Clinical evidence behind Avicenna.AI
Welcome to our clinical publications repository, where we showcase the evidence supporting the impact of Avicenna.AI’s advanced solutions in healthcare. Dive into our publications and explore how our CINA portfolio is improving radiology.
Validated performance
Every figure, traced to its study
Every peer-reviewed and conference-presented study behind CINA, grouped by product: the accuracy and clinical-impact figures each one reports, and a link to the paper. Filter by product to see only the evidence that applies to you.
Study & cohort
Sens.
Spec.
Clinical impact
CINA-ICH8 publications
Validation of a Deep Learning Tool in the Detection of Intracranial Hemorrhage and Large Vessel OcclusionMcLouth et al., Frontiers in Neurology, 2021 · n=814 · multicenter US, 44 states
91.4%
97.5%
—
Automated Identification of Intracranial Hemorrhage by Artificial Intelligence Improves Report Turnaround Time in Real Clinical PracticeAyobi et al., ESNR, 2024 · Real-world teleradiology deployment
—
—
Report turnaround time
146.8 → 88.8 min−39%
Assessment of an Artificial Intelligence Algorithm for Detection of Intracranial HemorrhageRava et al., World Neurosurgery, 2021 · Cohort not stated
93%
93%
—
Refining Deep Learning Application Diagnostic Accuracy in Intracerebral Hemorrhage (ICH)Tassy et al., Stroke, 2025 · Cohort not stated
89%
94%
—
Performance of an Artificial Intelligence Tool for Multi-Step Acute Stroke Imaging: A Multicenter Diagnostic StudyAgripnidis et al., European Journal of Radiology, 2025 · Cohort not stated
86%
98%
—
Real-world impact of an AI-driven teleradiology workflow for intracranial hemorrhage detection: Reducing diagnostic delays across a multicenter emergency networkBani-Sadr et al., Neuroscience Informatics, 2026 · Multicenter emergency network
97%
96%
Report turnaround time
−26 min
Patient turnaround time
−14 min
Performance of an AI-Based Automated Identification of Ischemic and Hemorrhagic Stroke in Clinical RoutineEl Ahmadi et al., ECR, 2024 · Clinical routine
—
—
Accuracy (Se/Sp not reported)
93%
Also published, no headline figureArtificial Intelligence and Acute Stroke Imaging - A Comprehensive Review (Soun et al., AJNR, 2021 (review))
CINA-LVO8 publications
Validation of a Deep Learning Tool in the Detection of Intracranial Hemorrhage and Large Vessel OcclusionMcLouth et al., Frontiers in Neurology, 2021 · n=378 CTA · multicenter US
98.1%
98.2%
—
Automated Large Vessel Occlusions Detection: Improved Diagnostic Accuracy in MCA M2 Segment Using Deep LearningFranciosini et al., Stroke, 2025 · n=557 · incl. M2 segment
93.8%[90.2–96.3]
91.2%[87.3–94.2]
—
Duration and Accuracy of Automated Stroke CT Workflow with AI-Supported Intracranial Large Vessel Occlusion DetectionTemmen et al., Scientific Reports, 2023 · Automated stroke CT workflow
—
—
Radiologist reading time
395 → 60 sec−85%
Validation of an Artificial Intelligence Driven Large Vessel Occlusion Detection Algorithm for Acute Ischemic Stroke PatientsRava et al., Neuroradiology Journal, 2021 · Cohort not stated
73%
98%
—
Head-to-Head Comparison of Commercial Artificial Intelligence Solutions for Detection of Large Vessel OcclusionSchlossman et al., Frontiers in Neurology, 2022 · Head-to-head comparison
76%
98%
—
Performance of an Artificial Intelligence Tool for Multi-Step Acute Stroke Imaging: A Multicenter Diagnostic StudyAgripnidis et al., European Journal of Radiology, 2025 · Cohort not stated
55%
97%
—
Performance of an AI-Based Automated Identification of Ischemic and Hemorrhagic Stroke in Clinical RoutineEl Ahmadi et al., ECR, 2024 · Clinical routine
—
—
Accuracy (Se/Sp not reported)
86%
Also published, no headline figureArtificial Intelligence and Acute Stroke Imaging - A Comprehensive Review (Soun et al., AJNR, 2021 (review))
CINA-ASPECTS6 publications
Validation of a Deep Learning AI-Based Software for Automated ASPECTS AssessmentAyobi et al., ECR, 2023 · n=139 · 3 readers
77%
89%
—
Deep Learning in Emergency Radiology: Evaluating a Tool for Automated ASPECTS Scoring in Acute Ischemic StrokeO’Connor et al., Clinical NeuroImaging, 2025 · n=327 · multinational
72.8%
91.8%
—
Deep Learning-Based ASPECTS Algorithm Enhances Reader Performance and Reduces Interpretation TimeAyobi et al., AJNR, 2024 · Multi-reader study
—
—
Reader accuracy (AUC)
0.78 → 0.82
Performance of an Artificial Intelligence Tool for Multi-Step Acute Stroke Imaging: A Multicenter Diagnostic StudyAgripnidis et al., European Journal of Radiology, 2025 · Cohort not stated
68%
89%
—
Performance of an AI-Based Automated Identification of Ischemic and Hemorrhagic Stroke in Clinical RoutineEl Ahmadi et al., ECR, 2024 · Clinical routine
—
—
Accuracy (Se/Sp not reported)
90%
Also published, no headline figureArtificial Intelligence and Acute Stroke Imaging - A Comprehensive Review (Soun et al., AJNR, 2021 (review))
CINA-PE8 publications
Performance and clinical utility of an artificial intelligence-enabled tool for pulmonary embolism detectionAyobi et al., Clinical Imaging, 2024 · n=1,204 CTPA
93.9%[89.3–96.9]
94.8%[93.3–96.1]
Missed-diagnosis rate
15.6% → 3.8%76% recovered
Deep Learning-Based Algorithm for Automatic Detection of Pulmonary Embolism in Chest CT AngiogramsGrenier et al., Diagnostics, 2023 · n=387 · multicenter
91.4%[86.4–95.0]
91.5%[86.8–95.0]
—
Shorter Time to Assessment and Anticoagulation with Decreased Mortality in Patients with Pulmonary Embolism Following Implementation of AI SoftwareShapiro et al., JVS Venous, 2024 · Tertiary referral center, pre/post AI
—
—
Scan-to-alert time
318 → 5.47 min−98%
In-hospital mortality
8.4% → 2.2%−74%
Validation of a Deep Learning Tool for Automatic Pulmonary Embolism DetectionSchlossman et al., ATS, 2023 · Cohort not stated
91%
92%
—
Deep Learning-based Automated Detection of Pulmonary EmbolismBabacan et al., J Comput Assist Tomogr, 2026 · Cohort not stated
>81%
>99%
—
Exploratory validation of an AI-based pulmonary embolism detection tool: Bridging technology and clinical practiceNobushima et al., Int. J. Cardiology Innovations, 2026 · Cohort not stated
95%
99%
—
A Deep Learning-Based Algorithm Improves Radiology Residents' Diagnoses of Acute PE on CT Pulmonary AngiogramsVallée et al., European Journal of Radiology, 2024 · Radiology residents, reader study
—
—
Reader accuracy
95% → 98%
Also published, no headline figureThe Use of Artificial Intelligence Technology in the Detection and Treatment of Pulmonary Embolism at a Tertiary Referral Center (Shapiro et al., Journal of Vascular Surgery, 2023)
CINA-iPE3 publications
Contribution of an Artificial Intelligence Tool in the Detection of Incidental Pulmonary Embolism on Oncology Assessment ScansAmmari et al., Life, 2024 · n=3,050 · oncology
97.3%[85.8–99.9]
97.7%[97.1–98.2]
—
Deep Learning-Based Algorithm for Automatic Detection of Incidental Pulmonary Embolism on Contrast-Enhanced CTFarzaneh et al., Radiology Advances, 2025 · n=381 · multicenter, multivendor
87.8%[82.2–92.2]
92.0%[87.3–95.4]
—
Performance Evaluation of an Artificial Intelligence (AI)-Based Algorithm for Incidental Findings of Pulmonary EmbolismAyobi et al., ATS, 2024 · Cohort not stated
92%
90%
—
CINA-AD3 publications
Diagnostic Performance of a Deep Learning-Powered Application for Aortic Dissection Triage Prioritization and ClassificationLaletin et al., Diagnostics, 2024 · n=1,303 CTA · multicenter
94.2%[88.8–97.5]
97.3%[96.2–98.1]
—
Enhancing Radiologist Efficiency with AI: A Multi-Reader Multi-Case Study on Aortic Dissection Detection and PrioritizationCotena et al., Diagnostics, 2024 · Multi-reader, multi-case study
—
—
Scan-to-assessment time
15.84 → 5.07 min−68%
Algorithmic Performance Consistency Across Patient Demographics and Scanner ManufacturersSalehi et al., Veith Symposium, 2022 · Across demographics and scanner vendors
94%
97%
—
CINA-CAC1 publication
Evaluation of a Deep Learning-Driven Incidental Detection of Coronary Artery Calcium on Non-Gated CT ScansKiewsky et al., ATS, 2025 · Non-gated CT
—
—
Overall agreement (Se/Sp not reported)
85.6%
CINA-VCF4 publications
Validation of a Deep Learning Tool for Detection of Incidental Vertebral Compression FracturesDai et al., J Comput Assist Tomogr, 2025 · n=474 · multicenter
95.2%[90.7–97.9]
92.9%[89.4–96.5]
—
Performance Evaluation of an AI-Based Application for Opportunistic Screening of Thoraco-Lumbar Vertebral Compression FractureAyobi et al., WCO-IOF, 2024 · n=317 · multicenter
92.5%[86.2–96.5]
95.4%[91.5–97.9]
Additional patients identified
+55.8%
Deep Learning-Driven Incidental Detection of Vertebral Fractures in Cancer Patients: Advancing Diagnostic Precision and Clinical ManagementMniai et al., La Radiologia Medica, 2025 · Opportunistic screening, cancer patients
—
—
Unreported prior to AI
80.9% of detections
Performance Evaluation of an AI-Based Application for Opportunistic Screening of Thoraco-Lumbar Vertebral Compression FractureQuemeneur et al., ECR, 2024 · Cohort not stated
92%
92%
—
CINA-VCF Quantix2 publications
Automated Vertebral Compression Fracture Detection and Quantification on Opportunistic CT ScansGuenoun et al., Clinical Radiology, 2025 · n=100 · opportunistic CT
92.3%[81.5–97.9]
91.7%[80.0–97.7]
—
Improving Early Diagnosis of Osteoporotic Vertebral Compression Fractures: A Deep Learning-Based Approach Using CT ImagingCastineira et al., ECR, 2025 · Labelling accuracy 95%
85%
91%
—
CINA-CSpine2 publications
Multicenter, Multinational, and Multivendor Validation of an Artificial Intelligence Application for Acute Cervical Spine Fracture Detection on CTSung et al., Diagnostics, 2026 · n=328 · multinational, multivendor
90.3%[84.5–94.5]
91.9%[86.8–95.5]
—
Also published, no headline figureDiagnostic Accuracy of an AI-powered application for Automated Detection and Prioritisation of Cervical Spine Fractures (Cotena et al., ECR, 2025)
No studies match your filters.