H Kuang1, M Najm1, D Chakraborty1, N Maraj1, S I Sohn2, M Goyal1,3,4, M D Hill1,3,5,4, A M Demchuk1,3,4, B K Menon1,3,5,4, W Qiu6. 1. From the Calgary Stroke Program (H.K., W.Q., M.N., D.C., N.M., M.G., M.D.H., A.M.D., B.K.M.). 2. Department of Neurology (S.I.S.), Keimyung University, Daegu, South Korea. 3. Department of Clinical Neurosciences, Department of Radiology (M.D.H., A.M.D., M.G., B.K.M.). 4. Hotchkiss Brain Institute, Calgary, Alberta, Canada (M.D.H., A.M.D., M.G., B.K.M.). 5. Department of Community Health Sciences (M.D.H., B.K.M.), University of Calgary, Calgary, Alberta, Canada. 6. From the Calgary Stroke Program (H.K., W.Q., M.N., D.C., N.M., M.G., M.D.H., A.M.D., B.K.M.) qiu.wu.ch@gmail.com.
Abstract
BACKGROUND AND PURPOSE: Alberta Stroke Program Early CT Score (ASPECTS) was devised as a systematic method to assess the extent of early ischemic change on noncontrast CT (NCCT) in patients with acute ischemic stroke (AIS). Our aim was to automate ASPECTS to objectively score NCCT of AIS patients. MATERIALS AND METHODS: We collected NCCT images with a 5-mm thickness of 257 patients with acute ischemic stroke (<8 hours from onset to scans) followed by a diffusion-weighted imaging acquisition within 1 hour. Expert ASPECTS readings on DWI were used as ground truth. Texture features were extracted from each ASPECTS region of the 157 training patient images to train a random forest classifier. The unseen 100 testing patient images were used to evaluate the performance of the trained classifier. Statistical analyses on the total ASPECTS and region-level ASPECTS were conducted. RESULTS: For the total ASPECTS of the unseen 100 patients, the intraclass correlation coefficient between the automated ASPECTS method and DWI ASPECTS scores of expert readings was 0.76 (95% confidence interval, 0.67-0.83) and the mean ASPECTS difference in the Bland-Altman plot was 0.3 (limits of agreement, -3.3, 2.6). Individual ASPECTS region-level analysis showed that our method yielded κ = 0.60, sensitivity of 66.2%, specificity of 91.8%, and area under curve of 0.79 for 100 × 10 ASPECTS regions. Additionally, when ASPECTS was dichotomized (>4 and ≤4), κ = 0.78, sensitivity of 97.8%, specificity of 80%, and area under the curve of 0.89 were generated between the proposed method and expert readings on DWI. CONCLUSIONS: The proposed automated ASPECTS scoring approach shows reasonable ability to determine ASPECTS on NCCT images in patients presenting with acute ischemic stroke.
BACKGROUND AND PURPOSE: Alberta Stroke Program Early CT Score (ASPECTS) was devised as a systematic method to assess the extent of early ischemic change on noncontrast CT (NCCT) in patients with acute ischemic stroke (AIS). Our aim was to automate ASPECTS to objectively score NCCT of AISpatients. MATERIALS AND METHODS: We collected NCCT images with a 5-mm thickness of 257 patients with acute ischemic stroke (<8 hours from onset to scans) followed by a diffusion-weighted imaging acquisition within 1 hour. Expert ASPECTS readings on DWI were used as ground truth. Texture features were extracted from each ASPECTS region of the 157 training patient images to train a random forest classifier. The unseen 100 testing patient images were used to evaluate the performance of the trained classifier. Statistical analyses on the total ASPECTS and region-level ASPECTS were conducted. RESULTS: For the total ASPECTS of the unseen 100 patients, the intraclass correlation coefficient between the automated ASPECTS method and DWI ASPECTS scores of expert readings was 0.76 (95% confidence interval, 0.67-0.83) and the mean ASPECTS difference in the Bland-Altman plot was 0.3 (limits of agreement, -3.3, 2.6). Individual ASPECTS region-level analysis showed that our method yielded κ = 0.60, sensitivity of 66.2%, specificity of 91.8%, and area under curve of 0.79 for 100 × 10 ASPECTS regions. Additionally, when ASPECTS was dichotomized (>4 and ≤4), κ = 0.78, sensitivity of 97.8%, specificity of 80%, and area under the curve of 0.89 were generated between the proposed method and expert readings on DWI. CONCLUSIONS: The proposed automated ASPECTS scoring approach shows reasonable ability to determine ASPECTS on NCCT images in patients presenting with acute ischemic stroke.
Authors: Simon Nagel; Devesh Sinha; Diana Day; Wolfgang Reith; René Chapot; Panagiotis Papanagiotou; Elizabeth A Warburton; Paul Guyler; Sharon Tysoe; Klaus Fassbender; Silke Walter; Marco Essig; Jens Heidenrich; Angelos A Konstas; Michael Harrison; Michalis Papadakis; Eric Greveson; Olivier Joly; Stephen Gerry; Holly Maguire; Christine Roffe; James Hampton-Till; Alastair M Buchan; Iris Q Grunwald Journal: Int J Stroke Date: 2016-12-01 Impact factor: 5.266
Authors: A Przelaskowski; K Sklinda; P Bargieł; J Walecki; M Biesiadko-Matuszewska; M Kazubek Journal: Comput Biol Med Date: 2006-09-25 Impact factor: 4.589
Authors: Albert J Yoo; Olvert A Berkhemer; Puck S S Fransen; Lucie A van den Berg; Debbie Beumer; Hester F Lingsma; Wouter J Schonewille; Marieke E S Sprengers; René van den Berg; Marianne A A van Walderveen; Ludo F M Beenen; Marieke J H Wermer; Geert J Lycklama À Nijeholt; Jelis Boiten; Sjoerd F M Jenniskens; Joseph C J Bot; Anna M M Boers; Henk A Marquering; Yvo B W E M Roos; Robert J van Oostenbrugge; Diederik W J Dippel; Aad van der Lugt; Wim H van Zwam; Charles B L M Majoie Journal: Lancet Neurol Date: 2016-05-09 Impact factor: 44.182
Authors: T Nezu; M Koga; K Kimura; Y Shiokawa; J Nakagawara; E Furui; H Yamagami; Y Okada; Y Hasegawa; K Kario; S Okuda; K Nishiyama; M Naganuma; K Minematsu; K Toyoda Journal: Neurology Date: 2010-08-10 Impact factor: 9.910
Authors: Mayank Goyal; Andrew M Demchuk; Bijoy K Menon; Muneer Eesa; Jeremy L Rempel; John Thornton; Daniel Roy; Tudor G Jovin; Robert A Willinsky; Biggya L Sapkota; Dar Dowlatshahi; Donald F Frei; Noreen R Kamal; Walter J Montanera; Alexandre Y Poppe; Karla J Ryckborst; Frank L Silver; Ashfaq Shuaib; Donatella Tampieri; David Williams; Oh Young Bang; Blaise W Baxter; Paul A Burns; Hana Choe; Ji-Hoe Heo; Christine A Holmstedt; Brian Jankowitz; Michael Kelly; Guillermo Linares; Jennifer L Mandzia; Jai Shankar; Sung-Il Sohn; Richard H Swartz; Philip A Barber; Shelagh B Coutts; Eric E Smith; William F Morrish; Alain Weill; Suresh Subramaniam; Alim P Mitha; John H Wong; Mark W Lowerison; Tolulope T Sajobi; Michael D Hill Journal: N Engl J Med Date: 2015-02-11 Impact factor: 91.245
Authors: Michael D Hill; Howard A Rowley; Felix Adler; Michael Eliasziw; Anthony Furlan; Randall T Higashida; Lawrence R Wechsler; Heidi C Roberts; William P Dillon; Nancy J Fischbein; Carolyn M Firszt; Gregory A Schulz; Alastair M Buchan Journal: Stroke Date: 2003-07-03 Impact factor: 7.914
Authors: V K Sundaram; J Goldstein; D Wheelwright; A Aggarwal; P S Pawha; A Doshi; J T Fifi; R De Leacy; J Mocco; J Puig; K Nael Journal: AJNR Am J Neuroradiol Date: 2019-11-14 Impact factor: 3.825
Authors: Bach Xuan Tran; Carl A Latkin; Giang Thu Vu; Huong Lan Thi Nguyen; Son Nghiem; Ming-Xuan Tan; Zhi-Kai Lim; Cyrus S H Ho; Roger C M Ho Journal: Int J Environ Res Public Health Date: 2019-07-29 Impact factor: 3.390
Authors: J E Soun; D S Chow; M Nagamine; R S Takhtawala; C G Filippi; W Yu; P D Chang Journal: AJNR Am J Neuroradiol Date: 2020-11-26 Impact factor: 3.825