Decision Criteria for Large Vessel Occlusion Using Transcranial Doppler Waveform Morphology

A Multi-Parametric Approach for Characterising Cerebral Haemodynamics in Acute Ischaemic and Haemorrhagic Stroke

BACKGROUND AND PURPOSE

Early differentiation between acute ischaemic (AIS) and haemorrhagic stroke (ICH), based on cerebral and peripheral hemodynamic parameters, would be advantageous to allow for pre-hospital interventions. In this preliminary study, we explored the potential of multiple parameters, including dynamic cerebral autoregulation, for phenotyping and differentiating each stroke sub-type.

METHODS

Eighty patients were included with clinical stroke syndromes confirmed by computed tomography within 48 h of symptom onset. Continuous recordings of bilateral cerebral blood velocity (transcranial Doppler ultrasound), end-tidal CO2 (capnography), electrocardiogram (ECG), and arterial blood pressure (ABP, Finometer) were used to derive 67 cerebral and peripheral parameters.

RESULTS

A total of 68 patients with AIS (mean age 66.8 ± SD 12.4 years) and 12 patients with ICH (67.8 ± 16.2 years) were included. The median ± SD NIHSS of the cohort was 5 ± 4.6. Statistically significant differences between AIS and ICH were observed for (i) an autoregulation index (ARI) that was higher in the unaffected hemisphere (UH) for ICH compared to AIS (5.9 ± 1.7 vs. 4.9 ± 1.8 p = 0.07); (ii) coherence function for both hemispheres in different frequency bands (AH, p < 0.01; UH p < 0.02); (iii) a baroreceptor sensitivity (BRS) for the low-frequency (LF) bands that was higher for AIS (6.7 ± 4.2 vs. 4.10 ± 2.13 ms/mmHg, p = 0.04) compared to ICH, and that the mean gain of the BRS in the LF range was higher in the AIS than in the ICH (5.8 ± 5.3 vs. 2.7 ± 1.8 ms/mmHg, p = 0.0005); (iv) Systolic and diastolic velocities of the affected hemisphere (AH) that were significantly higher in ICH than in AIS (82.5 ± 28.09 vs. 61.9 ± 18.9 cm/s), systolic velocity (p = 0.002), and diastolic velocity (p = 0.05).

CONCLUSION

Further multivariate modelling might improve the ability of multiple parameters to discriminate between AIS and ICH and warrants future prospective studies of ultra-early classification (<4 h post symptom onset) of stroke sub-types.

Cost-benefit analysis of intraoperative neuromonitoring for cardiac surgery

BACKGROUND

Intraoperative neuromonitoring (IONM) can detect large vessel occlusion (LVO) in real-time during surgery. The aim of this study was to conduct a cost-benefit analysis of utilizing IONM among patients undergoing cardiac surgery.

METHODS

A decision-analysis tree with terminal Markov nodes was constructed to model functional outcome, as measured via the modified Rankin Scale (mRS), among 65-year-old patients undergoing cardiac surgery. Our cost-benefit analysis compares the use of IONM (electroencephalography and somatosensory evoked potential) against no IONM in preventing neurological complications from perioperative LVO during cardiac surgery. The study was performed over a lifetime horizon from a societal perspective in the United States. Base case and one-way probabilistic sensitivity analyses were performed.

RESULTS

At a baseline LVO rate of 0.31%, the mean attributable lifetime expenditure for IONM-monitored cardiac surgeries relative to unmonitored cardiac surgeries was $1047.41 (95% CI, $742.12 - $1445.10). At a critical LVO rate of approximately 3.67%, the costs of both monitored and unmonitored cardiac surgeries were the same. Above this critical rate, implementing IONM became cost-saving. On one-way sensitivity analysis, variation in LVO rate from 0% - 10% caused lifetime costs attributable to receiving IONM to range from $1150.47 - $29404.61; variations in IONM cost, percentage of intervenable LVOs, IONM sensitivity, and mechanical thrombectomy cost exerted comparably minimal influence over lifetime costs.

DISCUSSION

We find considerable cost savings favoring the use of IONM under certain parameters corresponding to high-risk patients. This study will provide financial perspective to policymakers, clinicians, and patients alike on the appropriate use of IONM during cardiac surgery.

Clinical relevance of transcranial Doppler in a cardiac surgery setting: embolic load predicts difficult separation from cardiopulmonary bypass

BACKGROUND

During cardiac surgery, transcranial Doppler (TCD) represents a non-invasive modality that allows measurement of red blood cell flow velocities in the cerebral arteries. TCD can also be used to detect and monitor embolic material in the cerebral circulation. Detection of microemboli is reported as a high intensity transient signal (HITS). The importance of cerebral microemboli during cardiac surgery has been linked to the increased incidence of postoperative renal failure, right ventricular dysfunction, and hemodynamic instability. The objective of this study is to determine whether the embolic load is associated with hemodynamic instability during cardiopulmonary bypass (CPB) separation and postoperative complications.

METHODS

A retrospective single-centre cohort study of 354 patients undergoing cardiac surgery between December 2015 and March 2020 was conducted. Patients were divided in tertiles, where 117 patients had a low quantity of embolic material (LEM), 119 patients have a medium quantity of microemboli (MEM) and 118 patients who have a high quantity of embolic material (HEM). The primary endpoint was a difficult CPB separation. Multivariate logistic regression was used to determine the potential association between a difficult CPB separation and the number of embolic materials.

RESULTS

Patients who had a difficult CPB separation had more HITS compared to patients who had a successful CPB separation (p < 0.001). In the multivariate analysis, patients with MEM decreased their odds of having a difficult CPB weaning compared to patients in the HEM group (OR = 0.253, CI 0.111-0.593; p = 0.001). In the postoperative period patients in the HEM group have a higher Time of Persistent Organ Dysfunction (TPOD), a longer stay in the ICU, a longer duration under vasopressor drugs and a higher mortality rate compared to those in the MEM and LEM groups.

CONCLUSION

The result of this study suggests that a high quantity of cerebral embolic material increases the odds of having a difficult CPB separation. Also, it seems to be associated to more complex surgery, a longer CPB time, a higher TPOD and a longer stay in the ICU. Six out of eight patients who died in this cohort were in the HEM group.

A data-driven computational methodology towards a pre-hospital Acute Ischaemic Stroke screening tool using haemodynamics waveforms

BACKGROUND AND OBJECTIVE

Acute Ischaemic Stroke (AIS), a significant global health concern, results from occlusions in cerebral arteries, causing irreversible brain damage. Different type of treatments exist depending on the size and location of the occlusion. Challenges persist in achieving faster diagnosis and treatment, which needs to happen in the first hours after the onset of symptoms to maximize the chances of patient recovery. The current diagnostic pipeline, i.e. "drip and ship", involves diagnostic via advanced imaging tools, only available in large clinical facilities, which poses important delays. This study investigates the feasibility of developing a machine learning model to diagnose and locate occluding blood clots from velocity waveforms, which can be easily be obtained with portable devices such as Doppler Ultrasound. The goal is to explore this approach as a cost-effective and time-efficient alternative to advanced imaging techniques typically available only in large hospitals.

METHODS

Simulated haemodynamic data is used to conduct blood flow simulations representing healthy and different AIS scenarios using a population-based database. A Machine Learning classification model is trained to solve the inverse problem, this is, detect and locate a potentially occluding thrombus from measured waveforms. The classification process involves two steps. First, the region where the thrombus is located is classified into nine groups, including healthy, left or right large vessel occlusion, left or right anterior cerebral artery, and left or right posterior cerebral artery. In a second step, the bifurcation generation of the thrombus location is classified as small, medium, or large vessel occlusion.

RESULTS

The proposed methodology is evaluated for data without noise, achieving a true prediction rate exceeding 95% for both classification steps mentioned above. The inclusion of up to 20% noise reduces the true prediction rate to 80% for region detection and 70% for bifurcation generation detection.

CONCLUSIONS

This study demonstrates the potential effectiveness and efficiency of using haemodynamic data and machine learning to detect and locate occluding thrombi in AIS patients. Although the geometric and topological data used in this study are idealized, the results suggest that this approach could be applicable in real-world situations with appropriate adjustments. Source code is available in https://github.com/ahmetsenemse/Acute-Ischaemic-Stroke-screening-tool-.

Point-of-Care Ultrasound to Detect Acute Large Vessel Occlusions in Stroke Patients: A Proof-of-Concept Study

BACKGROUND AND PURPOSE

A primary admission of patients with suspected acute ischemic stroke and large vessel occlusion (LVO) to centers capable of providing endovascular stroke therapy (EVT) may induce shorter time to treatment and better functional outcomes. One of the limitations in this strategy is the need for accurately identifying LVO patients in the prehospital setting. We aimed to study the feasibility and diagnostic performance of point-of-care ultrasound (POCUS) for the detection of LVO in patients with acute stroke.

METHODS

We conducted a proof-of-concept study and selected 15 acute ischemic stroke patients with angiographically confirmed LVO and 15 patients without LVO. Duplex ultrasonography (DUS) of the common carotid arteries was performed, and flow profiles compatible with LVO were scored independently by one experienced and one junior neurologist.

RESULTS

Among the 15 patients with LVO, 6 patients presented with an occlusion of the carotid-T and 9 patients presented with an M1 occlusion. Interobserver agreement between the junior and the experienced neurologist was excellent (kappa = 0.813, p < 0.001). Flow profiles of the CAA allowed the detection of LVO with a sensitivity of 73%, a positive predictive value of 92 and 100%, and a c-statistics of 0.83 (95%CI = 0.65-0.94) and 0.87 (95%CI = 0.69-0.94) (experienced neurologist and junior neurologist, respectively). In comparison with clinical stroke scales, DUS was associated with better trade-off between sensitivity and specificity.

CONCLUSION

POCUS in acute stroke setting is feasible, it may serve as a complementary tool for the detection of LVO and is potentially applicable in the prehospital phase.

Prehospital technologies for early stroke detection - A review

The rate of neural circuitry loss in a typical large vessel occlusion well emphasizes that 'Time is Brain'. Every untreated minute in a large vessel ischaemic stroke results in loss of 1.9 million neurons and 13.8 billion synapses. As such, it is essential to optimize the flow-limiting steps in delivering the current standard of care. The current diagnostic model involves recognition of symptoms by patients, followed by access to Emergency Medical Services and subsequent physical examination and neuroimaging in the Emergency Department. With more than 50% of stroke patients using Emergency Medical Services as the first point of care contact, it can be deduced that the outcome of the 'stroke chain of survival' can be improved by addressing the bottleneck of prehospital stroke diagnosis. Here we present a review of the existing technologies.

Mechanisms of Transcranial Doppler Ultrasound phenotypes in paediatric cerebral malaria remain elusive

Background

Cerebral malaria (CM) results in significant paediatric death and neurodisability in sub-Saharan Africa. Several different alterations to typical Transcranial Doppler Ultrasound (TCD) flow velocities and waveforms in CM have been described, but mechanistic contributors to these abnormalities are unknown. If identified, targeted, TCD-guided adjunctive therapy in CM may improve outcomes.

Methods

This was a prospective, observational study of children 6 months to 12 years with CM in Blantyre, Malawi recruited between January 2018 and June 2021. Medical history, physical examination, laboratory analysis, electroencephalogram, and magnetic resonance imaging were undertaken on presentation. Admission TCD results determined phenotypic grouping following a priori definitions. Evaluation of the relationship between haemodynamic, metabolic, or intracranial perturbations that lead to these observed phenotypes in other diseases was undertaken. Neurological outcomes at hospital discharge were evaluated using the Paediatric Cerebral Performance Categorization (PCPC) score.

Results

One hundred seventy-four patients were enrolled. Seven (4%) had a normal TCD examination, 57 (33%) met criteria for hyperaemia, 50 (29%) for low flow, 14 (8%) for microvascular obstruction, 11 (6%) for vasospasm, and 35 (20%) for isolated posterior circulation high flow. A lower cardiac index (CI) and higher systemic vascular resistive index (SVRI) were present in those with low flow than other groups (p < 0.003), though these values are normal for age (CI 4.4 [3.7,5] l/min/m2, SVRI 1552 [1197,1961] dscm-5m2). Other parameters were largely not significantly different between phenotypes. Overall, 118 children (68%) had a good neurological outcome. Twenty-three (13%) died, and 33 (19%) had neurological deficits. Outcomes were best for participants with hyperaemia and isolated posterior high flow (PCPC 1–2 in 77 and 89% respectively). Participants with low flow had the least likelihood of a good outcome (PCPC 1–2 in 42%) (p < 0.001). Cerebral autoregulation was significantly better in children with good outcome (transient hyperemic response ratio (THRR) 1.12 [1.04,1.2]) compared to a poor outcome (THRR 1.05 [0.98,1.02], p = 0.05).

Conclusions

Common pathophysiological mechanisms leading to TCD phenotypes in non-malarial illness are not causative in children with CM. Alternative mechanistic contributors, including mechanical factors of the cerebrovasculature and biologically active regulators of vascular tone should be explored.

Portable stroke detection devices: a systematic scoping review of prehospital applications

BACKGROUND

The worldwide burden of stroke remains high, with increasing time-to-treatment correlated with worse outcomes. Yet stroke subtype determination, most importantly between stroke/non-stroke and ischemic/hemorrhagic stroke, is not confirmed until hospital CT diagnosis, resulting in suboptimal prehospital triage and delayed treatment. In this study, we survey portable, non-invasive diagnostic technologies that could streamline triage by making this initial determination of stroke type, thereby reducing time-to-treatment.

METHODS

Following PRISMA guidelines, we performed a scoping review of portable stroke diagnostic devices. The search was executed in PubMed and Scopus, and all studies testing technology for the detection of stroke or intracranial hemorrhage were eligible for inclusion. Extracted data included type of technology, location, feasibility, time to results, and diagnostic accuracy.

RESULTS

After a screening of 296 studies, 16 papers were selected for inclusion. Studied devices utilized various types of diagnostic technology, including near-infrared spectroscopy (6), ultrasound (4), electroencephalography (4), microwave technology (1), and volumetric impedance spectroscopy (1). Three devices were tested prior to hospital arrival, 6 were tested in the emergency department, and 7 were tested in unspecified hospital settings. Median measurement time was 3 minutes (IQR: 3 minutes to 5.6 minutes). Several technologies showed high diagnostic accuracy in severe stroke and intracranial hematoma detection.

CONCLUSION

Numerous emerging portable technologies have been reported to detect and stratify stroke to potentially improve prehospital triage. However, the majority of these current technologies are still in development and utilize a variety of accuracy metrics, making inter-technology comparisons difficult. Standardizing evaluation of diagnostic accuracy may be helpful in further optimizing portable stroke detection technology for clinical use.

Emerging Detection Techniques for Large Vessel Occlusion Stroke: A Scoping Review

Background: Large vessel occlusion (LVO) is the obstruction of large, proximal cerebral arteries and can account for up to 46% of acute ischaemic stroke (AIS) when both the A2 and P2 segments are included (from the anterior and posterior cerebral arteries). It is of paramount importance that LVO is promptly recognised to provide timely and effective acute stroke management. This review aims to scope recent literature to identify new emerging detection techniques for LVO. As a good comparator throughout this review, the commonly used National Institutes of Health Stroke Scale (NIHSS), at a cut-off of ≥11, has been reported to have a sensitivity of 86% and a specificity of 60% for LVO. Methods: Four electronic databases (Medline via OVID, CINAHL, Scopus, and Web of Science), and grey literature using OpenGrey, were systematically searched for published literature investigating developments in detection methods for LVO, reported from 2015 to 2021. The protocol for the search was published with the Open Science Framework (10.17605/OSF.IO/A98KN). Two independent researchers screened the titles, abstracts, and full texts of the articles, assessing their eligibility for inclusion. Results: The search identified 5,082 articles, in which 2,265 articles were screened to assess their eligibility. Sixty-two studies remained following full-text screening. LVO detection techniques were categorised into 5 groups: stroke scales (n = 30), imaging and physiological methods (n = 15), algorithmic and machine learning approaches (n = 9), physical symptoms (n = 5), and biomarkers (n = 3). Conclusions: This scoping review has explored literature on novel and advancements in pre-existing detection methods for LVO. The results of this review highlight LVO detection techniques, such as stroke scales and biomarkers, with good sensitivity and specificity performance, whilst also showing advancements to support existing LVO confirmatory methods, such as neuroimaging.

The Future of Endovascular Therapy

PURPOSE OF THE REVIEW

This article summarizes a broad range of the most recent advances and future directions in stroke diagnostics, endovascular robotics, and neuromodulation.

RECENT FINDINGS

In the past 5 years, the field of interventional neurology has seen major technological advances for the diagnosis and treatment of cerebrovascular diseases. Several new technologies became available to aid in complex prehospital stroke triage, stroke diagnosis, and interpretation of radiologic findings. Robotics and neuromodulation promise to expand access to established treatments and broaden neuroendovascular indications.

SUMMARY

Mobile applications offer a solution to simplify prehospital diagnostic and transfer decisions. Several prehospital devices are also under development to improve the accuracy of detection of large vessel occlusion (LVO). Artificial intelligence is now routinely used in early diagnosis of LVO and for detecting salvageability of the affected brain parenchyma. Technological advances have also paved the way to incorporate endovascular robotics and neuromodulation into practice. This may expand the deliverability of established treatments and facilitate the development of cutting-edge treatments for other complex neurologic diseases.

Comparison of transcranial doppler ultrasound indices in large and small vessel disease cerebral infarction

Background: Atherosclerotic involvement of large and small cerebral arteries leading to infarction is among the most prevalent subtypes of stroke worldwide. The hemodynamic changes due to these arterial pathologies can be studied non-invasively and in real-time by using transcranial Doppler (TCD) techniques. TCD indices of the studied arteries may guide the clinician in differentiating these two underlying arterial pathologies. Methods: A cross-sectional study of patients with small and large vessel types of cerebral infraction based on the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) stroke classification was undertaken in the inpatient population of neurology service of Razi Hospital, Tabriz, Iran, from October 2018 to October 2019. After clinical diagnosis, all cases underwent TCD studies, brain magnetic resonance imaging (MRI), and brain and cervical four-vessel magnetic resonance angiography (MRA). The results of TCD indices related to major arteries of the circle of Willis were tabulated and compared between large and small vessel subtypes of cerebral infarction. Results: A statistically significant difference between right middle cerebral artery (MCA) pulsatility index (PI), left MCA PI, right internal carotid artery (ICA) PI, end-diastolic velocity (EDV), left ICA PI, left ICA EDV, left anterior cerebral artery (ACA) PI, and right vertebral artery (VA) PI measures of the two groups was seen (P < 0.05). In comparison to the large vessel group, left ACA, right VA, and bilateral MCAs and ICAs in the small-vessel stroke group demonstrated an elevated PI. Conclusion: A significant increase of PI occurs in the majority of intracranial arteries of patients with small vessel stroke. This makes PI a valuable marker for differentiating strokes with different underlying pathophysiologies.

Large Vessel Occlusion Stroke Detection in the Prehospital Environment

Purpose of Review

Endovascular therapy for acute ischemic stroke secondary to large vessel occlusion (LVO) is time-dependent. Prehospital patients with suspected LVO stroke should be triaged directly to specialized stroke centers for endovascular therapy. This review describes advances in LVO detection among prehospital suspected stroke patients.

Recent Findings

Clinical prehospital stroke severity tools have been validated in the prehospital setting. Devices including EEG, SSEPs, TCD, cranial accelerometry, and volumetric impedance phase-shift-spectroscopy have recently published data regarding LVO detection in hospital settings. Mobile stroke units bring thrombolysis and vessel imaging to patients.

Summary

The use of a prehospital stroke severity tool for LVO triage is now widely supported. Ease of use should be prioritized as there are no meaningful differences in diagnostic performance amongst tools. LVO diagnostic devices are promising, but none have been validated in the prehospital setting. Mobile stroke units improve patient outcomes and cost-effectiveness analyses are underway.

A Review of the use of Transcranial Doppler Waveform Morphology for Acute Stroke Assessment

Acute ischemic stroke is a source of long-term disability in the United States, of which a large portion of cases are a result of large vessel occlusion (LVO). LVO strokes have high rates of morbidity and mortality due to difficulty of treatments in achieving recanalization. Recently, however, results of randomized clinical trials have shown that treatment options are expanding in both availability and efficacy. As these methods of intervention become more optimal, so must the preceding methods of assessment. Transcranial Doppler (TCD) ultrasound is a non-invasive method of evaluating cerebral hemodynamics, and has a long history in stroke assessment. Despite the importance of information provided by a TCD exam, its utilization in the acute stroke workflow has remained low because of its dependence on expert analysis. Here, we review the evolution of morphological analysis of TCD waveforms for the indication, localization, and monitoring of acute LVO.

Pathway Design for Acute Stroke Care in the Era of Endovascular Thrombectomy: A Critical Overview of Optimization Efforts

The efficacy of intravenous thrombolysis and endovascular thrombectomy (EVT) for acute ischemic stroke is highly time dependent. Optimal organization of acute stroke care is therefore important to reduce treatment delays but has become more complex after the introduction of EVT as regular treatment for large vessel occlusions. There is no singular optimal organizational model that can be generalized to different geographic regions worldwide. Current dominant organizational models for EVT include the drip-and-ship- and mothership model. Guidelines recommend routing of suspected patients with stroke to the nearest intravenous thrombolysis capable facility; however, the choice of routing to a certain model should depend on regional stroke service organization and individual patient characteristics. In general, design approaches for organizing stroke care are required, in which 2 key strategies could be considered. The first entails the identification of interventions within existing organizational models for optimizing timely delivery of intravenous thrombolysis and/or EVT. This includes adaptive patient routing toward a comprehensive stroke center, which focuses particularly on prehospital triage tools; bringing intravenous thrombolysis or EVT to the location of the patient; and expediting services and processes along the stroke pathway. The second strategy is to develop analytical or simulation model-based approaches enabling the design and evaluation of organizational models before their implementation. Organizational models for acute stroke care need to take regional and patient characteristics into account and can most efficiently be assessed and optimized through the application of model-based approaches.

Transcranial Doppler ultrasonography in neurological surgery and neurocritical care

Transcranial Doppler (TCD) ultrasonography is an inexpensive, noninvasive means of measuring blood flow within the arteries of the brain. In this review, the authors outline the technology underlying TCD ultrasonography and describe its uses in patients with neurosurgical diseases. One of the most common uses of TCD ultrasonography is monitoring for vasospasm following subarachnoid hemorrhage. In this setting, elevated blood flow velocities serve as a proxy for vasospasm and can herald the onset of ischemia. TCD ultrasonography is also useful in the evaluation and management of occlusive cerebrovascular disease. Monitoring for microembolic signals enables stratification of stroke risk due to carotid stenosis and can also be used to clarify stroke etiology. TCD ultrasonography can identify patients with exhausted cerebrovascular reserve, and after extracranial-intracranial bypass procedures it can be used to assess adequacy of flow through the graft. Finally, assessment of cerebral autoregulation can be performed using TCD ultrasonography, providing data important to the management of patients with severe traumatic brain injury. As the clinical applications of TCD ultrasonography have expanded over time, so has their importance in the management of neurosurgical patients. Familiarity with this diagnostic tool is crucial for the modern neurological surgeon.

A scoping review of pre-hospital technology to assist ambulance personnel with patient diagnosis or stratification during the emergency assessment of suspected stroke

BACKGROUND

Pre-hospital identification of key subgroups within the suspected stroke population could reduce delays to emergency treatment. We aimed to identify and describe technology with existing proof of concept for diagnosis or stratification of patients in the pre-hospital setting.

METHODS

A systematic electronic search of published literature (from 01/01/2000 to 06/06/2019) was conducted in five bibliographic databases. Two reviewers independently assessed eligibility of studies or study protocols describing diagnostic/stratification tests (portable imaging/biomarkers) or technology facilitating diagnosis/stratification (telemedicine) used by ambulance personnel during the assessment of suspected stroke. Eligible descriptions required use of tests or technology during the actual assessment of suspected stroke to provide information directly to ambulance personnel in the pre-hospital setting. Due to study, intervention and setting heterogeneity there was no attempt at meta-analysis.

RESULTS

2887 articles were screened for eligibility, 19 of which were retained. Blood biomarker studies (n = 2) were protocols of prospective diagnostic accuracy studies, one examining purines and the other a panel of known and novel biomarkers for identifying stroke sub-types (versus mimic). No data were yet available on diagnostic accuracy or patient health outcomes. Portable imaging studies (n = 2) reported that an infrared screening device for detecting haemorrhages yielded moderate sensitivity and poor specificity in a small study, whilst a dry-EEG study to detect large vessel occlusion in ischaemic stroke has not yet reported results. Fifteen evaluations of pre-hospital telemedicine were identified (12 observational and 3 controlled comparisons) which all involved transmission of stroke assessment data from the pre-hospital setting to the hospital. Diagnosis was generally comparable with hospital diagnosis and most telemedicine systems reduced time-to-treatment; however, it is unknown whether this time saving translated into more favourable clinical outcomes. Telemedicine systems were deemed acceptable by clinicians.

CONCLUSIONS

Pre-hospital technologies to identify clinically important subgroups amongst the suspected stroke population are in development but insufficient evidence precludes recommendations about routine use in the pre-hospital setting. Multi-centre diagnostic accuracy studies and clinical utility trials combining promising technologies are warranted.

Optimizing Stroke Care for Patients with Large Vessel Occlusions: Current State of the Art and Future Directions

Acute ischemic stroke (AIS) is a severely disabling disease. Endovascular therapy is a powerful and highly effective treatment option for these patients and has recently become standard of care. The benefits of endovascular treatment (EVT) are tremendous both from a patient and from an economic perspective, since it dramatically improves individual patient outcomes while reducing long-term healthcare costs at the same time. The effect of EVT is highly time-dependent. Thus, the overarching goal in AIS is to quickly transport and diagnose the patient to minimize treatment delays. In this review, we provide an overview about the current state of stroke care, propose a fast and simplified imaging protocol and management approach for AIS patients. We also highlight the challenges we are currently facing in endovascular stroke treatment and suggest possible solutions to overcome these.

Toward automated classification of pathological transcranial Doppler waveform morphology via spectral clustering

Cerebral Blood Flow Velocity waveforms acquired via Transcranial Doppler (TCD) can provide evidence for cerebrovascular occlusion and stenosis. Thrombolysis in Brain Ischemia (TIBI) flow grades are widely used for this purpose, but require subjective assessment by expert evaluators to be reliable. In this work we seek to determine whether TCD morphology can be objectively assessed using an unsupervised machine learning approach to waveform categorization. TCD beat waveforms were recorded at multiple depths from the Middle Cerebral Arteries of 106 subjects; 33 with Large Vessel Occlusion (LVO). From each waveform, three morphological features were extracted, quantifying onset of maximal velocity, systolic canopy length, and the number/prominence of peaks/troughs. Spectral clustering identified groups implicit in the resultant three-dimensional feature space, with gap statistic criteria establishing the optimal cluster number. We found that gap statistic disparity was maximized at four clusters, referred to as flow types I, II, III, and IV. Types I and II were primarily composed of control subject waveforms, whereas types III and IV derived mainly from LVO patients. Cluster morphologies for types I and IV aligned clearly with Normal and Blunted TIBI flows, respectively. Types II and III represented commonly observed flow-types not delineated by TIBI, which nonetheless deviate from normal and blunted flows. We conclude that important morphological variability exists beyond that currently quantified by TIBI in populations experiencing or at-risk for acute ischemic stroke, and posit that the observed flow-types provide the foundation for objective methods of real-time automated flow type classification.

Decision Analysis Model for Prehospital Triage of Patients With Acute Stroke

Background and Purpose- We used a decision analysis approach to analyze triage strategies for patients with acute stroke symptoms while accounting for prehospital large vessel occlusion (LVO) screening methods and key time metrics. Methods- Our decision analysis compared anticipated functional outcomes for patients within the IV-tPA (intravenous tissue-type plasminogen activator) treatment window in the mothership and drip-and-ship frameworks. Key branches of the model included IV-tPA eligibility, presence of an LVO, and endovascular therapy eligibility. Our decision analysis evaluated 2 prehospital LVO screening approaches: (1) no formal screening and (2) the use of clinical LVO screening scales. An excellent outcome was defined as modified Rankin Scale scores 0-1. Probabilities and workflow times were guideline-based or imputed from published studies. In sensitivity analyses, we individually and jointly varied transport time to the nearest primary stroke center, additional time required to transport directly to a comprehensive stroke center, and LVO screening scale predictive probabilities. We evaluated 2 separate scenarios: one in which ideal time metrics were achieved and one under current real-world metrics. Results- In the ideal metrics scenario, the drip-and-ship strategy was almost always favored in the absence of formal LVO screening. For patients screened positive for an LVO, mothership was favored if the additional transport time to the comprehensive stroke center was <3 to 23 minutes. Under real-world conditions, in which primary stroke center workflow is slower than ideal, the mothership strategy was favored in more scenarios, regardless of formal LVO screening. For example, mothership was favored with an additional transport time to the comprehensive stroke center of <32 to 99 minutes for patients screened positive for an LVO and <28 to 39 minutes in the absence of screening. Conclusions- Joint consideration of LVO probability, screening, workflow times, and transport times may improve prehospital stroke triage. Drip-and-ship was more favorable when more ideal primary stroke center workflow times were modeled.

Integrating Artificial and Human Intelligence: A Partnership for Responsible Innovation in Biomedical Engineering and Medicine

Historically, the term "artificial intelligence" dates to 1956 when it was first used in a conference at Dartmouth College in the US. Since then, the development of artificial intelligence has in part been shaped by the field of neuroscience. By understanding the human brain, scientists have attempted to build new intelligent machines capable of performing complex tasks akin to humans. Indeed, future research into artificial intelligence will continue to benefit from the study of the human brain. While the development of artificial intelligence algorithms has been fast paced, the actual use of most artificial intelligence (AI) algorithms in biomedical engineering and clinical practice is still markedly below its conceivably broader potentials. This is partly because for any algorithm to be incorporated into existing workflows it has to stand the test of scientific validation, clinical and personal utility, application context, and is equitable as well. In this context, there is much to be gained by combining AI and human intelligence (HI). Harnessing Big Data, computing power and storage capacities, and addressing societal issues emergent from algorithm applications, demand deploying HI in tandem with AI. Very few countries, even economically developed states, lack adequate and critical governance frames to best understand and steer the AI innovation trajectories in health care. Drug discovery and translational pharmaceutical research stand to gain from AI technology provided they are also informed by HI. In this expert review, we analyze the ways in which AI applications are likely to traverse the continuum of life from birth to death, and encompassing not only humans but also all animal, plant, and other living organisms that are increasingly touched by AI. Examples of AI applications include digital health, diagnosis of diseases in newborns, remote monitoring of health by smart devices, real-time Big Data analytics for prompt diagnosis of heart attacks, and facial analysis software with consequences on civil liberties. While we underscore the need for integration of AI and HI, we note that AI technology does not have to replace medical specialists or scientists and rather, is in need of such expert HI. Altogether, AI and HI offer synergy for responsible innovation and veritable prospects for improving health care from prevention to diagnosis to therapeutics while unintended consequences of automation emergent from AI and algorithms should be borne in mind on scientific cultures, work force, and society at large.

Objective Assessment of Beat Quality in Transcranial Doppler Measurement of Blood Flow Velocity in Cerebral Arteries

OBJECTIVE

Transcranial Doppler (TCD) ultrasonography measures pulsatile cerebral blood flow velocity in the arteries and veins of the head and neck. Similar to other real-time measurement modalities, especially in healthcare, the identification of high-quality signals is essential for clinical interpretation. Our goal is to identify poor quality beats and remove them prior to further analysis of the TCD signal.

METHODS

We selected objective features for this purpose including Euclidean distance between individual and average beat waveforms, cross-correlation between individual and average beat waveforms, ratio of the high-frequency power to the total beat power, beat length, and variance of the diastolic portion of the beat waveform. We developed an iterative outlier detection algorithm to identify and remove the beats that are different from others in a recording. Finally, we tested the algorithm on a dataset consisting of more than 15 h of TCD data recorded from 48 stroke and 34 in-hospital control subjects.

RESULTS

We assessed the performance of the algorithm in the improvement of estimation of clinically important TCD parameters by comparing them to that of manual beat annotation. The results show that there is a strong correlation between the two, that demonstrates the algorithm has successfully recovered the clinically important features. We obtained significant improvement in estimating the TCD parameters using the algorithm accepted beats compared to using all beats.

SIGNIFICANCE

Our algorithm provides a valuable tool to clinicians for automated detection of the reliable portion of the data. Moreover, it can be used as a pre-processing tool to improve the data quality for automated diagnosis of pathologic beat waveforms using machine learning.