{"\ufeff\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nAI\nIN\nTELESTROKE:\nAN\nEARLY\nSINGLE-CENTER\nEXPERIENCE IN EV ALUATING THE PERFORMANCE OF AI \nFOR DETECTION OF ICH, LVO, AND ASPECTS SCORE \nKamran U1*, Kanwal S1, Durrani M1, Yousaf A1, Abdul Qayum S2, Ur Rehman \nA3, Abbasi W3 \n1Department of Radiology, Rawalpindi Institute of Cardiology, Rawalpindi, Pakistan \n2Department of Neurology, Rawalpindi Institute of Cardiology, Rawalpindi, Pakistan \n3Department of Cardiology, Rawalpindi Institute of Cardiology, Rawalpindi, Pakistan \n*Corresponding author: \nKamran Ummarah, Department of Radiology, Rawalpindi Institute of Cardiology, Rawalpindi, Pakistan \nEmail: ummarahkamran1@gmail.com\nDOI: https://doi.org/10.32896/cvns.v8n1.1-10 \nReceived: 10.11.2025 \nRevised: 01.02.2026 \nAccepted: 05.02.2026 \nPublished: 31.03.2026 \nABSTRACT \nBackground \nStroke represents a critical health crisis, causing significant global mortality and long-term disability. Acute ischemic stroke treatment focuses on timely interventions such as tPA administration and mechanical thrombectomy, but the narrow time window and specialist shortages pose challenges. \nMethods \nThis retrospective study evaluated the efficacy of the Canon Automation Platform software in detecting intracranial hemorrhage (ICH), large vessel occlusion (LVO), and Alberta Stroke Program Early CT Scores (ASPECTS) in 20 stroke patients. \nResults \nAI demonstrated 100% sensitivity in identifying ICH and LVO, with high specificity and negative predictive value. While the AI matched radiologist findings perfectly in extreme ASPECTS categories (6 and 10), it showed variability in intermediate scores (specifically score 8). \nConclusion \nThese findings serve as a preliminary proof of concept. AI improves diagnostic speed and triage accuracy, though radiologist confirmation remains essential for intermediate ASPECTS and false-positive mitigation. \nKeywords: Artificial intelligence (AI)": null, " Deep learning (DL)": null, " Intracranial hemorrhage (ICH)": null, " Large vessel occlusion (LVO)": null, " Machine learning (ML) \n1 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nINTRODUCTION \nThe A stroke represents a critical health crisis characterized by the obstruction of blood flow to the brain, depriving it of oxygen and essential nutrients [1]. This condition is a major cause of global \nCT Score (ASPECTS) is a numerical rating system ranging from 0 to 10 that is utilized in assessing middle cerebral artery (MCA) stroke cases based on CT scan findings. This scoring method involves dividing the MCA vascular territory into segments, with \nmortality\nand\nlong-term\ndisability,\none point deducted from the total score of\nunderscoring\nthe\nneed\nfor\neffective\n10 for each region affected at both the\n\ndiagnostic and therapeutic strategies [2]. Treatment for acute ischemic stroke focuses \nganglionic and supraganglionic levels as shown in Table 1 and Figure 1. The \non \trestoring \tblood \tflow \tthrough inclusion criteria are as follows: \nmedications or procedures within strict time 1. \tPatients with an activated telestroke \nlimits. Intravenous thrombolysis with code. \nalteplase is most effective within 4.5 hours 2. \tCompletion of CTA head and neck \nof symptom onset, while endovascular scans. \ntreatment should occur within 6 hours for 3. \tAnalysis via the automated AI \nlarge vessel occlusions [3,4]. software. \nDue to the narrow diagnostic window and a 4. \tConfirmation \tof \tfindings \tby \nshortage of specialized \tmedical radiologists. \nprofessionals, innovative approaches are essential. Teleneurology has emerged as a significant advancement, offering remote consultations that enhance recovery rates [5]. Artificial intelligence (AI), specifically through machine learning (ML) and deep learning (DL), presents promising solutions to high workloads and slow assessments \nThe imaging data were processed using the Canon Automation Platform, utilizing the ICH, LVO, and ASPECTS modules. These DL-based algorithms analyze non-contrast CT and CTA scans to flag life-threatening neurovascular conditions. LVO specifically targets occlusions in the Internal Carotid Artery (ICA) and the M1/M2 segments of \n[7,8]. the Middle Cerebral Artery (MCA). \nThe integration of AI, such as the computer-assisted Alberta Stroke Program Early CT Score (ASPECTS), has demonstrated high \nASPECTS automatically segments the 10 defined regions of the MCA territory to calculate a score based on tissue density \naccuracy in detecting early ischemic changes. \nchanges [14,15]. This retrospective study \naimed to assess the effectiveness of the Canon Automation platform in detecting ICH, LVO, and ASPECTS in a clinical \nRESULTS AND DISCUSSION \nOut of 20 patients, 14 were male and 6 were female (Mean age: 43.11 \u00b1 20.12 years). \ntelestroke setting.\nSignificant\nco-morbidities\nincluded\ndiabetes mellitus (79.7%) and hypertension \nMETHODOLOGY (66.1%). \nWe conducted a retrospective analysis of The AI \tsystem \tdemonstrated \t100% \npatients presenting with stroke symptoms at sensitivity for both ICH and LVO, correctly \nthe RIC Emergency Department between identifying \tall \tcases \tconfirmed \tby \nJuly 2023 and November 2023. \nWe gathered data from clinical documents, imaging reports, and records from the AI software. We collected information about patients' age, gender and Alberta Stroke Program Early Computed Tomography Score (ASPECTS) for anterior circulation infarcts. The Alberta Stroke Program Early \n2 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nHowever, for the intermediate score of 8, the AI correctly identified only one out of four cases identified by radiologists, \nindicating\nlower\nperformance\nand\nvariability in intermediate ischemic change detection, illustrated in Table 2 and Figures 2-4. \nThe Diagnosing LVO is crucial for \nmechanical\nthrombectomy\neligibility.\nWhile literature reports AI sensitivity between 86% and 97.5%, our study using \nCanon\nAutomation\nPlatform,achieved\n100% sensitivity. This high sensitivity is vital for mobilizing stroke teams promptly [17,18]. \nHowever, our findings regarding ASPECTS indicate that AI performance fluctuates. The drop in accuracy for Score 8 suggests that subtle \"gray-sign\" changes\u2014such as early loss of the insular ribbon\u2014remain difficult for algorithms to distinguish from \nnormal\npatient\nvariation\nor\nchronic\nleukoencephalopathy [19]. \nWith a sample size of N": "20, these findings serve as a preliminary proof of concept. While results are encouraging, a single case can significantly distort percentages in a small cohort. Future multi-center studies with larger datasets are required to validate these performance metrics. Additionally, infrastructure obstacles such as internet interoperability remain a challenge for real-time AI integration. \nCONCLUSION \nAI technology in stroke imaging could significantly change future diagnostics. Our \nobservations\nsuggest\nthat\nautomated\nsoftware like Canon Automation Platform is highly beneficial for rapid triage and LVO detection, especially for non-expert staff. While AI does not currently substitute for an experienced radiologist, its evolution into deep learning suggests it will become an increasingly robust partner in precision medicine. \n3 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \n2018", "22(4):479\u2013484. \n6.Koenig MA, editor. Telemedicine in the ICU. Cham: Springer": null, " 2019. \n7.Krittanawong C, Zhang H, Wang Z, Aydar M, Kitai T. Artificial \nintelligence in precision cardiovascular medicine. J Am Coll Cardiol. \n2017": null, "69(21):2657\u20132664. \n8.Lee JG, Jun S, Cho YW, Lee H, Kim GB, Seo JB, et al. Deep learning in medical imaging: general overview. \nKorean J Radiol. 2017": null, "18(4):570\u2013584. \n9.Erickson BJ, Korfiatis P, Akkus Z, Kline TL. Machine learning for medical imaging. Radiographics. \n2017": null, "37(2):505\u2013515. \n10.Hamburg MA, Collins FS. The path to personalized medicine. N Engl J Med. 2010": null, "363(4):301\u2013304. \n14.Chawla M, Sharma S, Sivaswamy J, Kishore LT. A method for automatic detection and classification of stroke from brain CT images. In: 2009 \nAnnual International Conference of the IEEE Engineering in Medicine and Biology Society. Conf Proc IEEE Eng Med Biol Soc. 2009": null, "2009:3581\u20133584. \n15.Saeys Y, Inza I, Larra\u00f1aga P. A review of feature selection techniques in \nbioinformatics. Bioinformatics. \n2007": null, "23(19):2507\u20132517. \n16.Tang FH, Ng DK, Chow DH. An image feature approach for computer-aided detection of ischemic stroke. Comput Biol Med. 2011": null, "41(7):529\u2013536. \n17.Chatterjee A, Somayaji NR, Kabakis IM. AI detection of large vessel \nocclusion. Stroke. 2019": null, "50(Suppl \n11.Ovbiagele B, Nguyen-Huynh MN. \t1):WP428. \nStroke epidemiology: advancing our understanding of disease mechanism and therapy. Neurotherapeutics. \n2011": null, "8(3):319\u2013329. \n12.Donkor ES. Stroke in the 21st century: a snapshot of the burden, \nepidemiology, and quality of life. \nStroke Res Treat. 2018": null, "2018:3238165. \n13.Lee EJ, Kim YH, Kim N, Kang DW. \nDeep into the brain: artificial intelligence in stroke imaging. J Stroke. 2017": null, "19(3):277\u2013285. \n4 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nFIGURE LEGENDS: \n\nFigure 1: Basal Ganglia and supra ganglionic levels for ASPECT scoring. \n5 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \n\nFigure 2: Left MCA occlusion correctly identified by AI software. \n6 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \n\nFigure 3: Presence of ICH in right basal ganglia correctly identified by AI software. \n7 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \n\nFigure 4: Absence of Bleed and ASPECT score correctly identified by AI software. \n8 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nTABLE LEGEND: \nTable 1: Alberta Stroke Program Early CT Score (ASPECTS) ranging from 0 to 10 for \nassessing middle cerebral artery (MCA). \nRegion involved\nPoints\nGanglionic level:\n\nCaudate\n1\nputamen\n1\nAnterior or posterior limb of Internal capsule\n1\nInsular cortex\n1\nM1: cortex corresponding to frontal operculum\n1\nM2: cortex corresponding to anterior temporal lobe\n1\nM3: cortex corresponding to posterior temporal lobe\n1\nSupraganglionic level:\n\nM4: MCA territory superior to M1\n1\nM5:  MCA territory superior to M2\n1\nM6: MCA territory superior to M3\n1\n9 \n\nJournal of Cardiovascular, Neurovascular & Stroke \nhttps://mycvns.com \nTable 2: Performance Metrics of AI Systems for ICH, LVO, and ASPECT Scores. \n Metric\nICH - AI System\nLVO - AI System\nASPECT Scores - AI System \n(NA, 6, 10)\nASPECT \nScores - AI System (8)\nASPECT \nScores - AI System (9)\nSensitivity\n100% (5/5)\n100% (3/3)\n100%\nN/A\nN/A\nSpecificity\n86.7% \n(13/15)\n88.2% \n(15/17)\n100%\nN/A\nN/A\nPositive \nPredictive Value (PPV)\n71.4% (5/7)\n60.0% (3/5)\n100%\n50% (2/4)\n100% (3/3)\nNegative \nPredictive Value (NPV)\n100% \n(13/13)\n100% \n(15/15)\n100%\nN/A\nN/A\n10": null}