Limits...
Automated detection framework of the calcified plaque with acoustic shadowing in IVUS images.

Gao Z, Guo W, Liu X, Huang W, Zhang H, Tan N, Hau WK, Zhang YT, Liu H - PLoS ONE (2014)

Bottom Line: However, it is very time-consuming, and with high inter-observer and intra-observer variability between different doctors.The experimental results are quantitatively analyzed separately by three evaluation methods, the test of the sensitivity and specificity, the linear regression and the Bland-Altman analysis.High sensitivity (94.68%) and specificity (95.82%), good correlation and agreement (>96.82% results fall within the 95% confidence interval in the Student t-test) demonstrate the effectiveness of the proposed method in the detection of the calcified plaque with acoustic shadowing in IVUS images.

View Article: PubMed Central - PubMed

Affiliation: Shenzhen College of Advanced Technology, University of Chinese Academy of Sciences, Shenzhen, China; Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China; Key Laboratory of Biomedical information and Health Engineering, Chinese Academy of Sciences, Shenzhen, China.

ABSTRACT
Intravascular Ultrasound (IVUS) is one ultrasonic imaging technology to acquire vascular cross-sectional images for the visualization of the inner vessel structure. This technique has been widely used for the diagnosis and treatment of coronary artery diseases. The detection of the calcified plaque with acoustic shadowing in IVUS images plays a vital role in the quantitative analysis of atheromatous plaques. The conventional method of the calcium detection is manual drawing by the doctors. However, it is very time-consuming, and with high inter-observer and intra-observer variability between different doctors. Therefore, the computer-aided detection of the calcified plaque is highly desired. In this paper, an automated method is proposed to detect the calcified plaque with acoustic shadowing in IVUS images by the Rayleigh mixture model, the Markov random field, the graph searching method and the prior knowledge about the calcified plaque. The performance of our method was evaluated over 996 in-vivo IVUS images acquired from eight patients, and the detected calcified plaques are compared with manually detected calcified plaques by one cardiology doctor. The experimental results are quantitatively analyzed separately by three evaluation methods, the test of the sensitivity and specificity, the linear regression and the Bland-Altman analysis. The first method is used to evaluate the ability to distinguish between IVUS images with and without the calcified plaque, and the latter two methods can respectively measure the correlation and the agreement between our results and manual drawing results for locating the calcified plaque in the IVUS image. High sensitivity (94.68%) and specificity (95.82%), good correlation and agreement (>96.82% results fall within the 95% confidence interval in the Student t-test) demonstrate the effectiveness of the proposed method in the detection of the calcified plaque with acoustic shadowing in IVUS images.

Show MeSH

Related in: MedlinePlus

The results of the linear regression with respect to AL, AC, PA, PT and DC for all patients.The horizontal axis and vertical axis represent results acquired from our method and the manual drawing, respectively.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4220935&req=5

pone-0109997-g006: The results of the linear regression with respect to AL, AC, PA, PT and DC for all patients.The horizontal axis and vertical axis represent results acquired from our method and the manual drawing, respectively.

Mentions: PT and DC can measure the radial location (the location along the radial direction) of the calcified plaque, where DC equals the distance between the leading edge of the calcified plaque and the catheter's center. These measurements calculated from our method and the manual drawing method were compared separately by two different evaluation methods: linear regression and Bland-Altman analysis [39], in order to analyze their correlation and agreement, respectively. In every evaluation method, we firstly analyze the comparative results on all IVUS images in order to evaluate the overall performance of our method, and then on IVUS images from every patient in order to evaluate the between-patient difference of our method. In linear regression, the correlation coefficient and the root-mean-square error (RMSE) are used to investigate the overall correlation between our method and the manual drawing method. Figure 6 shows the overall correlation. For AL, AC, PA, PT and DC, the values of are 0.9557, 0.9747, 0.8311, 0.4461 and 0.9828, respectively, and the values of RMSE are 9.5055, 20.5260, 0.1964, 0.0674 and 0.0758, respectively. Figure 7 shows the between-patient difference for the five measurements. The standard deviation of with respect to the five measurements is smaller than 0.15. Table 2 shows the corresponding numerical results. In the Bland-Altman analysis, four indices, , , and CI, are used to investigate the agreement between our method and the manual drawing method. Let be the difference between same two measurements computed separately by our method and by manual, and let be the mean value of the two measurements. CI represents the 95% confidence interval of , formulated as , where and are the mean value and standard deviation of , respectively. represents the ratio of d to the mean value of . represents the ratio of the maximum absolute value of to the mean value of . represents the frequency of the points without CI in the Bland-Altman plot. Figure 8 shows the overall agreement between our method and the manual drawing method. With respect to the five measurements, 7.1% of results fall without CI in the Student t-test. Figure 7 shows the between-patient difference of our method. The standard deviation of the number of results within the 95% confidence interval with respective to the five measurements is smaller than 2.61%. Table 3 presents the corresponding numerical results. Additionally, the overlap between two calcified plaques detected by our method and drawn by manual separately can also represent the location accuracy of the calcified plaque, which is measured by the sensitivity and specificity. The sensitivity and specificity can also be calculated by Equation (6), where TP is the number of pixels inside the calcified plaque correctly identified, FP is the number of pixels inside the calcified plaque incorrectly identified, FN is the number of pixels outside the calcified plaque incorrectly identified, and TP is the number of pixels outside the calcified plaque correctly identified. Table 4 shows that the mean values of the sensitivity and specificity in the measurement of the overlap are 83.79% and 99.74%, respectively.


Automated detection framework of the calcified plaque with acoustic shadowing in IVUS images.

Gao Z, Guo W, Liu X, Huang W, Zhang H, Tan N, Hau WK, Zhang YT, Liu H - PLoS ONE (2014)

The results of the linear regression with respect to AL, AC, PA, PT and DC for all patients.The horizontal axis and vertical axis represent results acquired from our method and the manual drawing, respectively.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4220935&req=5

pone-0109997-g006: The results of the linear regression with respect to AL, AC, PA, PT and DC for all patients.The horizontal axis and vertical axis represent results acquired from our method and the manual drawing, respectively.
Mentions: PT and DC can measure the radial location (the location along the radial direction) of the calcified plaque, where DC equals the distance between the leading edge of the calcified plaque and the catheter's center. These measurements calculated from our method and the manual drawing method were compared separately by two different evaluation methods: linear regression and Bland-Altman analysis [39], in order to analyze their correlation and agreement, respectively. In every evaluation method, we firstly analyze the comparative results on all IVUS images in order to evaluate the overall performance of our method, and then on IVUS images from every patient in order to evaluate the between-patient difference of our method. In linear regression, the correlation coefficient and the root-mean-square error (RMSE) are used to investigate the overall correlation between our method and the manual drawing method. Figure 6 shows the overall correlation. For AL, AC, PA, PT and DC, the values of are 0.9557, 0.9747, 0.8311, 0.4461 and 0.9828, respectively, and the values of RMSE are 9.5055, 20.5260, 0.1964, 0.0674 and 0.0758, respectively. Figure 7 shows the between-patient difference for the five measurements. The standard deviation of with respect to the five measurements is smaller than 0.15. Table 2 shows the corresponding numerical results. In the Bland-Altman analysis, four indices, , , and CI, are used to investigate the agreement between our method and the manual drawing method. Let be the difference between same two measurements computed separately by our method and by manual, and let be the mean value of the two measurements. CI represents the 95% confidence interval of , formulated as , where and are the mean value and standard deviation of , respectively. represents the ratio of d to the mean value of . represents the ratio of the maximum absolute value of to the mean value of . represents the frequency of the points without CI in the Bland-Altman plot. Figure 8 shows the overall agreement between our method and the manual drawing method. With respect to the five measurements, 7.1% of results fall without CI in the Student t-test. Figure 7 shows the between-patient difference of our method. The standard deviation of the number of results within the 95% confidence interval with respective to the five measurements is smaller than 2.61%. Table 3 presents the corresponding numerical results. Additionally, the overlap between two calcified plaques detected by our method and drawn by manual separately can also represent the location accuracy of the calcified plaque, which is measured by the sensitivity and specificity. The sensitivity and specificity can also be calculated by Equation (6), where TP is the number of pixels inside the calcified plaque correctly identified, FP is the number of pixels inside the calcified plaque incorrectly identified, FN is the number of pixels outside the calcified plaque incorrectly identified, and TP is the number of pixels outside the calcified plaque correctly identified. Table 4 shows that the mean values of the sensitivity and specificity in the measurement of the overlap are 83.79% and 99.74%, respectively.

Bottom Line: However, it is very time-consuming, and with high inter-observer and intra-observer variability between different doctors.The experimental results are quantitatively analyzed separately by three evaluation methods, the test of the sensitivity and specificity, the linear regression and the Bland-Altman analysis.High sensitivity (94.68%) and specificity (95.82%), good correlation and agreement (>96.82% results fall within the 95% confidence interval in the Student t-test) demonstrate the effectiveness of the proposed method in the detection of the calcified plaque with acoustic shadowing in IVUS images.

View Article: PubMed Central - PubMed

Affiliation: Shenzhen College of Advanced Technology, University of Chinese Academy of Sciences, Shenzhen, China; Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China; Key Laboratory of Biomedical information and Health Engineering, Chinese Academy of Sciences, Shenzhen, China.

ABSTRACT
Intravascular Ultrasound (IVUS) is one ultrasonic imaging technology to acquire vascular cross-sectional images for the visualization of the inner vessel structure. This technique has been widely used for the diagnosis and treatment of coronary artery diseases. The detection of the calcified plaque with acoustic shadowing in IVUS images plays a vital role in the quantitative analysis of atheromatous plaques. The conventional method of the calcium detection is manual drawing by the doctors. However, it is very time-consuming, and with high inter-observer and intra-observer variability between different doctors. Therefore, the computer-aided detection of the calcified plaque is highly desired. In this paper, an automated method is proposed to detect the calcified plaque with acoustic shadowing in IVUS images by the Rayleigh mixture model, the Markov random field, the graph searching method and the prior knowledge about the calcified plaque. The performance of our method was evaluated over 996 in-vivo IVUS images acquired from eight patients, and the detected calcified plaques are compared with manually detected calcified plaques by one cardiology doctor. The experimental results are quantitatively analyzed separately by three evaluation methods, the test of the sensitivity and specificity, the linear regression and the Bland-Altman analysis. The first method is used to evaluate the ability to distinguish between IVUS images with and without the calcified plaque, and the latter two methods can respectively measure the correlation and the agreement between our results and manual drawing results for locating the calcified plaque in the IVUS image. High sensitivity (94.68%) and specificity (95.82%), good correlation and agreement (>96.82% results fall within the 95% confidence interval in the Student t-test) demonstrate the effectiveness of the proposed method in the detection of the calcified plaque with acoustic shadowing in IVUS images.

Show MeSH
Related in: MedlinePlus