Abstract
Purpose
Geometric features of the aortic valve play an important role in many applications, such as the clinical diagnostics, shape modeling and image-guided cardiac interventions, especially for the transcatheter aortic valve implantation (TAVI) procedure. However, few works have been reported on the topic of aortic valve segmentation from transesophageal echocardiography (TEE) sequences. To obtain accurate segmentation results and further provide valid support for TAVI, this paper presents a real-time method for segmenting the aortic valve from intraoperative, short-axis view TEE sequences, using an improved probability estimation and continuous max-flow (CMF) approach.
Methods
The proposed segmentation method includes two key stages: (1) In the probability estimation stage, five different prior frames spanning a cardiac circle are firstly selected with the aortic valve manually segmented by an expert. Then, the improved composite probability estimation (CPE) and single probability estimation (SPE) over the five prior frames are, respectively, constructed based on their radial average intensity and radial distance. (2) In the energy function construction stage, the similarity metric is calculated to find out the matching exponents between the current input TEE frame and the prior frames. The typical foreground and background intensities of prior images are therefore used to construct the corresponding energy function. Finally, the CMF approach, accelerated with a graphic processing unit (GPU), is employed to achieve the aortic valve contours in real time.
Results
The evaluation study contained 30 sequences, with each containing 62–146 short-axis TEE frames. The results were compared with the manual segmentation (ground truth). The average symmetric contour distance (ASCD), dice metric (DM) and the reliability of the algorithm reached 0.85 \(\pm \) 0.21 mm, 0.96 \(\pm \) 0.01 and 0.90 (\(d=0.95\)), respectively, and the computation time was 57.04 \(\pm \) 8.98 ms per frame.
Conclusion
The experiment results reveal that the proposed method can achieve accurate and real-time segmentation of aortic valve from TEE sequence of short-axis view.
Similar content being viewed by others
References
Nkomo VT, Gardin JM, Skelton TN, Gottdiener JS, Scott CG, Enriquez-Sarano M (2006) Burden of valvular heart diseases: a population-based study. Lancet 368(9540):1005–1011
Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR (2010) Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med 363(17):1597–1607
Luo Z, Cai J, Gu L (2013) A pilot study on magnetic navigation for transcatheter aortic valve implantation using dynamic aortic model and US image guidance. Int J Comput Assist Radiol Surg 8(4):677–690
Masuda K, Kimura E, Tateishi N, Ishihara K (2001) Three dimensional motion mechanism of ultrasound probe and its application for tele-echography system. In: Proceedings. 2001 IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 1112–1116
Huber CH, Nasratulla M, Augstburger M, Von Segesser LK (2004) Ultrasound navigation through the heart for off-pump aortic valved stent implantation: new tools for new goals. J Endovasc Ther 11(4):503–510
Schneider RJ, Perrin DP, Vasilyev NV, Marx GR, del Nido PJ, Howe RD (2010) Mitral annulus segmentation from 3D ultrasound using graph cuts. IEEE Trans Med Imaging 29(9):1676–1687
Wolf I, Hastenteufel M, De Simone R, Vetter M, Glombitza G, Mottl-Link S, Vahl C-F, Meinzer H-P (2002) ROPES: A semiautomated segmentation method for accelerated analysis of three-dimensional echocardiographic data. IEEE Trans Med Imaging 21(9):1091–1104
Luo Z, Cai J, Peters TM, Gu L (2013) Intra-operative 2-D ultrasound and dynamic 3-D aortic model registration for magnetic navigation of transcatheter aortic valve implantation. IEEE Trans Med Imaging 32(11):2152–2165
Lang P, Rajchl M, McLeod AJ, Chu MW, Peters TM (2012) Feature identification for image-guided transcatheter aortic valve implantation. In: SPIE Medical Imaging. International Society for Optics and Photonics, pp 83162X–83162X-83114
Dias JM, Leitao JM (1996) Wall position and thickness estimation from sequences of echocardiographic images. IEEE Trans Med Imaging 15(1):25–38
Chen Y, Huang F, Tagare HD, Rao M (2007) A coupled minimization problem for medical image segmentation with priors. Int J Comput Vis 71(3):259–272
Carneiro G, Nascimento J, Freitas A (2010) Robust left ventricle segmentation from ultrasound data using deep neural networks and efficient search methods. In: 2010 IEEE international symposium on biomedical imaging: from nano to macro. IEEE, pp 1085–1088
Boykov Y, Kolmogorov V (2004) An experimental comparison of min-cut/max-flow algorithms for energy minimization in vision. IEEE Trans Pattern Anal Mach Intell 26(9):1124–1137
Rajchl M, Yuan J, Ukwatta E, Peters P (2012) Fast interactive multi-region cardiac segmentation with linearly ordered labels. In: 2012 9th IEEE international symposium on biomedical imaging (ISBI). IEEE, pp 1409–1412
Yuan J, Bae E, Tai X-C, Boykov Y (2010) A continuous max-flow approach to potts model. In: Daniilidis K, Maragos P, Paragios N (eds) Computer Vision-ECCV 2010. Springer, Berlin, pp 379–392
Yuan J, Bae E, Tai X-C (2010) A study on continuous max-flow and min-cut approaches. In: 2010 IEEE conference on computer vision and pattern recognition (CVPR). IEEE, pp 2217–2224
Boykov YY, Jolly M-P (2001) Interactive graph cuts for optimal boundary & region segmentation of objects in ND images. In: Proceedings. Eighth IEEE international conference on computer vision. ICCV 2001. IEEE, pp 105–112
Noble JA, Boukerroui D (2006) Ultrasound image segmentation: a survey. IEEE Trans Med Imaging 25(8):987–1010
Lang P, Rajchl M, Li F, Peters TM (2011) Towards model-enhanced real-time ultrasound guided cardiac interventions. In: 2011 International conference on intelligent computation and bio-medical instrumentation (ICBMI). IEEE, pp 89–92
Nie Y, Luo Z, Cai J, Gu L (2013) A novel aortic valve segmentation from ultrasound image using continuous max-flow approach. In: 2013 35th annual international conference of the IEEE engineering in medicine and biology society (EMBC). IEEE, pp 3311–3314
Hatt M, Cheze le Rest C (2009) A fuzzy locally adaptive Bayesian segmentation approach for volume determination in PET. IEEE Trans Med Imaging 28(6):881–893
Kwak N, Choi C-H (2002) Input feature selection by mutual information based on Parzen window. IEEE Trans Pattern Anal Mach Intell 24(12):1667–1671
Heimann T, Van Ginneken B, Styner MA, Arzhaeva Y, Aurich V, Bauer C, Beck A, Becker C, Beichel R, Bekes G (2009) Comparison and evaluation of methods for liver segmentation from CT datasets. IEEE Trans Med Imaging 28(8):1251–1265
Ben Ayed I, H-m Chen (2012) Max-flow segmentation of the left ventricle by recovering subject-specific distributions via a bound of the Bhattacharyya measure. Med Image Anal 16(1):87–100
Arya S, Mount DM, Netanyahu NS, Silverman R, Wu A (1994) An optimal algorithm for approximate nearest neighbor searching. In: Proceedings of the fifth annual ACM-SIAM symposium on Discrete algorithms. Society for Industrial and Applied Mathematics, Philadelphia, pp 573–582
Acknowledgments
This research is partially supported by the Chinese NSFC research fund (81301283, 61190120, 61190124, 61271318), SRF for ROCS, SEM, the Shanghai municipal health bureau research fund (2011216) and Biomedical engineering fund of Shanghai Jiao Tong University (YG2012MS21).
Conflict of interest
Lixu Gu, Junfeng Cai, Xiahai Zhuang, Yuanyuan Nie and Zhe Luo declare that they have no conflict of interest.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Cai, J., Zhuang, X., Nie, Y. et al. Real-time aortic valve segmentation from transesophageal echocardiography sequence. Int J CARS 10, 447–458 (2015). https://doi.org/10.1007/s11548-014-1104-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11548-014-1104-y