2-step deep learning model for landmarks localization in spine radiographs

Abstract In this work we propose to use Deep Learning to automatically calculate the coordinates of the vertebral corners in sagittal x-rays images of the thoracolumbar spine and, from those landmarks, to calculate relevant radiological parameters such as L1–L5 and L1–S1 lordosis and sacral slope. F...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Andrea Cina, Tito Bassani, Matteo Panico, Andrea Luca, Youssef Masharawi, Marco Brayda-Bruno, Fabio Galbusera
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/797643136c8f4695bab1a7e61c6b3066
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:797643136c8f4695bab1a7e61c6b3066
record_format	dspace
spelling	oai:doaj.org-article:797643136c8f4695bab1a7e61c6b30662021-12-02T14:41:52Z2-step deep learning model for landmarks localization in spine radiographs10.1038/s41598-021-89102-w2045-2322https://doaj.org/article/797643136c8f4695bab1a7e61c6b30662021-05-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-89102-whttps://doaj.org/toc/2045-2322Abstract In this work we propose to use Deep Learning to automatically calculate the coordinates of the vertebral corners in sagittal x-rays images of the thoracolumbar spine and, from those landmarks, to calculate relevant radiological parameters such as L1–L5 and L1–S1 lordosis and sacral slope. For this purpose, we used 10,193 images annotated with the landmarks coordinates as the ground truth. We realized a model that consists of 2 steps. In step 1, we trained 2 Convolutional Neural Networks to identify each vertebra in the image and calculate the landmarks coordinates respectively. In step 2, we refined the localization using cropped images of a single vertebra as input to another convolutional neural network and we used geometrical transformations to map the corners to the original image. For the localization tasks, we used a differentiable spatial to numerical transform (DSNT) as the top layer. We evaluated the model both qualitatively and quantitatively on a set of 195 test images. The median localization errors relative to the vertebrae dimensions were 1.98% and 1.68% for x and y coordinates respectively. All the predicted angles were highly correlated with the ground truth, despite non-negligible absolute median errors of 1.84°, 2.43° and 1.98° for L1–L5, L1–S1 and SS respectively. Our model is able to calculate with good accuracy the coordinates of the vertebral corners and has a large potential for improving the reliability and repeatability of measurements in clinical tasks.Andrea CinaTito BassaniMatteo PanicoAndrea LucaYoussef MasharawiMarco Brayda-BrunoFabio GalbuseraNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Andrea Cina Tito Bassani Matteo Panico Andrea Luca Youssef Masharawi Marco Brayda-Bruno Fabio Galbusera 2-step deep learning model for landmarks localization in spine radiographs
description	Abstract In this work we propose to use Deep Learning to automatically calculate the coordinates of the vertebral corners in sagittal x-rays images of the thoracolumbar spine and, from those landmarks, to calculate relevant radiological parameters such as L1–L5 and L1–S1 lordosis and sacral slope. For this purpose, we used 10,193 images annotated with the landmarks coordinates as the ground truth. We realized a model that consists of 2 steps. In step 1, we trained 2 Convolutional Neural Networks to identify each vertebra in the image and calculate the landmarks coordinates respectively. In step 2, we refined the localization using cropped images of a single vertebra as input to another convolutional neural network and we used geometrical transformations to map the corners to the original image. For the localization tasks, we used a differentiable spatial to numerical transform (DSNT) as the top layer. We evaluated the model both qualitatively and quantitatively on a set of 195 test images. The median localization errors relative to the vertebrae dimensions were 1.98% and 1.68% for x and y coordinates respectively. All the predicted angles were highly correlated with the ground truth, despite non-negligible absolute median errors of 1.84°, 2.43° and 1.98° for L1–L5, L1–S1 and SS respectively. Our model is able to calculate with good accuracy the coordinates of the vertebral corners and has a large potential for improving the reliability and repeatability of measurements in clinical tasks.
format	article
author	Andrea Cina Tito Bassani Matteo Panico Andrea Luca Youssef Masharawi Marco Brayda-Bruno Fabio Galbusera
author_facet	Andrea Cina Tito Bassani Matteo Panico Andrea Luca Youssef Masharawi Marco Brayda-Bruno Fabio Galbusera
author_sort	Andrea Cina
title	2-step deep learning model for landmarks localization in spine radiographs
title_short	2-step deep learning model for landmarks localization in spine radiographs
title_full	2-step deep learning model for landmarks localization in spine radiographs
title_fullStr	2-step deep learning model for landmarks localization in spine radiographs
title_full_unstemmed	2-step deep learning model for landmarks localization in spine radiographs
title_sort	2-step deep learning model for landmarks localization in spine radiographs
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/797643136c8f4695bab1a7e61c6b3066
work_keys_str_mv	AT andreacina 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT titobassani 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT matteopanico 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT andrealuca 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT youssefmasharawi 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT marcobraydabruno 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs AT fabiogalbusera 2stepdeeplearningmodelforlandmarkslocalizationinspineradiographs
_version_	1718389878733930496

2-step deep learning model for landmarks localization in spine radiographs

Ejemplares similares