Validation of a machine learning software tool for automated large vessel occlusion detection in patients with suspected acute stroke.

PURPOSE: CT angiography (CTA) is the imaging standard for large vessel occlusion (LVO) detection in patients with acute ischemic stroke. StrokeSENS LVO is an automated tool that utilizes a machine learning algorithm to identify anterior large vessel occlusions (LVO) on CTA. The aim of this study was to test the algorithm's performance in LVO detection in an independent dataset. METHODS: A total of 400 studies (217 LVO, 183 other/no occlusion) read by expert consensus were used for retrospective analysis. The LVO was defined as intracranial internal carotid artery (ICA) occlusion and M1 middle cerebral artery (MCA) occlusion. Software performance in detecting anterior LVO was evaluated using receiver operator characteristics (ROC) analysis, reporting area under the curve (AUC), sensitivity, and specificity. Subgroup analyses were performed to evaluate if performance in detecting LVO differed by subgroups, namely M1 MCA and ICA occlusion sites, and in data stratified by patient age, sex, and CTA acquisition characteristics (slice thickness, kilovoltage tube peak, and scanner manufacturer). RESULTS: AUC, sensitivity, and specificity overall were as follows: 0.939, 0.894, and 0.874, respectively, in the full cohort; 0.927, 0.857, and 0.874, respectively, in the ICA occlusion cohort; 0.945, 0.914, and 0.874, respectively, in the M1 MCA occlusion cohort. Performance did not differ significantly by patient age, sex, or CTA acquisition characteristics. CONCLUSION: The StrokeSENS LVO machine learning algorithm detects anterior LVO with high accuracy from a range of scans in a large dataset.

Near-Infrared Spectroscopy Reveals Brain Hypoxia and Cerebrovascular Dysregulation in Primary Biliary Cholangitis.

BACKGROUND AND AIMS: Primary biliary cholangitis (PBC) is an autoimmune cholestatic liver disease linked to symptoms including fatigue and altered mood/cognition, indicating that chronic liver inflammation associated with PBC can impact brain function. We employed near-infrared spectroscopy (NIRS), a noninvasive neuroimaging technique, to determine whether patients with PBC exhibit reduced cerebral oxygen saturation (StO2 ) and altered patterns of microvascular cerebral blood perfusion and whether these alterations were associated with clinical phenotype. This observational case-control study was conducted at a tertiary hospital clinic (University of Calgary Liver Unit). APPROACH AND RESULTS: Thirteen female patients with noncirrhotic PBC, seven female patients with cirrhotic PBC, and 11 healthy female controls were recruited by physician referral and word of mouth, respectively. NIRS was used to measure cerebral hemoglobin and oxygen saturation. A wavelet phase coherence method was used to estimate the coherent frequency coupling of temporal changes in cerebral hemodynamics. The PBC group demonstrated significantly reduced cerebral StO2 (P = 0.01, d = 0.84), indicating cerebral hypoxia, significantly increased cerebral deoxygenated hemoglobin concentration (P < 0.01, d = 0.86), and significantly reduced hemodynamic coherence in the low-frequency band (0.08-0.15 Hz) for oxygenated hemoglobin concentration (P = 0.02, d = 0.99) and total hemoglobin (tHb) concentration (P = 0.02, d = 0.50), indicating alterations in cerebrovascular activity. Complete biochemical response to ursodeoxycholic acid (UDCA) therapy in early patients with PBC was associated with increased cerebral tHb concentration and decreased hemodynamic coherence. CONCLUSIONS: Using NIRS, patients with PBC were found to have hypoxia, increased cerebral hemoglobin concentration, and altered cerebrovascular activity, which were reversed in part in UDCA responders. In addition, symptoms and quality-of-life measures did not correlate with brain hypoxia or cerebrovascular dysregulation in patients with PBC.

Cerebral Hemodynamics and Microvasculature Changes in Relation to White Matter Microstructure After Pediatric Mild Traumatic Brain Injury: An A-CAP Pilot Study.

Advanced neuroimaging techniques show promise as a biomarker for mild traumatic brain injury (mTBI). However, little research has evaluated cerebral hemodynamics or its relation to white matter microstructure post-mTBI in children. This novel pilot study examined differences in cerebral hemodynamics, as measured using functional near-infrared spectroscopy (fNIRS), and its association with diffusion tensor imaging (DTI) metrics in children with mTBI or mild orthopedic injury (OI) to address these gaps. Children 8.00-16.99 years of age with mTBI (n = 9) or OI (n = 6) were recruited in a pediatric emergency department, where acute injury characteristics were assessed. Participants completed DTI twice, post-acutely (2-33 days) and chronically (3 or 6 months), and fNIRS ∼1 month post-injury. Automated deterministic tractography was used to compute DTI metrics. There was reduced absolute phase globally and coherence in the dorsolateral pre-frontal cortex (DLPFC) after mTBI compared to the OI group. Coherence in the DLPFC and absolute phase globally showed distinct associations with fractional anisotropy in interhemispheric white matter pathways. Two fNIRS metrics (coherence and absolute phase) differentiated mTBI from OI in children. Variability in cerebral hemodynamics related to white matter microstructure. The results provide initial evidence that fNIRS may have utility as a clinical biomarker of pediatric mTBI.

Evaluating nnU-Net for early ischemic change segmentation on non-contrast computed tomography in patients with Acute Ischemic Stroke.

Identifying the presence and extent of early ischemic changes (EIC) on Non-Contrast Computed Tomography (NCCT) is key to diagnosing and making time-sensitive treatment decisions in patients that present with Acute Ischemic Stroke (AIS). Segmenting EIC on NCCT is however a challenging task. In this study, we investigated a 3D CNN based on nnU-Net, a self-adapting CNN technique that has become the state-of-the-art in medical image segmentation, for segmenting EIC in NCCT of AIS patients. We trained and tested this model on a sizeable and heterogenous dataset of 534 patients, split into 438 for training and validation and 96 for testing. On this test set, we additionally assessed the inter-rater performance by comparing the proposed approach against two reference segmentation annotations by expert neuroradiologist readers, using this as the benchmark against which to compare our model. In terms of spatial agreement, we report median Dice Similarity Coefficients (DSCs) of 39.8% for the model vs. Reader-1, 39.4% for the model vs. Reader-2, and 55.6% for Reader-2 vs. Reader-1. In terms of lesion volume agreement, we report Intraclass Correlation Coefficients (ICCs) of 83.4% for model vs. Reader-1, 80.4% for model vs. Reader-2, and 94.8% for Reader-2 vs. Reader-1. Based on these results, we conclude that our model performs well relative to expert human performance and therefore may be useful as a decision-aid for clinicians.

Using Functional Near-Infrared Spectroscopy to Study the Effect of Repetitive Transcranial Magnetic Stimulation in Concussion: A Two-Patient Case Study.

Background: Approximately 25% of concussion patients experience persistent post-concussion symptoms (PPCS). Repetitive transcranial magnetic stimulation (rTMS) has been explored as a treatment, and functional near-infrared spectroscopy (fNIRS) may be a cost-effective method for assessing response. Objectives: Evaluate rTMS for the treatment of PPCS and introduce fNIRS as a method of assessing treatment response. Methods: Design: Two-patient case study. Setting: Calgary Brain Injury Program. Participants: 47 and 49 years. male, with PPCS for 1-2 years (headache, cognitive difficulties, nausea, visual difficulties, irritability, anxiety, poor mood, sleep, and fatigue). Intervention: 10 sessions of rTMS therapy to the left dorsolateral prefrontal cortex (DLPFC), at 10 Hz (600 pulses) and 70% of resting motor threshold amplitude. Participants completed an 8-week headache diary and a battery of clinical questionnaires prior to each fNIRS session. fNIRS: Hemodynamic changes were recorded over the frontoparietal cortex during rest, finger tapping, and a graded working memory test. fNIRS was completed pre-rTMS, following rTMS (day 14), and at 1-month post-rTMS (day 45). For comparison, two healthy, sex-matched controls were scanned with fNIRS once daily for five consecutive days. Results: Clinical scores improved (headache severity, MoCA, HIT-6, PHQ-9, GAD-7, QOLIBRI, RPSQ, BCPSI) or remained stable (PCL-5, headache frequency) post-rTMS, for both participants. Participant 1 reported moderate symptom burden, and a fNIRS task-evoked hemodynamic response showing increased oxyhemoglobin was observed following a working memory task, as expected. Participant 2 exhibited a high symptom burden pre-treatment, with abnormal fNIRS hemodynamic response where oxyhemoglobin declined, in response to task. One month following rTMS treatment, participant 2 had a normal fNIRS hemodynamic response to task, corresponding to significant improvements in clinical outcomes. Conclusion: This case study suggests fNIRS may be sensitive to physiological changes that accompany rTMS treatment. Further studies exploring fNIRS as a cost-effective technology for monitoring rTMS response in patients with PPCS are suggested.

Automated Processing of fNIRS Data-A Visual Guide to the Pitfalls and Consequences.

With the rapid increase in new fNIRS users employing commercial software, there is a concern that many studies are biased by suboptimal processing methods. The purpose of this study is to provide a visual reference showing the effects of different processing methods, to help inform researchers in setting up and evaluating a processing pipeline. We show the significant impact of pre- and post-processing choices and stress again how important it is to combine data from both hemoglobin species in order to make accurate inferences about the activation site.

Reduced Functional Connectivity in Adults with Persistent Post-Concussion Symptoms: A Functional Near-Infrared Spectroscopy Study.

Concussion, or mild traumatic brain injury (mTBI), accounts for ∼80% of all TBIs across North America. The majority of mTBI patients recover within days to weeks; however, 14-36% of the time, acute mTBI symptoms persist for months or even years and develop into persistent post-concussion symptoms (PPCS). There is a need to find biomarkers in patients with PPCS, to improve prognostic ability and to provide insight into the pathophysiology underlying chronic symptoms. Recent research has pointed toward impaired network integrity and cortical communication as a biomarker. In this study we investigated functional near-infrared spectroscopy (fNIRS) as a technique to assess cortical communication deficits in adults with PPCS. Specifically, we aimed to identify cortical communication patterns in prefrontal and motor areas during rest and task, in adult patients with persistent symptoms. We found that (1) the PPCS group showed reduced connectivity compared with healthy controls, (2) increased symptom severity correlated with reduced coherence, and (3) connectivity differences were best distinguishable during task and in particular during the working memory task (n-back task) in the right and left dorsolateral prefrontal cortex (DLPFC). These data show that reduced brain communication may be associated with the pathophysiology of mTBI and that fNIRS, with a relatively simple acquisition paradigm, may provide a useful biomarker of this injury.