Petra Cimflova1,2, Rotem Golan3, Johanna M Ospel4,5, Alireza Sojoudi3, Chris Duszynski3, Ibukun Elebute3, Houssam El-Hariri3, Seyed Hossein Mousavi3, Luis A Souto Maior Neto3, Najratun Pinky3, Benjamin Beland6, Fouzi Bala6, Nima R Kashani4, William Hu4, Manish Joshi4, Wu Qiu7, Bijoy K Menon7,8. 1. Department of Clinical Neurosciences and Radiology, Cumming School of Medicine, Foothills Medical Centre, University of Calgary, 1403 29th Street NW, Calgary, AB, T2N 2T9, Canada. petracimflova@gmail.com. 2. Department of Medical Imaging, St. Anne's University Hospital Brno and Faculty of Medicine, Masaryk University Brno, Brno, Czech Republic. petracimflova@gmail.com. 3. Circle Neurovascular Imaging Inc., Calgary, AB, Canada. 4. Department of Radiology, Cumming School of Medicine, University of Calgary, 1403 29th Street NW, Calgary, AB, T2N 2T9, Canada. 5. Department of Radiology, University Hospital of Basel, Basel, Switzerland. 6. Department of Clinical Neurosciences, Cumming School of Medicine, University of Calgary, 1403 29th Street NW, Calgary, AB, T2N 2T9, Canada. 7. Department of Clinical Neurosciences and Radiology, Cumming School of Medicine, Foothills Medical Centre, University of Calgary, 1403 29th Street NW, Calgary, AB, T2N 2T9, Canada. 8. Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Canada.
Abstract
PURPOSE: CT angiography (CTA) is the imaging standard for large vessel occlusion (LVO) detection in patients with acute ischemic stroke. StrokeSENS LVO is an automated tool that utilizes a machine learning algorithm to identify anterior large vessel occlusions (LVO) on CTA. The aim of this study was to test the algorithm's performance in LVO detection in an independent dataset. METHODS: A total of 400 studies (217 LVO, 183 other/no occlusion) read by expert consensus were used for retrospective analysis. The LVO was defined as intracranial internal carotid artery (ICA) occlusion and M1 middle cerebral artery (MCA) occlusion. Software performance in detecting anterior LVO was evaluated using receiver operator characteristics (ROC) analysis, reporting area under the curve (AUC), sensitivity, and specificity. Subgroup analyses were performed to evaluate if performance in detecting LVO differed by subgroups, namely M1 MCA and ICA occlusion sites, and in data stratified by patient age, sex, and CTA acquisition characteristics (slice thickness, kilovoltage tube peak, and scanner manufacturer). RESULTS: AUC, sensitivity, and specificity overall were as follows: 0.939, 0.894, and 0.874, respectively, in the full cohort; 0.927, 0.857, and 0.874, respectively, in the ICA occlusion cohort; 0.945, 0.914, and 0.874, respectively, in the M1 MCA occlusion cohort. Performance did not differ significantly by patient age, sex, or CTA acquisition characteristics. CONCLUSION: The StrokeSENS LVO machine learning algorithm detects anterior LVO with high accuracy from a range of scans in a large dataset.
PURPOSE: CT angiography (CTA) is the imaging standard for large vessel occlusion (LVO) detection in patients with acute ischemic stroke. StrokeSENS LVO is an automated tool that utilizes a machine learning algorithm to identify anterior large vessel occlusions (LVO) on CTA. The aim of this study was to test the algorithm's performance in LVO detection in an independent dataset. METHODS: A total of 400 studies (217 LVO, 183 other/no occlusion) read by expert consensus were used for retrospective analysis. The LVO was defined as intracranial internal carotid artery (ICA) occlusion and M1 middle cerebral artery (MCA) occlusion. Software performance in detecting anterior LVO was evaluated using receiver operator characteristics (ROC) analysis, reporting area under the curve (AUC), sensitivity, and specificity. Subgroup analyses were performed to evaluate if performance in detecting LVO differed by subgroups, namely M1 MCA and ICA occlusion sites, and in data stratified by patient age, sex, and CTA acquisition characteristics (slice thickness, kilovoltage tube peak, and scanner manufacturer). RESULTS: AUC, sensitivity, and specificity overall were as follows: 0.939, 0.894, and 0.874, respectively, in the full cohort; 0.927, 0.857, and 0.874, respectively, in the ICA occlusion cohort; 0.945, 0.914, and 0.874, respectively, in the M1 MCA occlusion cohort. Performance did not differ significantly by patient age, sex, or CTA acquisition characteristics. CONCLUSION: The StrokeSENS LVO machine learning algorithm detects anterior LVO with high accuracy from a range of scans in a large dataset.
Authors: Bijoy K Menon; Christopher D d'Esterre; Emmad M Qazi; Mohammed Almekhlafi; Leszek Hahn; Andrew M Demchuk; Mayank Goyal Journal: Radiology Date: 2015-01-29 Impact factor: 11.105
Authors: Mayank Goyal; Bijoy K Menon; Wim H van Zwam; Diederik W J Dippel; Peter J Mitchell; Andrew M Demchuk; Antoni Dávalos; Charles B L M Majoie; Aad van der Lugt; Maria A de Miquel; Geoffrey A Donnan; Yvo B W E M Roos; Alain Bonafe; Reza Jahan; Hans-Christoph Diener; Lucie A van den Berg; Elad I Levy; Olvert A Berkhemer; Vitor M Pereira; Jeremy Rempel; Mònica Millán; Stephen M Davis; Daniel Roy; John Thornton; Luis San Román; Marc Ribó; Debbie Beumer; Bruce Stouch; Scott Brown; Bruce C V Campbell; Robert J van Oostenbrugge; Jeffrey L Saver; Michael D Hill; Tudor G Jovin Journal: Lancet Date: 2016-02-18 Impact factor: 79.321
Authors: Myron D Ginsberg; Yuko Y Palesch; Michael D Hill; Renee H Martin; Claudia S Moy; William G Barsan; Bonnie D Waldman; Diego Tamariz; Karla J Ryckborst Journal: Lancet Neurol Date: 2013-09-27 Impact factor: 44.182
Authors: M A Almekhlafi; W G Kunz; B K Menon; R A McTaggart; M V Jayaraman; B W Baxter; D Heck; D Frei; C P Derdeyn; T Takagi; A H Aamodt; I M R Fragata; M D Hill; A M Demchuk; M Goyal Journal: AJNR Am J Neuroradiol Date: 2019-01-31 Impact factor: 3.825
Authors: Michael D Hill; Mayank Goyal; Bijoy K Menon; Raul G Nogueira; Ryan A McTaggart; Andrew M Demchuk; Alexandre Y Poppe; Brian H Buck; Thalia S Field; Dar Dowlatshahi; Brian A van Adel; Richard H Swartz; Ruchir A Shah; Eric Sauvageau; Charlotte Zerna; Johanna M Ospel; Manish Joshi; Mohammed A Almekhlafi; Karla J Ryckborst; Mark W Lowerison; Kathy Heard; David Garman; Diogo Haussen; Shawna M Cutting; Shelagh B Coutts; Daniel Roy; Jeremy L Rempel; Axel Cr Rohr; Daniela Iancu; Demetrios J Sahlas; Amy Y X Yu; Thomas G Devlin; Ricardo A Hanel; Volker Puetz; Frank L Silver; Bruce C V Campbell; René Chapot; Jeanne Teitelbaum; Jennifer L Mandzia; Timothy J Kleinig; David Turkel-Parrella; Donald Heck; Michael E Kelly; Aditya Bharatha; Oh Young Bang; Ashutosh Jadhav; Rishi Gupta; Donald F Frei; Jason W Tarpley; Cameron G McDougall; Staffan Holmin; Joung-Ho Rha; Ajit S Puri; Marie-Christine Camden; Götz Thomalla; Hana Choe; Stephen J Phillips; Joseph L Schindler; John Thornton; Simon Nagel; Ji Hoe Heo; Sung-Il Sohn; Marios-Nikos Psychogios; Ronald F Budzik; Sidney Starkman; Coleman O Martin; Paul A Burns; Seán Murphy; George A Lopez; Joey English; Michael Tymianski Journal: Lancet Date: 2020-02-20 Impact factor: 79.321
Authors: William J Powers; Alejandro A Rabinstein; Teri Ackerson; Opeolu M Adeoye; Nicholas C Bambakidis; Kyra Becker; José Biller; Michael Brown; Bart M Demaerschalk; Brian Hoh; Edward C Jauch; Chelsea S Kidwell; Thabele M Leslie-Mazwi; Bruce Ovbiagele; Phillip A Scott; Kevin N Sheth; Andrew M Southerland; Deborah V Summers; David L Tirschwell Journal: Stroke Date: 2019-10-30 Impact factor: 7.914