The CT perfusion mismatch ratio compares the total volume of brain tissue starved of blood flow during an acute stroke against the smaller volume of tissue already irreversibly damaged, and it has become one of the most consequential measurements in modern stroke care. A high ratio signals that a large zone of brain is threatened but not yet dead, meaning a patient could still benefit from clot-removal procedures hours after symptoms begin. The concept reshaped clinical trials and treatment guidelines over the past decade, extending the window for intervention well beyond the traditional few hours after onset.
What the Mismatch Ratio Actually Measures
When a major artery in the brain is blocked, blood flow drops across a wide region. At the center of that region, tissue dies quickly because perfusion falls to nearly zero. Surrounding that dead core is a zone called the penumbra, where cells are struggling but still alive, kept going by trickle flow from neighboring blood vessels. The mismatch ratio is the mathematical relationship between the total critically underperfused volume (core plus penumbra) and the core volume alone. A ratio of 3.0, for instance, means the total threatened territory is three times the size of the dead tissue, leaving a substantial penumbra that could be rescued if blood flow is restored in time.
CT perfusion imaging captures this by tracking a bolus of contrast dye as it passes through the brain’s blood vessels, generating maps of several parameters. Cerebral blood flow measures how much blood moves through a given volume of brain per minute. Cerebral blood volume captures the total amount of blood present in a tissue region. Mean transit time reflects how long blood takes to pass through. These parameters are linked: blood flow equals blood volume divided by transit time.1Insights into Imaging. CT Perfusion in Acute Stroke Together, they allow software to map which tissue is irreversibly injured, which is in jeopardy, and which is safely perfused.
Defining the core and penumbra relies on threshold values applied to these maps. Research has found that a delay time of at least 3 seconds most accurately identifies penumbral tissue, and within that delayed zone, a relative cerebral blood flow at or below 30% of normal best pinpoints the irreversible core.2Scientific Reports. Defining Core and Penumbra in Ischemic Stroke: A Voxel- and Volume-Based Analysis of Whole Brain CT Perfusion Most automated platforms use a related metric called Tmax, which reflects the time-to-peak of the residue function. A Tmax above 6 seconds is the widely adopted threshold for critically hypoperfused tissue, while the core is often defined by relative cerebral blood flow below 30%.
How the Mismatch Ratio Changed Stroke Treatment Timelines
For years, stroke treatment operated under a rigid clock. Clot-busting drugs were given within about 4.5 hours of symptom onset, and clot-retrieval procedures (thrombectomy) within about 6 hours. Beyond those windows, the assumption was that brain tissue had largely progressed from penumbra to core, leaving little to save. The mismatch ratio upended that thinking by showing that many patients retain salvageable tissue far longer than expected.
The DEFUSE 3 trial was pivotal. It enrolled patients with large-vessel blockages who arrived between 6 and 16 hours after they were last known to be well. Eligibility required a core volume under 70 milliliters and a mismatch ratio of 1.8 or greater. Patients randomized to thrombectomy plus standard medical care had dramatically better outcomes: roughly 45% achieved functional independence at 90 days versus 17% with medical care alone.3PubMed Central. Thrombectomy for Stroke at 6 to 16 Hours with Selection by Perfusion Imaging That trial, along with the DAWN trial, fundamentally changed practice. Today, perfusion-based mismatch imaging is the standard gatekeeper for late-window thrombectomy decisions around the world.
Perfusion mismatch selection also appears to identify more patients as eligible for treatment than older approaches based on clinical severity alone. A comparison of the two methods found that perfusion-based mismatch criteria captured a broader pool of treatable patients than clinical-core mismatch criteria used in DAWN.4Stroke. Abstract 113: CT Perfusion Mismatch Identifies More Thrombectomy Patients Than Clinical Core Mismatch Beyond thrombectomy, a systematic review and meta-analysis found that intravenous thrombolysis given in an extended window of 4.5 to 24 hours, guided by perfusion imaging, may improve functional outcomes as well.5PubMed. Intravenous thrombolysis for ischemic stroke in extended time window selected with CT perfusion: a systematic review and meta-analysis
Why Mismatch Profiles Do Not Always Shrink With Time
One of the more counterintuitive findings is that mismatch ratios do not reliably decline as more time passes from symptom onset. You might expect the penumbra to steadily convert to core as hours tick by, eventually erasing the mismatch. But a multicenter study of patients with large-vessel blockages found no significant drop in mismatch ratio over time. Patients who presented beyond 6 hours did have worse scores on standard noncontrast CT (a scoring system called ASPECTS), yet their perfusion-based core volumes were not significantly different from those who arrived earlier.6PubMed Central. Mismatch between automated CTP and ASPECTS score in patients with anterior large vessel occlusion
This finding reinforces why perfusion imaging is so valuable in the extended window. Standard CT may underperform at capturing what is really happening in the brain hours after a stroke starts. CT perfusion-based scoring, meanwhile, has shown stronger correlations with clinical outcomes than noncontrast CT scoring, with cerebral blood volume maps proving statistically better at distinguishing reversible from irreversible injury.7PubMed Central. CT Perfusion ASPECTS in the Evaluation of Acute Ischemic Stroke: Thrombolytic Therapy Perspective Perfusion-based ASPECTS also outperformed both manual and automated noncontrast CT ASPECTS in diagnostic accuracy for large-vessel occlusions.8PubMed Central. CT perfusion based ASPECTS improves the diagnostic performance of early ischemic changes in large vessel occlusion
The Ghost Infarct Core Problem
One of the more concerning pitfalls of mismatch-based decision-making is that CT perfusion can overestimate the infarct core, particularly in the early hours after symptom onset. This phenomenon has been termed the “ghost infarct core.” It occurs when a severely hypoperfused area on CT perfusion looks like dead tissue but later turns out to be viable once blood flow is restored. Follow-up imaging in these cases shows a final infarct that is substantially smaller than the core estimated at admission.9Journal of NeuroInterventional Surgery. Admission CT perfusion may overestimate initial infarct core: the ghost infarct core concept
The practical danger is straightforward: if a patient’s core is overestimated, their mismatch ratio looks worse than it truly is, and they could be denied a treatment that might have helped them. A systematic review confirmed that ghost infarct core is a recognized and recurring phenomenon across multiple retrospective studies, though its exact frequency and the variables that drive it are still being defined.10PubMed. Ghost infarct core: A systematic review of the frequency, magnitude, and variables of CT perfusion overestimation Clinicians increasingly factor this into their decisions, especially when a patient presents very early and the core estimate seems larger than expected for the clinical picture.
Benign Oligemia and Where the Penumbra Boundary Gets Blurry
On the other side of the equation, the penumbra can be overestimated too. Not all underperfused tissue is in danger. Some mildly slow flow, called benign oligemia, represents tissue that is getting enough blood to survive even without intervention. If the software counts this tissue as penumbra, the mismatch ratio gets inflated, making the situation seem more salvageable than it is.
The threshold used to define the penumbra boundary matters enormously. Early studies used a Tmax delay of just over 2 seconds, but research found that a stricter cutoff between 4 and 6 seconds better identifies truly at-risk tissue.11PubMed Central. Optimal Tmax Threshold for Predicting Penumbral Tissue in Acute Stroke Most current platforms have adopted the 6-second Tmax threshold for this reason. Even so, some overestimation of the penumbra persists. One observational study found that Tmax thresholds, while reliable for identifying global hypoperfusion, tended to overestimate both the at-risk territory and the mismatch ratio, and did not cleanly separate true penumbra from benign oligemia in every case.12PubMed. Predictive value of Tmax perfusion maps on final core in acute ischemic stroke: an observational single-center study
How Collateral Blood Vessels Shape the Mismatch
The mismatch ratio is not just a snapshot of the blockage itself. It reflects the brain’s backup blood supply, known as collateral circulation. When one major artery is blocked, neighboring arteries can partially compensate by rerouting blood through smaller connecting vessels. Patients with robust collaterals tend to have smaller infarct cores and larger penumbras, producing higher mismatch ratios and better treatment prospects.
Data from the IMS III trial showed a clear relationship: good collaterals on CT angiography correlated with smaller perfusion-estimated cores and higher mismatch ratios.13PubMed Central. Association between CT angiogram collaterals and CT perfusion in the Interventional Management of Stroke (IMS) III Trial This held true in late time windows as well. A study of patients presenting beyond 6 hours found that those with good collaterals had a median core of 0 milliliters and a median mismatch ratio of about 14, compared to a median core of about 41 milliliters and a mismatch ratio of only about 3 in those with poor collaterals.14PubMed. Association Between CT Angiogram Collaterals and CT Perfusion in Delayed Time Windows for Large Vessel Occlusion Ischemic Strokes This is partly why some patients maintain favorable mismatch profiles many hours after onset, while others with similar blockages do not. Collateral anatomy varies widely between individuals and is influenced by age, chronic high blood pressure, and pre-existing vascular disease.
Software Matters More Than You Might Think
The mismatch ratio is not simply “read” off a scan like a fracture on an X-ray. It is computed by automated software that applies algorithms to the raw perfusion data, and different software platforms can produce meaningfully different numbers from the same images. The two most commonly used systems, RAPID and Olea, have been directly compared.
A multicenter study found that Olea estimated core volumes roughly 8 milliliters larger on average than RAPID, despite a strong correlation between the two systems.15Journal of Stroke and Cerebrovascular Diseases. CTP-based estimated ischemic core: A comparative multicenter study between Olea and RAPID software An earlier comparison found that RAPID’s core estimates correlated more closely with the gold standard of diffusion-weighted MRI, while Olea showed slightly higher sensitivity for detecting any acute infarct but lower specificity.16PubMed. Comparison of Automated CT Perfusion Softwares in Evaluation of Acute Ischemic Stroke An 8-milliliter difference in estimated core can shift a patient’s mismatch ratio enough to change a treatment decision, especially near the eligibility thresholds used in clinical trials. Clinicians are increasingly aware that the software choice is not a neutral technical detail.
Technical Pitfalls That Can Distort the Ratio
Beyond software differences, the raw data acquisition itself introduces opportunities for error. One critical step is choosing the arterial input function, essentially the reference point for how contrast dye enters the brain. If that reference is measured from an artery downstream of the blockage rather than upstream, the contrast curve is flattened and delayed, which can distort all the perfusion maps. In a study of arterial input function placement, measuring from an artery distal to the clot reduced the peak contrast signal from about 438 Hounsfield units to about 151, a more than twofold drop, and introduced bolus-arrival delays of up to 5 seconds in the majority of cases.17PubMed Central. Arterial Input Function Placement for Accurate CT Perfusion Map Construction in Acute Stroke These distortions can artificially enlarge or shrink the estimated core and penumbra.
Other common pitfalls include patient motion during the scan, incomplete brain coverage on older scanners, and low contrast-bolus quality in patients with poor cardiac output. A review of CT perfusion artifacts in acute stroke noted that pitfalls span every stage of the process, from data acquisition to map calculation to interpretation.18PubMed. CT perfusion in acute stroke: know the mimics, potential pitfalls, artifacts, and technical errors Awareness of these issues is part of why stroke centers emphasize standardized protocols and quality assurance for perfusion imaging.
Applying Mismatch Criteria to Medium Vessel Occlusions
The mismatch concept was developed and validated primarily for large-vessel occlusions, blockages in the major arteries feeding the brain. But strokes caused by more distal or medium-sized vessel occlusions are common, and the question of whether mismatch-guided treatment works the same way for these patients is being actively studied.
A recent analysis of patients with distal-medium vessel occlusions who received thrombectomy found that a modified set of target mismatch criteria predicted better outcomes. These criteria used a lower mismatch ratio threshold of 1.2 (instead of 1.8), a core volume at or below 30 milliliters, and a penumbra of at least 10 milliliters. Meeting these criteria was an independent predictor of functional independence, with roughly seven-fold higher odds of a good outcome.19PubMed Central. Target mismatch criteria in acute ischemic stroke patients with distal-medium vessel occlusion Interestingly, the study found that successful recanalization alone did not predict outcomes, and only helped patients who also had a favorable mismatch profile.20European Stroke Journal. Target mismatch criteria in acute ischemic stroke patients with distal-medium vessel occlusion This suggests that for smaller-territory strokes, the imaging selection profile may matter even more than whether the clot is physically removed.
The Posterior Circulation Gap
Most CT perfusion research has focused on strokes in the front of the brain, supplied by the carotid arteries and middle cerebral artery. Strokes in the posterior circulation, supplied by the vertebral and basilar arteries and feeding the brainstem and cerebellum, are a different story. These regions are harder to image with CT perfusion due to bone artifacts, smaller vessel caliber, and lower tissue volumes.
A study evaluating CT perfusion in posterior circulation strokes found that perfusion maps showed abnormalities in only about 65% of confirmed cases, with mean transit time being the most sensitive parameter.21PubMed Central. Effectiveness of CT perfusion in posterior circulation stroke: evaluation of perfusion abnormalities and associated clinical signs That means roughly a third of posterior strokes were missed by CT perfusion entirely. Applying mismatch ratio concepts developed in anterior-circulation trials to posterior strokes remains problematic, and many stroke teams rely more heavily on CT angiography and clinical presentation when the posterior circulation is suspected.
Predicting Hemorrhagic Transformation
CT perfusion data is not just used to decide whether to treat. It also helps predict complications, the most feared of which is hemorrhagic transformation, where brain tissue that has been starved of blood begins to bleed once flow is restored. A systematic review and meta-analysis found that perfusion CT parameters, particularly measures of blood-brain barrier leakiness and the severity of underperfusion, predicted hemorrhagic transformation with pooled sensitivity of about 84% and specificity of about 74%.22PubMed. Perfusion CT for prediction of hemorrhagic transformation in acute ischemic stroke: a systematic review and meta-analysis Among the standard perfusion parameters, lower relative cerebral blood volume was the strongest individual predictor.23PubMed Central. Association of CT perfusion parameters with hemorrhagic transformation in acute ischemic stroke Tissue with very low blood volume has lost its microvascular integrity, making it prone to hemorrhage once reperfusion occurs.
What Happens After the Clot Is Removed
Getting the clot out does not always mean the brain is rescued. Perfusion imaging performed after thrombectomy has revealed a phenomenon called no-reflow, where tissue-level blood flow remains poor despite the large artery being reopened. No-reflow is defined as a greater than 15% reduction in relative cerebral blood flow or volume within the infarct territory compared to the opposite side of the brain, even when angiography shows the artery is fully open.24PubMed. Persistent Tissue-Level Hypoperfusion (No-Reflow) Negates the Clinical Benefit of Successful Thrombectomy A scoping review found that post-recanalization hypoperfusion, though reported in only a minority of clinical studies, was consistently associated with worse outcomes. Post-recanalization hyperperfusion, by contrast, was more commonly reported and showed mixed associations, sometimes linked to good recovery and sometimes to complications like edema.25PubMed Central. Implications of Post-recanalization Perfusion Deficit After Acute Ischemic Stroke: a Scoping Review of Clinical and Preclinical Imaging Studies
When White Matter Disease Muddies the Picture
Older patients commonly have leukoaraiosis, white matter damage visible on brain scans that reflects chronic small-vessel disease. This pre-existing damage complicates CT perfusion interpretation because it lowers baseline blood flow in white matter, making it harder to distinguish genuinely infarcting tissue from chronically underperfused tissue that has looked that way for years.
A study of patients treated with thrombectomy found that as the severity of white matter disease increased, the accuracy of CT perfusion core estimates fell sharply. In patients with no white matter disease, agreement between CT perfusion core and follow-up infarct volume was excellent. In patients with severe leukoaraiosis, that agreement was poor. The standard core threshold of relative cerebral blood flow below 30% only worked well in patients with mild or no white matter disease; patients with moderate disease needed a stricter 25% cutoff, and those with severe disease needed 20%.26PubMed Central. Leukoaraiosis May Confound the Interpretation of CT Perfusion in Patients Treated with Mechanical Thrombectomy for Acute Ischemic Stroke Another study confirmed that low cerebral blood flow in non-infarcted white matter, especially in the setting of leukoaraiosis, reduces specificity when using perfusion maps to define the core.27PubMed. Cerebral blood flow is the optimal CT perfusion parameter for assessing infarct core Since leukoaraiosis is common in the older population most likely to have strokes, this is not an edge case but a routine challenge.
Radiation Exposure and Practical Trade-Offs
CT perfusion adds radiation dose to the stroke workup because it requires repeated imaging of the brain over the course of the contrast-dye pass, typically about 40 to 60 seconds of continuous or near-continuous scanning. One study of 320 stroke patients found a mean effective dose of about 13 millisieverts for the full multimodal CT stroke protocol, though individual doses varied widely depending on scanner settings and whether multiple acquisitions were performed.28Radiation Physics and Chemistry. Evaluation of radiation exposure for patients undergoing computed tomography perfusion procedure for acute ischemic stroke When protocol deviations occur, such as extended scanning times or repeated acquisitions, doses can climb steeply. One analysis found that certain extended protocols delivered peak skin doses exceeding 3 Gy, with significant scatter to the eye lenses as well.29PubMed. Estimation of radiation exposure for brain perfusion CT: standard protocol compared with deviations in protocol
For the typical stroke patient, who is in their 60s or older, the lifetime cancer risk from a single perfusion scan is very low, and the benefit of accurate treatment selection clearly outweighs it. The concern is more relevant for younger patients and for anyone who might undergo repeat perfusion studies. Stroke centers generally advocate judicious use of the full multimodal protocol, particularly in younger patients, and careful adherence to standardized acquisition parameters to avoid unnecessary dose accumulation.30PubMed Central. Radiation exposure of computed tomography imaging for the assessment of acute stroke
Economic Value and Access Disparities
CT perfusion is not universally available. Smaller hospitals and rural facilities often lack the scanners, software licenses, and trained personnel to run and interpret perfusion studies in real time. This creates a two-tier system where patients who happen to have their stroke near a comprehensive stroke center get the benefit of mismatch-guided selection, while others do not.
A health economic modeling study estimated that using CT perfusion for detecting large-vessel occlusions, and then routing those patients to thrombectomy, saved roughly €2,700 per patient and added a meaningful gain in quality-adjusted life years compared to protocols without perfusion imaging.31PubMed Central. Cost-effectiveness of CT perfusion for the detection of large vessel occlusion acute ischemic stroke followed by endovascular treatment: a model-based health economic evaluation study Cloud-based processing has helped somewhat. Services like RAPID can receive perfusion data from a remote scanner, process it in the cloud, and send back the core and penumbra maps within minutes, allowing hospitals without on-site software to participate in mismatch-based triage. But this still requires a scanner capable of performing the perfusion acquisition, stable internet connectivity, and a software subscription.
Deep Learning Approaches to Perfusion Analysis
Emerging research is exploring whether artificial intelligence can sidestep some of the technical complexity of traditional perfusion analysis. One approach uses deep neural networks to predict final infarct volume directly from the raw CT perfusion images, without the intermediate step of generating the conventional perfusion parameter maps through deconvolution mathematics.32PubMed. Prediction of final infarct volume from native CT perfusion and treatment parameters using deep learning By training on paired data of admission perfusion scans and follow-up imaging showing the actual infarct, these models learn to incorporate treatment information and time variables alongside the imaging data itself. The promise is a more personalized prediction that accounts for the messy real-world factors conventional threshold-based approaches cannot easily handle, though clinical validation of these tools remains in progress.