Introduction: The Large Vessel Occlusion Identification Problem

Acute ischemic stroke is a time-critical medical emergency in which minutes of delayed treatment translate directly into millions of irreversibly lost neurons. The colloquial aphorism "time is brain" reflects a stark mathematical reality: approximately 1.9 million neurons, 14 billion synapses, and 12 kilometers of axonal fibers are lost for every minute that a large vessel occlusion (LVO) stroke goes untreated. Against this backdrop, the ability to rapidly and accurately identify which stroke patients harbor a large vessel occlusion, rather than a smaller branch occlusion or lacunar infarct, has become one of the most consequential decisions in the prehospital and emergency setting.

The emergence of mechanical thrombectomy as a definitive treatment for LVO stroke transformed this identification problem from an academic interest into a clinical imperative. Multiple landmark randomized controlled trials published between 2015 and 2018, including MR CLEAN, ESCAPE, SWIFT PRIME, EXTEND-IA, and DAWN, demonstrated that mechanical thrombectomy at comprehensive stroke centers (CSCs) with neurointerventional capability produced dramatic improvements in functional outcomes for patients with LVO affecting the internal carotid artery (ICA), middle cerebral artery (MCA M1 and M2 segments), basilar artery, and posterior cerebral artery. The number needed to treat to achieve one independent survivor was as low as 2.6 in some trials, making thrombectomy one of the most effective acute interventions in all of medicine.

The challenge is delivery. In most regions, comprehensive stroke centers with around-the-clock neurointerventional capability are distributed sparsely. A patient with LVO stroke transported to the nearest primary stroke center (PSC) may receive thrombolysis but will require a secondary transfer for thrombectomy, adding crucial hours to treatment delay. If that same patient could be identified as an LVO candidate in the prehospital setting by emergency medical services (EMS) personnel and triaged directly to a CSC ("drip and ship" versus "mothership" model), the time to groin puncture and reperfusion could be dramatically shortened.

The Rapid Arterial oCclusion Evaluation (RACE) Scale was developed precisely to enable this prehospital triage decision. It is a brief, structured neurological assessment tool designed to be administered by EMS personnel at the scene or en route to hospital, taking approximately 60 to 90 seconds, and yielding a score that reliably identifies patients likely to have an LVO who should be transported directly to a thrombectomy-capable center.

Epidemiology of Large Vessel Occlusion Stroke

Ischemic stroke accounts for approximately 87 percent of all strokes, with hemorrhagic stroke comprising the remainder. Among ischemic strokes, large vessel occlusions account for approximately 24 to 46 percent of cases, depending on the definition and imaging modality used. The incidence of LVO stroke in the United States is estimated at approximately 150,000 to 200,000 cases per year, representing a massive potential pool of thrombectomy candidates if they can be identified and triaged appropriately.

LVO strokes disproportionately cause severe neurological deficits and poor functional outcomes compared to strokes from small vessel disease or branch occlusions. The occlusion of a proximal artery deprives an entire vascular territory of perfusion, creating a large core of irreversibly infarcted tissue surrounded by a penumbra of hypoperfused but potentially salvageable tissue. The ratio of core to penumbra, and therefore the potential benefit of reperfusion, is highly time-dependent: the penumbra shrinks progressively as ischemia continues, until eventual infarction of the entire territory eliminates the possibility of meaningful recovery from revascularization.

The most common sites of LVO causing anterior circulation strokes amenable to thrombectomy are the terminal internal carotid artery (ICA terminus), the proximal middle cerebral artery (MCA M1 segment), and to a lesser extent the proximal MCA M2 segment. Posterior circulation LVO primarily affecting the basilar artery carries an extremely high mortality and morbidity without treatment and benefits from thrombectomy in appropriately selected patients, though outcome data are less robust than for anterior circulation LVO.

Historical Development of the RACE Scale

The RACE Scale was developed by Natalia Pérez de la Ossa and colleagues in Catalonia, Spain, and published in Stroke in 2014. The Catalonian stroke system provided an ideal development environment: a geographically defined regional network with a single tertiary comprehensive stroke center (Hospital Germans Trias i Pujol in Badalona) serving a large catchment area, with well-organized EMS and a prospective stroke registry enabling systematic data collection.

The development process began with a retrospective analysis of prospectively collected data from consecutive acute stroke activations in the Catalan stroke network. The researchers identified which prehospital and early clinical variables were independently predictive of LVO confirmed by CT angiography (CTA) and correlated with a National Institutes of Health Stroke Scale (NIHSS) score of 14 or greater, which had been previously validated as a surrogate for LVO likelihood. Through logistic regression analysis, they identified facial palsy, arm motor function, leg motor function, gaze deviation, and cortical signs (aphasia or agnosia) as the strongest independent predictors of LVO, and constructed a weighted scoring system from these five domains.

The prospective validation of RACE was performed in the same Catalonian stroke network, with EMS personnel trained to administer the scale to consecutive acute stroke activations before hospital arrival. The prospective validation confirmed that a RACE score of 5 or greater, applied by trained EMS, accurately identified LVO with clinically useful diagnostic performance and was feasible to implement in the prehospital setting without significant additional training burden or time delay.

RACE Scale Components and Scoring Methodology

The RACE Scale consists of five neurological assessment domains, collectively scored from 0 to 9. Each domain can be assessed through brief standardized examination maneuvers that EMS personnel can perform without specialized equipment. The examination is performed on all patients with suspected acute stroke, regardless of the initial clinical impression of severity.

Domain 1: Facial Palsy (0–2 points)

Assessment technique: Ask the patient to show their teeth, smile broadly, and puff out both cheeks simultaneously. Observe for asymmetry of the lower face. Assess the upper face by asking the patient to raise their eyebrows and close both eyes tightly against resistance. In LVO stroke, lower facial palsy (contralateral to the lesion) is more prominent because the upper face has bilateral cortical representation and therefore partial sparing even with unilateral hemispheric injury.

Pathophysiological basis: Facial palsy in LVO stroke results from damage to the corticobulbar tracts descending from the primary motor cortex through the posterior limb of the internal capsule to the contralateral facial nerve nucleus. The degree of palsy reflects the extent of cortical and subcortical damage to the motor pathway serving the face. Severe facial palsy is strongly associated with large cortical and subcortical infarcts characteristic of LVO rather than small vessel or branch occlusion, which tend to produce milder or absent facial involvement.

Domain 2: Arm Motor Function (0–2 points)

Assessment technique: Ask the patient to extend both arms horizontally with palms facing upward, close their eyes, and hold the position for 10 seconds. Observe for pronation, downward drift, or complete inability to maintain the position. In unconscious or uncooperative patients, apply a noxious stimulus (sternal rub or trapezius pinch) and observe the motor response of both upper extremities.

Pathophysiological basis: Upper extremity motor deficits in LVO reflect disruption of the corticospinal tract, which descends from the hand and arm areas of the primary motor cortex (in the precentral gyrus and the superior portion of the motor strip), through the corona radiata, the posterior limb of the internal capsule, the cerebral peduncles, and into the spinal cord. Proximal MCA or ICA occlusions that involve the posterior limb of the internal capsule or large portions of the motor cortex produce severe contralateral arm weakness, while more distal or branch occlusions may produce milder or more selective deficits.

Domain 3: Leg Motor Function (0–2 points)

Assessment technique: In a supine patient, ask them to raise one leg to approximately 30 degrees off the bed or stretcher and hold it for 5 seconds. Repeat on the opposite side. Assess for downward drift or inability to maintain position. Note that the leg motor examination uses a 5-second hold rather than the 10 seconds used for the arm, consistent with the NIHSS methodology and reflecting the greater gravitational load on the lower extremity.

Pathophysiological basis: Leg motor representation in the motor cortex is located in the medial aspect of the hemisphere (the paracentral lobule), which is typically supplied by the anterior cerebral artery (ACA) rather than the MCA. However, the corticospinal fibers from the leg area course through the internal capsule immediately adjacent to the arm fibers, and large MCA or ICA occlusions that produce dense hemiplegia often cause proportional arm and leg weakness through internal capsule involvement. Leg weakness adds incrementally to the RACE score by providing evidence of a large stroke involving the corticospinal pathway at a subcortical level, which correlates with LVO rather than small vessel disease.

Domain 4: Head and Gaze Deviation (0–1 point)

Assessment technique: Observe the patient at rest and with the eyelids open. Note whether the eyes are conjugately deviated to one side without the patient being able to bring them back to midline with voluntary effort. Ask the patient to follow your finger through the full range of horizontal eye movement. Assess whether there is a forced gaze preference that the patient cannot voluntarily overcome. Also note head turning, which may accompany gaze deviation or may be the more prominent finding in patients with impaired cooperation.

Pathophysiological basis: Conjugate gaze deviation (eyes deviated toward the side of the infarct in hemispheric stroke, away from the side of the infarct in pontine stroke) occurs because the frontal eye field (FEF) in the posterior frontal lobe normally drives conjugate gaze toward the contralateral side. When the FEF is damaged by a large hemispheric stroke (typically involving the MCA territory), the contralateral FEF drives the eyes ipsilaterally unopposed, producing forced gaze deviation toward the infarcted hemisphere. The FEF is supplied by the MCA, and its involvement is strongly predictive of a proximal MCA or ICA occlusion rather than a branch or lacunar infarct. Gaze deviation is one of the single most specific clinical signs of LVO stroke.

Domain 5: Cortical Signs (Aphasia or Agnosia) (0–2 points)

This domain uses a different assessment depending on which hemisphere appears to be primarily affected based on the laterality of the motor deficits observed in domains 1 through 4. If the motor deficits are right-sided (suggesting left hemisphere involvement), aphasia is assessed. If the motor deficits are left-sided (suggesting right hemisphere involvement), agnosia is assessed. If laterality is unclear, aphasia testing is typically used.

Left Hemisphere (Aphasia Assessment):

Ask the patient to perform two simple language tasks:

• Task 1: "Close your eyes" (tests comprehension)
• Task 2: "Make a fist" (tests motor command comprehension)

Right Hemisphere (Agnosia Assessment):

Present the patient with their own paretic arm and ask two questions:

• Question 1: "Is this your arm?" (tests somatognosia: recognition of the limb as belonging to self)
• Question 2: "Does this arm have a problem?" (tests anosognosia: awareness of the motor deficit)

Pathophysiological basis (Aphasia): Language function is almost exclusively localized to the left hemisphere in right-handed individuals and in the majority of left-handed individuals. Broca's area (inferior frontal gyrus, typically MCA territory) mediates speech production, and Wernicke's area (posterior superior temporal gyrus, also MCA territory) mediates language comprehension. Large MCA or ICA occlusions typically produce global aphasia (impaired both production and comprehension), which is one of the most highly specific clinical signs of LVO stroke. The inability to follow simple motor commands reflects impaired comprehension, consistent with involvement of Wernicke's area or the perisylvian language network.

Pathophysiological basis (Agnosia): The right hemisphere, particularly the right parietal and temporoparietal junction, subserves awareness of the body schema (somatognosia) and awareness of neurological deficits (anosognosia). Large right hemisphere MCA or ICA strokes that involve the right parietal cortex produce striking neglect syndromes in which patients may deny ownership of their left arm, deny that it is paralyzed, or be entirely unaware of the left side of space (hemispatial neglect). These cortical signs are highly specific for large cortical lesions from LVO and are virtually never seen with small vessel lacunar infarcts, which do not produce cortical dysfunction.

Score Interpretation and LVO Prediction

The threshold of RACE ≥5 was derived from receiver operating characteristic (ROC) curve analysis in the original validation cohort, representing the optimal balance between sensitivity and specificity for LVO detection. In the original prospective validation study, a RACE score ≥5 achieved a sensitivity of 85%, specificity of 68%, positive predictive value (PPV) of 42%, and negative predictive value (NPV) of 94% for LVO confirmed on CTA. The high NPV is clinically important: it means that a RACE score below 5 reliably excludes LVO in most cases, supporting transport to the nearest stroke center without bypassing for thrombectomy-specific capability.

The area under the ROC curve (AUC) for RACE in predicting LVO was 0.82 in the original validation, indicating good overall discriminative ability. This AUC was comparable to the NIHSS (AUC 0.83) for LVO detection, which is remarkable given that the NIHSS is a 15-item scale requiring trained neurological examiners and taking 5 to 10 minutes to administer, while RACE has 5 items and takes approximately 60 to 90 seconds.

Anatomical Targets: What LVO Vessels Does RACE Detect Best?

The RACE Scale was validated against any anterior circulation LVO (ICA, MCA M1, or MCA M2) as the primary reference standard. Its diagnostic performance varies by occlusion location:

• ICA terminus occlusion: Typically produces the highest RACE scores due to the large infarct volume and the combination of dense hemiplegia, hemianopia, gaze deviation, and global aphasia or severe neglect. RACE sensitivity for ICA terminus LVO is highest.
• Proximal MCA (M1) occlusion: The classic LVO causing dense hemiplegia, gaze deviation, and cortical signs. RACE performs very well for M1 LVO, which is the most common and most important target for thrombectomy.
• Distal MCA (M2) occlusion: M2 occlusions produce more variable deficits, sometimes involving only one limb or producing isolated aphasia or isolated neglect without motor deficits. RACE sensitivity for M2 LVO is lower than for M1, reflecting the more limited deficits that may not meet the RACE ≥5 threshold.
• Posterior circulation LVO (basilar artery): The RACE Scale was not specifically designed or validated for posterior circulation stroke. Basilar artery occlusion produces brainstem and cerebellar signs (diplopia, dysarthria, dysphagia, ataxia, crossed sensorimotor findings, coma) that are not well captured by the RACE domains, which are focused on anterior circulation cortical and motor findings. RACE should not be relied upon to identify posterior circulation LVO, and clinicians should use supplementary posterior circulation screening (sudden onset of coma, bilateral motor signs, prominent brainstem findings) in conjunction with RACE.

Prehospital Application: Practical Implementation

EMS Training Requirements

The RACE Scale was specifically designed for administration by EMS personnel, including emergency medical technicians (EMTs) and paramedics, without specialized neurological training. The original validation study demonstrated that EMS personnel could be trained to administer the scale reliably after a brief training session. Studies evaluating the inter-rater reliability of RACE administered by EMS versus emergency physicians have shown moderate to good agreement (Cohen's kappa 0.60–0.75), comparable to other prehospital stroke scales.

Key practical considerations for prehospital RACE administration include:

• The scale is administered after initial patient stabilization (airway, breathing, circulation assessment and management)
• Blood glucose must be checked and hypoglycemia excluded before stroke triage, as hypoglycemia can mimic stroke with focal deficits
• The scale should be administered at scene, with the results communicated to receiving hospitals during transport to pre-activate the stroke team
• EMS personnel should document the RACE score, the time of assessment, the time of symptom onset (or last known well), and the blood glucose value
• Transport decisions based on RACE score must account for local geography: if the nearest CSC is more than 30 to 45 minutes farther than the nearest PSC, the transport time benefit of direct CSC transport may be offset by the risk of delayed thrombolysis

The Mothership vs. Drip-and-Ship Decision

The central operational question that RACE scoring supports is the mothership versus drip-and-ship triage decision. In the mothership model, EMS transports suspected LVO patients directly to a CSC with thrombectomy capability, bypassing closer primary stroke centers. In the drip-and-ship model, patients receive thrombolysis at the nearest PSC and are then transferred to a CSC for thrombectomy if indicated.

The optimal strategy depends critically on geography and transport times. Modeling studies and observational data suggest that the mothership model is superior when the time difference between the nearest PSC and the nearest CSC is less than approximately 30 to 45 minutes. When geographic distances are greater, the delay in thrombolysis imposed by bypassing the nearest PSC may outweigh the benefit of faster thrombectomy access. RACE scoring enables EMS to make this triage decision at the point of care, directing the highest-risk patients (RACE ≥5) to the mothership pathway while routing lower-risk patients to the nearest appropriate center.

Advance Hospital Notification

A high RACE score communicated during transport should trigger a standardized hospital pre-notification protocol that activates the comprehensive stroke team before the patient arrives. This typically includes emergency physician notification, neurology or neurovascular attending availability, CT scanner preparation (including CT angiography protocol), neurointerventional team notification, and angiography suite preparation. Studies demonstrate that pre-notification with a structured severity metric reduces door-to-CT time, door-to-groin-puncture time, and overall onset-to-reperfusion time by meaningful margins.

Diagnostic Performance: Sensitivity, Specificity, and Clinical Utility Metrics

Multiple validation studies across different health systems, EMS configurations, and geographic settings have evaluated RACE performance. Key findings from these external validation studies are summarized below:

The relatively modest positive predictive value (approximately 42%) means that fewer than half of patients triaged to a CSC based on RACE ≥5 will ultimately have an LVO confirmed on vascular imaging. This is by design: in the prehospital setting, it is far more acceptable to over-triage (send non-LVO patients to a CSC unnecessarily) than to under-triage (fail to send an LVO patient to a CSC and deny them thrombectomy access). The high NPV ensures that under-triage is minimized.

The modest specificity also reflects the fundamental clinical challenge: the prehospital setting does not permit vascular imaging, and no clinical examination alone can perfectly replicate the diagnostic accuracy of CTA. RACE performs as well as any clinical scale for LVO detection and substantially better than unstructured clinical impression.

Comparison with Other LVO Detection Scales

RACE is one of several prehospital LVO detection scales developed since 2011. Understanding the relative strengths and limitations of each tool provides context for choosing among them in different clinical systems.

Los Angeles Motor Scale (LAMS)

The LAMS uses three domains (facial droop, arm drift, and grip strength), scored 0 to 5. A score of 4 or more predicts LVO with sensitivity of approximately 81% and specificity of 89%. LAMS is even simpler to administer than RACE but sacrifices the cortical sign assessment (aphasia/agnosia and gaze deviation), which are among the most specific signs of LVO. In head-to-head comparisons, RACE and LAMS show comparable sensitivity but RACE achieves higher sensitivity at the cost of slightly lower specificity in most studies.

Stroke Vision, Aphasia, Neglect (VAN) Assessment

The VAN assessment focuses specifically on cortical signs: vision loss or double vision, aphasia, and neglect. It was developed at a single center and achieves very high sensitivity for LVO (approximately 100% in the development study) by capturing the cortical features most specific for LVO. However, its high sensitivity comes at the cost of very low specificity (approximately 40%), meaning that a very large proportion of patients assessed as VAN-positive will not have LVO on imaging. VAN has not been as broadly externally validated as RACE or LAMS.

Cincinnati Prehospital Stroke Severity (CPSS) Scale

The CPSS uses three items: conjugate gaze deviation, inability to follow commands, and incorrect arm movement. It is scored 0 to 4, with a score of 2 or more predicting LVO and NIHSS ≥15. Its simplicity is a strength, but it does not assess aphasia or agnosia separately and does not provide a leg motor assessment. It has been validated primarily in the Cincinnati regional stroke system.

FAST-ED (Field Assessment Stroke Triage for Emergency Destination)

FAST-ED is an extension of the original FAST (Face, Arm, Speech, Time) public awareness campaign into a clinical triage tool for LVO detection. It adds denial of deficit, gaze deviation, and extinction/neglect to the three FAST components, with a score ≥4 predicting LVO. It achieves sensitivity of approximately 60% and specificity of 89%, with a higher specificity but lower sensitivity profile compared to RACE. FAST-ED is most useful in systems prioritizing minimization of over-triage.

Prehospital Acute Stroke Severity (PASS) Scale

The PASS Scale assesses three items: level of consciousness, gaze deviation, and arm weakness. A score of 2 or more predicts NIHSS ≥15 and LVO with sensitivity of approximately 66% and specificity of 83%. Its three-item structure makes it very fast to administer, and it has been validated in multiple external cohorts.

Summary Comparison

No single scale dominates all others across all metrics. The choice between scales in a given EMS system should be guided by the priority placed on sensitivity versus specificity, the geographic relationship between PSC and CSC facilities, local EMS training resources, and institutional validation data. RACE represents a good balance between diagnostic completeness (5 items capturing both motor and cortical signs) and practical usability in the prehospital setting.

In-Hospital Applications of the RACE Scale

While RACE was designed and validated as a prehospital tool, it has utility in in-hospital settings as well. Emergency department nurses and physicians can use RACE as a rapid structured assessment to triage stroke patients within the ED and to prioritize who receives immediate CT angiography. In facilities without 24/7 neurology coverage, RACE can provide a standardized severity metric for telemedicine consultations, allowing the remote neurologist to rapidly contextualize the deficit severity and LVO probability reported by on-site staff.

In stroke units and neurology floors, RACE can serve as a rapid reassessment tool for monitoring neurological deterioration. A sudden increase in RACE score in a patient previously assessed as having a moderate deficit may signal proximal clot propagation, hemorrhagic transformation, or malignant edema development, each of which warrants immediate reassessment and escalation of care.

RACE in the Context of Extended Thrombectomy Windows

The DAWN and DEFUSE 3 trials demonstrated that carefully selected patients with salvageable penumbra could benefit from thrombectomy up to 24 hours after stroke onset or last known well time. This extended window substantially increased the population of potentially eligible thrombectomy candidates. RACE scoring in the extended window setting requires careful consideration: a high RACE score in a patient presenting beyond 6 hours from onset supports urgent CT perfusion imaging to assess the core-penumbra mismatch ratio that determines eligibility for late-window thrombectomy. The principle is that RACE identifies the clinical severity and LVO likelihood that makes advanced imaging worthwhile, not that it establishes thrombectomy eligibility in isolation.

RACE and Wake-Up Stroke

Approximately 25 percent of ischemic strokes are discovered upon awakening ("wake-up strokes"), in whom the precise onset time is unknown. MRI diffusion-weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) mismatch (DWI positive, FLAIR negative) has been used as an imaging surrogate for recent stroke onset within the 4.5-hour thrombolysis window. In wake-up LVO strokes, a high RACE score should trigger urgent brain and vascular imaging with CT perfusion and CTA (or MRI/MRA if available) to assess for LVO and salvageable tissue, rather than defaulting to exclusion from acute treatment based on unknown onset time alone.

RACE in Mobile Stroke Unit Programs

Mobile stroke units (MSUs), specially equipped ambulances with point-of-care CT scanners, telemedicine capabilities, and on-board thrombolysis capability, are being deployed in an increasing number of metropolitan areas. In the MSU context, RACE scoring guides the initial clinical assessment and determines the urgency of CT scanning and thrombolysis initiation. A high RACE score in an MSU patient who receives thrombolysis can immediately prompt notification to the receiving CSC for thrombectomy preparation, optimizing the bridge between prehospital treatment and definitive endovascular therapy. The RACE score documented at MSU assessment also provides a reference point for monitoring neurological change during MSU transport.

Important Limitations of the RACE Scale

• Not validated for posterior circulation LVO: RACE was designed and validated exclusively for anterior circulation LVO. Its five domains (facial palsy, arm weakness, leg weakness, gaze deviation, aphasia/agnosia) do not capture the cardinal features of posterior circulation stroke (vertigo, diplopia, dysphagia, limb or gait ataxia, bilateral motor signs, altered consciousness). A patient with basilar artery occlusion presenting with coma, bilateral Babinski signs, and pinpoint pupils would score very low on RACE despite having a catastrophic and immediately life-threatening LVO. Posterior circulation stroke should be identified through clinical suspicion and supplementary assessment rather than RACE alone.
• Reduced performance in patients with pre-existing neurological deficits: Patients with baseline hemiplegia from prior stroke, aphasia from a previous infarct, or dementia affecting language and awareness will score higher on RACE at baseline than neurologically intact individuals, potentially mimicking LVO even in the absence of acute stroke. In these patients, comparison with documented baseline neurological status (ideally from bystanders, medical records, or prior EMS encounters) is essential before using the RACE score to guide triage decisions.
• Impaired assessment in intubated or sedated patients: Several RACE domains (aphasia, agnosia, gaze assessment with command following) require patient cooperation and verbal communication. In patients who are intubated, sedated, or have severely depressed consciousness, these domains cannot be reliably assessed, and the RACE score may be artificially low despite severe neurological impairment from LVO.
• Inter-rater variability in the cortical sign domain: The aphasia and agnosia assessments, while straightforward in concept, have somewhat more inter-rater variability than the purely motor components. Patients with mild or moderate aphasia may partially perform the language tasks, and the distinction between a score of 1 and 2 requires judgment. Similarly, anosognosia may be subtle or inconsistently expressed in patients with incomplete right parietal lesions. Training and experience reduce but do not eliminate this variability.
• Optimal threshold may vary by population and setting: The RACE ≥5 threshold was derived from a Spanish EMS population and may not be universally optimal across different stroke systems with different LVO prevalence rates. In populations with higher LVO prevalence, a higher threshold might improve specificity without sacrificing clinically meaningful sensitivity. System-specific validation and threshold calibration is ideally performed before widespread implementation.
• Does not replace vascular imaging: RACE is a clinical screening tool, not a substitute for CT angiography, MR angiography, or transcranial Doppler for definitive LVO confirmation. A positive RACE screen should lead to expedited vascular imaging at the receiving facility; it should not independently trigger thrombectomy without imaging confirmation of the occlusion site, thrombus characteristics, and core-penumbra assessment.
• Time pressure may affect assessment quality: EMS personnel operate under significant time pressure at stroke scenes, with competing demands of patient stabilization, transport preparation, and family communication. Rushed or incomplete RACE assessments may reduce diagnostic accuracy. Embedded assessment reminders in EMS protocols and electronic documentation systems can mitigate this risk.