MRIKardiyak MR Görüntülerinde Sol Kulakçık Bölütlenmesi İçin Derin Öğrenme Tabanlı Hibrit Model

Kardiyak manyetik rezonans görüntülerinde (MRG) başarılı sol kulakçık (SK) bölütlemesi daha yüksek hassasiyetli atriyal fibrilasyon ablasyonu, SA fibroz ölçümü ve kardiyak elektromekanik modelleme elde etmek için önemli zorluklardan biridir. Bununla birlikte, kardiyak MR dan otomatik SK bölütlemesi kalbin dinamik hareketi ve sınır noktalarındaki yeğinlik değeri farkları, görüntü çözünürlüğü, denekler arasındaki anatomik yapılarda önemli değişkenlik nedeniyle zordur. Bu çalışmada, SK'nın kardiyak volumetrik MRG'den tam otomatik bölütlenmesi için 3-boyutlu birleşik derin öğrenme ve rastgele alanlar yaklaşımı önerilmiştir. İlk önce U-net mimarisini kullanarak kaba bir SK bölütlmesi elde edilmiştir. Daha sonra son SK bölütlemesi için kardiyak hacimlerdeki bağlamsal ve görünüm bilgilerini etkili bir şekilde entegre etmek için koşullu rastgele alanlar uygulanmıştır. Yöntem rastgele alanlar sayesinde 3-boyutlu düzenlenmiş ve pürüzsüz bir sol karıncık yüzeyi sonucu vermektedir. STACOM 2018 veritabanında 100 adet kardiyak volumetrik MRG veri kümesi üzerinde önerilen yöntemi doğrulanmıştır. Önerilen yöntem sadece U-net yöntemini uygulamaktandaha iyi performans göstermiştir.Ayrıca 2-boyutlu ve 3-boyutlu yöntem karşılaştırılması yapılarak, 3-boyutlu yaklaşımın SK yüzeyi için sağladığı iyi performans nicel ve nitel olarak gösterilmiştir. Doğrulama ölçütü olarak, ortalama Dice katsayısı, Jaccard Benzerlik Katsayısı skoru ve Hausdorff Mesafesi kullanılmıştır. Önerlien yöntemin test verileri üzerinde Dice Katsayısı 0.912 ± 0.068, Jaccard Benzerlik Katsayısı skoru 0.927±0.097ve Hausdorff Mesafesisi 18.54±4.21 olarak hesaplanmıştır.

Deep Learning Based Hybrid Model for Left Atrial Segmentation in Cardiac

ccurate left atrial (LA) segmentation from cardiac magnetic resonance imaging (MRI) is one of the key challenges for achieving higher precision atrial fibrillation ablation, quantification of fibrosis and enabling electromechanical modelling ofthe heart. However, due to motion blur caused by heart movement and inconsistent intensity patterns near the atrial border, automating left atrial segmentation is a challenging task. In this work, we propose a hybrid approach, where we combine a convolutional neural network with conditional random fields to fully automatically delineate the LA from cardiac volumetric MRI. We first generate a coarse segmentation using U-net architecture and then apply conditional random fields to effectively integrate contextual and appearance information on LA cardiac volumes. Our method is capable to generate a smooth 3D atrial geometry with the regularization of conditional random fields. We trained and tested the proposed method on 100 3D cardiac MRI datasets from the STACOM 2018 atrial segmentation challenge and showed that it outperforms single U-net and its variants both in 2D and 3D. Dice coefficient, Jaccard Similarity Index and Hausdorf metricwereused as validation metricsto evaulate the segmentation. For testingcases the Dice metric, Jaccard Similarity Index and Hausdorf metricwerecomputed as 0.912 ± 0.0680.927±0.097and 18.54±4.21 respectively.

PDF

___

Hansen, B. J., Zhao, J., & Fedorov, V. V. (2017). Fibrosis and atrial fibrillation: computerized and optical mapping: a view into the human atria at submillimeter resolution. JACC: Clinical Electrophysiology,3(6), 531-546.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770- 778).
Higuchi, K., Cates, J., Gardner, G., Morris, A., Burgon, N. S., Akoum, N., & Marrouche, N. F. (2018). The spatial distribution of late gadolinium enhancement of left atrial magnetic resonance imaging in patients with atrial fibrillation. JACC: Clinical Electrophysiology, 4(1), 49-58.
Ho, T. K. (1995, August). Random decision forests. In Proceedings of 3rd international conference on document analysis and recognition (Vol. 1, pp. 278-282). IEEE.
Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
Krähenbühl, P., & Koltun, V. (2011). Efficient inference in fully connected crfs with gaussian edge potentials. In Advances in neural information processing systems (pp. 109-117).
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444. Narayan, S. M., Rodrigo, M., Kowalewski, C. A., Shenasa, F., Meckler, G. L., Vishwanathan, M.
N., ... & Wang, P. J. (2017). Ablation of focal impulses and rotational sources: What can be learned from differing procedural outcomes?. Current Cardiovascular Risk Reports, 11(9), 27.
Narayan, S. M., & Zaman, J. A. (2016). Mechanistically based mapping of human cardiac fibrillation. The Journal of physiology, 594(9), 2399-2415.
Njoku, A., Kannabhiran, M., Arora, R., Reddy, P., Gopinathannair, R., Lakkireddy, D., & Dominic, P. (2018). Left atrial volume predicts atrial fibrillation recurrence after radiofrequency ablation: a meta-analysis. Ep Europace, 20(1), 33-42.
Noh, H., Hong, S., & Han, B. (2015). Learning deconvolution network for semantic segmentation. In Proceedings of the IEEE international conference on computer vision (pp. 1520-1528).
Papik, K., Molnar, B., Schaefer, R., Dombovari, Z., Tulassay, Z., & Feher, J. (1998). Application of neural networks in medicine-a review. Medical Science Monitor, 4(3), MT538-MT546.
Park, C., Took, C. C., & Seong, J. K. (2018). Machine learning in biomedical engineering.
Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in neural information processing systems (pp. 91- 99).
Ronneberger, O., Fischer, P., & Brox, T. (2015, October). U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention (pp. 234-241). Springer, Cham.
Shahid, N., Rappon, T., & Berta, W. (2019). Applications of artificial neural networks in health care organizational decision-making: A scoping review. PloS one, 14(2), e0212356.
Suykens, J. A., & Vandewalle, J. (1999). Least squares support vector machine classifiers. Neural processing letters, 9(3), 293-300.
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).
Tao, Q., Ipek, E. G., Shahzad, R., Berendsen, F. F., Nazarian, S., & van der Geest, R. J. (2016). Fully automatic segmentation of left atrium and pulmonary veins in late gadolinium‐enhanced MRI: Towards objective atrial scar assessment. Journal of magnetic resonance imaging, 44(2), 346-354.
Tobon-Gomez, C., Geers, A. J., Peters, J., Weese, J., Pinto, K., Karim, R., ... & Stender, B. (2015). Benchmark for algorithms segmenting the left atrium from 3D CT and MRI datasets. IEEE transactions on medical imaging, 34(7), 1460-1473.
Veni, Gopalkrishna, et al. "Bayesian segmentation of atrium wall using globally-optimal graph cuts on 3D meshes." International Conference on Information Processing in Medical Imaging. Springer, Berlin, Heidelberg, 2013.
Veni, G., Elhabian, S. Y., & Whitaker, R. T. (2017). ShapeCut: Bayesian surface estimation using shape-driven graph. Medical image analysis, 40, 11-29.
Xiong, Z., Fedorov, V. V., Fu, X., Cheng, E., Macleod, R., & Zhao, J. (2018). Fully automatic left atrium segmentation from late gadolinium enhanced magnetic resonance imaging using a dual fully convolutional neural network. IEEE transactions on medical imaging, 38(2), 515- 524.
Xiong, Z., Xia, Q., Hu, Z., Huang, N., Vesal, S., Ravikumar, N., ... & Puybareau, E. (2020). A Global Benchmark of Algorithms for Segmenting Late Gadolinium-Enhanced Cardiac Magnetic Resonance Imaging. arXiv preprint arXiv:2004.12314.
Zhao, J., Butters, T. D., Zhang, H., Pullan, A. J., LeGrice, I. J., Sands, G. B., & Smaill, B. H. (2012). An image-based model of atrial muscular architecture: effects of structural anisotropy on electrical activation. Circulation: Arrhythmia and Electrophysiology, 5(2), 361-370.
Zhao, J., Hansen, B. J., Wang, Y., Csepe, T. A., Sul, L. V., Tang, A., ... & Kilic, A. (2017). Threedimensional integrated functional, structural, and computational mapping to define the structural “fingerprints” of heart‐specific atrial fibrillation drivers in human heart ex vivo. Journal of the American Heart Association, 6(8), e005922.

Turkish Studies - Information Technologies and Applied Sciences-Cover

ISSN: 2667-5633
Yayın Aralığı: Yılda 4 Sayı
Başlangıç: 2006
Yayıncı: ASOS Eğitim Bilişim Danışmanlık Otomasyon Yayıncılık Reklam Sanayi ve Ticaret LTD ŞTİ