Carlos Cano-Espinosa1, Germán González2, George R Washko3, Miguel Cazorla1, Raúl San José Estépar4. 1. University of Alicante. Department of Computer Science & Artificial Intelligence. Alicante. Spain. 2. Sierra Research SL; Avda. Costa Blanca 132. Alicante. Spain. 3. Brigham and Women's Hospital, Pulmonary and Critical Care Medicine. 75 Francis St. Boston. Massachusetts. United States. 4. Applied Chest Imaging Laboratory. Brigham and Women's Hospital. 1249 Boylston St. Boston. Massachusetts. United States.
Abstract
INTRODUCTION: The Agatston score is a well-established metric of cardiovascular disease related to clinical outcomes. It is computed from CT scans by a) measuring the volume and intensity of the atherosclerotic plaques and b) aggregating such information in an index. OBJECTIVE: To generate a convolutional neural network that inputs a non-contrast chest CT scan and outputs the Agatston score associated with it directly, without a prior segmentation of Coronary Artery Calcifications (CAC). MATERIALS AND METHODS: We use a database of 5973 non-contrast non-ECG gated chest CT scans where the Agatston score has been manually computed. The heart of each scan is cropped automatically using an object detector. The database is split in 4973 cases for training and 1000 for testing. We train a 3D deep convolutional neural network to regress the Agatston score directly from the extracted hearts. RESULTS: The proposed method yields a Pearson correlation coefficient of r = 0.93; p ≤ 0.0001 against manual reference standard in the 1000 test cases. It further stratifies correctly 72.6% of the cases with respect to standard risk groups. This compares to more complex state-of-the-art methods based on prior segmentations of the CACs, which achieve r = 0.94 in ECG-gated pulmonary CT. CONCLUSIONS: A convolutional neural network can regress the Agatston score from the image of the heart directly, without a prior segmentation of the CACs. This is a new and simpler paradigm in the Agatston score computation that yields similar results to the state-of-the-art literature.
INTRODUCTION: The Agatston score is a well-established metric of cardiovascular disease related to clinical outcomes. It is computed from CT scans by a) measuring the volume and intensity of the atherosclerotic plaques and b) aggregating such information in an index. OBJECTIVE: To generate a convolutional neural network that inputs a non-contrast chest CT scan and outputs the Agatston score associated with it directly, without a prior segmentation of Coronary Artery Calcifications (CAC). MATERIALS AND METHODS: We use a database of 5973 non-contrast non-ECG gated chest CT scans where the Agatston score has been manually computed. The heart of each scan is cropped automatically using an object detector. The database is split in 4973 cases for training and 1000 for testing. We train a 3D deep convolutional neural network to regress the Agatston score directly from the extracted hearts. RESULTS: The proposed method yields a Pearson correlation coefficient of r = 0.93; p ≤ 0.0001 against manual reference standard in the 1000 test cases. It further stratifies correctly 72.6% of the cases with respect to standard risk groups. This compares to more complex state-of-the-art methods based on prior segmentations of the CACs, which achieve r = 0.94 in ECG-gated pulmonary CT. CONCLUSIONS: A convolutional neural network can regress the Agatston score from the image of the heart directly, without a prior segmentation of the CACs. This is a new and simpler paradigm in the Agatston score computation that yields similar results to the state-of-the-art literature.
Entities:
Keywords:
Agatston score; computed aided detection; deep learning; pulmonary CT
Authors: Ivana Isgum; Marius Staring; Annemarieke Rutten; Mathias Prokop; Max A Viergever; Bram van Ginneken Journal: IEEE Trans Med Imaging Date: 2009-01-06 Impact factor: 10.048
Authors: Ruediger E Schernthaner; Alfred Stadler; Dietrich Beitzke; Peter Homolka; Michael Weber; Johannes Lammer; Martin Czerny; Christian Loewe Journal: Eur J Radiol Date: 2011-08-05 Impact factor: 3.528
Authors: Jelmer M Wolterink; Tim Leiner; Bob D de Vos; Robbert W van Hamersvelt; Max A Viergever; Ivana Išgum Journal: Med Image Anal Date: 2016-04-21 Impact factor: 8.545
Authors: Farbod N Rahaghi; Gonzalo Vegas-Sanchez-Ferrero; Jasleen K Minhas; Carolyn E Come; Isaac De La Bruere; James M Wells; Germán González; Surya P Bhatt; Brett E Fenster; Alejandro A Diaz; Puja Kohli; James C Ross; David A Lynch; Mark T Dransfield; Russel P Bowler; Maria J Ledesma-Carbayo; Raúl San José Estépar; George R Washko Journal: Acad Radiol Date: 2017-02-15 Impact factor: 3.173
Authors: Shin Matsuoka; Tsuneo Yamashiro; Alejandro Diaz; Raúl San José Estépar; James C Ross; Edwin K Silverman; Yasuyuki Kobayashi; Mark T Dransfield; Brian J Bartholmai; Hiroto Hatabu; George R Washko Journal: Acad Radiol Date: 2010-10-13 Impact factor: 3.173
Authors: Ivana Isgum; Mathias Prokop; Meindert Niemeijer; Max A Viergever; Bram van Ginneken Journal: IEEE Trans Med Imaging Date: 2012-09-03 Impact factor: 10.048
Authors: Rahil Shahzad; Theo van Walsum; Michiel Schaap; Alexia Rossi; Stefan Klein; Annick C Weustink; Pim J de Feyter; Lucas J van Vliet; Wiro J Niessen Journal: Acad Radiol Date: 2012-09-13 Impact factor: 3.173
Authors: Carlos Cano-Espinosa; Germán González; George R Washko; Miguel Cazorla; Raúl San José Estépar Journal: Image Anal Mov Organ Breast Thorac Images (2018) Date: 2018-09-12
Authors: Sanne G M van Velzen; Bob D de Vos; Julia M H Noothout; Helena M Verkooijen; Max A Viergever; Ivana Išgum Journal: J Med Imaging (Bellingham) Date: 2022-05-31
Authors: George Koulaouzidis; Tomasz Jadczyk; Dimitris K Iakovidis; Anastasios Koulaouzidis; Marc Bisnaire; Dafni Charisopoulou Journal: J Clin Med Date: 2022-07-05 Impact factor: 4.964
Authors: Carlos Cano-Espinosa; German Gonzalez; George R Washko; Miguel Cazorla; Raul San Jose Estepar Journal: IEEE Trans Med Imaging Date: 2020-01-09 Impact factor: 10.048