Wall Sheer Stress (WSS) Analysis in Human In-Vivo Coronary Plaqued Artery – Turbulence Vs. Laminar Flow Modeling: An Investigational Study

    Arun Mahalingam, Udhav Ulhas Gawandalkar, Girish Kini,Abdulrajak Buradi, Tadashi Araki,Nobutaka Ikeda,Andrew Nicolaides, John R. Laird, Luca Saba,Jasjit S. Suri. .

    Abstract: It is established that the local hemodynamics plays a significant role in atherogenisis and the progression of atherosclerosis.The Reynolds number of blood flow through arteries typically ranges from 1 to 4000 [3-Ku] where blood flow is laminar in small arteries and turbulent large arteries. The predisposition of the coronary and carotid arteries to develop atherosclerosis has led many investigators to assess the nature of flow in these vessels. In coronary arteries, the Reynolds number ranges from 188-1671 and hence it is assumed as laminar flow in WSS calculations by several investigators. However, the blood flow through coronary arteries under stenosed condition along with periodic nature of the blood flow can result in transition from laminar to turbulent flow condition. Turbulence in blood flow affects the Wall Shear Stress (WSS) acting on the endothelial cell membrane, mass transport from the blood to the vessel wall, atherogenisis, plaque progression, blood rheology through the formation of rouleaux by RBCs, the hematocrit values etc. Biomechanical forces such as Wall pressure and Wall Shear Stress (WSS) influences vascular remodelling and physiological changes of endothelial cells. In the present work , the price of assuming laminar flow through in-vivo coronary with stenosis has been quantified by comparing hemodynamic parameters like WSS and Oscillatory Shear Index (OSI) for laminar and turbulent flow models. It is observed that the local WSS can differ up to 15% between the laminar and turbulent models and that the laminar model under-predicts the WSS and OSI acting on the vessel wall.

83.   Relationship between white matter hyperintensities volume and the Circle of Willis configurations in patients with carotid artery pathology Stroke

    Luca Saba, Michele di Martino, Paolo Siotto, Pierleone Lucatelli, Michele Anzidei, Jasjit Suri, Carlo Catalano, Mario Piga.

    Stroke. 2015. (Under Review) [Impact Factor: 6.158]

    Abstract: We aimed to assess if there is a difference of distribution and volume of white matter hyperintensities (WMH) in the brain according to the Circle of Willis (CoW) configuration in patients with carotid artery pathology. One-hundred consecutive patients (79 males, 21 females; mean age 70 years; age range 46-84 years) that underwent brain MRI before carotid endarterectomy were included. FLAIR-WMH lesion volume was performed using a semi-automated segmentation technique and the status of the Circle of Willis was assessed by two neuroradiologists in consensus. We found a prevalence of 55% of variants in the CoW configuration; 22 cases had one variants (40%); 25 cases had two variants (45.45%) and 8 cases showed 3 variants (14.55%). The configuration that was associated with the biggest WMH volume and number of lesions was the A1 + PcoA + PcoA. The PcoA variants was the most prevalent and there was no statistically significant difference in number of lesions and WMH for each vascular territory assessed and the same results were found for PcoA and A1 variants. Results of our study suggest that the more common CoW variants are not associated with the presence of an increased WMH or number of lesions whereas uncommon configurations, in particular when 2 or more segment are missing increase the WMH volume and number of lesions. The WHM volume of the MCA territory seems to be more affected by the CoW configuration.

84.   Improved Correlation between Carotid and Coronary Atherosclerosis SYNTAX Score Using Automated Ultrasound Carotid Bulb Plaque IMT Measurement.

    Ikeda N, Gupta A, Dey N, Bose S, Shafique S, Arak T, Godia EC, Saba L, Laird JR, Nicolaides A, Suri JS.

    Ultrasound in Medicine & Biology. 2015 Jan 28. pii: S0301-5629(14)00845-X. doi: 10.1016/j.ultrasmedbio.2014.12.024.  [Impact Factor: 2.416]

    Abstract: Described here is a detailed novel pilot study on whether the SYNTAX (Synergy between percutaneous coronary intervention with TAXUS and cardiac surgery) score, a measure of coronary artery disease complexity, could be better predicted with carotid intima-media thickness (cIMT) measures using automated IMT all along the common carotid and bulb plaque compared with manual IMT determined by sonographers. Three hundred seventy consecutive patients who underwent carotid ultrasound and coronary angiography were analyzed. SYNTAX score was determined from coronary angiograms by two experienced interventional cardiologists. Unlike most methods of cIMT measurement commonly used by sonographers, our method involves a computerized automated cIMT measurement all along the carotid artery that includes the bulb region and the region proximal to the bulb (under the class of AtheroEdge systems from AtheroPoint, Roseville, CA, USA). In this study, the correlation between automated cIMT that includes bulb plaque and SYNTAX score was found to be 0.467 (p < 0.0001), compared with 0.391 (p < 0.0001) for the correlation between the sonographer's IMT reading and SYNTAX score. The correlation between the automated cIMT and the sonographer's IMT was 0.882. When compared against the radiologist's manual tracings, automated cIMT system performance had a lumen-intima error of 0.007818 ± 0.0071 mm, media-adventitia error of 0.0179 ± 0.0125 mm and automated cIMT error of 0.0099 ± 0.00988 mm. The precision of automated cIMT against the manual radiologist's reading was 98.86%. This current automated algorithm revealed a significantly stronger correlation between cIMT and coronary SYNTAX score as compared with the sonographer's cIMT measurements with multiple cardiovascular risk factors. We benchmarked our correlation between the automated cIMT that includes bulb plaque and SYNTAX score against a previously published (Ikeda et al. 2013) AtheroEdgeLink (AtheroPoint) correlation between the automated cIMT that does not include bulb plaque and SYNTAX score and had an improvement of 44.58%. By sampling cIMT in the bulb region, the automated cIMT technique improves the degree of correlation between coronary artery disease lesion complexity and carotid atherosclerosis characteristics.

Back to Top
Home

2014 Journal Publications

61.   Volumetric Distribution of the White Matter Hyper-intensities. Does the side play a role?

    Luca Saba, Michele di Martino, Paolo Siotto Michele Anzidei, Jasjit S Suri, Carlo Catalano, Mario Piga.

    Stroke. 2014. (Under Review) [Impact Factor: 6.158]

    Abstract: Purpose of this paper is to assess if there is a difference of distribution of white matter hyperintensities (WMH) in the brain hemispheres according to the left/right side. Eighty consecutive patients (mean age 71 ± 6 years, males 66) with carotid artery stenosis were prospectively recruited. FLAIR-WMH lesion volume was performed using a semi-automated segmentation technique (Jim, Xinapse System, Leicester, UK). A statistically significant difference was found in the MCA territory for the volume of the lesions (median volume of WMH in the left side = 889.5 mm3; median volume of WMH in the right side = 580.5 mm3; p value = 0.0416); no statistically significant difference were found with the other territories and by considering the number of lesions. A statistically significant difference between the volume of the DWMH and PVWMH (p value = 0.001) was found as well as for the PVWMH volume (p value = 0.0088) between right and left side. Results of our study suggest that there is a difference of distribution of WMH in the brain hemispheres according to the left/right side in the MCA territories and for the peri-ventricular white matter.

62.   Accuracy of Automated Calcium Burden Assessment in Calcified Atherosclerotic Plaque of the Cervical Internal Carotid Artery

    Ajay Gupta, Arnav V. Bhavsar, Edward E. Mtui, Hediyeh Baradaran, Francesco Lavra, John R. Laird, Luca Saba, Jasjit S. Suri.

    International Angiology. 2014. (Under Review) [Impact Factor: 1.462]

    Abstract: Studying calcium burden and its measurements in internal carotid artery (ICA) is imperative for assessing stroke risk and stenosis quantification in the ICA. In this study, an automatic method which builds upon classification of CT slices is proposed and evaluated for labelling calcified plaques in ICA. Computed tomography (CT) imaging of diseased internal carotid arteries were acquired from 85 patients. Manual data for calcified plaque measurement was collected using vascular radiologist. An automatic measurement based on mean-shift method is employed to label calcified plaque in single slices from each of the 85 CT images. The calcium burden was measured as the area of the labelled calcified plaque. The data spanned 1 to 3 number of calcified regions, and a wide variety of calcium burden plaque sizes. A variety of statistical analysis was carried out to compare results of the automated labelled area and manually labelled area. This included regression plot, Bland-Altman (BA) plot and Cumulative frequency of error. The regression plot showed a very high correlation between the manual and automated area with a Correlation Coefficient (CC) of 0.978. The BA plot and the cumulative frequency of error plot also depicted a small range of error between manual and automated methods. The Precision of Merit (POM), was also high (POM = 0.901). The proposed calcium burden measurement which builds upon the mean-shift classifier, can generate reasonably accurate labelling of calcified plaque as compared to manual delineation, and can work similar to a human reader. The approach is independent of the number of calcium regions in the ICA. The prototype design shows encouraging results and can be applied to clinical practice.

63.   Robust, Automated and Fast Technique for Lumen Measurement in Common Carotid Artery

    Nobu Ikeda, Asheed Kumar, Francesco Lavra, Jeny Rajan, Luca Saba, John R. Laird, Andrew Nicolaides, Jasjit S. Suri.

    Ultrasound in Medicine and Biology. 2014. (Under Review) [Impact Factor: 2.455]

    Abstract: Lumen measurement has always very important diagnostic parameter for measuring the severity of the carotid disease. Ultrasound scanners yield carotid scans which shows image quality where automated lumen segmentation and diameter measurement carry clinically value. The quantification of lumen diameter is subjective when measured manually. Further, the lumen diameter measurement all along the carotid artery has never been performed automatically. This paper presents a robust two stage process for lumen segmentation in common carotid scans using Toshiba scanner. Stage one estimates the region of interest where lumen is present using comprehensive vertical spectral analysis. Stage two uses classification paradigm combined morphology. The system was designed keep accuracy and speed of processing into consideration. Our database consists of 250 patients for left and right CCA artery (500 images) using B-mode imaging. The performance of the system was computed using standard polyline distance and centerline distance methods. Our accuracy show nearly 100% accuracy. The feature of lumen measurement will soon be available in AtheroEdge™ 2.0 system (AtheroPoint™, Roseville, CA, USA).

64.   Atherosclerotic Carotid Wall Denoising/Enhancement in MRI on Retrospective Stoke Patients: 3D Adaptive Non-Local Approach

    Luca Saba, Ajay Gupta, Sanchi Rajput, Jeny Rajan, Krishna Kumar P., John R. Laird, Jasjit S. Suri.

    J Magn Reson Imaging. 2014. (Under Review) [Impact Factor: 2.7]

    Abstract: MR images can have Gaussian and Rician distributed noise with uniform or non uniform variance across the image. These noise levels affects many components in Atherosclerotic management such as carotid disease analysis, plaque build-up monitoring and therapeutic paradigms for drug evaluation and benchmarking against placebo. This paper presents a comprehensive study to analysis plaque build-up in the walls of the carotid walls in MRI volumes. Unlike traditional methods of 2-D denoising, this paper presents 3-D denoising method and performance evaluation. This method takes into consideration both the Rician nature of MR data and spatially varying noise pattern. Further, the paper presents benchmarking strategy against two other methods. We use the metric of SNR and show that overall performance shows a 20% improvement using our proposed method.

65.   Atherosclerotic Tissue Morphology and Characterization for Early Stroke Risk Assessment: A Review

    Aditya Sharma, Heather Gornick, Krishna Kumar P, Jeny Rajan, Luca Saba, Ikeda Nobutaka, John R Laird, Andrew Nicolades, Jasjit S. Suri.

    Vascular Medicine . 2014. (Under Review) [Impact Factor: 1.732]

    Abstract: Stroke and heart attack are identified as the leading cause of death in today’s world. The major cause of this is the plaque build-up in the artery and subsequent atherosclerosis. Studies shows that atherosclerosis in the carotid artery is an indication of coronary artery disease. This illustrates the significance of carotid atherosclerosis diagnosis. An automatic study of stages in the plaque development and its composition will highly helpful to the clinicians in predicting the risk of cardio vascular diseases (CVD). This article reviews the machine learning techniques used so far in the literature for classification of patients into symptomatic and asymptomatic. In addition, the article presents a general architecture for the classification task and discusses different feature extraction techniques. Echogenecity of the carotid plaque produces variation in greyscale or intensity values in B-mode ultrasound images. Linear and non-linear features along with a good classifier will be able to differentiate between plaque types. Through more advanced techniques like virtual histology IVUS, accurate evaluation of plaque morphology is possible. For validation, four parameters namely, sensitivity specificity, positive predictive value (PPV) and negative predictive value and negative predictive value (NPV) are discussed.

66.   Is there an association between cerebral micro-bleeds and leukoaraiosis?

    Luca Saba, Eytan Raz, Pier Paolo Bassareo, Michele di Martino, Carlo Nicola de Cecco, Roberto Grassi, Jasjit S Suri, Mario Piga.

    Journal of Stroke and Cerebrovascular Diseases. 2014 Nov 22. pii: S1052-3057(14)00365-6. doi: 10.1016/j.jstrokecerebrovasdis.2014.07.035. [Impact Factor: 1.984]

    Abstract: Cerebral microbleeds (CMBs) are small dot-like lesions appearing as hypo-signals on gradient echo (GRE) T2* magnetic resonance (MR) sequences whereas the leukoaraiosis (LA) indicates the presence of patchy areas of hyper-signal on FLAIR MR sequences in the periventricular white matter. The purpose of this work was to evaluate the association between LA and CMBs. 85 consecutive (males 55; median age 64) patients were retrospectively analyzed using a 1.5 Tesla system; CMBs were studied using a T2*-weighted GRE sequence and CMBs were classified as absent (grade 1), mild (grade 2; total number of microbleeds, 1–2), moderate (grade 3; total number of micro-bleeds, 3–10), and severe (grade 4; total number of microbleeds, >10). LA was assessed with FLAIR MR sequences and was graded based on the European Task Force on Age-Related White Matter Changes as follows: 1 (no lesions); 2 (focal lesions > 5 mm); 3 (early confluent lesions); and 4 (diffuse involvement of an entire brain region).We considered 170 cerebral hemispheres. The prevalence of CMBs was 24.7% (42170) whereas the prevalence of LA was 27.1% (46170). A statistically significant correlation was observed between LA and CMBs (correlation rho = 0.495, p value = 0.001). Multiple logistic regression analysis showed an association between CMBs and cerebrovascular symptoms (p = 0.0023). Results of this study suggest an association between CMBs and LA. Moreover we found that LA is associated with the presence of cerebrovascular symptoms.

67.   Computed Tomography Volume and Composition Assessment of Radiation-induced Carotid Atherosclerosis

    Luca Saba, Michele Anzidei, Roberto Sanfilippo, Jasjit S. Suri, Eytan Raz, Roberto Montisci, Giovanni Maria Argiolas, Mario Piga.

    Radiology . 2014. (Under Review) [Impact Factor: 6.339]

    Abstract: To study the variations in carotid artery composition and volume using multi-detector-row Computed Tomography (MDCT) in a group of patients that underwent 2 years of the head and neck radiotherapy treatment ( HNXRT). In this retrospective study, 62 patients (41 males and 21 females; mean age 63 age range 52-81 years) that underwent HNXRT and 40 patients (24 males and 16 females; mean age 65 age range 47-85 years) that underwent surgical resection of neoplasm of the neck who did not undergo HNXRT. In both cohorts, MDCT of neck was performed before the treatment and in follow up phase (2 years). The volume of the carotid artery plaques, as well as the volume of the sub-components (fatty-mixed-calcified), were semi-automatically quantified according to the Hounsfield unit thresholds (fatty plaque < 60 HU, mixed tissue between 60 and 130 HU, and calcium tissue > 130 HU). In the HNXRT group there was a statistically significant increase in the total volume of the carotid artery plaque (from 533 to 746 cubic mm; p value = 0.001), in the fatty (103 vs. 202 cubic mm; p = 0.001) and mixed component volume (328 vs. 419 cubic mm ; p = 0.034); moreover, a statistically significant variation (from 21.8% to 27.6%) in the percentage of the fatty tissue was found. In surgery group no statistically significant variation in the total plaque volume was found in the 2 years follow-up (463 vs. 527 cubic mm) and the only volume of calcified tissue showed a statistically significant increase (114 vs. 153 cubic mm). Results of this preliminary study suggest that the HNXRT promotes the increase of the carotid artery plaque volume by increasing in particular the fatty components that shows a statistically significant increase in its absolute percentage.

68.   Carotid Artery Dissection on Non-Contrast CT: Does Color Improve the Diagnostic Confidence?

    Luca Saba, Giovanni Maria Argiolas, Eytan Raz, Stefano Sannia, Jasjit S Suri, Paolo Siotto, Roberto Sanfilippo,Roberto Montisci, Mario Piga, Max Wintermark.

    European Journal of Radiology. 2014 Sep 21. pii: S0720-048X(14)00429-X. doi: 10.1016/j.ejrad.2014.09.001. [Impact Factor: 2.304]

    Abstract: The purpose of this work was to evaluate if the use of color maps, instead of conventional grayscale images, would improve the observer’s diagnostic confidence in the noncontrast CT evaluation of internal carotid artery dissection (ICAD). One hundred patients (61 men, 39 women; mean age, 51 years; range, 25 – 78 years), 40 with and 60 without ICAD, underwent noncontrast CT and were included in this the retrospective study. In this study, three groups of patients were considered: patients with MR confirmation of ICAD, n = 40; patients with MR confirmation of ICAD absence, n = 20; patients who underwent CT of the carotid arteries because of atherosclerotic disease, n = 40). Four blinded observers with different level of expertise (expert, intermediate A, intermediate B and trainee) analyzed the noncontrast CT datasets using a cross-model (one case grayscale and the following case using the color scale). The presence of ICAD was scored on a 5 point scale in order to assess the observer’s diagnostic confidence. After 3 months the four observers evaluated the same datasets by using the same cross-model for the alternate readings (one case color scale and the following case using the grayscale). Statistical analysis included Receiver Operating Characteristics (ROC) curve analysis, the Cohen weighted test and sensitivity, specificity, PPV, NPV, accuracy, LR+ and LR–. The ROC curve analysis showed that, for all observers, the use of color scale resulted in an improved diagnostic confidence with AUC values increasing from 0.896 to 0.936; from 0.823 to 0.849; from 0.84 to 0.909 and from 0.749 to 0.861 for expert, intermediate A, intermediate B and trainee observers, respectively. The increase in diagnostic confidence (between the AUC areas) was statistically significant (p = 0.036) for the trainee. Accuracy, as well as sensitivity, specificity, PPV, NPV, LR+ and LR–, were improved using the color scale.

69.    Correlation between Atherosclerosis Plaque Score and Automated Bulb Plaque IMT using Carotid Ultrasound

    Nobutaka Ikeda, Ajay Gupta, Nilanjan De, Soumyo Bose, Shoaib Shafique, Tadashi Araki, Elisa Cuadrado Godia, Luca Saba, John Laird, Andrew Nicolaides, Jasjit S Suri.

    Journal of the American Society of Echocardiography. 2014. (Under Review) [Impact Factor: 4.277]

    Abstract: We studied whether the plaque score, a measure of carotid artery disease complexity is correlated with carotid intima-media thickness measure using automated sampling from the common carotid and bulb plaque. Three-hundred-seventy consecutive patients who underwent carotid ultrasound were analysed. Plaque score was determined using a validated score based on sonographic plaque characteristics. Carotid intima-media thickness was determined using a computerized automated measurement technique all along the carotid artery that includes sampling from the bulb region and the region proximal to the bulb. Using our computerized automated system for carotid intima-media thickness measurement, the correlation between automated carotid intima-media thickness and plaque score was 0.7610 (p-value < 0.0001). Using receiver operating characteristics (ROC) curves, the areas under the curve (AUC) for plaque score was 0.865 using an automated cartoid intima-media thickness value of greater than 1 mm as the classification variable. When compared against manual tracings, automated carotid intima-media thickness system performance was as follows: lumen-intima error: 0.007818±0.0071 mm; media-adventitia error: 0.0179±0.0125 mm; and automated carotid intima-media thickness error: 0.0099±0.00988 mm. The precision of automated carotid intima-media thickness against the manual radiologist's tracings was 98.86%. Our automated algorithm for carotid intima-media thickness measurement was strongly correlated with carotid plaque score in a cohort of 370 patients with multiple cardiovascular risk factors. By sampling carotid intima-media thickness in the bulb region, our automated carotid intima-media thickness technique can demonstrate a high degree of correlation between carotid artery disease lesion complexity and carotid atherosclerosis characteristics.

70.   Is there an association between Asymmetry of Carotid Artery Wall thickness (ACAWT) and cerebrovascular symptoms?

    Luca Saba, Eytan Raz, Michele di Martino, Roberto Montisci, Jasjit Suri, Mario Piga.

    International Journal of Neuroscience, .2014 Aug 21. [Epub ahead of print] [Impact Factor: 1.273]

    Abstract: Previous publications demonstrated that MDCTA can evaluate the carotid artery wall thickness (CAWT). The purpose of this work was to compare the asymmetry of CAWT between carotids in symptomatic and asymptomatic patients. Sixty consecutive symptomatic (males 44; median age 64) and sixty asymptomatic sex- and age-matched patients, were analyzed by using a 40-detector-row CT system. CAWT was calculated for both carotids in each patient and the ratio between the thicker CAWT and the contra-lateral was calculated to obtain the ACAWT index. Bland-Altman, logistic regression and ROC curve analysis were calculated. The Bland-Altman plot demonstrates a very good agreement between measurements with a mean difference value of 3.4% and 95% CI from -8% to 14.8% The ACAWT was significantly different between symptomatic and asymptomatic patients (with a P value of 0.0001). The ROC area under the curve was 0.742 (P = 0.001). Logistic regression model indicated that ACAWT, stenosis degree and fatty plaques were independent variables associated with cerebrovascular symptoms (P value respectively 0.0057, 0.0003 and 0.0178). Results of our study indicated that the index of asymmetry in the CAWT may be used as a further parameter to stratify the risk of symptoms related to carotid artery.

71.    Segmentation of Ventricles in Alzheimer MR Brain Images: its Validation, Performance Evaluation, Reproducibility, and Classification: A Reaction Diffusion-based Level Set Paradigm

    K. R. Anandh, C. M. Sujatha, Michele Porcu, S. S. Suganthi, Luigi Barberini, Luca Saba, S. Ramakrishnan , Jasjit S. Suri. .

    Neuroscience Letters. 2014. (Under Review) [Impact Factor: 2.146]

    Abstract: Alzheimer’s Disease (AD) is a complex and devastating neurodegenerative disorder that causes dilation of ventricles, shrinking of whole brain and resulting in slow progression of dementia. Ventricle enlargement is a significant biomarker for the diagnosis of AD. Due to its high spatial resolution, Magnetic Resonance (MR) imaging clearly demonstrates the pathology of brain and shown to improve the diagnosis of AD. In this work, a modified level set method is adopted to segment the ventricles in AD and normal MR images. The modified diffusion rate equation used in this method overcomes the complex re-initialization procedure of level set. Further, shape-based features is extracted from the segmented ventricle to assess in understanding shape change in AD vs. normal. Further, we developed a software system for analysing the reproducibility of segmentation. OASIS database taken from NIH was selected, consisting of 184 patient T1-weighted MR images equally divided between normal and AD cases. Using the Reaction Diffusion (RD)-based level set method, method. the computer-based segmentation of ventricle area for normal and AD subjects shows an precision of accuracy of 99% (p<0.0001) when compared against manual tracings. Our reproducibility study shows coefficient of correlation of 0.99 between three different users. Taking one slice criteria, and using the Jacard index and Dice similarity measures, we show that AD has enlarged ventricles compared to normal by 7.43%. Our pilot study was comprehensive, validated, reproducible, and its performance evaluated against the manual tracings by expert Neurologists.

72.   Improved Correlation between Carotid and Coronary Atherosclerosis Plaque Score Using Automated Ultrasound Carotid Bulb Plaque IMT Measurement

    Nobutaka Ikeda, Ajay Gupta, Nilanjan Dey, Soumyo Bose, Shoaib Shafique, Tadashi Arak, Elisa Cuadrado Godia, Luca Saba, John Laird, Andrew Nicolaides, Jasjit S Suri.

    Journal of the American Society of Echocardiography. 2014. (Under Review) [Impact Factor: 4.277]

    Abstract: We studied whether the Plaque score, a measure of coronary artery disease (CAD) complexity, could be better predicted with carotid intima media (IMT) measures using automated IMT with sampling from the common carotid and bulb plaque compared with manual IMT determined by sonographers. Three-hundred-seventy consecutive patients who underwent carotid-ultrasound and coronary angiography were analyzed. Plaque score was determined by two experienced interventional cardiologist taken from coronary angiograms. Our method involves a computerized automated IMT measurement all along the carotid artery that includes bulb region and region proximal to bulb. Using our automated system of cIMT measurement, the correlation between automated cIMT and Plaque score (PS) was: 0.7610 (P < 0.0001), between automated cIMT and SYNTAX score (SC) was: 0.467 (P < 0.0001), and between automated cIMT and HbA1c is 0.1933 (P=0.0003). Obtained Area under the ROC curve (AUC) for PS, SS and HbA1c is respectively 0.865, 0.696 and 0.585 when the classification variable is automated cIMT. When compared against the Radiologist's manual tracings, automated cIMT system performance was: lumen-intima error: 0.007818±0.0071 mm, media-adventitia error: 0.0179±0.0125 mm and automated cIMT error: 0.0099±0.00988 mm. The precision of automated cIMT against the manual Radiologist's reading showed a precision of: 98.86%. Our automated algorithm for cIMT measurement which includes both the common carotid artery and bulb regions showed a strong correlation between cIMT and coronary Plaque and SYNTAX score and moderate correlation between cIMT and HbA1 in a cohort of 370 patients with multiple cardiovascular risk factors. By sampling cIMT in the bulb region, our automated technique improves the degree of correlation between CAD lesion complexity and carotid atherosclerosis characteristics.

73.   Is there an Association between Leukoaraiosis Volume and Diabetes

    Pierleone Lucatelli, Roberto Montisci, Roberto Sanfilippo, Jasjit Suri, Carlo Catalano, Luca Saba.

    Stroke. 2014. (Under Review) [Impact Factor: 6.158]

    Abstract: The association between diabetes mellitus and leukoaraiosis is still debated while the association with hypertension, age and carotid artery stenosis is widely demonstrated. In this work our purpose is to evaluate the association of white matter hyper intense lesion volume and diabetes in a cohort of patients who underwent an MRI of the brain at our Institution. Ninety-three consecutive patients (mean age 71 ± 9 years, males 71) were prospectively recruited at our Institution before undergoing carotid endarterectomy. The patients were assessed for the presence of vascular risk factors, including diabetes mellitus, hypertension, hyperlipidemia. The patients underwent an admission MRI of the brain which included T2-FLAIR, T2-SE and DWI sequences. FLAIR-leukoaraiosis lesion volume was performed by a neuroradiologist, blinded to the presence of risk factors, using a semi-automated segmentation technique (Jim, Xinapse System, Leicester, UK). ROC analysis was performed to evaluate the AUC of FLAIR-leukoaraiosis lesion volume and number of lesions with respect to the vascular risk factors. The ROC curve analysis showed an AUC of 0.812 (SE = 0.058 and 95% CI from 0.711 to 0.891) with a statistically significant p-value of 0.001 between the presence of diabetes mellitus and the FLAIR-leukoaraiosis lesion volume. In the ROC curve analysis between the presence of diabetes mellitus and number of lesion the AUC was 0.797 (SE = 0.056 and 95% CI from 0.696 to 0.878) with a p-value of 0.001. Our data demonstrate an association between diabetes mellitus and the volume of the cerebral white matter abnormality (leukoaraiosis). Leukoaraiosis, which is a commonly identified lesion pattern in the MRI of the brain, is associated with different vascular risk factors, including diabetes mellitus.

74.   Plaque Echolucency and Stroke Risk in Asymptomatic Carotid Stenosis: A Systematic Review and Meta-Analysis

    Ajay Gupta, Kartik Kesavabhotla, Hediyeh Baradaran, Edward E. Mtui, Hooman Kamel, Ankur Pandya, Ashley Giamborone, Drew Wright, Kevin Pain, Jasjit S. Suri, Pina C. Sanelli, Alvin I. Mushlin.

    Stroke. 2014. (Accepted) [Impact Factor: 6.158]

    Abstract: Ultrasound echolucency has been studied as a potential tool to predict stroke risk in carotid atherosclerotic disease. We performed a systematic review and meta-analysis to summarize the association between carotid plaque echolucency and future ipsilateral stroke risk. We systematically searched the medical literature for studies evaluating the association between ultrasound-determined carotid plaque echolucency and future stroke in asymptomatic patients. We included prospective observational studies with a mean follow-up of ≥1 month and outcome ascertainment of stroke after baseline carotid plaque echolucency assessment. We performed a meta-analysis using a random-effects model with assessment of study heterogeneity and publication bias. We also performed subgroup analyses limited to patients with stenosis ≥50%, studies in which plaque echolucency was determined qualitatively, studies with a relatively lower risk of bias, and studies published after the year 2000. We were able to perform meta-analysis from 7 studies on 7557 subjects with an approximate mean follow up of 37.2 months. We found a significant positive relationship between plaque echolucency and the risk of future ipsilateral stroke with a random effects RR of 2.31 (95% confidence interval [CI], 1.58-3.39, P<.001). A statistically significant increased RR for future stroke was preserved in all subgroup analyses. No statistically significant heterogeneity of publication bias was present in the main or subgroup meta-analyses. The presence of ultrasonographic carotid plaque echolucency is a predictor of stroke in asymptomatic carotid artery stenosis. Using standard clinical ultrasound equipment, evaluation for plaque echolucency can offer stroke risk information beyond stenosis severity assessment in carotid atherosclerotic disease.

75.   A Comparative Approach of Four Different Image Registration Techniques for Quantitative Assessment of Coronary Artery Calcium using Intravascular Ultrasound

    Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Sayan Chakraborty, Luca Saba, Dinesh Kumar, Elisa Cuadrado Godia, Xiaoyi Jiang, Ajay Gupta, Petia Radevia, Andrew Nicolaides, Jasjit S Suri.

    Computer Methods and Programs in Biomedicine. 2014 Dec 2. pii: S0169-2607(14)00385-X. doi: 10.1016/j.cmpb.2014.11.006. [Impact Factor: 1.483]

    Abstract: In IVUS imaging, linear constant velocity and a constant angular velocity of 1800 rev/min causes displacement of the calcium in subsequent image frames. To overcome this error in intravascular ultrasound video, IVUS image frames must be registered prior to the lesion quantification. This paper presents a comprehensive comparison of four registration methods, namely: Rigid, Affine, B-Splines and Demons on five set of calcium quantification parameters namely: (i) the mean lesion area, (ii) mean lesion arc, (iii) mean lesion span, (iv) mean lesion length, and (v) mean lesion distance from catheter. Using our IRB approved data of 100 patient volumes, our results shows that all four registrations showed a decrease in five lesion parameters as follows: for Rigid registration, the values were: 4.92%, 5.84%, 5.89%, 5.27%, and 4.57%, respectively, for Affine registration the values were: 6.06%, 6.51%, 7.28%, 6.50%, and 5.94%, respectively, for B-Splines registration the values were: 7.35%, 8.03%, 9.54%, 8.18%, and 7.62%, respectively, and for Demons registration the five parameters were 7.32%, 8.02%, 10.11%, 7.94%, and 8.92% respectively. The relative overlap of identified lesions decreased by 5.91% in case of Rigid registration, 6.23% in case of Affine registration, 4.48% for Demons registration, whereas it increased by 3.05% in case of B-Splines registration. Rigid and Affine transformation based registration took only 0.1936 and 0.2893 seconds per frame, respectively. Demons and B-Splines framework took only 0.5705 and 0.9405 seconds per frame, respectively, which were significantly slower than Rigid and Affine transformation based image registration.

76.   A Comparative Approach of Four different Image Registration Techniques in Atherosclerotic IVUS Video

    Tadashi Arak, Nobutaka Ikeda, Nilanjan De, Sayan Chakraborty, Luca Saba, Dinesh Kumar, Elisa Cuadrado Godia, Tianzi Jiang, Jasjit S Suri.

    Pattern Recognition. 2014. (Under Review) [Impact Factor: 2.632]

    Abstract: In IVUS imaging, linear constant velocity and a constant angular velocity of 1800 rev/min causes displacement of the calcium in subsequent image frames. To overcome this error in intravascular ultrasound video, IVUS image frames must be registered prior to the lesion quantification.This paper presents the effect of image registration on quantification of calcium parameters and its volume using two different paradigms using four different methods: Under global transformation category: (i) rigid transformation-based and (ii) affine registration-based, and under local transformation category: (iii) B-Splines-based and (iv) Demons-based. Largest calcium lesion is computed before and after registration using both paradigms. The following five quantification parameters: (i) the mean lesion area, (ii) mean lesion arc, (iii) mean lesion span, (iv) mean lesion length and (v) mean lesion distance from catheter (vi) overlapped percentage were studied when comparing the four registration paradigms on 100 patients. Our results shows that rigid transformation based registration resulted in decrease in these five coronary calcium lesion parameters by 4.92%, 5.84%, 5.89%, 5.27%, and 4.57%, respectively. Affine transformation based registration resulted in decrease in these five coronary calcium lesion parameters by 6.06%, 6.51%, 7.28%, 6.50%, and 5.94%, respectively. B-Splines registration resulted in decrease in these five coronary calcium lesion parameters by 7.35%, 8.03%, 9.54%, 8.18%, and 7.62%, respectively. Demons registration provided similar results where the corresponding reduction in these five parameters were 7.32%, 8.02%, 10.11%, 7.94%, and 8.92% respectively. The relative overlap of identified lesions increased by 5.91% in case of rigid transformation based, 6.23% in case of affine transformation based registration, whereas it increased by 3.05% in case of B-Splines, and by 4.48% for Demons. However, rigid and affine transformation based registration took only 0.1936 and 0.2893 seconds per frame, whereas, Demons framework took only 0.5705 seconds per frame and B-Splines framework took which took 0.9405 seconds per frame which were significantly slower than rigid and affine transformation based image registration.

77.   Automated and accurate carotid bulb detection, its verification and validation in low quality frozen frames and motion video.

    Nobutaka Ikeda, Tadashi Araki, Nilanjan Dey, Soumyo Bose, Shoaib Shafique, Ayman El-Baz, Elisa Cuadrado-Godia, Anzidei M, Luca Saba, Jasjit S. Suri.

    INTERNATIONAL ANGIOLOGY. 2014 March 21. [Impact Factor: 1.462]

    Abstract: cIMT measurements during clinical trials need to have a fixed reference point (also called as bulb edge points) in the anatomy from which the carotid intimamedia thickness (cIMT) can be measured. Identification of the bulb edge points in carotid ultrasound images faces the challenge to be detected automatically due to low image quality and variations in ultrasound images, motion artefacts, image acquisition protocols, position of the patient, and orientation of the linear probe with respect to bulb and ultrasound gain controls during acquisition. This paper presents a patented comprehensive methodology for carotid bulb localization and bulb edge detection as a reference point. The method consists of estimating the lumenintima borders accurately using classification paradigm. Transition points are located automatically based on curvature characteristics. Further we verify and validate the locations of bulb edge points using combination of several local image processing methods such as (i) lumenintima shapes, (ii) bulb slopes, (iii) bulb curvature, (iv) mean lumen thickness and its variations, and (v) geometric shape fitting. Our database consists of 155 ultrasound bulb images taken from various ultrasound machines with varying resolutions and imaging conditions. Further we run our automated system blindly to spot out the bulbs in a mixture database of 336 images consisting of bulbs and nobulbs. We are able to detect the bulbs in the bulb database with 100% accuracy having 92% as close as to a Neurologists’s bulb location. Our mean lumenintima error is 0.0133 mm with precision against the manual tracings to be 98.92%. Our bulb detection system is fast and takes on an average 9 seconds per image for detection for the bulb edge points and 4 seconds for verification/validation of the bulb edge points.

78.   Novel, Automated and Robust Site-IMT Measurement in Carotid Ultrasound with Bulb Presence: A more accurate Stroke Risk Assessment Tool

    Nobutaka Ikeda, Tadashi Araki, Nilanjan Dey, Ajay Gupta, Soumyo Bose, Suvojit Acharjee, Shoaib Shafique, Elisa Cuadrado-Godia, Luca Saba, Andrew Nicolaides, Jasjit S. Suri.

    IEEE Transactions on Instrumentation and Measurement. 2014. (Under Review) [Impact Factor: 1.357]

    Abstract: Segmental-IMT (sIMT) measures the IMT proximal to the bulb edge of the carotid artery in 10 mm segments (s1, s2, and s3). Bulb edge is considered the reference marker for measurements of the cIMT and is difficult to locate due to (a) low SNR compared to the CCA region (b) uncertainty of its presences in either E-W direction and (c) not always present in full shape and size. This paper presents an automated measurement of sIMT. This paper presents a novel automated system for sIMT measurement using an integrated approach which combines geometry and classification paradigm. The system first finds the bulb edge and then measures the sIMT proximal to the bulb edge. We further measure the IMT in bulb region (bIMT). Our database consists of two kinds: (a) Bulb database of 172 images and (b) Mixed database of 725 images with varying image resolutions and different ethnic backgrounds. We show our contribution of different IMTs in bulb database as: bIMT 34%, s1-IMT 29.46%, s2-IMT 11.48%, and s3-IMT 12.75%. In mixed database of 725 images, our automated system shows the following performance against manual tracings as: LI error mean (sd) is 0.011644 (xxx) mm, MA error mean (sd) is 0.017851 (yyy) mm, GT IMT mean (sd) 0.833558854 (zzz) mm, auto IMT mean (sd) 0.833934976 (kkk) mm. Our FOM was: 98.47%. These numbers are improved compared to previous publications by Suri’s group which is conventional-based IMT. Our bulb detection system shows 100% accuracy when compared against clinical output. The system is fully automated.

79.   Automated Bulb Detection, Verification and Validation in Carotid Ultrasound

    Luca Saba, Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Suvojit Acharjee, Ayman El-Baz, Filippo Molinari, Elisa Cuadrado Godia, Andrew Nicolaides, Jasjit S Suri.

    Int Angiol. 2014. (Under Review) [Impact Factor: 1.462]

    Abstract: cIMT measurements during clinical trials need to have a fixed reference point in the anatomy from which the IMT can be measured. Further, benchmarking different cIMT measurement techniques/algorithms also require a reference point. Carotid bulb is an anatomic location which is most pronounced in the image. We thus hypothesize that carotid bulb is an excellent source for a reference point. Though it is easily visible, it faces the challenge to be detected automatically due to low image quality and variations in ultrasound images, motion artifacts, image acquisition protocols, position of the patient and the linear probe with respect to bulb. This paper presents a comprehensive methodology for carotid bulb localization and the bulb edge detection as a reference point. The method consists of estimating the lumen-intima borders accurately using classification paradigm. We then locate the intersection point between the carotid bulb and the common carotid artery automatically, called bulb edge or reference point. Further we verify and validate it using combination of several local image processing methods such as (i) lumen-intima shapes, (ii) bulb slopes, (iii) bulb curvature, (iv) mean lumen thickness and its variations, and (v) geometric shape fitting. Our database consists of ultrasound images taken from various ultrasound machines with varying resolutions and imaging conditions consisting of a database of 155 known bulbs. Further we run our automated system blindly to spot out the bulbs in a mixture database of 336 images consisting of bulbs and no-bulbs. We are able to detect the bulbs in the bulb database with 100% accuracy. Our bulb edge location shows that we are 99% in limits within 0.2 mm from the ideal bulb location. Our mean lumen-intima error is less than 0.0801 mm. (99.9% accurate with the ground truth (GT)). It is fast and takes on an average 9 seconds for detection of the reference point and 4 seconds for verification/validation the reference point. AtheroPointTM, Roseville, CA, USA has started using this in there beta sites which is undergoing clinical evaluation for its commercialization. We demonstrated our bulb detection method under several scenarios during varying acquisition protocols.

80.   Comparative Approach between Demons and B-Spline Registration for Studying its effect on Quantification of Coronary Calcium in IVUS Video Frames

    Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Sayan Chakraborty, Luca Saba, Dinesh Kumar, Elisa Cuadrado Godia, Filippo Molinari, Petia Radevia, Andrew Nicolaides, Jasjit S Suri.

    Computer Methods and Programs in Biomedicine. 2014. (Under Review) [Impact Factor: 1.555]

    Abstract: During IVUS image acquisition pullback strategy and the quasi-periodic movement of heart, the images acquired can affect by these movements. This can cause the small calcium lesions to displace laterally, longitudinally and rotationally. Thus, for accurate lesion quantification in intravascular ultrasound video, neighbouring frames must be registered before the calcium volume is computed. This paper presents two comparative approaches for image registration techniques: Damon vs. B-spline for IVUS image registration, and further quantifies the calcium volumes and computes the cardiological risk with respect the neurological risk factors like cIMT. An automated system was developed which can automatically register and locate and further quantify detected culprit calcium. Using B-spline registration, there was a decrease in the following lesion quantification parameters: (i) the mean lesion area, (ii) mean lesion arc, (iii) mean lesion span, (iv) mean lesion length and (v) mean lesion distance from catheter with following values: 8.40%, 8.67%, 9.33%, 8.8%, 8.0%. Similar behaviour was obseverd when using Demon registration where the same parameters showed reduction in values as: 6.75%, 8.02%, 10.11%, 7.94%, and 8.92% respectively. The performance of registration methods were also compatible: the overlapping ratio index increased by 3.05 % in case of B-spline, whereas 4.48% for Demons. There is an improvement of 7.3% in area under curve (AUC) for computed normalized volume vs. cIMTs for B-spline registration, whereas 6% for Demons registration. It is reported that the time taken to register each frame using Demon framework was (0.6961 seconds/frame), much lesser than the time taken by B-spline framework (1.3431 seconds/frame).

81.   Effect of Watermarking on Carotid Bulb Bifurcation Detection in Atherosclerotic Ultrasound Video

    Nilanjan Dey, Soumya Bose, Achintya Das,Sheli Sinha Chaudhuri, Luca Saba, Shoaib Shafique, Jasjit S. Suri. .

    Journal of Medical Systems. 2014. (Under Review) [Impact Factor: 1.783]

    Abstract: Embedding of diagnostic and health care information requires secure encryption and watermarking. This research paper presents new methodology for devalorization of diagnostic parameters in Atherosclerotic plaque ultrasound video namely: (a) bulb identification and recognition thereby identifying the bulb edge points in far and near carotid walls; (b) carotid bulb diameter; (c) carotid lumen thickness all along the carotid artery from the bulb edge locations. Further, embedding these clinical parameters in watermarking paradigm using two different sets of watermarking algorithms to study the effect of these watermarking methods on these diagnostic parameters. They are: two correlation-based (binary logo hiding) and two singular value decomposition (SVD)-based (gray logo hiding) watermarking algorithms. Our data sets consist of carotid atherosclerotic movies taken under IRB protocol from University of Indiana Hospital, USA-AtheroPoint™ (Roseville, CA, USA) joint pilot study. The protocol adapted is conventional ultrasound image acquisition using carotid linear probe. The ROC analysis of the detected bulb has been done with the ground truth (GT). High sensitivity (100%), low specificity (0%), and high accuracy (100%) shows the robustness of the proposed bulb detection method. DWT-DCT-SVD based approach having high PSNR value (41.26) and high payload, ensures the retention of diagnostic parameter devalorization as an effect of watermarking. The system is fully automated.

82.   A Online Grayscale Feature Classification Paradigm for Cardiological Risk Assessment on Diabetic Cohort using IVUS

    Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Sourav Samanta, Ayman El-Baz, Filippo Molinari, Elisa Cuadrado Godia, Luca Saba, Andrew Nicolaides, Jasjit S Suri.

    Physics in Medicine & Biology. 2014. (Under Review) [Impact Factor: 2.701]

    Abstract: The study of automatic determination of cardiological risk assessment is based on Support Vector Machine (SVM) classification approach which uses shape-based coronary calcium lesion features and neurological risk biomarker cIMT. The study was undertaken on a cohort of 100 patients. Acquisition of IVUS data was done by using a 40 MHz Atlantis SR Pro, Boston Scientific intravascular ultrasound catheter. A reverse engineering strategy was adopted for automatic detection and segmentation of coronary calcium lesions in IVUS video frames. Different grayscale based features of the detected lesion such as (i) Haralic texture features, (ii) Gabor features, (iii) fractal features (Total 17 features), were computed for risk classification. A novel, automated and 510 (K) FDA cleared software; AtheroEdge™ (AtheroPoint LLC, Roseville, CA, USA) was used for computation of cIMT on B-mode carotid ultrasound. Our results demonstrate that the shape-based coronary calcium lesion characteristic features are capable of accurately classifying 85% of neurological risky patients. This analysis reports classification of neurological risk biomarker through carotid window (carotid imaging) called intima-media thickness (IMT).

83.   Effect of Image Registration on Quantification of Coronary Calcium from Intravascular Ultrasound (IVUS) Video Frames

    Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Filippo Molinari , Suvojit Acharjee, Sayan Chakraborty,Luca Saba, Dinesh Kumar, Elisa Cuadrado Godia,Andrew Nicolaides, Petia Radevia, Jasjit S Suri.

    Computers in Biology and Medicine. 2014. (Under Review) [Impact Factor: 1.162]

    Abstract: The process of interventional cardiology image acquisition is highly affected by the quasi-periodic movement of heart and pullback strategy adapted for image acquisition. In IVUS imaging, catheter pullback is made at linear constant velocity, usually of 0.5 mm/s, and at a constant angular velocity of 1800 rev/min. These may result in displacement of the calcium laterally, longitudinally and/or rotationally. As a consequence the shape and the size of attached calcium get altered. For accurate lesion quantification from IVUS video, subsequent frames must follow the same paradigm. Image registration is a solution for resolving the issue of spatio-temporal localization changes of lesion by increasing the correlation between two consecutive frames. An automated computer-based application is developed and tested on 100 patient IVUS video volumes. IVUS data was acquired using 40 MHz IVUS catheter, Atlantis SR Pro, Boston Scientific. Segmentation of the calcium region requires adaptation of a reverse engineering strategy, i.e., removal of non-region of interest which consists of media/adventitia high gradient regions along with catheter regions. Once these non-regional interest regions are removed the calcium is automatically detected. We have adopted a B-spline registration technique by taking a bin of specific size IVUS video images. Each bin has its own reference for registration. The bin-width and reference is optimized under B-spline frame work leading to accurate lesion quantification from IVUS video. The software is capable of automatic localization of the calcium lesion frames, further quantification of the calcium volume before and after registration of IVUS video. Registration effect on 100 patient IVUS videos shows the following parameters decreases: the mean lesion area, mean lesion arc, mean lesion span, mean lesion length and mean lesion distance from catheter, mean normalized volume is decreased by 8.4079% , 8.6754%, 9.3379%, 8.8084%, 7.995 % and 8.0708% respectively. Our performance overlapping ratio index increased by 3.05% after registration. IVUS image registration is essential for the accurate quantification of lesion.

84.   Cardiological Risk Assessment using Support Vector Machine by combining Coronary Calcium Lesions and cIMT as a Neurological Risk Biomarker

    Luca Saba, Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Suvojit Acharjee, Ayman El-Baz, Filippo Molinari, Elisa Cuadrado Godia, Andrew Nicolaides, Jasjit S Suri.

    Ultrasound in Medicine and Biology. 2014. (Under Review) [Impact Factor: 2.455]

    Abstract: The study of automatic determination of cardiological risk assessment based on SVM classification approach using shape-based coronary calcium lesion features and neurological risk biomarker cIMT. A study was undertaken on a cohort of 100 patients. Acquisition of IVUS data was done by using a 40 MHz Atlantis SR Pro, Boston Scientific intravascular ultrasound catheter. A reverse engineering strategy was adopted for automatic detection and segmentation of coronary calcium lesions in IVUS video frames. Different shape-based characteristic features such as (i) mean lesion thickness, (ii) mean standard deviation of lesion thickness, (iii) mean central line length, (iv) mean length of all branches, (v) mean span (arc angle) of the detected lesion, (vi) mean lesion irregularity and (vii) lesion distance to catheter center were computed for risk classification. A novel, automated and 510 (K) FDA cleared software; AtheroEdge™ (AtheroPoint LLC, Roseville, CA, USA) was used for computation of cIMT on B-mode carotid ultrasound. Our results demonstrate that the shape-based coronary calcium lesion characterization and classification software is capable of accurately classifying 80% of cardiological risky patients (sensitivity is 0%, specificity is 80%). This analysis reports classification of cardiological risk patients by combining coronary calcium lesions from IVUS and neurological risk biomarker of carotid IMT is able to classify coronary artery disease patients with high success rate.

85.   Hunting for Calcium Lesions in Intravascular Ultrasound: an automated and accurate calcium volume quantification and its correlation to automated cIMT using AtheroEdge

    Tadashi Araki, Nobutaka Ikeda, Nilanjan Dey, Suvojit Acharjee, Filippo Molinari, Luca Saba, Ayman El-Baz, Andrew Nicolaides, Jasjit S Suri.

    Medical Physics. 2014. (Under Review) [Impact Factor: 2.83]

    Abstract: The carotid intima-media thickness (cIMTs) using AtheroEdge™ is a strong predictor of cerebrovascular complications. Our recent studies showed a strong correlation of cIMT with Ankle Brackle Index (ABI) and Syntax Score. In our previous research work we have reported that the normalized volume of the detected coronary calcium from IVUS videos using shape-based approach correlates to cIMT. We here demonstrate an improved method of false positive calcium elimination thereby improving the automated calcium volume estimation followed by the correlation with cIMTs. An automated computer-based application is developed and tested on 92 IVUS patient volume videos. The concept is to estimate the calcium regional area by counting all the pixels of the region corresponding to the coronary calcium in each detected frame. To segment the calcium region, a reverse engineering strategy is adapted, that is removal of non-region of interest which consists of media/adventitia high gradient regions along with catheter regions. The calcium is automatically detected once these non-regional interest regions are removed. False positive calcium elimination is corrected using a combination of textured entropy and distance constraint. Volume is then quantified by summing all the pixels in the corrected segmented lesion in automatically detected video frames. Normalized calcium volume is then compared to AtheroEdge™-cIMTs automatically taken from the B-mode carotid ultrasound using Pearson correlation coefficient.The software is able to detect the calcium lesion frames accurately and automatically, and also computes the corresponding calcium volume. Further, it establishes a relationship between calcium volume and cIMTs estimated by AtheroEdge™ from AtheroPoint™. We show a 6.16% improvement in elimination of false positive using our textured entropy strategy. Further, our results show a mean improvement in correlation coefficient between calcium volume and AtheroEdge™-cIMT by 45.42% when taking both the left and right carotids using our new method of textured entropy and distance constraint. Textured entropy strategy improves the correlation between coronary calcium volume computed from IVUS using shape-based approach and cIMTs computed using AtheroEdge™ software from carotid ultrasound. The system is fully automated and links cardio and neuro-fields having same genesis as atherosclerosis genesis.

86.   VesselOMeasure: User Friendly Tools and Techniques for Ultrasound Carotid Artery Analysis

    Luca Saba, Stefano Sannia, Filippo Molinari, G. Swapna, Vinitha Sree, Kristen M. Meiburger, U. Rajendra Acharya, S. William Liboni, Giorgio Mallarini, Shoaib Shafique, Masataka Nakano, Andrew Nicolaides, Jasjit S. Suri.

    Italian Journal of Vascular and Endovascular Surgery. 2013 December;20(4):225-49. [Impact Factor: 0.043]

    Abstract: Atherosclerosis is a disease in which plaque builds up inside the arteries resulting in several complications such as narrowing of the lumen (stenosis), plaque rupture, and consequently heart attacks and stroke. Our group has been actively working in this area for nearly a decade and has developed several proprietary Computer Aided Diagnostic (CAD) tools that assist in every stage of carotid artery analysis. The goal of this paper is to review CAD based carotid artery analysis in general, and to present brief descriptions of the techniques that we developed in this area. Specifically, we first describe the CAD techniques that we developed to handle the pre-processing of B-mode ultrasound images of the carotid artery. We then summarize the most advanced technical solutions that we developed for automated Intima-Media Thickness (IMT) measurement. We also review the most widely used and accepted performance evaluation metrics for IMT measurement, and clinical validation. Subsequently, we discuss the possibility of plaque delineation using contrast-enhanced ultrasound imaging and present a tool for differentiation of symptomatic and asymptomatic plaques. We also present a technique for atherosclerosis monitoring, and finally review the capabilities and limitations of our algorithms.

87.   Effect of Geometric-based Coronary Calcium Volume as a feature along with its Shape-based Attributes for Cardiological Risk Prediction using IVUS and B-mode Automated Carotid Ultrasound

    Nobutaka Ikeda, Tadashi Araki, Filippo Molinari, Nilanjan Dey, Suvojit Acharjee, Luca Saba, Andrew Nicolaides, Jasjit S Suri.

    Journal of Medical Imaging and Health Informatics. 2014. (Accepted) [Impact Factor: 0.642]

    Abstract: Cardiological risk assessment based on coronary arterial lesions characteristics uses neurological paradigm such as carotid intima-media thickness (cIMT), a biomarker for neurological risk. In order to measure cIMTs from a population of 92 patients, AtheroEdge™, (AtheroPoint™ LLC, Roseville, CA, USA), a novel and dedicated automated software analysis was adapted. Our results demonstrate that the area under the curve (AUC) was 0.616 when combining volume feature with shape-based features in comparison to 0.58 when using shape-based features alone. Receiver operating characteristic (ROC) curve analysis shows that the statistical significance of the established association helps in analysing the sensitivity of coronary calcium quantification in neurological risk patients. The analysis reported showed a sizeable improvement in AUC by 6.2% when using the combination of volume and shape-based features alone compared to shape-based features alone.

88.   Multi-modality CT scanning in the evaluation of Cerebrovascular disease patients.

    Luca Saba, Michele Anzidei, Mario Piga, Federica Ciolina, Lorenzo Manneli, Carlo Catalano, Jasjit S. Suri, Eytan Raz.

    The International Journal of Cardiovascular Imaging. 2014. Jun;4(3):245-62. doi: 10.3978/j.issn.2223-3652.2014.06.05. [Impact Factor: 2.648]

    Abstract: Ischemic stroke nowadays represents one of the leading causes of severe disability and mortality in the Western World. In the past years angiography was the imaging technique most used for the detection of the extra-cranial and intracranial vessel pathology but nowadays non invasive imaging tool like Ultrasound, Magnetic Resonance and Computed Tomography (CT) have demonstrated that can offer a detailed analysis of the vascular system. In particular CT represents an advanced system to explore the pathology of carotid arteries and intracranial vessel and also offers tools like CT perfusion that provides valuable information of the of brain vascular physiology by increasing the stroke diagnostic. In this review our purpose is to discuss stroke risk prediction and detection using CT.

89.   Relationship between leukoaraiosis, carotid intima-media thickness (IMT) and intimamedia thickness variability (IMTV)

    Jasjit S Suri.

    Radiology. 2014. (Under Review) [Impact Factor: 6.339]

    Abstract: The purpose of this study was to assess the relationship between the degree of leukoaraiosis (LA), carotid intima-media thickness (IMT) and intima-media thickness variability (IMTV). IMTV significantly correlates with LA volume. Further studies are warranted to verify whether this parameter can be used clinically as a marker of cerebrovascular risk.

90.   Link between Automated Coronary Calcium Volumes from Intravascular Ultrasound (IVUS) to Automated Carotid IMT from B-mode Ultrasound in Coronary Artery Disease Population

    Tadashi Araki, Nobutaka Ikeda, Filippo Molinari, Nilanjan Dey, Suvojit Acharjee, Luca Saba, Jasjit S Suri.

    INTERNATIONAL ANGIOLOGY. 2014. (Accepted) [Impact Factor: 1.462]

    Abstract: Establishing relationship between coronary calcium volumes from Intravascular Ultrasound (IVUS) and automated carotid intima-media thickness (cIMT) helps in understanding the genetic nature of atherosclerosis disease. In this research, we have quantified the detected calcium from IVUS video frames and associated a relationship between coronary calcium volumes computed and automated cIMT from B-mode ultrasound.The computer-based coronary calcium volume (from IVUS) showed a correlation coefficient with respect to cIMT for left and right carotids as 9.1% and 13.9%, respectively. Coronary calcium volume computed from IVUS and auto cIMT are moderately correlated. The association between auto cIMT (right side) vs. computer-based coronary calcium volume (IVUS) is stronger than the association between auto cIMT (left side) vs. computer-based coronary calcium volume.

91.   Shape-based approach for Calcium Lesion Volume Measurements in Coronary Artery using IVUS and its association to Carotid IMT using B-mode Ultrasound

    Tadashi Araki, Nobutaka Ikeda, Filippo Molinari, Nilanjan Dey, Suvojit Acharjee, Luca Saba, Andrew Nicolaides, Jasjit S Suri

    IEEE Transactions on Instrumentation and Measurement. 2014.(Under Review) [Impact Factor: 1.357]

    Abstract: Coronary calcification plays a significant role in diagnostic classification of lesion subsets. According to histopathological studies, vulnerable atherosclerotic plaques contain calcific deposits. Even though, there can be a considerable variation in the extent and degree of calcification. Intravascular ultrasound (IVUS) performs a noteworthy role in the identification and quantification of coronary calcification. Volumetric measurement and analysis to quantify the deposited calcium is a challenging task because of the irregularity and randomness of the shape of the detected calcium per frame per volume. A need was felt for an automated system, as the manual quantification is one of the most complex and tedious processes. We have developed a system which can automatically detect the calcium detected frames, identify the calcium and can also perform a shape-based measurement of volume from IVUS. We have validated our results obtained from shape-based study (shape-based method) with simple integration method. To our knowledge, the association between ultrasound carotid intima-media thickness and IVUS coronary calcium volume is yet not reported, which can help in understanding the genetic nature of atherosclerosis disease. In this current study, we have tried to establish the relationship between coronary calcium volume from IVUS and carotid intimal medial thickness (cIMT). We here demonstrate the relationship between automated computed IMT using AtheroEdge™ software and automated calcium volume computed by IVUS using integration approach and shape-based approach. We have also compared the results obtained from both the approaches.

92.   Coronary calcium lesion attributes for cardiovascular risk using Intravascular Ultrasound and Carotid B-mode Ultrasound

    Tadashi Araki, Nobutaka Ikeda, Filippo Molinari, Nilanjan Dey, Suvojit Acharjee, Luca Saba, Andrew Nicolaides, Jasjit S Suri.

    Circulation Journal (Japanese Circulation Society). 2014.(Under Review) [Impact Factor: 3.578]

    Abstract: The study of sensitivity analysis of coronary calcium quantification in neurological risk patients is significantly improved by the linear combination of shape based lesion characteristic features with normalized volume.Receiver Operating Characteristic (ROC) curve analysis illustrates the statistical significance of the established association which helps in analyzing the sensitivity of coronary calcium quantification in neurological risk patients. The analysis reported significant improvement of obtained AUC by 16% after incorporation of the linear combination of shape based features along with the normalized volume using integration method.

Back to Top
Home

29.   Ankle Brachial Index (ABI) and its link to Automated Carotid Ultrasound IMT Variability (IMTV) in 500 Japanese Coronary Artery Disease Patients.

    Nobutaka Ikeda, Tadashi Araki, Kaoru Sugi, Masatako Nakamura, Martino Deidda, Filippo Molinari, Kristen M. Meiburger, U Rajendra Acharya, Luca Saba, Pier Paolo Bassareo, Michele Di Martino, Yoshinori Nagashima, Giuseppe Mercuro, Masataka Nakano, Andrew Nicolaides, Jasjit S. Suri.

    Current Atherosclerosis Reports. 2013. [Impact Factor: 2.923]

    Abstract:   The purpose of this study was to evaluate whether the carotid intima-media thickness (cIMT) and its variability (IMTV) along the artery was correlated to ankle-brachial index (ABI) in Japanese Coronary Artery Disease Patients. 500 consecutive patients (males 312; median age 69 years ± 11) who underwent carotid-US and first coronary angiography were prospectively analyzed. By using an automated software (AtheroEdge), cIMT and IMTV were obtained. Pearson correlation analysis was performed to calculate the association between automated ABI, cIMT, IMTV, and SYNTAX score. The mean value of cIMT was 0.881 ± 0.334 mm and the mean IMTV value was 0.141 ± 0.112. IMTV was negatively and significantly correlated to ABI (rho = -0.147; p = 0.001), whereas cIMT was not (rho = -0.075; p = 0.097). IMTV and cIMT had same significant correlation with SYNTAX score. When we considered patients with higher risk factor (ABI £ 0.9), we found higher values of IMTV, PS, and SYNTAX score, but not of cIMT. Logistic regression analysis showed that IMTV was independently associated to the complexity of the CAD (as assessed by the SYNTAX score). In Conclusion we showed that automatically measured IMTV using AtheroEdge was associated to ABI, whereas the intima-media thickness alone was not. IMTV could be integrated with cIMT measurement to improve the assessment of CVD.

30.   Semiautomated analysis of carotid artery wall thickness in MRI.

    Saba L, Gao H, Raz E, Sree SV, Mannelli L, Tallapally N, Molinari F, Bassareo PP, Acharya UR, Poppert H, Suri JS.

    J Magn Reson Imaging. 2013 Oct 22. doi: 10.1002/jmri.24307. [Impact Factor: 2.7]

    Abstract:   To develop a semiautomatic method based on level set method (LSM) for carotid arterial wall thickness (CAWT) measurement. Magnetic resonance imaging (MRI) of diseased carotid arteries was acquired from 10 patients. Ground truth (GT) data for arterial wall segmentation was collected from three experienced vascular clinicians. The semiautomatic variational LSM was employed to segment lumen and arterial wall outer boundaries on 102 MR images. Two computer-based measurements, arterial wall thickness (WT) and arterial wall area (AWA), were computed and compared with GT. Linear regression, Bland-Altman, and bias correlation analysis on WT and AWA were applied for evaluating the performance of the semiautomatic method. The proposed LSM can generate reasonably accurate lumen and outer wall boundaries compared to manual segmentation, and can work similar to a human reader. However, it tends to overestimate CAWT and AWA compared to the manual segmentation for arterial wall with small area.

31.   Automated classification of patients with coronary artery disease using grayscale features from left ventricle echocardiographic images.

    Acharya UR, Sree SV, Muthu Rama Krishnan M, Krishnananda N, Ranjan S, Umesh P, Suri JS.

    Comput Methods Programs Biomed. 2013 Aug 16. pii: S0169-2607(13)00247-2. doi: 10.1016/j.cmpb.2013.07.012. [Impact Factor: 1.555]

    Abstract:   Coronary Artery Disease (CAD), caused by the buildup of plaque on the inside of the coronary arteries, has a high mortality rate. To efficiently detect this condition from echocardiography images, with lesser inter-observer variability and visual interpretation errors, computer based data mining techniques may be exploited. We have developed and presented one such technique in this paper for the classification of normal and CAD affected cases. A multitude of grayscale features (fractal dimension, entropies based on the higher order spectra, features based on image texture and local binary patterns, and wavelet based features) were extracted from echocardiography images belonging to a huge database of 400 normal cases and 400 CAD patients. Only the features that had good discriminating capability were selected using t-test. Several combinations of the resultant significant features were used to evaluate many supervised classifiers to find the combination that presents a good accuracy. We observed that the Gaussian Mixture Model (GMM) classifier trained with a feature subset made up of nine significant features presented the highest accuracy, sensitivity, specificity, and positive predictive value of 100%. We have also developed a novel, highly discriminative HeartIndex, which is a single number that is calculated from the combination of the features, in order to objectively classify the images from either of the two classes. Such an index allows for an easier implementation of the technique for automated CAD detection in the computers in hospitals and clinics.

32.   Asymptomatic Carotid Disease—A New Tool for Assessing Neurological Risk.

    Luís M. Pedro, J. Miguel Sanches, José Seabra, Jasjit S. Suri, José Fernandes e Fernandes.

    Echocardiography. 2013 Sep 30. 00:1–9. doi: 10.1111/echo.12348. [Impact Factor: 1.261]

    Abstract:   Active carotid plaques are associated with atheroembolism and neurological events; its identification is crucial for stroke prevention. High-definition ultrasound (HDU) can be used to recognize plaque structure in carotid bifurcation stenosis associated with plaque vulnerability and occurrence of brain ischemic events. A new computer-assisted HDU method to study the echomorphology of the carotid plaque and to determine a risk score for developing appropriate symptoms is proposed in this study. Plaque echomorphology characteristics such as presence of ulceration at the plaque surface, juxta-luminal location of echolucent areas, echoheterogeneity were obtained from B-mode ultrasound scans using several image processing algorithms and were combined with measurement of severity of stenosis to obtain a clinical score—enhanced activity index (EAI)—which was correlated with the presence or absence of ipsilateral appropriate ischemic symptoms. An optimal cutoff value of EAI was determined to obtain the best separation between symptomatic (active) from asymptomatic (inactive) plaques and its diagnostic yield was compared to other 2 reference methods by means of Receiver-Operating Characteristic (ROC) analysis. Classification performance was evaluated by leave-one-patient-out cross-validation applied to a cohort of 146 carotid plaques from 99 patients. The proposed method was benchmarked against (a) degree of stenosis criteria and (b) earlier proposed activity index (AI) and demonstrated that EAI yielded the highest accuracy up to an accuracy of 77% to predict asymptomatic plaques that developed symptoms in a prospective cross-sectional study. Enhanced activity index is a noninvasive, easy to obtain parameter, which provided accurate estimation of neurological risk of carotid plaques.

33.   Inter- and Intra-Observer Variability Analysis of Completely Automated cIMT Measurement Software (AtheroEdge™) and its Benchmarking against Commercial Ultrasound Scanner and Expert Readers.

    Saba L, Molinari F, Meiburger KM, Acharya UR, Nicolaides A, Suri JS.

    Computers in Biology and Medicine. 2013 Sep; 43(9) pp. 1261-1272. doi: 10.1016/j.compbiomed.2013.06.012. [Impact Factor: 1.162]

    Abstract:   The purpose of this study was to evaluate the measurement error and inter- and intra- observer variability of completely off-line automated and semi-automated carotid intima-media thickness (cIMT) measurement software (AtheroEdge™). Two hundred carotid ultrasound images from 50 asymptomatic women were analyzed. AtheroEdge™ was bench marked against a commercial system(Syngo, Siemens) using automated and semi-automated modes. The measurement error and inter- and intra- observer variability of AtheroEdge™ were tested using three readings. The measurement error of AtheroEdge™ compared to the commercial software was 0.002±0.019 mm (r = 0.99) in the automated mode and -0.001±0.004 mm in the semi-automated mode (r = 0.99). The measurement error of AtheroEdge™ compared to the mean value of the three expert Readers (cIMT bias)for the automated and semi-automated methods was -0.0004±0.158 mm and -0.008±0.157 mm, respectively.The Figure-of-Merit was 99.8% and 99.9% when compared to the commercial ultrasound scanner (using the automated and semi-automated method, respectively) and was 99.9% and 98.9% when compared to the mean value of the three expert Readers.Regarding inter- and intra- observer variability, the intra-class correlation coefficient of the three independent users using the semi-automated AtheroEdge™ was 0.98. AtheroEdge™ showed a measurement performance comparable to the commercial ultrasound scanner software and the expert Readers’ tracings. AtheroEdge™ belongs to a class of automated systems that could find application in processing large datasets for common carotid arteries, avoiding subjectivity in cIMT measurements.

34.   Carotid artery wall thickness measured using CT: inter- and intraobserver agreement analysis.

    Saba L, Sanfilippo R, Montisci R, Suri JS, Mallarini G.

    AJNR Am J Neuroradiol. 2013 Feb;34(2):E13-8. doi: 10.3174/ajnr.A2796. Epub 2011 Nov 11. [Impact Factor: 3.167]

    Abstract:   The purpose of this work was to compare inter- and intraobserver agreement in the analysis of CAWT by using MDCTA. The CAWT in 35 patients was quantified by 4 observers. Bland-Altman statistics were used to measure the agreement between observers. The results of our study demonstrated that the CAWT measured by using MDCTA shows a good reproducibility between observers by considering inter- and intraobserver agreement.

35.   Automated carotid intima-media thickness and its link for prediction of SYNTAX score in Japanese coronary artery disease patients.

    Ikeda N, Saba L, Molinari F, Piga M, Meiburger K, Sugi K, Porcu M, Bocchiddi L, Acharya UR, Nakamura M, Nakano M, Nicolaides A, Suri JS.

    Int Angiol. 2013 Jun;32(3):339-48. [Impact Factor: 1.462]

    Abstract: The purpose of this study was to evaluate whether the automated carotid intima-media thickness (CIMT) identified by using automated software could predict the SYNTAX score for coronary artery disease (CAD) patients. Results of our study using an automated algorithm showed a statistical significant association between CIMT and SYNTAX score and indicated that CIMT may be considered a reliable parameter for prediction of SYNTAX score in Coronary Artery Disease patient population from Japan.

36.   Computed tomography carotid wall plaque characterization using a combination of discrete wavelet transform and texture features: A pilot study.

    Acharya U, Sree SV, Mookiah M, Saba L, Gao H, Mallarini G, Suri JS.

    Proc Inst Mech Eng H. 2013 Jun;227(6):643-54. doi: 10.1177/0954411913480622. Epub 2013 Mar 22. [Impact Factor: 1.419]

    Abstract: In 30% of stroke victims, the cause of stroke has been found to be the stenosis caused by plaques in the carotid artery. Early detection of plaque and subsequent classification of the same into symptomatic and asymptomatic can help the clinicians to choose only those patients who are at a higher risk of stroke for risky surgeries and stenosis treatments. Therefore, in this work, we have proposed a non-invasive computer-aided diagnostic technique to classify the detected plaque into the two classes. Computed tomography (CT) images of the carotid artery images were used to extract Local Binary Pattern (LBP) features and wavelet energy features. Significant features were then used to train and test several supervised learning algorithm based classifiers. Our proposed technique enables automatic classification of plaque into asymptomatic and symptomatic with high accuracy, and hence, it can be used for deciding the course of treatment. We have also proposed a single-valued integrated index (Atheromatic Index) using the significant features which can provide a more objective and faster prediction of the class.

37.   Semiautomated and automated algorithms for analysis of the carotid artery wall on computed tomography and sonography: a correlation study.

    Saba L, Tallapally N, Gao H, Molinari F, Anzidei M, Piga M, Sanfilippo R, Suri JS.

    J Ultrasound Med. 2013 Apr;32(4):665-74. [Impact Factor: 1.402]

    Abstract: The purpose of this study was to compare automated and semiautomated algorithms for analysis of carotid artery wall thickness and intima- media thickness on multidetector row computed tomographic (CT) angiography and sonography, respectively, and to study the correlation between them. The results of this preliminary study showed that carotid artery wall thickness and intima-media thickness can be studied with automated software, although the CT analysis needs to be further improved.

38.   Atherosclerotic plaque tissue characterization in 2D ultrasound longitudinal carotid scans for automated classification: a paradigm for stroke risk assessment.

    Acharya UR, Mookiah MR, Vinitha Sree S, Afonso D, Sanches J, Shafique S, Nicolaides A, Pedro LM, Fernandes E Fernandes J, Suri JS.

    Med Biol Eng Comput. 2013 May;51(5):513-23. doi: 10.1007/s11517-012-1019-0. Epub 2013 Jan 6. [Impact Factor: 1.790]

    Abstract: In the case of carotid atherosclerosis, to avoid unnecessary surgeries in asymptomatic patients, it is necessary to develop a technique to effectively differentiate symptomatic and asymptomatic plaques. In this paper, we have presented a data mining framework that characterizes the textural differences in these two classes using several grayscale features based on a novel combination of trace transform and fuzzy texture.

39.   Understanding symptomatology of atherosclerotic plaque by image-based tissue characterization.

    Acharya UR, Faust O, S VS, Alvin AP, Krishnamurthi G, Seabra JC, Sanches J, Suri JS.

    Comput Methods Programs Biomed. 2013 Apr;110(1):66-75. doi: 10.1016/j.cmpb.2012.09.008. Epub 2012 Nov 1. [Impact Factor: 1.555]

    Abstract:  Characterization of carotid atherosclerosis and classification into either symptomatic or asymptomatic is crucial in terms of diagnosis and treatment planning for a range of cardiovascular diseases. This paper presents a computer-aided diagnosis (CAD) system (Atheromatic) that analyzes ultrasound images and classifies them into symptomatic and asymptomatic. Each analyzed ultrasound image yields on SACI (symptomatic asymptomatic carotid index)  number. A high SACI value indicates that the image shows symptomatic and low value indicates asymptomatic plaques.

40.   Association of automated carotid IMT measurement and HbA1c in Japanese patients with coronary artery disease.

    Saba L, Ikeda N, Deidda M, Araki T, Molinari F, Meiburger KM, Acharya UR, Nagashima Y, Mercuro G, Nakano M,Nicolaides A, Suri JS.

    Diabetes Res Clin Pract. 2013 Apr 20. pii: S0168-8227(13)00131-9. doi: 10.1016/j.diabres.2013.03.032. [Epub ahead of print] [Impact Factor: 2.741]

    Abstract:  The purpose of this study was to evaluate whether carotid IMT (cIMT) identified using automated software is associated with HbA1c in Japanese patients with coronary artery disease. The results of our study confirm that automated cIMT values and levels of HbA1c in Japanese patients with coronary artery disease are correlated whereas the plaque score does not show a statistically significant correlation.

Back to Top
Home

17.   What is the correct distance measurement metric when measuring carotidultrasound intima-media thickness automatically?

    Saba L, Molinari F, Meiburger KM, Piga M, Zeng G, Rajendra Achraya U, Nicolaides A, Suri JS.

    Int Angiol. 2012 Oct;31(5):483-9. [Impact Factor: 1.462]

    Abstract:  The aims of this study were: 1) to analyze the interobserver carotid intima-media thickness (CIMT) variability using three different measurement metrics on large multi-institutional databases; 2) to evaluate the three kinds of metrics when comparing completely automated CIMT measurement (Auto Edge CIMT) to two manually derived CIMT (manual CIMT). Centerline and Polyline yield very close results and are clinically suitable distance measurement techniques for computing the CIMT from LI/MA profiles.

18.   Carotid IMT variability (IMTV) and its validation in symptomatic versus asymptomatic Italian population: can this be a useful index for studying symptomaticity?

    Saba L, Meiburger KM, Molinari F, Ledda G, Anzidei M, Acharya UR, Zeng G, Shafique S, Nicolaides A, Suri JS.

    Echocardiography. 2012 Oct;29(9):1111-9. doi: 10.1111/j.1540-8175.2012.01763.x. Epub 2012 Jul 2. [Impact Factor: 1.261]

    Abstract:  The carotid intima-media thickness (IMT) is a validated marker of cerebrovascular disease risk. This paper presents a new parameter, the IMT variability (IMTV), and compares it between symptomatic and asymptomatic patients taken from a cohort of Italian population. We conclude that the IMT and IMTV values were very similar between Reader and AutoEdge software when studying symptomatic and asymptomatic patients in Italian population.

19.   Association between carotid artery plaque volume, composition, and ulceration: a retrospective assessment with MDCT.

    Saba L, Sanfilippo R, Sannia S, Anzidei M, Montisci R, Mallarini G, Suri JS.

    AJR Am J Roentgenol. 2012 Jul;199(1):151-6. doi: 10.2214/AJR.11.6955. [Impact Factor: 2.897]

    Abstract:  The purpose of this study was to evaluate the relationship between MDCT angiography-assessed carotid artery plaque volume and composition and the presence of ulceration. The results of our retrospective study suggest that there is no correlation between total carotidatherosclerotic plaque volume and ulcerations, whereas plaque relative lipid volume (using attenuation of < 60 HU) is associated with the presence of ulceration. This finding could indicate vulnerable plaques and increased risk for cerebrovascular events.

20.   Ultrasound IMT measurement on a multi-ethnic and multi-institutional database: our review and experience using four fully automated and one semi-automated methods.

    Molinari F, Meiburger KM, Saba L, Acharya UR, Ledda G, Zeng G, Ho SY, Ahuja AT, Ho SC, Nicolaides A, Suri JS.

    Comput Methods Programs Biomed. 2012 Dec;108(3):946-60. doi: 10.1016/j.cmpb.2012.05.008. Epub 2012 Jun 1. [Impact Factor: 1.555]

    Abstract:  Automated and high performance carotid intima-media thickness (IMT) measurement is gaining increasing importance in clinical practice to assess the cardiovascular risk of patients. In this paper, we compare four fully automated IMT measurement techniques (CALEX, CAMES, CARES and CAUDLES) and one semi-automated technique (FOAM). This is the first time that completely automated and user-driven techniques have been compared on a multi-ethnic dataset, acquired using multiple original equipment manufacturer (OEM) machines with different gain settings, representing normal and pathologic cases.

21.   Understanding symptomatology of atherosclerotic plaque by image-based tissue characterization.

    Acharya UR, Faust O, S VS, Alvin AP, Krishnamurthi G, Seabra JC, Sanches J, Suri JS.

    Comput Methods Programs Biomed.2013 Apr;110(1):66-75. doi: 10.1016/j.cmpb.2012.09.008. Epub 2012 Nov 1. [Impact Factor: 1.555]

    Abstract:  Characterization of carotid atherosclerosis and classification into either symptomatic or asymptomatic is crucial in terms of diagnosis and treatment planning for a range of cardiovascular diseases. This paper presents a computer-aided diagnosis (CAD) system (Atheromatic) that analyzes ultrasound images and classifies them into symptomatic and asymptomatic. The classification result is based on a combination of discrete wavelet transform, higher order spectra (HOS) and textural features. In this study, we compare support vector machine (SVM) classifiers with different kernels. The classifier with a radial basis function (RBF) kernel achieved an average accuracy of 91.7% as well as a sensitivity of 97%, and specificity of 80%. Thus, it is evident that the selected features and the classifier combination can efficiently categorize plaques into symptomatic and asymptomatic classes. Moreover, a novel symptomatic asymptomatic carotid index (SACI), which is an integrated index that is based on the significant features, has been proposed in this work. Each analyzed ultrasound image yields on SACI number. A high SACI value indicates that the image shows symptomatic and low value indicates asymptomatic plaques. We hope this SACI can support vascular surgeons during routine screening for asymptomatic plaques.

22.   Association between carotid artery plaque type and cerebral microbleeds.

    Saba L, Montisci R, Raz E, Sanfilippo R, Suri JS, Piga M.

    AJNR Am J Neuroradiol. 2012 Dec;33(11):2144-50. doi: 10.3174/ajnr.A3133. Epub 2012 May 24. [Impact Factor: 3.167]

    Abstract:  CMBs have become increasingly recognized with the widespread use of MR imaging techniques that are sensitive to iron deposits. The purpose of this study was to correlate the presence of CMBs and carotid plaque characteristics.The results of this study suggest an association between the presence of carotid artery fatty plaque, symptoms, and CMBs. Moreover, we found that the presence (and entity) of CMBs may represent an indicator of cerebrovascular symptom severity.

23.   Analysis of carotid artery plaque and wall boundaries on CT images by using a semi-automatic method based on level set model.

    Saba L, Gao H, Acharya UR, Sannia S, Ledda G, Suri JS.

    Neuroradiology. 2012 Nov;54(11):1207-14. doi: 10.1007/s00234-012-1040-x. Epub 2012 May 6. [Impact Factor: 2.700]

    Abstract:  The purpose of this study was to evaluate the potentialities of a semi-automated technique in the detection and measurement of the carotid artery plaque. Results of our study indicate that LSM (level-set-method) method can automatically measure the thickness of the plaque and that the best results are obtained with the RDM (radial distance method). Our results suggest that advanced computer-based algorithms can identify and trace the plaque boundaries like an experienced human reader.

24.   Atherosclerotic risk stratification strategy for carotid arteries using texture-based features.

    Acharya UR, Sree SV, Krishnan MM, Molinari F, Saba L, Ho SY, Ahuja AT, Ho SC, Nicolaides A, Suri JS.

    Ultrasound Med Biol. 2012 Jun;38(6):899-915. doi: 10.1016/j.ultrasmedbio.2012.01.015. Epub 2012 Apr 21. [Impact Factor: 2.455]

    Abstract:  Plaques in the carotid artery result in stenosis, which is one of the main causes for stroke. Patients have to be carefully selected for stenosis treatments as they carry some risk. We present a computer aided diagnostic (CAD) based ultrasound characterization methodology (a class of Atheromatic systems) that classifies the patient into symptomatic and asymptomatic classes using two kinds of datasets: (1) plaque regions in ultrasound carotids segmented semi-automatically and (2) far wall gray-scale intima-media thickness (IMT) regions along the common carotid artery segmented automatically. We have also proposed an index for each type of datasets: AtheromaticPi, forcarotid plaque region, and AtheromaticWi, for IMT carotid wall region, based on the combination of the respective significant features. These indices show a separation between symptomatic and asymptomatic by 4.53 units and 4.42 units, respectively, thereby supporting the texture hypothesis classification.

25.   Automated carotid IMT measurement and its validation in low contrast ultrasound database of 885 patient Indian population epidemiological study: results of AtheroEdge Software.

    Molinari F, Meiburger KM, Zeng G, Saba L, Rajendra Acharya U, Famiglietti L, Georgiou N, Nicolaides A, Sriswan Mamidi R, Kuper H, Suri JS.

    Int Angiol. 2012 Feb;31(1):42-53. [Impact Factor: 1.462]

    Abstract:  The aim of this paper was to demonstrate the usage of an automated computer-based IMT measurement system called – CALEX 3.0 (a class of patented AtheroEdge software) on a low contrast and low resolution image database acquired during an epidemiological study from India. The image contrast was very low with pixel density of 12.7 pixels/mm. Further, to demonstrate the accuracy and reproducibility of the AtheroEdge™ software system we compared it with the manual tracings of a vascular surgeon–considered as a gold standard.CALEX 3.0 had a 100% success in processing low contrast/low-resolution images. CALEX 3.0 is the first technique, which has led to high accuracy and reproducibility on low-resolution images acquired during an epidemiological study. We propose CALEX 3.0 as a generalized framework for IMT measurement on large datasets.

26.   Intima Media Thickness Variability (IMTV) and its association with cerebrovascular events: A novel marker of carotid atherosclerosis?

    Luca Saba, Giorgio Mallarini, Roberto Sanfilippo, Roberto Montisci, Jasjit S. Suri.

    Cardiovascular Diagnosis and Therapy.2012;2(1):10-18. doi: 10.3978/j.issn.2223-3652.2011.11.01.

    Abstract:  Increased intima-media thickness (IMT) is an earlier marker of atherosclerotic disease and several prospective studies have demonstrated that IMT is a strong predictor of cerebrovascular complications. In this paper we propose a novel method to assess IMT, called “intima media thickness variability” (IMTV) and evaluate its relationship with the development of cerebrovascular events. The study was approved by the local IRB. Twenty consecutive patients underwent ultrasound analysis of the carotid arteries (mean age 68 years; age range 59-81 years). The IMT and IMTV of the 40 carotid vessels was assessed. Bland–Altman statistics were employed to measure the inter-observer variability, and ROC analysis was used to assess the association with cerebrovascular events. In all cases the ROC area under the curve was higher for IMTV than IMT. There was a statistical significant association between IMTV and cerebrovascular events (P=0.018), but no relationship between IMT and events. The results for inter-observer variability showed a systematic error between 0.04 mm and 0.08 mm. In the Person Rho correlation analysis in no case a statistical association between IMT and symptoms was detected whereas in 2 cases (observer 1 and observer 2), a statistically significant association between IMTV and symptoms was found. Data of this preliminary study suggest that IMTV may represent a powerful method to assess carotid atherosclerotic disease with a significant association to cerebrovascular events.

27.   Comparison between manual and automated analysis for the quantification of carotid wall by using sonography. A validation study with CT.

    Saba L, Montisci R, Molinari F, Tallapally N, Zeng G, Mallarini G, Suri JS.

    Eur J Radiol. 2012 May;81(5):911-8. doi: 10.1016/j.ejrad.2011.02.047. Epub 2011 Mar 23. [Impact Factor: 2.512]

    Abstract:  The purpose of this paper was to compare manual and automated analysis for the quantification of carotid wall obtained with sonography by using the computed tomography as validation technique. On comparing AtheroEdge (using Ultrasound) with CAWT (using CT), the results suggested a very good concordance of 84.96%.Data of this preliminary study indicate that automated software AtheroEdge™ can analyze with precision the IMT of carotid arteries and that the concordance with CT is optimal.

28.   A state-of-the-art review on segmentation algorithms in intravascular ultrasound (IVUS) images.

    Katouzian A, Angelini ED, Carlier SG, Suri JS, Navab N, Laine AF.

    IEEE Trans Inf Technol Biomed.2012 Sep;16(5):823-34. Epub 2012 Feb 28. [Impact Factor: 1.978]

    Abstract:  Over the past two decades, intravascular ultrasound (IVUS) image segmentation has remained a challenge for researchers while the use of this imaging modality is rapidly growing in catheterization procedures and in research studies. IVUS provides cross-sectional grayscale images of the arterial wall and the extent of atherosclerotic plaques with high spatial resolution in real time. In this paper, we review recently developed image processing methods for the detection of media-adventitia and luminal borders in IVUS images acquired with different transducers operating at frequencies ranging from 20 to 45 MHz. We discuss methodological challenges, lack of diversity in reported datasets, and weaknesses of quantification metrics that make IVUS segmentation still an open problem despite all efforts. In conclusion, we call for a common reference database, validation metrics, and ground-truth definition with which new and existing algorithms could be benchmarked.

29.   Symptomatic vs. asymptomatic plaque classification in carotid ultrasound.

    Acharya RU, Faust O, Alvin AP, Sree SV, Molinari F, Saba L, Nicolaides A, Suri JS.

    J Med Syst. 2012 Jun;36(3):1861-71. doi: 10.1007/s10916-010-9645-2. Epub 2011 Jan 18. [Impact Factor: 1.783]

    Abstract:  Quantitative characterization of carotid atherosclerosis and classification into symptomatic or asymptomatic type is crucial in both diagnosis and treatment planning. This paper describes a computer-aided diagnosis (CAD) system which analyzes ultrasound images and classifies them into symptomatic and asymptomatic based on the textural features. The proposed CAD system consists of three modules. The first module is preprocessing, which conditions the images for the subsequent feature extraction. The feature extraction stage uses image texture analysis to calculate Standard deviation, Entropy, Symmetry, and Run Percentage. Finally, classification is performed using AdaBoost and Support Vector Machine for automated decision making. An Integrated Index, called symptomatic asymptomatic carotid index (SACI), is proposed using texture features to discriminate symptomatic and asymptomatic carotid ultrasound images using just one index or number. We hope this SACI can be used as an adjunct tool by the vascular surgeons for daily screening.

Back to Top
Home

4.   Completely automated robust edge snapper for carotid ultrasound IMT measurement on a multi-institutional database of 300 images.

   Molinari F, Rajendra Acharya U, Zeng G, Meiburger KM, Suri JS.

   Med Biol Eng Comput. 2011 Aug;49(8):935-45. doi: 10.1007/s11517-011-0781-8. Epub 2011 Apr 21. [Impact Factor: 1.790]

   Abstract:  The carotid intima-media thickness (IMT) is the most used marker for the progression of atherosclerosis and onset of cardiovascular diseases. Computer-aided measurements improve accuracy and precision, but usually require user interaction. In this paper we characterized a new and completely automated technique for carotid segmentation and IMT measurement based on the merits of two previously developed techniques. We used an integrated approach of intelligent image feature extraction and line fitting for automatically locating the carotid artery in the image frame, followed by wall interfaces extraction based on a Gaussian edge operator. We called our system-CARES. CARES could be a useful research tool for processing large datasets in multi-center studies involving atherosclerosis.

Back to Top
Home

1.   Intima-media thickness: setting a standard for a completely automated method of ultrasound measurement.

   Molinari F, Zeng G, Suri JS.

   IEEE Trans Ultrason Ferroelectr Freq Control. 2010 May;57(5):1112-24. doi: 10.1109/TUFFC.2010.1522. [Impact Factor: 1.822]

   Abstract:  The intima-media thickness (IMT) of the common carotid artery is a widely used clinical marker of severe cardiovascular diseases. IMT is usually manually measured on longitudinal B-mode ultrasound images. Many computer-based techniques for IMT measurement have been proposed to overcome the limits of manual segmentation. Most of these, however, require a certain degree of user interaction. In this paper we describe a new, completely automated layer extraction technique (named CALEXia) for the segmentation and IMT measurement of the carotid wall in ultrasound images. CALEXia is based on an integrated approach consisting of feature extraction, line fitting, and classification that enables the automated tracing of the carotid adventitial walls. IMT is then measured by relying on a fuzzy K-means classifier.

2.   An integrated approach to computer-based automated tracing and its validation for 200 common carotid arterial wall ultrasound images: a new technique.

   Molinari F, Zeng G, Suri JS.

   J Ultrasound Med. 2010 Mar;29(3):399-418. [Impact Factor: 1.402]

   Abstract:  Most of the algorithms for the segmentation of the common carotid artery (CCA) wall require human interaction to locate the vessel in the ultrasound image. The aim of this article is to show an accurate algorithm for the computer-based automated tracing of the CCA in longitudinal B-mode ultrasound images. Although the CALEXia algorithm automatically detects the CCA, it is also robust and validated over a large database. This can constitute a general basis for a completely automated segmentation procedure widely applicable to other anatomies.

3.   A state of the art review on intima-media thickness (IMT) measurement and wall segmentation techniques for carotid ultrasound.

   Molinari F, Zeng G, Suri JS.

   Comput Methods Programs Biomed.2010 Dec;100(3):201-21. doi: 10.1016/j.cmpb.2010.04.007. Epub 2010 May 15. [Impact Factor: 1.555]

   Abstract:  Last 10 years have witnessed the growth of many computer applications for the segmentation of the vessel wall in ultrasound imaging. Epidemiological studies showed that the thickness of the major arteries is an early and effective marker of onset of cardiovascular diseases. Ultrasound imaging, being real-time, economic, reliable, safe, and now seems to become a standard in vascular assessment methodology. This review is an attempt to discuss the most performing methodologies that have been developed so far to perform computer-based segmentation and intima-media thickness (IMT) measurement of the carotid arteries in ultrasound images. First we will present the rationale and the clinical relevance of computer-based measurements in clinical practice, followed by the challenges that one has to face when approaching the segmentation of ultrasound vascular images. The core of the paper is the presentation, discussion, benchmarking and evaluation of different segmentation techniques, including: edge-detection, active contours, dynamic programming, local statistics, Hough transform, statistical modeling, and integration of these approaches. Also, we will discuss and compare the different performance metrics that have been proposed and used to perform the validation. Best performing user-dependent techniques show an average IMT measurement error of about 1μm when compared to human tracings [57], whereas completely automated techniques show errors of about 10μm. The review ends with a discussion about the current standards in carotid wall segmentation and in an overview of the future perspectives, which may include the adoption of advanced and intelligent strategies to let the computer technique measure the IMT in the image portion where measurement is more reliable.

Back to Top
Home