Danis Alukaev1, Semen Kiselev1, Tamerlan Mustafaev1,2, Ahatov Ainur3, Bulat Ibragimov4,5, Tomaž Vrtovec6. 1. AI Lab, Innopolis University, Universitetskaya St 1, 420500, Innopolis, Republic of Tatarstan, Russian Federation. 2. Kazan Public Hospital, Chekhova 1A, 42000, Kazan, Republic of Tatarstan, Russian Federation. 3. Barsmed Diagnostic Center, Daurskaya 12, 42000, Kazan, Republic of Tatarstan, Russian Federation. 4. Department of Computer Science, University of Copenhagen, Universitetsparken 1, 2100, Copenhagen, Denmark. 5. Laboratory of Imaging Technologies, Faculty of Electrical Engineering, University of Ljubljana, Tržaška cesta 25, 1000, Ljubljana, Slovenia. 6. Laboratory of Imaging Technologies, Faculty of Electrical Engineering, University of Ljubljana, Tržaška cesta 25, 1000, Ljubljana, Slovenia. tomaz.vrtovec@fe.uni-lj.si.
Abstract
PURPOSE: To propose a fully automated deep learning (DL) framework for the vertebral morphometry and Cobb angle measurement from three-dimensional (3D) computed tomography (CT) images of the spine, and validate the proposed framework on an external database. METHODS: The vertebrae were first localized and segmented in each 3D CT image using a DL architecture based on an ensemble of U-Nets, and then automated vertebral morphometry in the form of vertebral body (VB) and intervertebral disk (IVD) heights, and spinal curvature measurements in the form of coronal and sagittal Cobb angles (thoracic kyphosis and lumbar lordosis) were performed using dedicated machine learning techniques. The framework was trained on 1725 vertebrae from 160 CT images and validated on an external database of 157 vertebrae from 15 CT images. RESULTS: The resulting mean absolute errors (± standard deviation) between the obtained DL and corresponding manual measurements were 1.17 ± 0.40 mm for VB heights, 0.54 ± 0.21 mm for IVD heights, and 3.42 ± 1.36° for coronal and sagittal Cobb angles, with respective maximal absolute errors of 2.51 mm, 1.64 mm, and 5.52°. Linear regression revealed excellent agreement, with Pearson's correlation coefficient of 0.943, 0.928, and 0.996, respectively. CONCLUSION: The obtained results are within the range of values, obtained by existing DL approaches without external validation. The results therefore confirm the scalability of the proposed DL framework from the perspective of application to external data, and time and computational resource consumption required for framework training.
PURPOSE: To propose a fully automated deep learning (DL) framework for the vertebral morphometry and Cobb angle measurement from three-dimensional (3D) computed tomography (CT) images of the spine, and validate the proposed framework on an external database. METHODS: The vertebrae were first localized and segmented in each 3D CT image using a DL architecture based on an ensemble of U-Nets, and then automated vertebral morphometry in the form of vertebral body (VB) and intervertebral disk (IVD) heights, and spinal curvature measurements in the form of coronal and sagittal Cobb angles (thoracic kyphosis and lumbar lordosis) were performed using dedicated machine learning techniques. The framework was trained on 1725 vertebrae from 160 CT images and validated on an external database of 157 vertebrae from 15 CT images. RESULTS: The resulting mean absolute errors (± standard deviation) between the obtained DL and corresponding manual measurements were 1.17 ± 0.40 mm for VB heights, 0.54 ± 0.21 mm for IVD heights, and 3.42 ± 1.36° for coronal and sagittal Cobb angles, with respective maximal absolute errors of 2.51 mm, 1.64 mm, and 5.52°. Linear regression revealed excellent agreement, with Pearson's correlation coefficient of 0.943, 0.928, and 0.996, respectively. CONCLUSION: The obtained results are within the range of values, obtained by existing DL approaches without external validation. The results therefore confirm the scalability of the proposed DL framework from the perspective of application to external data, and time and computational resource consumption required for framework training.
Authors: A Neubert; J Fripp; C Engstrom; D Walker; M-A Weber; R Schwarz; S Crozier Journal: J Am Med Inform Assoc Date: 2013-06-27 Impact factor: 4.497
Authors: Bulat Ibragimov; Robert Korez; Bostjan Likar; Franjo Pernus; Lei Xing; Tomaz Vrtovec Journal: IEEE Trans Med Imaging Date: 2017-02-13 Impact factor: 10.048