Slides
Page 1
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -1 PORTABLE RADIOFREQUENCY BRAIN SCANNER EARLY CLINICAL EXPERIENCE EARLY IMPLEMENTATION EXPERIENCE INTO EMERGENCY CODE STROKE WORKFLOW Presented By: Yelim Lee Co-PIs: Dr. David Freeman Dr. Leslie Simon ©2026 Mayo Foundation for Medical Education and Research | WF1197406 -1
Page 2
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -2 INTRODUCTION/HYPOTHESIS Brain Scanner Status: Investigational Device ~220 lbs Portable In-hospital Non-ionizing Trained healthcare professional Technology Non-ionizing radiation using ultra-high frequency radio waves (500-2000MHz) Dielectric property contrast (conductivity and permittivity) cause scan signals to reflect and phase shift Stroke type classification enabled by dielectric contrast between hemorrhage, ischemia and healthy tissue AI Algorithm recognizes scan characteristics of hemorrhage, ischemia and healthy tissue to generate neurodiagnostic output
Page 3
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -3 PRELIMINARY PERFORMANCE TESTING • Earlier Multi-centre Australian study (EMView) enrolled 307 participants including 277 suspected strokes (48 ICH and 140 confirmed ischaemic strokes), ~75% of data used for training and ~25% used for unseen testing Haemorrhage Not Haemorrhage Total Test Cases 13 55 Correctly Identified Cases 12 47 Performance 92% Sensitivity 95% CI: 66.1%-98.2% 85% Specificity 95% CI: 73.8%-92.4% ‘Ischemia (clot) or not’ Including 20 ischaemic, 15 mimics, 20 healthy Ischemic Not Ischemic Total Test Cases 20 50 Correctly Identified Cases 19 40 Performance 95% Sensitivity 95% CI: 75.1%-99.9% 80% Specificity 95% CI: 66.3%-90.0% Including 10 ICH, 18 mimics, 2 TIAs, and 20 healthy Left thalamic intracerebral haemorrhage (0.7 mL) successfully detected and classified as haemorrhage LR Hemorrhage Case Study Nouna Khandan, Zili Xu, Mojtaba Khosravi Farsani et al. Portable RF brain scanner enables deep-learning-based stroke type detection and ordering of brain-event dielectric permittivity, NPJ Digit Med. 2026 Jul 13. doi: 10.1038/s41746-026-02996-6. ‘Hemorrhage (bleed) or not’ Left M2 occlusion ischaemic lesion on follow-up MRI DWI (0.10 mL) successfully detected and classified as ischaemia LR Ischemia Case Study LR LR LR
Page 4
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -4 METHODS Suspected Stroke Clinical Assessment Imaging emu scan Ongoing Diagnostic Evaluation Standard of Care Treatment Biostatistician Hypothesis Testing Central Adjudicating Panel Reference Standard Determination Within 60 mins of baseline imaging Design Multi-Centre, Prospective, Consecutive, Paired Diagnosis, Diagnostic Performance Study Objective Demonstrate sensitivity and specificity >80% Participants ~300 acute suspected stroke patients within 24 hours of symptom onset, and prior to treatment (such as thrombolysis) Additional ‘Training Cohort’ participants enrolled at each site prior to commencing enrolment in above ‘Analysis Cohort’ United States Sites Memorial Hermann-Texas Medical Center Mayo Clinic Florida Mount Sinai Hospital Ronald Reagan UCLA Medical Center Network Sites Mount Sinai West Memorial Hermann Memorial City Medical Center Australian Sites Royal Melbourne Hospital Liverpool Hospital Princess Alexandra HospitalIntegrate into emergency code stroke workflow Currently onboarding additional academic centers Currently onboarding additional academic centers
Page 5
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -5 METHODS - During stroke team assessment - Waiting CT availability Common Scan Opportunities: Each scan 10 Minutes
Page 6
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -6 RESULTS Scanner placed in ED alcove in front of CT suite - Compatible with busy ED workflow - Operated by nurses, clinicians - Operator training ≅ 1 hour - Scan success in majority of cases - No device related adverse events - Agitated patients due to their condition - Inability to remain still for scan - Limited time window Key Challenges:
Page 7
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -7 CONCLUSION Compatible with emergency code stroke workflow Prospective Multicenter Study Study progressing rapidly with active enrolment across participating centers 33+ total participants enrolled at Mayo Clinic Jacksonville, FL 230+ total participants enrolled internationally (the majority in the primary analysis cohort) Obtain n ≅300 Analysis Phase Participants Goal: Demonstrate diagnostic performance to support FDA De Novo clearance
Page 8
©2026 Mayo Foundation for Medical Education and Research | WF1197406 -8 QUESTIONS & DISCUSSION