Fully automatic carotid arterial stiffness assessment from ultrasound videos based on machine learning
Tóm tắt
Arterial stiffness (AS) refers to the loss of arterial compliance and alterations in vessel wall properties. The study of local carotid stiffness (CS) is particularly important since carotid artery stiffening raises the risk of stroke, cognitive impairment, and dementia. So, stiffness measurement as a screening tool approach is crucial because it can reduce mortality and facilitate therapy planning. This study aims to evaluate the stiffness of the CCA using machine learning (ML) through the features of diameter change (ΔD) and stiffness parameters. This study was conducted in seven stages: data collection, preprocessing, CCA segmentation, CCA lumen diameter (DCCA) computing during cardiac cycles, denoising signals of DCCA, computational of AS parameters, and stiffness assessment using ML. The 51 videos (with 25 s) of CCA B-mode ultrasound (US) were used and analyzed. Each US video yielded approximately 750 sequential frames spanning about 24 cardiac cycles. Firstly, US preset settings with time gain compensation with a U-pattern were employed to enhance CCA segmentations. The study showed that auto region-growing, employed three times, is appropriate for segmenting walls with a short running time (4.56 s/frame). The diameter computed for frames constructs a signal (diameter signal) with noisy parts in the shape of peak variance and an un-smooth side. Among the 12 employed smoothing methods, spline fitting with a mean peak difference per cycle (MPDCY) of 0.58 pixels was the most effective for the diameter signal. The authors propose the MPDCY as a new selection criterion for smoothing methods with highly preserved peaks. The ΔD (Dsys–Ddia) determined in this study was validated by statistical analysis as a viable replacement for manual ΔD measurement. Statistical analysis was carried out by Mann–Whitney t-test with a p-value of 0.81, regression line R2 = 0.907, and there was no difference in means between the two groups for box plots. The stiffness parameters of the carotid arteries were calculated based on auto-ΔD and pulse pressure. Five ML models, including K-nearest neighbor (KNN), support vector machine (SVM), decision tree (DT), logistic regression (LR), and random forest (RF), fed by distension (ΔD) and five stiffness parameters, were used to distinguish between the stiffened and un-stiffened CCA. Except for SVM, all models performed excellently in terms of specificity, sensitivity, precision, and area under the curve (AUC). In addition, the scatterplot and statistical analysis of the fed features confirm these remarkable outcomes. The scatter plot demonstrates that a linear hyperline can easily distinguish between the two classes. The statistical analysis shows that the stiffness parameters computed from the database of this work were statistically (p < 0.05) distributed into the non-stiffness and stiffness groups. The presented models are validated by applying them to additional datasets. Applying models to other datasets reveals a model performance of 100%. The proposed ML models could be applied in clinical practice to detect CS early, which is essential for preventing stroke.
Tài liệu tham khảo
Laurent S, Boutouyrie P (2020) Arterial stiffness and hypertension in the elderly. Front Cardiovasc Med 7:544302
Uejima T et al (2020) Age-specific reference values for carotid arterial stiffness estimated by ultrasonic wall tracking. J Hum Hypertens 34(3):214–222
Mackenzie I, Wilkinson I, Cockcroft J (2002) Assessment of arterial stiffness in clinical practice. QJM 95(2):67–74
Hall JL et al (2012) A review of genetics, arterial stiffness, and blood pressure in African Americans. J Cardiovasc Transl Res 5(3):302–308
Ramirez JL et al (2019) Radial artery tonometry is associated with major adverse cardiac events in patients with peripheral artery disease. J Surg Res 235:250–257
Zhang J et al (2020) Comparison of the characteristics and risk factors of carotid atherosclerosis in high stroke risk populations between urban and rural areas in North China. Front Neurol 11:554778
Zhang Y et al (2017) Features and risk factors of carotid atherosclerosis in a population with high stroke incidence in China. Oncotarget 8(34):57477
Fukuda T et al (2014) Association between serum γ-glutamyltranspeptidase and atherosclerosis: a population-based cross-sectional study. BMJ Open 4(10):e005413
Helleputte S et al (2022) Arterial stiffness in patients with type 1 diabetes and its comparison to cardiovascular risk evaluation tools. Cardiovasc Diabetol 21(1):97
Kim Y, Kim G-T, Kang J (2021) Carotid arterial stiffness and cardiometabolic profiles in women with fibromyalgia. Biomedicines 9(12):1786
Vriz O et al (2019) Local carotid arterial stiffness is an independent determinant of left ventricular remodeling in never-treated hypertensive patients. Blood Press 28(1):23–33
Gao Z et al (2017) Motion tracking of the carotid artery wall from ultrasound image sequences: a nonlinear state-space approach. IEEE Trans Med Imaging 37(1):273–283
Lloyd KD et al (2012) Common carotid artery diameter and cardiovascular risk factors in overweight or obese postmenopausal women. Int J Vasc Med 2012:169323
Krejza J et al (2006) Carotid artery diameter in men and women and the relation to body and neck size. Stroke 37(4):1103–1105
Krishna Kumar P et al (2017) Accurate lumen diameter measurement in curved vessels in carotid ultrasound: an iterative scale-space and spatial transformation approach. Med Biol Eng Compu 55(8):1415–1434
Mao F et al (2000) Segmentation of carotid artery in ultrasound images: method development and evaluation technique. Med Phys 27(8):1961–1970
Hamou AK, El-Sakka MR (2004) A novel segmentation technique for carotid ultrasound images. 2004 IEEE international conference on acoustics, speech, and signal processing. IEEE
Abdel-Dayem AR, El-Sakka MR, Fenster A (2005) Watershed segmentation for carotid artery ultrasound images. The 3rd ACS/IEEE international conference on computer systems and applications, 2005. IEEE
Golemati S et al (2007) Using the Hough transform to segment ultrasound images of longitudinal and transverse sections of the carotid artery. Ultrasound Med Biol 33(12):1918–1932
Santos AMF et al (2013) A novel automatic algorithm for the segmentation of the lumen of the carotid artery in ultrasound B-mode images. Exp Syst Appl 40(16):6570–6579
Loizou CP et al (2014) Manual and automated intima-media thickness and diameter measurements of the common carotid artery in patients with renal failure disease. Comput Biol Med 53:220–229
Loizou CP et al (2015) A comparison of ultrasound intima-media thickness measurements of the left and right common carotid artery. IEEE J Transl Eng Health Med 3:1–10
Matsakou AI et al (2011) Automated detection of the carotid artery wall in longitudinal B-mode images using active contours initialized by the Hough transform. 2011 annual international conference of the IEEE engineering in medicine and biology society. IEEE
Kumar JH et al (2019) Automatic segmentation of common carotid artery in longitudinal mode ultrasound images using active oblongs. ICASSP 2019–2019 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE
Gutierrez MA et al (2002) Automatic measurement of carotid diameter and wall thickness in ultrasound images. Computers in Cardiology. IEEE
Selzer RH et al (2001) Improved common carotid elasticity and intima-media thickness measurements from computer analysis of sequential ultrasound frames. Atherosclerosis 154(1):185–193
Mokhtari-Dizaji M, Montazeri M, Saberi H (2006) Differentiation of mild and severe stenosis with motion estimation in ultrasound images. Ultrasound Med Biol 32(10):1493–1498
Patel AK, Jain SK (2019) Arterial parameters and elasticity estimation in common carotid artery using deep learning approach. Int J Image Gr Signal Process (IJIGSP) 11(11):18–28
Menchón-Lara RM, Bueno-Crespo A, Sancho-Gómez JL (2015) Estimation of the arterial diameter in ultrasound images of the common carotid artery. International work-conference on the interplay between natural and artificial computation. Springer
Li Q et al (2021) Development and validation of a prediction model for elevated arterial stiffness in Chinese patients with diabetes using machine learning. Front Physiol 12:714195
Elumalai A et al (2021) Optimal prediction of attacks and arterial stiffness effects on heart disease by hybrid machine learning algorithm. J Ambient Intell Humanized Comput 2021:1–11
Miao F et al (2018) A wearable sensor for arterial stiffness monitoring based on machine learning algorithms. IEEE Sens J 19(4):1426–1434
Beaussier H et al (2011) Mechanical and structural characteristics of carotid plaques by combined analysis with echotracking system and MR imaging. JACC Cardiovasc Imaging 4(5):468–477
Yule CE et al (2016) Does short-term whole-body vibration training affect arterial stiffness in chronic stroke? A preliminary study. J Phys Ther Sci 28(3):996–1002
Liao D et al (1999) Arterial stiffness and the development of hypertension: the ARIC study. Hypertension 34(2):201–206
Fernberg U, Fernström M, Hurtig-Wennlöf A (2019) Body composition is a strong predictor of local carotid stiffness in Swedish, young adults–the cross sectional Lifestyle, biomarkers, and atherosclerosis study. BMC Cardiovasc Disord 19(1):1–10
Wang M et al (2014) A new image denoising method based on Gaussian filter. 2014 international conference on information science, electronics and electrical engineering. IEEE
Ding F et al (2020) Real-time estimation for the parameters of Gaussian filtering via deep learning. J Real-Time Image Proc 17(1):17–27
Qiao Y et al (2007) Thresholding based on variance and intensity contrast. Pattern Recogn 40(2):596–608
Odat A, Otair M, Shehadeh F (2015) Image denoising by comprehensive median filter. Int J Appl Eng Res 10(15):36016–36022
Pramitha V, Kumar KA (2017) Image denoising techniques: a review. Int J Adv Res Methodol Eng Technol 2017:66–72
Murillo-Bracamontes EA et al (2012) Implementation of Hough transform for fruit image segmentation. Procedia Engineering 35:230–239
Nahar M, Ali MS, Rahman MM (2018) Improvement of single seeded region growing algorithm on image segmentation. Global J Comput Sci Technol 2018:1–10
Merzougui M, El Allaoui A (2019) Region growing segmentation optimized by evolutionary approach and maximum entropy. Proced Comput Sci 151:1046–1051
Chaturvedi A, Khanna R, Kumar V (2016) An analysis of region growing image segmentation schemes. Int J Comput Trends Technol (IJCTT) 34(1):1–6
Shrivastava N, Bharti J (2020) Automatic seeded region growing image segmentation for medical image segmentation: a brief review. Int J Image Gr 20(03):2050018
Chakraborty M, Das S (2012) Determination of signal to noise ratio of electrocardiograms filtered by band pass and Savitzky-Golay filters. Proced Technol 4:830–833
J Conder (2022) Gaussfilt. https://www.mathworks.com/matlabcentral/fileexchange/43182-gaussfilt-t-z-sigma. Accessed 10 May 2022
Hodrick RJ, Prescott EC (1997) Postwar US business cycles: an empirical investigation. J Money Credit Bank 1997:1–16
T O'Haver (2022) Fast smoothing function. https://www.mathworks.com/matlabcentral/fileexchange/19998-fast-smoothing-function. Accessed 10 May 2022
P Seibold (2022) Sine fitting. https://www.mathworks.com/matlabcentral/fileexchange/66793-sine-fitting. Accessed 10 May 2022
V Hampiholi (2022) Moving average filter. https://www.mathworks.com/matlabcentral/fileexchange/35925-moving-average-filter?s_tid=srchtitle. 10 May 2022
Łoboz-Rudnicka M et al (2018) Gender-related differences in the progression of carotid stiffness with age and in the influence of risk factors on carotid stiffness. Clin Interv Aging 13:1183
Núñez F et al (2010) Carotid artery stiffness as an early marker of vascular lesions in children and adolescents with cardiovascular risk factors. Rev Esp Cardiol (Engl Ed) 63(11):1253–1260
Boesen ME et al (2015) A systematic literature review of the effect of carotid atherosclerosis on local vessel stiffness and elasticity. Atherosclerosis 243(1):211–222
Gaye B, Zhang D, Wulamu A (2021) Improvement of support vector machine algorithm in big data background. Math Probl Eng. https://doi.org/10.1155/2021/5594899
Latha S, Samiappan D, Kumar R (2020) Carotid artery ultrasound image analysis: a review of the literature. Proc Inst Mech Eng [H] 234(5):417–443
Archana K, Vanithamani R (2021) Classification of plaque in carotid artery using intravascular ultrasound images (IVUS) by machine learning techniques. Ann Roman Soc Cell Biol 2021:4751–4760
Christodoulou C et al (2010) Image retrieval and classification of carotid plaque ultrasound images. Open Cardiovasc Imagin J 2(1):1–6
Seera M, Lim CP (2013) Online motor fault detection and diagnosis using a hybrid FMM-CART model. IEEE Trans Neural Netw Learn Syst 25(4):806–812
Qawqzeh YK et al (2019) A proposed decision tree classifier for atherosclerosis prediction and classification. IJCSNS 19(12):197
Lai KW, Khalil A, Samiappan D (2022) Performance analysis of machine learning and deep learning architectures on early stroke detection using carotid artery ultrasound images. Front Aging Neurosci 2022:1013
Lin SY et al (2009) Applying machine learning to carotid sonographic features for recurrent stroke in patients with acute stroke. Front Cardiovasc Med 9:804410
Ergün U et al (2004) Classification of carotid artery stenosis of patients with diabetes by neural network and logistic regression. Comput Biol Med 34(5):389–405
Göksan B et al (2001) Diabetes as a determinant of high-grade carotid artery stenosis: evaluation of 1,058 cases by Doppler sonography. J Stroke Cerebrovasc Dis 10(6):252–256
Cheng S et al (1999) Screening for asymptomatic carotid stenosis in patients with peripheral vascular disease: a prospective study and risk factor analysis. Cardiovasc Surg 7(3):303–309
Jun TJ et al (2019) Automated detection of vulnerable plaque in intravascular ultrasound images. Med Biol Eng Compu 57(4):863–876
Tanaka H et al (2009) Comparison between carotid-femoral and brachial-ankle pulse wave velocity as measures of arterial stiffness. J Hypertens 27(10):2022–2027
Yu W et al (2008) Brachial-ankle vs carotid-femoral pulse wave velocity as a determinant of cardiovascular structure and function. J Hum Hypertens 22(1):24–31
Engelen L et al (2015) Reference values for local arterial stiffness. Part A: carotid artery. J Hypertens 33(10):1981–1996
Spronck B et al (2021) Heart rate and blood pressure dependence of aortic distensibility in rats: comparison of measured and calculated pulse wave velocity. J Hypertens 39(1):117
Comelli A et al (2021) Deep learning-based methods for prostate segmentation in magnetic resonance imaging. Appl Sci 11(2):782
Meiburger KM et al (2021) Carotid ultrasound boundary study (CUBS): an open multicenter analysis of computerized intima-media thickness measurement systems and their clinical impact. Ultrasound Med Biol 47(8):2442–2455
Murray CS et al (2018) Ultrasound assessment of carotid arteries: current concepts, methodologies, diagnostic criteria, and technological advancements. Echocardiography 35(12):2079–2091
Menchón-Lara RM, Sancho-Gómez JL (2015) Fully automatic segmentation of ultrasound common carotid artery images based on machine learning. Neurocomputing 151:161–167
Zhou Y et al (2019) Contrast enhancement of medical images using a new version of the world cup optimization algorithm. Quant Imagin Med Surg 9(9):1528–1547
Loizou CP et al (2014) Despeckle filtering software toolbox for ultrasound imaging of the common carotid artery. Comput Methods Programs Biomed 114(1):109–124
Rafati M et al (2016) Assessment of noise reduction in ultrasound images of common carotid and brachial arteries. IET Comput Vision 10(1):1–8
Stanziola A et al (2018) Motion artifacts and correction in multipulse high-frame rate contrast-enhanced ultrasound. IEEE Trans Ultrason Ferroelectr Freq Control 66(2):417–420
Nock L, Trahey G (1990) Synthetic aperture imaging in medical ultrasound with correction for motion artifacts. IEEE symposium on ultrasonics. IEEE
Boehler T, Peitgen HO (2008) Reducing motion artifacts in 3-D breast ultrasound using non-linear registration. International conference on medical image computing and computer-assisted intervention. Springer
Gao Z et al (2017) Robust estimation of carotid artery wall motion using the elasticity-based state-space approach. Med Image Anal 37:1–21
Wang J et al (2014) Fractional zero-phase filtering based on the Riemann-Liouville integral. Signal Process 98:150–157
Tang G, Yan X, Wang X (2020) Chaotic signal denoising based on adaptive smoothing multiscale morphological filtering. Complexity 2020:1–14
Lv Z et al (2022) A multi-peak detection algorithm for FBG based on WPD-HT. Opt Fiber Technol 68:102805
Tulyakova N, Trofymchuk O (2022) Real-time filtering adaptive algorithms for non-stationary noise in electrocardiograms. Biomed Signal Process Control 72:103308
Charoenpong J et al (2019) A comparison of machine learning algorithms and their applications. Int J Simul Syst Sci Technol 20(4):1787