Yuki Shinohara1, Noriyuki Takahashi2, Yongbum Lee3, Tomomi Ohmura2, Toshibumi Kinoshita2. 1. Department of Radiology and Nuclear Medicine, Research Institute for Brain and Blood Vessels-Akita, 6-10 Senshu-kubota-machi, Akita, 010-0874, Japan. shino-y@akita-noken.jp. 2. Department of Radiology and Nuclear Medicine, Research Institute for Brain and Blood Vessels-Akita, 6-10 Senshu-kubota-machi, Akita, 010-0874, Japan. 3. Graduate School of Health Science, Niigata University, 2-746 Asahimachi-dori, Chuou-ku, Niigata, 951-8518, Japan.
Abstract
PURPOSE: The aim of this study was to develop an interactive deep learning-assisted identification of the hyperdense middle cerebral artery (MCA) sign (HMCAS) on non-contrast computed tomography (CT) among patients with acute ischemic stroke. MATERIALS AND METHODS: 35 HMCAS-positive and 39 HMCAS-negative samples extracted by 50-pixel-diameter circular regions of interest were obtained as training and validation datasets according to the consensus decisions of two experienced neuroradiologists. Data augmentation was performed to increase the number of training samples. A deep convolutional neural network (DCNN) (Xception) was used to classify input images as HMCAS-positive or -negative. Leave-one-case-out cross-validation was achieved to estimate sensitivity, specificity, and accuracy of the deep learning-based training model for identifying HMCAS. RESULTS: In terms of diagnostic performance, DCNN for HMCAS offered 82.9% sensitivity, 89.7% specificity, and 86.5% accuracy in leave-one-case-out cross-validation. Area under the receiver operating characteristic curve for HMCAS was 0.947 (95% confidence interval 0.895-0.998; P < 0.05). CONCLUSION: The deep learning method appears potentially beneficial for identifying HMCAS on non-contrast CT in patients with acute ischemic stroke.
PURPOSE: The aim of this study was to develop an interactive deep learning-assisted identification of the hyperdense middle cerebral artery (MCA) sign (HMCAS) on non-contrast computed tomography (CT) among patients with acute ischemic stroke. MATERIALS AND METHODS: 35 HMCAS-positive and 39 HMCAS-negative samples extracted by 50-pixel-diameter circular regions of interest were obtained as training and validation datasets according to the consensus decisions of two experienced neuroradiologists. Data augmentation was performed to increase the number of training samples. A deep convolutional neural network (DCNN) (Xception) was used to classify input images as HMCAS-positive or -negative. Leave-one-case-out cross-validation was achieved to estimate sensitivity, specificity, and accuracy of the deep learning-based training model for identifying HMCAS. RESULTS: In terms of diagnostic performance, DCNN for HMCAS offered 82.9% sensitivity, 89.7% specificity, and 86.5% accuracy in leave-one-case-out cross-validation. Area under the receiver operating characteristic curve for HMCAS was 0.947 (95% confidence interval 0.895-0.998; P < 0.05). CONCLUSION: The deep learning method appears potentially beneficial for identifying HMCAS on non-contrast CT in patients with acute ischemic stroke.
Entities:
Keywords:
Acute ischemic stroke; CT; Deep learning; Hyperdense MCA sign
Authors: Gregory W Albers; Michael P Marks; Stephanie Kemp; Soren Christensen; Jenny P Tsai; Santiago Ortega-Gutierrez; Ryan A McTaggart; Michel T Torbey; May Kim-Tenser; Thabele Leslie-Mazwi; Amrou Sarraj; Scott E Kasner; Sameer A Ansari; Sharon D Yeatts; Scott Hamilton; Michael Mlynash; Jeremy J Heit; Greg Zaharchuk; Sun Kim; Janice Carrozzella; Yuko Y Palesch; Andrew M Demchuk; Roland Bammer; Philip W Lavori; Joseph P Broderick; Maarten G Lansberg Journal: N Engl J Med Date: 2018-01-24 Impact factor: 91.245
Authors: Jeffrey L Saver; Mayank Goyal; Alain Bonafe; Hans-Christoph Diener; Elad I Levy; Vitor M Pereira; Gregory W Albers; Christophe Cognard; David J Cohen; Werner Hacke; Olav Jansen; Tudor G Jovin; Heinrich P Mattle; Raul G Nogueira; Adnan H Siddiqui; Dileep R Yavagal; Blaise W Baxter; Thomas G Devlin; Demetrius K Lopes; Vivek K Reddy; Richard du Mesnil de Rochemont; Oliver C Singer; Reza Jahan Journal: N Engl J Med Date: 2015-04-17 Impact factor: 91.245
Authors: Olvert A Berkhemer; Puck S S Fransen; Debbie Beumer; Lucie A van den Berg; Hester F Lingsma; Albert J Yoo; Wouter J Schonewille; Jan Albert Vos; Paul J Nederkoorn; Marieke J H Wermer; Marianne A A van Walderveen; Julie Staals; Jeannette Hofmeijer; Jacques A van Oostayen; Geert J Lycklama à Nijeholt; Jelis Boiten; Patrick A Brouwer; Bart J Emmer; Sebastiaan F de Bruijn; Lukas C van Dijk; L Jaap Kappelle; Rob H Lo; Ewoud J van Dijk; Joost de Vries; Paul L M de Kort; Willem Jan J van Rooij; Jan S P van den Berg; Boudewijn A A M van Hasselt; Leo A M Aerden; René J Dallinga; Marieke C Visser; Joseph C J Bot; Patrick C Vroomen; Omid Eshghi; Tobien H C M L Schreuder; Roel J J Heijboer; Koos Keizer; Alexander V Tielbeek; Heleen M den Hertog; Dick G Gerrits; Renske M van den Berg-Vos; Giorgos B Karas; Ewout W Steyerberg; H Zwenneke Flach; Henk A Marquering; Marieke E S Sprengers; Sjoerd F M Jenniskens; Ludo F M Beenen; René van den Berg; Peter J Koudstaal; Wim H van Zwam; Yvo B W E M Roos; Aad van der Lugt; Robert J van Oostenbrugge; Charles B L M Majoie; Diederik W J Dippel Journal: N Engl J Med Date: 2014-12-17 Impact factor: 91.245
Authors: P D Chang; E Kuoy; J Grinband; B D Weinberg; M Thompson; R Homo; J Chen; H Abcede; M Shafie; L Sugrue; C G Filippi; M-Y Su; W Yu; C Hess; D Chow Journal: AJNR Am J Neuroradiol Date: 2018-07-26 Impact factor: 3.825