Douglas T Hidlay1, Ryan A McTaggart2, Grayson Baird3, Shadi Yaghi4, Morgan Hemendinger5, Eric L Tung1, Eleanor L Dibiasio1, Richard A Haas2, Mahesh V Jayaraman6. 1. Department of Diagnostic Imaging, Warren Alpert School of Medicine at Brown University, United States. 2. Department of Diagnostic Imaging, Warren Alpert School of Medicine at Brown University, United States; Department of Neurology, Warren Alpert School of Medicine at Brown University, United States; Department of Neurosurgery, Warren Alpert School of Medicine at Brown University, United States; Norman Prince Neuroscience Institute, Rhode Island Hospital, United States. 3. Department of Diagnostic Imaging, Warren Alpert School of Medicine at Brown University, United States; Lifespan Biostatistics Core, Rhode Island Hospital, United States. 4. Department of Neurology, Warren Alpert School of Medicine at Brown University, United States; Norman Prince Neuroscience Institute, Rhode Island Hospital, United States. 5. Department of Neurology, Warren Alpert School of Medicine at Brown University, United States. 6. Department of Diagnostic Imaging, Warren Alpert School of Medicine at Brown University, United States; Department of Neurology, Warren Alpert School of Medicine at Brown University, United States; Department of Neurosurgery, Warren Alpert School of Medicine at Brown University, United States; Norman Prince Neuroscience Institute, Rhode Island Hospital, United States. Electronic address: mjayaraman@lifespan.org.
Abstract
OBJECTIVES: Mechanical thrombectomy is the standard of care for stroke caused by an emergent large vessel occlusion in the anterior circulation, and the ability to rapidly review CTA is one hurdle in minimizing the time from diagnosis to intervention. We evaluated the diagnostic accuracy and confidence in review of stroke CTA for ELVO via a smartphone-based application as compared to PACS workstation. PATIENTS AND METHODS: Seventy-six head and neck CTA studies performed for stroke from one comprehensive and seven primary stroke centers were independently reviewed remotely on smartphone by two blinded interventional neuroradiologists in actual-use circumstances. The presence and location of large vessel occlusion(s), diagnostic quality, and confidence in interpretation were recorded. Comparison was made to blinded PACS workstation review performed at a delayed interval. Weighted Kappa and Kendall's Tau statistics were calculated to evaluate intra- and inter-observer reliability. RESULTS: Of the 76 studies, 20 (26%) had a large vessel occlusion. 14 M1 segment occlusions (18%); 2 ICA terminus (3%); 2 tandem carotid and M1 (3%); and 2 basilar artery (3%). There was 100% diagnostic accuracy by both PACS workstation and smartphone review (p = .9999) with high inter- and intra-rater reliability for assessments of both image quality and diagnostic confidence. CONCLUSION: In actual-use circumstances, experienced neuroradiologists can diagnose ELVOs on CTA using a smartphone application as accurately as on PACS workstation without degradation of confidence. These findings support the use of mobile electronic devices by stroke centers to rapidly triage patients for mechanical thrombectomy.
OBJECTIVES: Mechanical thrombectomy is the standard of care for stroke caused by an emergent large vessel occlusion in the anterior circulation, and the ability to rapidly review CTA is one hurdle in minimizing the time from diagnosis to intervention. We evaluated the diagnostic accuracy and confidence in review of stroke CTA for ELVO via a smartphone-based application as compared to PACS workstation. PATIENTS AND METHODS: Seventy-six head and neck CTA studies performed for stroke from one comprehensive and seven primary stroke centers were independently reviewed remotely on smartphone by two blinded interventional neuroradiologists in actual-use circumstances. The presence and location of large vessel occlusion(s), diagnostic quality, and confidence in interpretation were recorded. Comparison was made to blinded PACS workstation review performed at a delayed interval. Weighted Kappa and Kendall's Tau statistics were calculated to evaluate intra- and inter-observer reliability. RESULTS: Of the 76 studies, 20 (26%) had a large vessel occlusion. 14 M1 segment occlusions (18%); 2 ICA terminus (3%); 2 tandem carotid and M1 (3%); and 2 basilar artery (3%). There was 100% diagnostic accuracy by both PACS workstation and smartphone review (p = .9999) with high inter- and intra-rater reliability for assessments of both image quality and diagnostic confidence. CONCLUSION: In actual-use circumstances, experienced neuroradiologists can diagnose ELVOs on CTA using a smartphone application as accurately as on PACS workstation without degradation of confidence. These findings support the use of mobile electronic devices by stroke centers to rapidly triage patients for mechanical thrombectomy.