When a stroke patient arrives in the ED, the clock is ticking, but speed without precision can be just as dangerous as delay. That was the central message from a recent RapidAI webinar featuring three internationally recognized experts: Dr. Greg Albers,Stanford University and RapidAI co-founder, Prof. Pierre Seners,Rothschild Foundation Hospital, Paris, and Prof. Marios Psychogios,University Hospital Basel. The session covered everything from the fundamentals of CT perfusion (CTP) interpretation to the frontier of medium vessel occlusion (MeVO) detection and treatment.
Dr. Albers opened with a striking case that set the tone for the entire discussion: a patient with a devastating clinical presentation — dense MCA occlusion, near-complete hemiplegia — whose non-contrast CT looked surprisingly benign. The CTP told an entirely different story: an estimated ischemic core approaching 300 mL, an HIR (Hypoperfusion Intensity Ratio) of 0.9, and near-zero collateral flow. The patient underwent thrombectomy but died within hours from herniation. The point wasn't that perfusion imaging would have changed the outcome,it's that CTP provided critical prognostic clarity that non-con CT simply cannot. In early-window patients especially, perfusion maps can expose catastrophic hemodynamic collapse invisible to conventional imaging.
Of course, that clarity is only possible when the scan itself is technically sound. Two factors can invalidate a study entirely:
To quickly confirm scan quality, clinicians should always review the arterial input function (AIF) and venous output function before interpreting perfusion maps.
Prof. Seners dove into the advanced biomarkers that make CTP so powerful for predicting what happens next in stroke care pathways. The HIR — the ratio of severely delayed tissue (Tmax >10s) to moderately delayed tissue (Tmax >6s) — provides a quantitative window into collateral quality. A high HIR signals poor collaterals and rapid infarct expansion. The CBV (Cerebral Blood Volume) Index complements this by measuring cerebrovascular reserve in the affected hemisphere; a low score correlates with absent collateral flow and poor prognosis.
Together, these metrics are particularly valuable during inter-hospital transfers, where a patient with a high HIR is likely to experience rapid core growth during transport; information that should directly inform both the urgency and appropriateness of the transfer decision. Importantly, both tools apply across all time windows, not just early presentations.
This raised a broader point the panelists were keen to make: perfusion imaging belongs in the early window too, not just the late one. Current guidelines often reserve CTP for patients beyond 4.5–6 hours, but Dr. Albers cited 25 years of Stanford data showing that early-window patients with poor collaterals rarely benefit from EVT. Perfusion imaging also plays an underappreciated role in differentiating true strokes from mimics — a critical triage function regardless of time from onset.
Prof. Psychogios made a compelling case that perfusion imaging is essential for detecting and treating medium vessel occlusions (MeVOs), which can be easily missed on CTA alone. Perfusion deficits make distal M2, M3, and posterior circulation occlusions visible and measurable. He also presented a remarkable case in which real-time CTP identified active hemorrhagic transformation during the scan itself, enabling immediate endovascular intervention.
Looking ahead, the panel agreed on several priorities for future MeVO trials:
Advanced CTP biomarkers can transform how clinicians evaluate stroke patients — not just confirming occlusions, but quantifying tissue viability, collateral reserve, and infarct trajectory. As Prof. Psychogios put it, perfusion is crucial today and will be even more so for the clinical trials that define tomorrow's standard of care. The message from all three experts was consistent: faster acquisition, smarter interpretation, and biomarker-guided decision-making are the path forward in acute stroke care.
Want to go deeper? Watch the full webinar recording or connect with our team to discuss how RapidAI's perfusion tools support these workflows at your institution.
Why should perfusion imaging be used in early-window patients if guidelines only recommend it for late-window cases?
Guidelines reserve CTP for late-window patients because only a small percentage of early arrivals have severely compromised collaterals, but that minority matters. Dr. Albers noted that Stanford's 25-year experience showed early-window patients with poor collateral status almost never benefit from EVT. Perfusion imaging also helps distinguish true strokes from mimics, a critical triage function at any time window.
How should HIR influence the decision to transfer a patient to a comprehensive stroke center?
HIR should be one input in the transfer decision, not the sole determining factor. A high HIR signals rapid infarct growth, which raises urgency,but clinical context matters. A young patient with no prior disability presenting early should almost certainly be transferred despite a high HIR, while an older patient with baseline limitations warrants more careful consideration. Perfusion data should be integrated with the full clinical picture, not used as a binary decision rule.
Does HIR retain its predictive value across different time windows?
Yes. Prof. Seners was clear that HIR applies regardless of time from onset. Patients with poor collaterals experience faster infarct growth, whether they present at two hours or twelve,the biology of inadequate collateral flow doesn't change. The metric is clinically relevant across the full range of time windows.
Is there a relationship between a patient's age and their collateral status?
No consistent one. Prof. Seners noted that the published literature is mixed, with some studies suggesting a link between older age and poorer collaterals and others showing no significant association. Age should not be used as a proxy for collateral quality — imaging-based biomarkers like HIR and the CBV Index are the reliable measure.