Comparison of 10 perfusion MRI parameters in 97 sub-6-hour stroke patients using voxel-based receiver operating characteristics analysis

Assessing the diagnostic accuracy of CT perfusion: a systematic review

Background and purpose

Computed tomography perfusion (CTP) has successfully extended the time window for reperfusion therapies in ischemic stroke. However, the published perfusion parameters and thresholds vary between studies. Using Preferred Reporting Items for Systematic Reviews and Meta-Analyses of Diagnostic Test Accuracy Studies (PRISMA-DTA) guidelines, we conducted a systematic review to investigate the accuracy of parameters and thresholds for identifying core and penumbra in adult stroke patients.

Methods

We searched Medline, Embase, the Cochrane Library, and reference lists of manuscripts up to April 2022 using the following terms "computed tomography perfusion," "stroke," "infarct," and "penumbra." Studies were included if they reported perfusion thresholds and undertook co-registration of CTP to reference standards. The quality of studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool and Standards for Reporting of Diagnostic Accuracy (STARD) guidelines.

Results

A total of 24 studies were included. A meta-analysis could not be performed due to insufficient data and significant heterogeneity in the study design. When reported, the mean age was 70.2 years (SD+/-3.69), and the median NIHSS on admission was 15 (IQR 13-17). The perfusion parameter identified for the core was relative cerebral blood flow (rCBF), with a median threshold of <30% (IQR 30, 40%). However, later studies reported lower thresholds in the early time window with rapid reperfusion (median 25%, IQR 20, 30%). A total of 15 studies defined a single threshold for all brain regions irrespective of collaterals and the gray and white matter.

Conclusion

A single threshold and parameter may not always accurately differentiate penumbra from core and oligemia. Further refinement of parameters is needed in the current era of reperfusion therapy.

Predicting treatment-specific lesion outcomes in acute ischemic stroke from 4D CT perfusion imaging using spatio-temporal convolutional neural networks

For the diagnosis and precise treatment of acute ischemic stroke, predicting the final location and volume of lesions is of great clinical interest. Current deep learning-based prediction methods mainly use perfusion parameter maps, which can be calculated from spatio-temporal (4D) CT perfusion (CTP) imaging data, to estimate the tissue outcome of an acute ischemic stroke. However, this calculation relies on a deconvolution operation, an ill-posed problem requiring strong regularization and definition of an arterial input function. Thus, improved predictions might be achievable if the deep learning models were applied directly to acute 4D CTP data rather than perfusion maps. In this work, a novel deep spatio-temporal convolutional neural network is proposed for predicting treatment-dependent stroke lesion outcomes by making full use of raw 4D CTP data. By merging a U-Net-like architecture with temporal convolutional networks, we efficiently process the spatio-temporal information available in CTP datasets to make a tissue outcome prediction. The proposed method was evaluated on 147 patients using a 10-fold cross validation, which demonstrated that the proposed 3D+time model (mean Dice=0.45) significantly outperforms both a 2D+time variant of our approach (mean Dice=0.43) and a state-of-the-art method that uses perfusion maps (mean Dice=0.38). These results show that 4D CTP datasets include more predictive information than perfusion parameter maps, and that the proposed method is an efficient approach to make use of this complex data.

MRI software for diffusion-perfusion mismatch analysis may impact on patients' selection and clinical outcome

Objective

Impact of different MR perfusion software on selection and outcome of patients with acute ischemic stroke (AIS) and large vessel occlusion (LVO) treated by endovascular thrombectomy (EVT) is unclear. We aimed at comparing two commercial MRI software, semi-automated with unadjusted (method A) and adjusted mask (method B), and fully automated (method C) in this setting.

Methods

MRI from 144 consecutive AIS patients with anterior circulation LVO was retrospectively analysed. All diffusion- and perfusion-weighted images (DWI-PWI) were post-processed with the three methods using standard thresholds. Concordance for core and hypoperfusion volumes was assessed with Lin’s test. Clinical outcome was compared between groups in patients who underwent successful EVT in the early and late time window.

Results

Mean core volume was higher and mean hypoperfusion volume was lower in method C than in methods A and B. In the early time window, methods A and B found fewer patients with a mismatch ratio ≤ 1.2 than method C (1/67 [1.5%] vs. 12/67 [17.9%], p = 0.0013). In the late time window, methods A and B found fewer patients with a mismatch ratio < 1.8 than method C (3/46 [6.5%] and 2/46 [4.3%] vs. 18/46 [39.1%], p ≤ 0.0002). More patients with functional independence at 3 months would not have been treated using method C versus methods A and B in the early (p = 0.0063) and late (p ≤ 0.011) time window.

Conclusions

MRI software for DWI-PWI analysis may influence patients’ selection before EVT and clinical outcome.

Key Points

• Method C detects fewer patients with favourable mismatch profile.

• Method C might underselect more patients with functional independence at 3 months.

• Software used before thrombectomy may influence patients’ outcome.

Integrating regional perfusion CT information to improve prediction of infarction after stroke

Physiological evidence suggests that neighboring brain regions have similar perfusion characteristics (vascular supply, collateral blood flow). It is largely unknown whether integrating perfusion CT (pCT) information from the area surrounding a given voxel (i.e. the receptive field (RF)) improves the prediction of infarction of this voxel. Based on general linear regression models (GLMs) and using acute pCT-derived maps, we compared the added value of cuboid RF to predict the final infarct. To this aim, we included 144 stroke patients with acute pCT and follow-up MRI, used to delineate the final infarct. Overall, the performance of GLMs to predict the final infarct improved when using RF for all pCT maps (cerebral blood flow, cerebral blood volume, mean transit time and time-to-maximum of the tissue residual function (Tmax)). The highest performance was obtained with Tmax (glm(Tmax); AUC = 0.89 ± 0.03 with RF vs. 0.78 ± 0.02 without RF; p < 0.001) and with a model combining all perfusion parameters (glm(multi); AUC 0.89 ± 0.02 with RF vs. 0.79 ± 0.02 without RF; p < 0.001). These results suggest that prediction of infarction improves by integrating perfusion information from adjacent tissue. This approach may be applied in future studies to better identify ischemic core and penumbra thresholds and improve patient selection for acute stroke treatment.

Impact of the reperfusion status for predicting the final stroke infarct using deep learning

BACKGROUND

Predictive maps of the final infarct may help therapeutic decisions in acute ischemic stroke patients. Our objectives were to assess whether integrating the reperfusion status into deep learning models would improve their performance, and to compare them to current clinical prediction methods.

METHODS

We trained and tested convolutional neural networks (CNNs) to predict the final infarct in acute ischemic stroke patients treated by thrombectomy in our center. When training the CNNs, non-reperfused patients from a non-thrombectomized cohort were added to the training set to increase the size of this group. Baseline diffusion and perfusion-weighted magnetic resonance imaging (MRI) were used as inputs, and the lesion segmented on day-6 MRI served as the ground truth for the final infarct. The cohort was dichotomized into two subsets, reperfused and non-reperfused patients, from which reperfusion status specific CNNs were developed and compared to one another, and to the clinically-used perfusion-diffusion mismatch model. Evaluation metrics included the Dice similarity coefficient (DSC), precision, recall, volumetric similarity, Hausdorff distance and area-under-the-curve (AUC).

RESULTS

We analyzed 109 patients, including 35 without reperfusion. The highest DSC were achieved in both reperfused and non-reperfused patients (DSC = 0.44 ± 0.25 and 0.47 ± 0.17, respectively) when using the corresponding reperfusion status-specific CNN. CNN-based models achieved higher DSC and AUC values compared to those of perfusion-diffusion mismatch models (reperfused patients: AUC = 0.87 ± 0.13 vs 0.79 ± 0.17, P < 0.001; non-reperfused patients: AUC = 0.81 ± 0.13 vs 0.73 ± 0.14, P < 0.01, in CNN vs perfusion-diffusion mismatch models, respectively).

CONCLUSION

The performance of deep learning models improved when the reperfusion status was incorporated in their training. CNN-based models outperformed the clinically-used perfusion-diffusion mismatch model. Comparing the predicted infarct in case of successful vs failed reperfusion may help in estimating the treatment effect and guiding therapeutic decisions in selected patients.

Automated MRI perfusion-diffusion mismatch estimation may be significantly different in individual patients when using different software packages

Objective

To compare two established software applications in terms of apparent diffusion coefficient (ADC) lesion volumes, volume of critically hypoperfused brain tissue, and calculated volumes of perfusion-diffusion mismatch in brain MRI of patients with acute ischemic stroke.

Methods

Brain MRI examinations of 81 patients with acute stroke due to large vessel occlusion of the anterior circulation were analyzed. The volume of hypoperfused brain tissue, ADC volume, and the volume of perfusion-diffusion mismatch were calculated automatically with two different software packages. The calculated parameters were compared quantitatively using formal statistics.

Results

Significant difference was found for the volume of hypoperfused tissue (median 91.0 ml vs. 102.2 ml; p < 0.05) and the ADC volume (median 30.0 ml vs. 23.9 ml; p < 0.05) between different software packages. The volume of the perfusion-diffusion mismatch differed significantly (median 47.0 ml vs. 67.2 ml; p < 0.05). Evaluation of the results on a single-subject basis revealed a mean absolute difference of 20.5 ml for hypoperfused tissue, 10.8 ml for ADC volumes, and 27.6 ml for mismatch volumes, respectively. Application of the DEFUSE 3 threshold of 70 ml infarction core would have resulted in dissenting treatment decisions in 6/81 (7.4%) patients.

Conclusion

Volume segmentation in different software products may lead to significantly different results in the individual patient and may thus seriously influence the decision for or against mechanical thrombectomy.

Key Points

• Automated calculation of MRI perfusion-diffusion mismatch helps clinicians to apply inclusion and exclusion criteria derived from randomized trials.

• Infarct volume segmentation plays a crucial role and lead to significantly different result for different computer programs.

• Perfusion-diffusion mismatch estimation from different computer programs may influence the decision for or against mechanical thrombectomy.

Non-invasive monitoring of longitudinal changes in cerebral hemodynamics in acute ischemic stroke using BOLD signal delay

Relative delays in blood-oxygen-level-dependent (BOLD) signal oscillations can be used to assess cerebral perfusion without using contrast agents. However, little is currently known about the utility of this method in detecting clinically relevant perfusion changes over time. We investigated the relationship between longitudinal BOLD delay changes, vessel recanalization, and reperfusion in 15 acute stroke patients with vessel occlusion examined within 24 h of symptom onset (D0) and one day later (D1). We created BOLD delay maps using time shift analysis of resting-state functional MRI data and quantified perfusion lesion volume changes (using the D1/D0 volume ratio) and severity changes (using a linear mixed model) over time. Between baseline and follow-up, BOLD delay lesions shrank (median D1/D0 ratio = 0.2, IQR = 0.03-0.7) and BOLD delay severity decreased (b = -4.4 s) in patients with recanalization, whereas they grew (median D1/D0 ratio = 1.47, IQR = 1.1-1.7) and became more severe (b = 4.3 s) in patients with persistent vessel occlusion. Clinically relevant changes in cerebral perfusion in early stroke can be detected using BOLD delay, making this non-invasive method a promising option for detecting tissue at risk of infarction and monitoring stroke patients following recanalization therapy.

Technical considerations of multi-parametric tissue outcome prediction methods in acute ischemic stroke patients

Decisions regarding acute stroke treatment rely heavily on imaging, but interpretation can be difficult for physicians. Machine learning methods can assist clinicians by providing tissue outcome predictions for different treatment approaches based on acute multi-parametric imaging. To produce such clinically viable machine learning models, factors such as classifier choice, data normalization, and data balancing must be considered. This study gives comprehensive consideration to these factors by comparing the agreement of voxel-based tissue outcome predictions using acute imaging and clinical parameters with manual lesion segmentations derived from follow-up imaging. This study considers random decision forest, generalized linear model, and k-nearest-neighbor machine learning classifiers in conjunction with three data normalization approaches (non-normalized, relative to contralateral hemisphere, and relative to contralateral VOI), and two data balancing strategies (full dataset and stratified subsampling). These classifier settings were evaluated based on 90 MRI datasets from acute ischemic stroke patients. Distinction was made between patients recanalized using intraarterial and intravenous methods, as well as those without successful recanalization. For primary quantitative comparison, the Dice metric was computed for each voxel-based tissue outcome prediction and its corresponding follow-up lesion segmentation. It was found that the random forest classifier outperformed the generalized linear model and the k-nearest-neighbor classifier, that normalization did not improve the Dice score of the lesion outcome predictions, and that the models generated lesion outcome predictions with higher Dice scores when trained with balanced datasets. No significant difference was found between the treatment groups (intraarterial vs intravenous) regarding the Dice score of the tissue outcome predictions.

Infarct Evolution in a Large Animal Model of Middle Cerebral Artery Occlusion

Mechanical thrombectomy for the treatment of ischemic stroke shows high rates of recanalization; however, some patients still have a poor clinical outcome. A proposed reason for this relates to the fact that the ischemic infarct growth differs significantly between patients. While some patients demonstrate rapid evolution of their infarct core (fast evolvers), others have substantial potentially salvageable penumbral tissue even hours after initial vessel occlusion (slow evolvers). We show that the dog middle cerebral artery occlusion model recapitulates this key aspect of human stroke rendering it a highly desirable model to develop novel multimodal treatments to improve clinical outcomes. Moreover, this model is well suited to develop novel image analysis techniques that allow for improved lesion evolution prediction; we provide proof-of-concept that MRI perfusion-based time-to-peak maps can be utilized to predict the rate of infarct growth as validated by apparent diffusion coefficient-derived lesion maps allowing reliable classification of dogs into fast versus slow evolvers enabling more robust study design for interventional research.

Cerebral Blood Flow Predicts the Infarct Core: New Insights From Contemporaneous Diffusion and Perfusion Imaging

Background and Purpose- The aim of this study is to determine the spatial and volumetric accuracy of infarct core estimates from relative cerebral blood flow (rCBF) by comparison with near-contemporaneous diffusion-weighted imaging (DWI), and evaluate whether it is sufficient for patient triage to reperfusion therapies. Methods- One hundred ninety-three patients enrolled in the DEFUSE 2 (Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution) and SENSE 3 (Sensitivity Encoding) stroke studies were screened, and 119 who underwent acute magnetic resonance imaging with DWI and perfusion imaging within 24 hours of onset were included. Infarct core was estimated using reduced rCBF at 12 thresholds (<0.20-<0.44) and compared against DWI (apparent diffusion coefficient <620 10^-6mm2/s). For each threshold, volumetric agreement between the rCBF and DWI core estimates was assessed using Bland-Altman, correlation, and linear regression analyses; spatial agreement was assessed using receiver operating characteristic analysis. Results- An rCBF threshold of 0.32 yielded the smallest mean absolute volume difference (14.7 mL), best linear regression fit (R²=0.84), and best spatial agreement (Youden index, 0.38; 95% CI, 0.34-0.41) between rCBF and DWI, with high correlation (r=0.91, P<0.05), a small mean volume difference (1.3 mL) and no fixed bias (P<0.05). At this threshold, 110 of 119 (92.4%) patients were correctly triaged when applying 70 mL as the volume limit for thrombectomy. Spatial agreement was better for prediction of large infarcts (>70 mL) than small infarcts (≤70 mL), with Youden indices of 0.53 (95% CI, 0.49-0.56) and 0.34 (95% CI, 0.30-0.37), respectively. Conclusions- Strong correlation and agreement with near-contemporaneous DWI indicate that infarct core estimates obtained using rCBF are sufficiently accurate for patient triage to reperfusion therapies. The identified optimal rCBF threshold of 0.32 closely approximates the threshold currently used in clinical practice.

Multimodal magnetic resonance imaging to identify stroke onset within 6 h in patients with large vessel occlusions

Introduction

Mechanical thrombectomy within 6 h after stroke onset improves the outcome in patients with large vessel occlusions. The aim of our study was to establish a model based on diffusion weighted and perfusion weighted imaging to provide an accurate prediction for the 6 h time-window in patients with unknown time of stroke onset.

Patients and methods

A predictive model was designed based on data from the DEFUSE 2 study and validated in a subgroup of patients with large vessel occlusions from the AXIS 2 trial.

Results

We constructed the model in 91 patients from DEFUSE 2. The following parameters were independently associated with <6 h time-window and included in the model: interquartile range and median relative diffusion weighted imaging, hypoperfusion intensity ratio, core volume and the interaction between median relative diffusion weighted imaging and hypoperfusion intensity ratio as predictors of the 6 h time-window. The area under the curve was 0.80 with a positive predictive value of 0.90 (95%CI 0.79-0.96). In the validation cohort (N = 90), the area under the curve was 0.73 (P for difference = 0.4) with a positive predictive value of 0.85 (95%CI 0.69-0.95).

Discussion

After validation in a larger independent dataset the model can be considered to select patients for endovascular treatment in whom stroke onset is unknown.

Conclusion

In patients with large vessel occlusion and unknown time of stroke onset an automated multivariate imaging model is able to select patients who are likely within the 6 h time-window.

Penumbra detection in acute stroke with perfusion magnetic resonance imaging: Validation with 15 O-positron emission tomography

Objective

Accurate identification of the ischemic penumbra, the therapeutic target in acute clinical stroke, is of critical importance to identify patients who might benefit from reperfusion therapies beyond the established time windows. Therefore, we aimed to validate magnetic resonance imaging (MRI) mismatch–based penumbra detection against full quantitative positron emission tomography (¹⁵O‐PET), the gold standard for penumbra detection in acute ischemic stroke.

Methods

Ten patients (group A) with acute and subacute ischemic stroke underwent perfusion‐weighted (PW)/diffusion‐weighted MRI and consecutive full quantitative ¹⁵O‐PET within 48 hours of stroke onset. Penumbra as defined by ¹⁵O‐PET cerebral blood flow (CBF), oxygen extraction fraction, and oxygen metabolism was used to validate a wide range of established PW measures (eg, time‐to‐maximum [Tmax]) to optimize penumbral tissue detection. Validation was carried out using a voxel‐based receiver‐operating‐characteristic curve analysis. The same validation based on penumbra as defined by quantitative ¹⁵O‐PET CBF was performed for comparative reasons in 23 patients measured within 48 hours of stroke onset (group B).

Results

The PW map Tmax (area‐under‐the‐curve = 0.88) performed best in detecting penumbral tissue up to 48 hours after stroke onset. The optimal threshold to discriminate penumbra from oligemia was Tmax >5.6 seconds with a sensitivity and specificity of >80%.

Interpretation

The performance of the best PW measure Tmax to detect the upper penumbral flow threshold in ischemic stroke is excellent. Tmax >5.6 seconds–based penumbra detection is reliable to guide treatment decisions up to 48 hours after stroke onset and might help to expand reperfusion treatment beyond the current time windows. ANN NEUROL 2019;85:875–886.

Predicting ischemic stroke tissue fate using a deep convolutional neural network on source magnetic resonance perfusion images

Predicting infarct volume from magnetic resonance perfusion-weighted imaging can provide helpful information to clinicians in deciding how aggressively to treat acute stroke patients. Models have been developed to predict tissue fate, yet these models are mostly built using hand-crafted features (e.g., time-to-maximum) derived from perfusion images, which are sensitive to deconvolution methods. We demonstrate the application of deep convolution neural networks (CNNs) on predicting final stroke infarct volume using only the source perfusion images. We propose a deep CNN architecture that improves feature learning and achieves an area under the curve of 0.871 ± 0.024 , outperforming existing tissue fate models. We further validate the proposed deep CNN with existing 2-D and 3-D deep CNNs for images/video classification, showing the importance of the proposed architecture. Our work leverages deep learning techniques in stroke tissue outcome prediction, advancing magnetic resonance imaging perfusion analysis one step closer to an operational decision support tool for stroke treatment guidance.

Extension of therapeutic window in ischemic stroke by selective mismatch imaging

The concept of the ischemic penumbra was formulated on the basis of animal experiments showing functional impairment and electrophysiologic disturbances with decreasing flow to the brain below defined values (the threshold for function) and irreversible tissue damage with blood supply further decreased (the threshold for infarction). The perfusion range between these thresholds was termed the “penumbra,” and restitution of flow above the functional threshold was able to reverse the deficits without permanent damage. In further experiments, the dependency of the development of irreversible lesions on the interaction of the severity and the duration of critically reduced blood flow was established, proving that the lower the flow, the shorter the time for efficient reperfusion. As a consequence, infarction develops from the core of ischemia to the areas of less severe hypoperfusion. The translation of this experimental concept as the basis for the efficient treatment of stroke requires methods by which regional flow and energy metabolism can be repeatedly investigated to demonstrate penumbra tissue, which can benefit from therapeutic interventions. Positron emission tomography allows the quantification of regional cerebral blood flow, the regional oxygen extraction fraction, and the regional metabolic rate for oxygen. With these variables, clear definitions of irreversible tissue damage and of critically hypoperfused but potentially salvageable tissue (i.e. the penumbra) in stroke patients can be achieved. However, positron emission tomography is a research tool, and its complex logistics limit clinical routine applications. Perfusion-weighted or diffusion-weighted magnetic resonance imaging is a widely applicable clinical tool, and the “mismatch” between perfusion-weighted and diffusion-weighted abnormalities serves as an indicator of the penumbra. Also computed tomography angiography and computed tomography perfusion imaging can be used to detect areas suspicious of penumbra. The findings with both methods should be validated by positron emission tomography measurements. Several studies included the selection of patients for intravenous thrombolysis on the basis of a perfusion-weighted imaging–diffusion-weighted imaging mismatch or computed tomography perfusion studies. A meta-analysis of several mismatch-based thrombolysis studies of delayed treatment from the DIAS, DIAS-2, DEDAS, EPITHET, and DEFUSE trials revealed increased recanalization. However, this analysis did not confirm an improvement in clinical outcome with delayed thrombolysis. Randomized controlled trials that did enroll patients based on the presence of a target mismatch on multimodal imaging demonstrated a higher benefit of revascularization treatment by comparison with those who did not and demonstrated for the first time that revascularization treatment for occlusion of an internal carotid artery (ICA) or a proximal middle cerebral artery (MCA) was still beneficial from 6 to 24 h after onset among patients in whom the clinical examination and the multimodal brain imaging indicate a persistent penumbra. On this background, the yield of imaging for the selection of patients for a revascularization therapy will be discussed.

A Quantitative Comparison of Clinically Employed Parameters in the Assessment of Acute Cerebral Ischemia Using Dynamic Susceptibility Contrast Magnetic Resonance Imaging

Purpose: Perfusion magnetic resonance imaging (P-MRI) is part of the mismatch concept employed for therapy decisions in acute ischemic stroke. Using dynamic susceptibility contrast (DSC) MRI the time-to-maximum (Tmax) parameter is quite popular, but its inconsistently defined computation, arterial input function (AIF) selection, and the applied deconvolution method may introduce bias into the assessment. Alternatively, parameter free methods, namely, standardized time-to-peak (stdTTP), z_f-score, and standardized-z_f (stdZ) are also available, offering consistent calculation procedures without the need of an AIF or deconvolution.

Methods: Tmax was compared to stdTTP, z_f-, and stdZ to evaluate robustness of infarct volume estimation in 66 patients, using data from two different sites and MR systems (i.e., 1.5T vs. 3T; short TR (= 689 ms) vs. medium TR (= 1,390 ms); bolus dose 0.1 or 0.2 ml/kgBW, respectively).

Results: Quality factors (QF) for Tmax were 0.54 ± 0.18 (sensitivity), 0.90 ± 0.06 (specificity), and 0.87 ± 0.05 (accuracy). Though not significantly different, best specificity (0.93 ± 0.05) and accuracy (0.90 ± 0.04) were found for stdTTP with a sensitivity of 0.56 ± 0.17. Other tested parameters performed not significantly worse than Tmax and stdTTP, but absolute values of QFs were slightly lower, except for z_f showing the highest sensitivity (0.72 ± 0.16). Accordingly, in ROC-analysis testing the parameter performance to predict the final infarct volume, stdTTP and z_f showed the best performance. The odds for stdTTP to obtain the best prediction of the final infarct size, was 6.42 times higher compared to all other parameters (odds-ratio test; p = 2.2^*10–16).

Conclusion: Based on our results, we suggest to reanalyze data from large cohort studies using the parameters presented here, particularly stdTTP and zf-score, to further increase consistency of perfusion assessment in acute ischemic stroke.

Advances in arterial spin labelling MRI methods for measuring perfusion and collateral flow

With the publication in 2015 of the consensus statement by the perfusion study group of the International Society for Magnetic Resonance in Medicine (ISMRM) and the EU-COST action 'ASL in dementia' on the implementation of arterial spin labelling MRI (ASL) in a clinical setting, the development of ASL can be considered to have become mature and ready for clinical prime-time. In this review article new developments and remaining issues will be discussed, especially focusing on quantification of ASL as well as on new technological developments of ASL for perfusion imaging and flow territory mapping. Uncertainty of the achieved labelling efficiency in pseudo-continuous ASL (pCASL) as well as the presence of arterial transit time artefacts, can be considered the main remaining challenges for the use of quantitative cerebral blood flow (CBF) values. New developments in ASL centre around time-efficient acquisition of dynamic ASL-images by means of time-encoded pCASL and diversification of information content, for example by combined 4D-angiography with perfusion imaging. Current vessel-encoded and super-selective pCASL-methodology have developed into easily applied flow-territory mapping methods providing relevant clinical information with highly similar information content as digital subtraction angiography (DSA), the current clinical standard. Both approaches seem therefore to be ready for clinical use.

Consensus statement on current and emerging methods for the diagnosis and evaluation of cerebrovascular disease

Cerebrovascular disease (CVD) remains a leading cause of death and the leading cause of adult disability in most developed countries. This work summarizes state-of-the-art, and possible future, diagnostic and evaluation approaches in multiple stages of CVD, including (i) visualization of sub-clinical disease processes, (ii) acute stroke theranostics, and (iii) characterization of post-stroke recovery mechanisms. Underlying pathophysiology as it relates to large vessel steno-occlusive disease and the impact of this macrovascular disease on tissue-level viability, hemodynamics (cerebral blood flow, cerebral blood volume, and mean transit time), and metabolism (cerebral metabolic rate of oxygen consumption and pH) are also discussed in the context of emerging neuroimaging protocols with sensitivity to these factors. The overall purpose is to highlight advancements in stroke care and diagnostics and to provide a general overview of emerging research topics that have potential for reducing morbidity in multiple areas of CVD.

Prediction of Tissue Outcome and Assessment of Treatment Effect in Acute Ischemic Stroke Using Deep Learning

Background and Purpose—

Treatment options for patients with acute ischemic stroke depend on the volume of salvageable tissue. This volume assessment is currently based on fixed thresholds and single imagine modalities, limiting accuracy. We wish to develop and validate a predictive model capable of automatically identifying and combining acute imaging features to accurately predict final lesion volume.

Methods—

Using acute magnetic resonance imaging, we developed and trained a deep convolutional neural network (CNN deep ) to predict final imaging outcome. A total of 222 patients were included, of which 187 were treated with rtPA (recombinant tissue-type plasminogen activator). The performance of CNN deep was compared with a shallow CNN based on the perfusion-weighted imaging biomarker Tmax (CNN Tmax ), a shallow CNN based on a combination of 9 different biomarkers (CNN shallow ), a generalized linear model, and thresholding of the diffusion-weighted imaging biomarker apparent diffusion coefficient (ADC) at 600×10 −6 mm 2 /s (ADC thres ). To assess whether CNN deep is capable of differentiating outcomes of ±intravenous rtPA, patients not receiving intravenous rtPA were included to train CNN deep, −rtpa to access a treatment effect. The networks’ performances were evaluated using visual inspection, area under the receiver operating characteristic curve (AUC), and contrast.

Results—

CNN deep yields significantly better performance in predicting final outcome (AUC=0.88±0.12) than generalized linear model (AUC=0.78±0.12; P =0.005), CNN Tmax (AUC=0.72±0.14; P <0.003), and ADC thres (AUC=0.66±0.13; P <0.0001) and a substantially better performance than CNN shallow (AUC=0.85±0.11; P =0.063). Measured by contrast, CNN deep improves the predictions significantly, showing superiority to all other methods ( P ≤0.003). CNN deep also seems to be able to differentiate outcomes based on treatment strategy with the volume of final infarct being significantly different ( P =0.048).

Conclusions—

The considerable prediction improvement accuracy over current state of the art increases the potential for automated decision support in providing recommendations for personalized treatment plans.

Boosted Tree Model Reforms Multimodal Magnetic Resonance Imaging Infarct Prediction in Acute Stroke

BACKGROUND AND PURPOSE

Stroke imaging is pivotal for diagnosis and stratification of patients with acute ischemic stroke to treatment. The potential of combining multimodal information into reliable estimates of outcome learning calls for robust machine learning techniques with high flexibility and accuracy. We applied the novel extreme gradient boosting algorithm for multimodal magnetic resonance imaging-based infarct prediction.

METHODS

In a retrospective analysis of 195 patients with acute ischemic stroke, fluid-attenuated inversion recovery, diffusion-weighted imaging, and 10 perfusion parameters were derived from acute magnetic resonance imaging scans. They were integrated to predict final infarct as seen on follow-up T2-fluid-attenuated inversion recovery using the extreme gradient boosting and compared with a standard generalized linear model approach using cross-validation. Submodels for recanalization and persistent occlusion were calculated and were used to identify the important imaging markers. Performance in infarct prediction was analyzed with receiver operating characteristics. Resulting areas under the curve and accuracy rates were compared using Wilcoxon signed-rank test.

RESULTS

The extreme gradient boosting model demonstrated significantly higher performance in infarct prediction compared with generalized linear model in both cross-validation approaches: 5-folds (P<10e-16) and leave-one-out (P<0.015). The imaging parameters time-to-peak, mean transit time, time-to-maximum, and diffusion-weighted imaging were indicated as most valuable for infarct prediction by the systematic algorithm rating. Notably, the performance improvement was higher with 5-folds cross-validation approach than leave-one-out.

CONCLUSIONS

We demonstrate extreme gradient boosting as a state-of-the-art method for clinically applicable multimodal magnetic resonance imaging infarct prediction in acute ischemic stroke. Our findings emphasize the role of perfusion parameters as important biomarkers for infarct prediction. The effect of cross-validation techniques on performance indicates that the intrapatient variability is expressed in nonlinear dynamics of the imaging modalities.

A Comparison of Relative Time to Peak and Tmax for Mismatch-Based Patient Selection

BACKGROUND AND PURPOSE

The perfusion-weighted imaging (PWI)/diffusion-weighted imaging (DWI) mismatch profile is used to select patients for endovascular treatment. A PWI map of Tmax is commonly used to identify tissue with critical hypoperfusion. A time to peak (TTP) map reflects similar hemodynamic properties with the added benefit that it does not require arterial input function (AIF) selection and deconvolution. We aimed to determine if TTP could substitute Tmax for mismatch categorization.

METHODS

Imaging data of the DEFUSE 2 trial were reprocessed to generate relative TTP (rTTP) maps. We identified the rTTP threshold that yielded lesion volumes comparable to Tmax > 6 s and assessed the effect of reperfusion according to mismatch status, determined based on Tmax and rTTP volumes.

RESULTS

Among 102 included cases, the Tmax > 6 s lesion volumes corresponded most closely with rTTP > 4.5 s lesion volumes: median absolute difference 6.9 mL (IQR: 2.3-13.0). There was 94% agreement in mismatch classification between Tmax and rTTP-based criteria. When mismatch was assessed by Tmax criteria, the odds ratio (OR) for favorable clinical response associated with reperfusion was 7.4 (95% CI 2.3-24.1) in patients with mismatch vs. 0.4 (95% CI 0.1-2.6) in patients without mismatch. When mismatch was assessed with rTTP criteria, these ORs were 7.2 (95% CI 2.3-22.2) and 0.3 (95% CI 0.1-2.2), respectively.

CONCLUSION

rTTP yields lesion volumes that are comparable to Tmax and reliably identifies the PWI/DWI mismatch profile. Since rTTP is void of the problems associated with AIF selection, it is a suitable substitute for Tmax that could improve the robustness and reproducibility of mismatch classification in acute stroke.

MRI-based mismatch detection in acute ischemic stroke: Optimal PWI maps and thresholds validated with PET

Perfusion-weighted (PW) magnetic resonance imaging (MRI) is used to detect penumbral tissue in acute stroke, but the selection of optimal PW-maps and thresholds for tissue at risk detection remains a matter of debate. We validated the performance of PW-maps with 15O-water-positron emission tomography (PET) in a large comparative PET-MR cohort of acute stroke patients. In acute and subacute stroke patients with back-to-back MRI and PET imaging, PW-maps were validated with 15O-water-PET. We pooled two different cerebral blood flow (CBF) PET-maps to define the critical flow (CF) threshold, (i) quantitative (q)CBF-PET with the CF threshold <20 ml/100 g/min and (ii) normalized non-quantitative (nq)CBF-PET with a CF threshold of <70% (corresponding to <20 ml/100 g/min according to a previously published normogram). A receiver operating characteristic (ROC) curve analysis was performed to specify the accuracy and the optimal critical flow threshold of each PW-map as defined by PET. In 53 patients, (stroke to imaging: 9.8 h; PET to MRI: 52 min) PW-time-to-maximum (Tmax) with a threshold >6.1 s (AUC = 0.94) and non-deconvolved PW-time-to-peak (TTP) >4.8 s (AUC = 0.93) showed the best performance to detect the CF threshold as defined by PET. PW-Tmax with a threshold >6.1 s and TTP with a threshold >4.8 s are the most predictive in detecting the CF threshold for MR-based mismatch definition.

Comparison of perfusion signal acquired by arterial spin labeling-prepared intravoxel incoherent motion (IVIM) MRI and conventional IVIM MRI to unravel the origin of the IVIM signal

PURPOSE

Applications of intravoxel incoherent motion (IVIM) imaging in the brain are scarce, whereas it has been successfully applied in other organs with promising results. To better understand the cerebral IVIM signal, the diffusion properties of the arterial blood flow within different parts of the cerebral vascular tree (i.e., different generations of the branching pattern) were isolated and measured by employing an arterial spin labeling (ASL) preparation module before an IVIM readout.

METHODS

ASL preparation was achieved by T₁ -adjusted time-encoded pseudo-continuous ASL (te-pCASL). The IVIM readout module was achieved by introducing bipolar gradients immediately after the excitation pulse. The results of ASL-IVIM were compared with those of conventional IVIM to improve our understanding of the signal generation process of IVIM.

RESULTS

The pseudo-diffusion coefficient D* as calculated from ASL-IVIM data was found to decrease exponentially for postlabeling delays (PLDs) between 883 ms and 2176 ms, becoming relatively stable for PLDs longer than 2176 ms. The fast compartment of the conventional IVIM-experiment shows comparable apparent diffusion values to the ASL signal with PLDs between 1747 ms and 2176 ms. At the longest PLDs, the observed D* values (4.0 ± 2.8 × 10^-3 mm² /s) are approximately 4.5 times higher than the slow compartment (0.90 ± 0.05 × 10^-3 mm² /s) of the conventional IVIM experiment.

CONCLUSION

This study showed much more complicated diffusion properties of vascular signal than the conventionally assumed single D* of the perfusion compartment in the two-compartment model of IVIM (biexponential behavior). Magn Reson Med 79:723-729, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Comparison of Different Post-Processing Algorithms for Dynamic Susceptibility Contrast Perfusion Imaging of Cerebral Gliomas

Purpose:

The purpose of the present study was to compare different software algorithms for processing DSC perfusion images of cerebral tumors with respect to i) the relative CBV (rCBV) calculated, ii) the cutoff value for discriminating low- and high-grade gliomas, and iii) the diagnostic performance for differentiating these tumors.

Methods:

Following approval of institutional review board, informed consent was obtained from all patients. Thirty-five patients with primary glioma (grade II, 9; grade III, 8; and grade IV, 18 patients) were included. DSC perfusion imaging was performed with 3-Tesla MRI scanner. CBV maps were generated by using 11 different algorithms of four commercially available software and one academic program. rCBV of each tumor compared to normal white matter was calculated by ROI measurements. Differences in rCBV value were compared between algorithms for each tumor grade. Receiver operator characteristics analysis was conducted for the evaluation of diagnostic performance of different algorithms for differentiating between different grades.

Results:

Several algorithms showed significant differences in rCBV, especially for grade IV tumors. When differentiating between low- (II) and high-grade (III/IV) tumors, the area under the ROC curve (Az) was similar (range 0.85–0.87), and there were no significant differences in Az between any pair of algorithms. In contrast, the optimal cutoff values varied between algorithms (range 4.18–6.53).

Conclusions:

rCBV values of tumor and cutoff values for discriminating low- and high-grade gliomas differed between software packages, suggesting that optimal software-specific cutoff values should be used for diagnosis of high-grade gliomas.

Relationship Between Changes in the Temporal Dynamics of the Blood-Oxygen-Level-Dependent Signal and Hypoperfusion in Acute Ischemic Stroke

Background and Purpose—

Changes in the blood-oxygen-level-dependent (BOLD) signal provide a noninvasive measure of blood flow, but a detailed comparison with established perfusion parameters in acute stroke is lacking. We investigated the relationship between BOLD signal temporal delay and dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) in stroke patients.

Methods—

In 30 patients with acute (<24 hours) ischemic stroke, we performed Pearson correlation and multiple linear regression between DSC-MRI parameters (time to maximum [Tmax], mean transit time, cerebral blood flow, and cerebral blood volume) and BOLD-based parameters (BOLD delay and coefficient of BOLD variation). Prediction of severe hypoperfusion (Tmax >6 seconds) was assessed using receiver–operator characteristic (ROC) analysis.

Results—

Correlation was highest between Tmax and BOLD delay (venous sinus reference; time shift range 7; median r =0.60; interquartile range=0.49–0.71). Coefficient of BOLD variation correlated with cerebral blood volume (median r = 0.37; interquartile range=0.24–0.51). Mean R 2 for predicting BOLD delay by DSC-MRI was 0.54 (SD=0.2) and for predicting coefficient of BOLD variation was 0.37 (SD=0.17). BOLD delay (whole-brain reference, time shift range 3) had an area under the curve of 0.76 for predicting severe hypoperfusion (sensitivity=69.2%; specificity=80%), whereas BOLD delay (venous sinus reference, time shift range 3) had an area under the curve of 0.76 (sensitivity=67.3%; specificity=83.5%).

Conclusions—

BOLD delay is related to macrovascular delay and microvascular hypoperfusion, can identify severely hypoperfused tissue in acute stroke, and is a promising alternative to gadolinium contrast agent–based perfusion assessment in acute stroke.

Clinical Trial Registration—

URL: http://www.clinicaltrials.gov . Unique identifier: NCT00715533 and NCT02077582.

A PET-Guided Framework Supports a Multiple Arterial Input Functions Approach in DSC-MRI in Acute Stroke

BACKGROUND AND PURPOSE

In acute stroke, arterial-input-function (AIF) determination is essential for obtaining perfusion estimates with dynamic susceptibility-weighted contrast-enhanced magnetic resonance imaging (DSC-MRI). Standard DSC-MRI postprocessing applies single AIF selection, ie, global AIF. Physiological considerations, however, suggest that a multiple AIFs selection method would improve perfusion estimates to detect penumbral flow. In this study, we developed a framework based on comparable DSC-MRI and positron emission tomography (PET) images to compare the two AIF selection approaches and assess their performance in penumbral flow detection in acute stroke.

METHODS

In a retrospective analysis of 17 sub(acute) stroke patients with consecutive MRI and PET scans, voxel-wise optimized AIFs were calculated based on the kinetic model as derived from both imaging modalities. Perfusion maps were calculated based on the optimized-AIF using two methodologies: (1) Global AIF and (2) multiple AIFs as identified by cluster analysis. Performance of penumbral-flow detection was tested by receiver-operating characteristics (ROC) curve analysis, ie, the area under the curve (AUC).

RESULTS

Large variation of optimized AIFs across brain voxels demonstrated that there is no optimal single AIF. Subsequently, the multiple-AIF method (AUC range over all maps: .82-.90) outperformed the global AIF methodology (AUC .72-.85) significantly.

CONCLUSIONS

We provide PET imaging-based evidence that a multiple AIF methodology is beneficial for penumbral flow detection in comparison with the standard global AIF methodology in acute stroke.

Distal Middle Cerebral Artery Branch Recanalization Does Not Affect Final Lesion Volume in Ischemic Stroke

Background: Early recanalization in ischemic stroke is associated with favorable outcome. However, limited data are available on the effect of recanalization on infarct growth and functional outcome in stroke with distal middle cerebral artery (MCA) pathology. This study was aimed at determining the effect of recanalization in acute stroke patients with perfusion-diffusion mismatch and occlusion or high-grade stenosis of a distal MCA branch. Methods: We prospectively examined 34 consecutive stroke patients with perfusion-diffusion mismatch and M3 or M4 pathology within 24 h of symptom onset. The MRI protocol consisted of diffusion-weighted images (DWI), fluid-attenuated inversion recovery (FLAIR), T2*, perfusion-weighted imaging, time-of-flight magnetic resonance angiography at days 0, 1, and 4-6. Volume measurements were performed with MRIcron. Infarct growth was defined as the difference between lesion volumes on FLAIR at days 4-6 and DWI at day 0. Certified raters assessed modified Rankin Scale scores at discharge and day 90. Results: Twenty-four patients (71%) showed recanalization at day 1. Infarct growth was modest (median 2.4 mL, 95% CI 0.8-6.7) and not significantly different between patients with and without recanalization (p = 0.87). Functional outcome at discharge was good with 70% of patients suffering no significant disabilities. There was no association between functional outcome at discharge and recanalization (OR 2.1, 95% CI 0.4-13.0, p = 0.40) or infarct volume at days 4-6 (p = 0.40). Conclusions: The high rate of spontaneous recanalization and good functional outcome in patients with distal MCA pathology might obscure a potential benefit from recanalization in this population.

Validation of MRI Determination of the Penumbra by PET Measurements in Ischemic Stroke

The concept of the ischemic penumbra was formulated on the basis of animal experiments showing functional impairment and electrophysiologic disturbances with decreasing flow to the brain below defined values (the threshold for function) and irreversible tissue damage with blood supply further decreased (the threshold for infarction). The perfusion range between these thresholds was termed the "penumbra," and restitution of flow above the functional threshold was able to reverse the deficits without permanent damage. In further experiments, the dependency of the development of irreversible lesions on the interaction of the severity and the duration of critically reduced blood flow was established, proving that the lower the flow, the shorter the time for efficient reperfusion. As a consequence, infarction develops from the core of ischemia to the areas of less severe hypoperfusion. The translation of this experimental concept as the basis for the efficient treatment of stroke requires noninvasive methods with which regional flow and energy metabolism can be repeatedly investigated to demonstrate penumbra tissue, which can benefit from therapeutic interventions. PET allows the quantification of regional cerebral blood flow, the regional oxygen extraction fraction, and the regional metabolic rate for oxygen. With these variables, clear definitions of irreversible tissue damage and of critically hypoperfused but potentially salvageable tissue (i.e., the penumbra) in stroke patients can be achieved. However, PET is a research tool, and its complex logistics limit clinical routine applications. Perfusion-weighted or diffusion-weighted MRI is a widely applicable clinical tool, and the "mismatch" between perfusion-weighted and diffusion-weighted abnormalities serves as an indicator of the penumbra. However, comparative studies of perfusion-weighted or diffusion-weighted MRI and PET have indicated overestimation of the core of irreversible infarction as well as of the penumbra by the MRI modalities. Some of these discrepancies can be explained by the nonselective application of relative perfusion thresholds, which might be improved by more complex analytic procedures. The heterogeneity of the MRI signatures used for the definition of the mismatch are also responsible for disappointing results in the application of perfusion-weighted or diffusion-weighted MRI to the selection of patients for clinical trials. As long as validation of the mismatch selection paradigm is lacking, the use of this paradigm as a surrogate marker of outcome is limited.

Quantification of Serial Cerebral Blood Flow in Acute Stroke Using Arterial Spin Labeling

Supplemental Digital Content is available in the text.

Background and Purpose—

Perfusion-weighted imaging is used to select patients with acute ischemic stroke for intervention, but knowledge of cerebral perfusion can also inform the understanding of ischemic injury. Arterial spin labeling allows repeated measurement of absolute cerebral blood flow (CBF) without the need for exogenous contrast. The aim of this study was to explore the relationship between dynamic CBF and tissue outcome in the month after stroke onset.

Methods—

Patients with nonlacunar ischemic stroke underwent ≤5 repeated magnetic resonance imaging scans at presentation, 2 hours, 1 day, 1 week, and 1 month. Imaging included vessel-encoded pseudocontinuous arterial spin labeling using multiple postlabeling delays to quantify CBF in gray matter regions of interest. Receiver–operator characteristic curves were used to predict tissue outcome using CBF. Repeatability was assessed in 6 healthy volunteers and compared with contralateral regions of patients. Diffusion-weighted and T2-weighted fluid attenuated inversion recovery imaging were used to define tissue outcome.

Results—

Forty patients were included. In contralateral regions of patients, there was significant variation of CBF between individuals, but not between scan times (mean±SD: 53±42 mL/100 g/min). Within ischemic regions, mean CBF was lowest in ischemic core (17±23 mL/100 g/min), followed by regions of early (21±26 mL/100 g/min) and late infarct growth (25±35 mL/100 g/min; ANOVA P<0.0001). Between patients, there was marked overlap in presenting and serial CBF values.

Conclusions—

Knowledge of perfusion dynamics partially explained tissue fate. Factors such as metabolism and tissue susceptibility are also likely to influence tissue outcome.

Expanding the concept of neuroprotection for acute ischemic stroke: The pivotal roles of reperfusion and the collateral circulation

This review surveys the efforts taken to achieve clinically efficacious protection of the ischemic brain and underscores the necessity of expanding our purview to include the essential role of cerebral perfusion and the collateral circulation. We consider the development of quantitative strategies to measure cerebral perfusion at the regional and local levels and the application of these methods to elucidate flow-related thresholds of ischemic viability and to characterize the ischemic penumbra. We stress that the modern concept of neuroprotection must consider perfusion, the necessary substrate upon which ischemic brain survival depends. We survey the major mechanistic approaches to neuroprotection and review clinical neuroprotection trials, focusing on those phase 3 multicenter clinical trials for acute ischemic stroke that have been completed or terminated. We review the evolution of thrombolytic therapies; consider the lessons learned from the initial, negative multicenter trials of endovascular therapy; and emphasize the highly successful positive trials that have finally established a clinical role for endovascular clot removal. As these studies point to the brain's collateral circulation as key to successful reperfusion, we next review the anatomy and pathophysiology of collateral perfusion as it relates to ischemic infarction, as well as the molecular and genetic influences on collateral development. We discuss the current MR and CT-based diagnostic methods for assessing the collateral circulation and the prognostic significance of collaterals in ischemic stroke, and we consider past and possible future therapeutic directions.

Imaging acute ischemic stroke

Acute ischemic stroke is common and often treatable, but treatment requires reliable information on the state of the brain that may be provided by modern neuroimaging. Critical information includes: the presence of hemorrhage; the site of arterial occlusion; the size of the early infarct "core"; and the size of underperfused, potentially threatened brain parenchyma, commonly referred to as the "penumbra." In this chapter we review the major determinants of outcomes in ischemic stroke patients, and the clinical value of various advanced computed tomography and magnetic resonance imaging methods that may provide key physiologic information in these patients. The focus is on major strokes due to occlusions of large arteries of the anterior circulation, the most common cause of a severe stroke syndrome. The current evidence-based approach to imaging the acute stroke patient at the Massachusetts General Hospital is presented, which is applicable for all stroke types. We conclude with new information on time and stroke evolution that imaging has revealed, and how it may open the possibilities of treating many more patients.

Arterial input function and gray matter cerebral blood volume measurements in children

Purpose

To investigate how arterial input functions (AIFs) vary with age in children and compare the use of individual and population AIFs for calculating gray matter CBV values. Quantitative measures of cerebral blood volume (CBV) using dynamic susceptibility contrast (DSC) magnetic resonance imaging (MRI) require measurement of an AIF. AIFs are affected by numerous factors including patient age. Few data presenting AIFs in the pediatric population exists.

Materials and Methods

Twenty‐two previously treated pediatric brain tumor patients (mean age, 6.3 years; range, 2.0–15.3 years) underwent DSC‐MRI scans on a 3T MRI scanner over 36 visits. AIFs were measured in the middle cerebral artery. A functional form of an adult population AIF was fitted to each AIF to obtain parameters reflecting AIF shape. The relationship between parameters and age was assessed. Correlations between gray matter CBV values calculated using the resulting population and individual patient AIFs were explored.

Results

There was a large variation in individual patient AIFs but correlations between AIF shape and age were observed. The center (r = 0.596, P < 0.001) and width of the first‐pass peak (r = 0.441, P = 0.007) were found to correlate significantly with age. Intrapatient coefficients of variation were significantly lower than interpatient values for all parameters (P < 0.001). Differences in CBV values calculated with an overall population and age‐specific population AIF compared to those calculated with individual AIFs were 31.3% and 31.0%, respectively.

Conclusion

Parameters describing AIF shape correlate with patient age in line with expected changes in cardiac output. In pediatric DSC‐MRI studies individual patient AIFs are recommended. J. Magn. Reson. Imaging 2016;43:981–989

Analysis of perfusion MRI in stroke: To deconvolve, or not to deconvolve

PURPOSE

There is currently controversy regarding the benefits of deconvolution-based parameters in stroke imaging, with studies suggesting a similar infarct prediction using summary parameters. We investigate here the performance of deconvolution-based parameters and summary parameters for dynamic-susceptibility contrast (DSC) MRI analysis, with particular emphasis on precision.

METHODS

Numerical simulations were used to assess the contribution of noise and arterial input function (AIF) variability to measurement precision. A realistic AIF range was defined based on in vivo data from an acute stroke clinical study. The simulated tissue curves were analyzed using two popular singular value decomposition (SVD) based algorithms, as well as using summary parameters.

RESULTS

SVD-based deconvolution methods were found to considerably reduce the AIF-dependency, but a residual AIF bias remained on the calculated parameters. Summary parameters, in turn, show a lower sensitivity to noise. The residual AIF-dependency for deconvolution methods and the large AIF-sensitivity of summary parameters was greatly reduced when normalizing them relative to normal tissue.

CONCLUSION

Consistent with recent studies suggesting high performance of summary parameters in infarct prediction, our results suggest that DSC-MRI analysis using properly normalized summary parameters may have advantages in terms of lower noise and AIF-sensitivity as compared to commonly used deconvolution methods. Magn Reson Med 76:1282-1290, 2016. © 2015 Wiley Periodicals, Inc.

Infarct Volume Prediction by Early Magnetic Resonance Imaging in a Murine Stroke Model Depends on Ischemia Duration and Time of Imaging

BACKGROUND AND PURPOSE

Despite standardization of experimental stroke models, final infarct sizes after middle cerebral artery occlusion (MCAO) vary considerably. This introduces uncertainties in the evaluation of drug effects on stroke. Magnetic resonance imaging may detect variability of surgically induced ischemia before treatment and thus improve treatment effect evaluation.

METHODS

MCAO of 45 and 90 minutes induced brain infarcts in 83 mice. During, and 3 and 6 hours after MCAO, we performed multiparametric magnetic resonance imaging. We evaluated time courses of cerebral blood flow, apparent diffusion coefficient (ADC), T1, T2, accuracy of infarct prediction strategies, and impact on statistical evaluation of experimental stroke studies.

RESULTS

ADC decreased during MCAO but recovered completely on reperfusion after 45 and partially after 90-minute MCAO, followed by a secondary decline. ADC lesion volumes during MCAO or at 6 hours after MCAO largely determined final infarct volumes for 90 but not for 45 minutes MCAO. The majority of chance findings of final infarct volume differences in random group allocations of animals were associated with significant differences in early ADC lesion volumes for 90, but not for 45-minute MCAO.

CONCLUSIONS

The prediction accuracy of early magnetic resonance imaging for infarct volumes depends on timing of magnetic resonance imaging and MCAO duration. Variability of the posterior communicating artery in C57Bl6 mice contributes to differences in prediction accuracy between short and long MCAO. Early ADC imaging may be used to reduce errors in the interpretation of post MCAO treatment effects on stroke volumes.

C-Arm Flat Detector CT Parenchymal Blood Volume Thresholds for Identification of Infarcted Parenchyma in the Neurointerventional Suite

BACKGROUND AND PURPOSE

C-arm flat detector CT parenchymal blood volume imaging allows functional assessment of the brain parenchyma in the neurointerventional suite. This study aimed to determine the optimal C-arm flat detector CT parenchymal blood volume thresholds for demarcating irreversibly infarcted brain parenchyma by using areas of restricted diffusion on MR imaging as a surrogate marker for infarction.

MATERIALS AND METHODS

Twenty-six patients with delayed cerebral ischemia following aneurysmal SAH underwent research C-arm CT parenchymal blood volume scans by using a biplane angiography system and contemporaneous MR imaging. Infarct and peri-infarct tissue VOIs and their homologous VOIs in the contralateral uninvolved hemisphere were delineated on the basis of the review of DWI, PWI, and ADC images. Voxel-based receiver operating characteristic curve analysis was performed to estimate the optimal absolute and normalized parenchymal blood volume values for demarcating the infarct voxels.

RESULTS

For 12 patients with areas of restricted diffusion (infarct volume, 6.38 ± 7.09 mL; peri-infarct tissue volume, 22.89 ± 21.76 mL) based on the voxel-based receiver operating characteristic curve analysis, optimal absolute and normalized parenchymal blood volume thresholds for infarction were 2.49 mL/100 g (area under curve, 0.76; sensitivity, 0.69; specificity, 0.71) and 0.67 (area under curve, 0.77; sensitivity, 0.69; specificity, 0.72), respectively (P value < .01). For the moderate-to-severely ischemic peri-infarct zone, mean parenchymal blood volume values of the involved hemisphere VOIs were lower compared with the uninvolved hemisphere VOIs (P value < .01). However, for the mild-to-moderately ischemic peri-infarct zone, there was no statistically significant difference between the mean parenchymal blood volume values of the involved and uninvolved hemisphere VOIs (P value > .05).

CONCLUSIONS

C-arm flat detector CT parenchymal blood volume maps in conjunction with optimal thresholds are sensitive and specific for the estimation of irreversibly infarcted parenchyma. Parenchymal blood volume maps allow reliable detection of moderate-to-severe ischemia; however, the potential for underestimation of mild-to-moderate ischemia exists.

Comparison of the 2 Most Popular Deconvolution Techniques for the Detection of Penumbral Flow in Acute Stroke

BACKGROUND AND PURPOSE

Dynamic susceptibility-weighted contrast-enhanced (DSC) magnetic resonance imaging (MRI) is used to identify the tissue-at-risk in acute stroke, but the choice of optimal DSC postprocessing in the clinical setting remains a matter of debate. Using 15O-water positron emission tomography (PET), we validated the performance of 2 common deconvolution methods for DSC-MRI.

METHODS

In (sub)acute stroke patients with consecutive MRI and PET imaging, DSC maps were calculated applying 2 deconvolution methods, standard and block-circulant single value decomposition. We used 2 standardized analysis methods, a region of interest-based and a voxel-based analysis, where PET cerebral blood flow masks of <20 mL/100 g per minute (penumbral flow) and gray matter masks were overlaid on DSC parameter maps. For both methods, receiver operating characteristic curve analysis was performed to identify the accuracy of each DSC-MR map for the detection of PET penumbral flow.

RESULTS

In 18 data sets (median time after stroke onset: 18 hours; median time PET to MRI: 101 minutes), block-circulant single value decomposition showed significantly better performance to detect PET penumbral flow only for mean transit time maps. Time-to-maximum (Tmax) had the highest performance independent of the deconvolution method.

CONCLUSIONS

Block-circulant single value decomposition seems only significantly beneficial for mean transit time maps in (sub)acute stroke. Tmax is likely the most stable deconvolved parameter for the detection of tissue-at-risk using DSC-MRI.

Evaluation of Early Reperfusion Criteria in Acute Ischemic Stroke

BACKGROUND AND PURPOSE

Though still debated, early reperfusion is increasingly used as a biomarker for clinical outcome. However, the lack of a standard definition hinders the assessment of reperfusion therapies and study comparisons. The objective was to determine the optimal early reperfusion criteria that predicts clinical outcome in ischemic stroke.

METHODS

Early reperfusion was assessed voxel-wise in 57 patients within 6 hours of symptom onset. The performance of the time to peak (TTP), the mean transit time (MTT), and the time to maximum of residue function (Tmax ) at various delays thresholds in predicting the neurological response (based on the National Institutes of Health Stroke Scale) and the functional outcome (modified Rankin scale ≤1) at 1 month were compared. A receiver operating characteristics (ROC) analysis determined the optimal extent of reperfusion. A novel unsupervised classification of reperfusion using group-based trajectory modeling (GBTM) was evaluated.

RESULTS

MTT had a lower performance than TTP and Tmax in predicting the neurological response (P = .008 vs. TTP and P = .006 vs. Tmax ) or the functional outcome (P = .0006 vs. TTP; P = .002 vs. Tmax ). No delay threshold had a significantly higher predictive value than another. The optimal percentage of reperfusion was dependent on the outcome scale (P < .001). The GBTM-based classification of reperfusion was closely associated with the clinical outcome and had a similar accuracy compared to ROC-based classification.

CONCLUSIONS

TTP and Tmax should be preferred to MTT in defining early reperfusion. GBTM provided a clinically relevant reperfusion classification that does not require prespecified delay thresholds or clinical outcomes.

Biased visualization of hypoperfused tissue by computed tomography due to short imaging duration: improved classification by image down-sampling and vascular models

OBJECTIVES

Lesion detection in acute stroke by computed-tomography perfusion (CTP) can be affected by incomplete bolus coverage in veins and hypoperfused tissue, so-called bolus truncation (BT), and low contrast-to-noise ratio (CNR). We examined the BT-frequency and hypothesized that image down-sampling and a vascular model (VM) for perfusion calculation would improve normo- and hypoperfused tissue classification.

METHODS

CTP datasets from 40 acute stroke patients were retrospectively analysed for BT. In 16 patients with hypoperfused tissue but no BT, repeated 2-by-2 image down-sampling and uniform filtering was performed, comparing CNR to perfusion-MRI levels and tissue classification to that of unprocessed data. By simulating reduced scan duration, the minimum scan-duration at which estimated lesion volumes came within 10% of their true volume was compared for VM and state-of-the-art algorithms.

RESULTS

BT in veins and hypoperfused tissue was observed in 9/40 (22.5%) and 17/40 patients (42.5%), respectively. Down-sampling to 128 × 128 resolution yielded CNR comparable to MR data and improved tissue classification (p = 0.0069). VM reduced minimum scan duration, providing reliable maps of cerebral blood flow and mean transit time: 5 s (p = 0.03) and 7 s (p < 0.0001), respectively).

CONCLUSIONS

BT is not uncommon in stroke CTP with 40-s scan duration. Applying image down-sampling and VM improve tissue classification.

KEY POINTS

• Too-short imaging duration is common in clinical acute stroke CTP imaging. • The consequence is impaired identification of hypoperfused tissue in acute stroke patients. • The vascular model is less sensitive than current algorithms to imaging duration. • Noise reduction by image down-sampling improves identification of hypoperfused tissue by CTP.

Low dose dynamic CT myocardial perfusion imaging using a statistical iterative reconstruction method

PURPOSE

Dynamic CT myocardial perfusion imaging has the potential to provide both functional and anatomical information regarding coronary artery stenosis. However, radiation dose can be potentially high due to repeated scanning of the same region. The purpose of this study is to investigate the use of statistical iterative reconstruction to improve parametric maps of myocardial perfusion derived from a low tube current dynamic CT acquisition.

METHODS

Four pigs underwent high (500 mA) and low (25 mA) dose dynamic CT myocardial perfusion scans with and without coronary occlusion. To delineate the affected myocardial territory, an N-13 ammonia PET perfusion scan was performed for each animal in each occlusion state. Filtered backprojection (FBP) reconstruction was first applied to all CT data sets. Then, a statistical iterative reconstruction (SIR) method was applied to data sets acquired at low dose. Image voxel noise was matched between the low dose SIR and high dose FBP reconstructions. CT perfusion maps were compared among the low dose FBP, low dose SIR and high dose FBP reconstructions. Numerical simulations of a dynamic CT scan at high and low dose (20:1 ratio) were performed to quantitatively evaluate SIR and FBP performance in terms of flow map accuracy, precision, dose efficiency, and spatial resolution.

RESULTS

Forin vivo studies, the 500 mA FBP maps gave -88.4%, -96.0%, -76.7%, and -65.8% flow change in the occluded anterior region compared to the open-coronary scans (four animals). The percent changes in the 25 mA SIR maps were in good agreement, measuring -94.7%, -81.6%, -84.0%, and -72.2%. The 25 mA FBP maps gave unreliable flow measurements due to streaks caused by photon starvation (percent changes of +137.4%, +71.0%, -11.8%, and -3.5%). Agreement between 25 mA SIR and 500 mA FBP global flow was -9.7%, 8.8%, -3.1%, and 26.4%. The average variability of flow measurements in a nonoccluded region was 16.3%, 24.1%, and 937.9% for the 500 mA FBP, 25 mA SIR, and 25 mA FBP, respectively. In numerical simulations, SIR mitigated streak artifacts in the low dose data and yielded flow maps with mean error <7% and standard deviation <9% of mean, for 30 × 30 pixel ROIs (12.9 × 12.9 mm(2)). In comparison, low dose FBP flow errors were -38% to +258%, and standard deviation was 6%-93%. Additionally, low dose SIR achieved 4.6 times improvement in flow map CNR(2) per unit input dose compared to low dose FBP.

CONCLUSIONS

SIR reconstruction can reduce image noise and mitigate streaking artifacts caused by photon starvation in dynamic CT myocardial perfusion data sets acquired at low dose (low tube current), and improve perfusion map quality in comparison to FBP reconstruction at the same dose.

Diffusion-perfusion mismatch: an opportunity for improvement in cortical function

OBJECTIVE

There has been controversy over whether diffusion-perfusion mismatch provides a biomarker for the ischemic penumbra. In the context of clinical stroke trials, regions of the diffusion-perfusion mismatch that do not progress to infarct in the absence of reperfusion are considered to represent "benign oligemia." However, at least in some cases (particularly large vessel stenosis), some of this hypoperfused tissue may remain dysfunctional for a prolonged period without progressing to infarct and may recover function if eventually reperfused. We hypothesized that patients with persistent diffusion-perfusion mismatch using a hypoperfusion threshold of 4-5.9 s delay on time-to-peak (TTP) maps at least sometimes have persistent cognitive deficits relative to those who show some reperfusion of this hypoperfused tissue.

METHODS

We tested this hypothesis in 38 patients with acute ischemic stroke who had simple cognitive tests (naming or line cancelation) and MRI with diffusion and perfusion imaging within 24 h of onset and again within 10 days, most of whom had large vessel stenosis or occlusion.

RESULTS

A persistent perfusion deficit of 4-5.9 s delay in TTP on follow up MRI was associated with a persistent cognitive deficit at that time point (p < 0.001). When we evaluated only patients who did not have infarct growth (n = 14), persistent hypoperfusion (persistent mismatch) was associated with a lack of cognitive improvement compared with those who had reperfused. The initial volume of hypoperfusion did not correlate with the later infarct volume (progression to infarct), but change in volume of hypoperfusion correlated with change in cognitive performance (p = 0.0001). Moreover, multivariable regression showed that the change in volume of hypoperfused tissue of 4-5.9 s delay (p = 0.002), and change in volume of ischemic tissue on diffusion weighted imaging (p = 0.02) were independently associated with change in cognitive function.

CONCLUSION

Our results provide additional evidence that non-infarcted tissue with a TTP delay of 4-5.9 s may be associated with persistent deficits, even if it does not always result in imminent progression to infarct. This tissue may represent the occasional opportunity to intervene to improve function even days after onset of symptoms.

ANTONIA perfusion and stroke. A software tool for the multi-purpose analysis of MR perfusion-weighted datasets and quantitative ischemic stroke assessment

OBJECTIVES

The objective of this work is to present the software tool ANTONIA, which has been developed to facilitate a quantitative analysis of perfusion-weighted MRI (PWI) datasets in general as well as the subsequent multi-parametric analysis of additional datasets for the specific purpose of acute ischemic stroke patient dataset evaluation.

METHODS

Three different methods for the analysis of DSC or DCE PWI datasets are currently implemented in ANTONIA, which can be case-specifically selected based on the study protocol. These methods comprise a curve fitting method as well as a deconvolution-based and deconvolution-free method integrating a previously defined arterial input function. The perfusion analysis is extended for the purpose of acute ischemic stroke analysis by additional methods that enable an automatic atlas-based selection of the arterial input function, an analysis of the perfusion-diffusion and DWI-FLAIR mismatch as well as segmentation-based volumetric analyses.

RESULTS

For reliability evaluation, the described software tool was used by two observers for quantitative analysis of 15 datasets from acute ischemic stroke patients to extract the acute lesion core volume, FLAIR ratio, perfusion-diffusion mismatch volume with manually as well as automatically selected arterial input functions, and follow-up lesion volume. The results of this evaluation revealed that the described software tool leads to highly reproducible results for all parameters if the automatic arterial input function selection method is used.

CONCLUSION

Due to the broad selection of processing methods that are available in the software tool, ANTONIA is especially helpful to support image-based perfusion and acute ischemic stroke research projects.

Comparison of perfusion- and diffusion-weighted imaging parameters in brain tumor studies processed using different software platforms

RATIONALE AND OBJECTIVES

To compare quantitative imaging parameter measures from diffusion- and perfusion-weighted imaging magnetic resonance imaging (MRI) sequences in subjects with brain tumors that have been processed with different software platforms.

MATERIALS AND METHODS

Scans from 20 subjects with primary brain tumors were selected from the Comprehensive Neuro-oncology Data Repository at Washington University School of Medicine (WUSM) and the Swedish Neuroscience Institute. MR images were coregistered, and each subject's data set was processed by three software packages: 1) vendor-specific scanner software, 2) research software developed at WUSM, and 3) a commercially available, Food and Drug Administration-approved, processing platform (Nordic Ice). Regions of interest (ROIs) were chosen within the brain tumor and normal nontumor tissue. The results obtained using these methods were compared.

RESULTS

For diffusion parameters, including mean diffusivity and fractional anisotropy, concordance was high when comparing different processing methods. For perfusion-imaging parameters, a significant variance in cerebral blood volume, cerebral blood flow, and mean transit time (MTT) values was seen when comparing the same raw data processed using different software platforms. Correlation was better with larger ROIs (radii ≥ 5 mm). Greatest variance was observed in MTT.

CONCLUSIONS

Diffusion parameter values were consistent across different software processing platforms. Perfusion parameter values were more variable and were influenced by the software used. Variation in the MTT was especially large suggesting that MTT estimation may be unreliable in tumor tissues using current MRI perfusion methods.