scholarly journals Deep Learning Based Software to Identify Large Vessel Occlusion on Noncontrast Computed Tomography

Stroke ◽  
2020 ◽  
Vol 51 (10) ◽  
pp. 3133-3137
Author(s):  
Marta Olive-Gadea ◽  
Carlos Crespo ◽  
Cristina Granes ◽  
Maria Hernandez-Perez ◽  
Natalia Pérez de la Ossa ◽  
...  

Background and Purpose: Reliable recognition of large vessel occlusion (LVO) on noncontrast computed tomography (NCCT) may accelerate identification of endovascular treatment candidates. We aim to validate a machine learning algorithm (MethinksLVO) to identify LVO on NCCT. Methods: Patients with suspected acute stroke who underwent NCCT and computed tomography angiography (CTA) were included. Software detection of LVO (MethinksLVO) on NCCT was tested against the CTA readings of 2 experienced radiologists (NR-CTA). We used a deep learning algorithm to identify clot signs on NCCT. The software image output trained a binary classifier to determine LVO on NCCT. We studied software accuracy when adding National Institutes of Health Stroke Scale and time from onset to the model (MethinksLVO+). Results: From 1453 patients, 823 (57%) had LVO by NR-CTA. The area under the curve for the identification of LVO with MethinksLVO was 0.87 (sensitivity: 83%, specificity: 71%, positive predictive value: 79%, negative predictive value: 76%) and improved to 0.91 with MethinksLVO+ (sensitivity: 83%, specificity: 85%, positive predictive value: 88%, negative predictive value: 79%). Conclusions: In patients with suspected acute stroke, MethinksLVO software can rapidly and reliably predict LVO. MethinksLVO could reduce the need to perform CTA, generate alarms, and increase the efficiency of patient transfers in stroke networks.

Stroke ◽  
2017 ◽  
Vol 48 (suppl_1) ◽  
Author(s):  
Prasannna Tadi ◽  
Pranav Reddy ◽  
Sravanthi Pemmasani ◽  
Nasir Fakhri ◽  
Matthew Siket ◽  
...  

Introduction: Stroke is a common medical emergency. The outcome depends upon access to stroke specialists, rapid scanning, assessment, and treatments. Identification of large vessel occlusion (LVO) is critical in the selection of patients for emergency embolectomy (clot removal) in patients with acute ischemic stroke. A pre-hospital stroke severity scale, such as the Los Angeles Motor Scale (LAMS) may have utility in selecting appropriate patients for CTA, while minimizing radiation exposure risk to the population as a whole. Methods: This was a retrospective analysis of 249 consecutive code stroke activations at a comprehensive stroke center during a 3.5 month period using a LAMS cutoff of ≥4 to trigger CTA acquisition. We determined the sensitivity, specificity, positive predictive value, and negative predictive value of using LAMS to detect large vessel occlusion (LVO). Gold standard was any vessel imaging within 24 hours. Inter-rater reliability of LAMS scoring was determined by blinded scoring of physical exam data from the chart by 3 neurovascular physicians. Results: There were 249 code stroke activations during the study period: 91 acute CTAs were recommended based on LAMS scoring. 20 large vessel occlusions were detected. 158 patients did not have a CTA acutely; none had a LVO during subsequent vessel imaging. The sensitivity is 100%, negative predictive value 100%, specificity 69%, positive predictive value 22% of the LAMS triage method. Inter-rater Reliability: Shrout-Fleiss pairwise weighted kappa coefficients between the three raters on LAMS scores were 0.67, 0.55, and 0.62. Kappa coefficients for pairs of raters when LAMS were dichotomized as <3 vs 4-5 were 0.64, 0.50, and 0.71. Clinically meaningful disagreements were evident. Conclusions: Accuracy and ease-of-use makes LAMS an ideal clinical tool to rapidly assess acute stroke patients for LVO and emergency mechanical thrombectomy. LAMS demonstrated excellent sensitivity in excluding patients who did not have a LVO. Appropriate training is required to ensure accuracy of LAMS scoring by providers.


Stroke ◽  
2014 ◽  
Vol 45 (suppl_1) ◽  
Author(s):  
Alexander Venizelos ◽  
Sherman Chen ◽  
Ryan Gianatasio ◽  
Stewart Coffman ◽  
Mark Gamber ◽  
...  

Introduction: A pre-hospital stroke severity scale that correlates well with an NIHSS of 10 or greater as well as with large vessel occlusions, but is easier and faster to perform than full NIHSS, would be a very useful triaging tool to emergency medical services (EMS). The LEGS score (Lower extremity strength, Eyes/visual fields, Gaze deviation, Speech difficulty) is a 16-point pre-hospital stroke severity scale that is a shortened NIHSS-5. Hypothesis: We assessed the hypothesis that the LEGS score was a useful pre-hospital stroke severity scoring system to identify large vessel acute ischemic strokes. Methods: The LEGS score (0-16) and NIHSS (0-42) were performed in the emergency department over a 6-month period. We retrospectively reviewed those charts for correlation to an NIHSS of 10 or greater and evidence of large-vessel occlusion on either CT or MR Angiography within 48 hours of last known normal. Results: A total of 181 consecutive ischemic stroke patients were evaluated. LEGS score 4 or greater was a good predictor of an NIHSS of 10 or greater (59/181; positive predictive value 92%; and specificity 95%) and false positives noted was 5/181. LEGS score of less than 4 was a good predictor of an NIHSS of less than 10 (108/181; negative predictive value 91%; and sensitivity 95%) and false negatives noted was 10/181. Of those patients 155 underwent intracranial vascular imaging. LEGS score of less than 4 was a good test to rule-out large vessel occlusion (negative predictive value of 86%; 89/103), but had modest sensitivity (69%; 31/45) and positive predictive value (60%; 31/52). The LEGS score of 4 or greater, however, was fairly specific for determining large-vessel occlusions at 81% (89/110). Conclusions: LEGS score of 4 or greater has good correlation with an NIHSS of 10 or greater as well as modest correlation with large vessel occlusion on CT or MR Angiography. This shortened NIHSS-5 may be a useful pre-hospital indicator of patients who may benefit from endovascular intervention.


Stroke ◽  
2021 ◽  
Author(s):  
Sriharsha Voleti ◽  
Johnathan Vidovich ◽  
Brendan Corcoran ◽  
Bin Zhang ◽  
Vivek Khandwala ◽  
...  

Background and Purpose: The Alberta Stroke Program Early Computed Tomography (CT) Score (ASPECTS) and CT perfusion (CTP) are commonly used to predict the ischemic core in acute ischemic strokes. CT angiography source images (CTA-SI) can also provide additional information to identify the extent of ischemia. Our objective was to investigate the correlation of noncontrast CT (NCCT) ASPECTS and CTA-SI ASPECTS with CTP core volumes. Methods: We utilized a single institutional, retrospective registry of consecutive patients with acute ischemic stroke with large vessel occlusion between May 2016 and May 2018. We graded ASPECTS both on baseline NCCT and CTA-SI and measured CTP core using automated RAPID software (cerebral blood flow <30%). We used Spearman’s correlation coefficients to evaluate the correlation between continuous variables. Results: A total of 52 patients fit the inclusion criteria of large vessel occlusion in 6 to 24 hours and baseline imaging work up of NCCT, CTA, and CTP. The median age was 63 (interquartile range=53.5–75) and 38.46% were female. The median NCCT ASPECTS was 7 (interquartile range=6–9), CTA-SI ASPECTS was 5 (interquartile range=4–7), and CTP core was 14.5 mL (interquartile range=0–46 mL). There was a moderate correlation between NCCT ASPECTS and CTP core (r s =−0.55, P <0.0001) and between CTA-SI ASPECTS and CTP core (r s =−0.50, P =0.0002). The optimal NCCT ASPECTS cutoff score to detect CTP core ≤70 mL was ≥6 (sensitivity, 0.84; specificity, 0.57; positive predictive value, 0.93; negative predictive value, 0.36) and the optimal CTA-SI ASPECTS was ≥5 (sensitivity, 0.76; specificity, 0.71; positive predictive value, 0.94; negative predictive value, 0.31). Conclusions: There was a moderate correlation between NCCT and CTA-SI ASPECTS in predicting CTP defined ischemic core in delayed time windows. Further studies are needed to determine if NCCT and CTA imaging could be used for image-based patient selection when CTP imaging is not available.


2019 ◽  
Vol 90 (e7) ◽  
pp. A39.1-A39
Author(s):  
Jonathan JD Baird-Gunning ◽  
Shaun Zhai ◽  
Brett Jones ◽  
Neha Nandal ◽  
Chandi Das ◽  
...  

Introduction25%-30% of patients admitted with acute stroke are stroke mimics. Clinical assessment plays a major role in diagnosis in the hyperacute clinical setting. Identifying physical signs that correctly identify stroke is therefore important. A retrospective study1 suggested that the presence of sensory inattention (or neglect) was seen exclusively in stroke patients, suggesting that inattention might be a reliable discriminator between stroke and mimics. This study aimed to test that hypothesis.MethodsProspective assessment of suspected stroke patients for the presence of neglect (NIHSS definition). Neglect could be visual and/or somatosensory. The presence of neglect was then correlated with eventual diagnosis at 48 hours. Sensitivity, specificity and predictive values were calculated. A post-hoc analysis evaluated the correlation of neglect with large vessel occlusion in patients who underwent angiography.Results115 patients were recruited, 70 ultimately with stroke and 45 with other diagnoses. Neglect was present in 27 patients (of whom 23 had stroke) and absent in 88. This yielded: sensitivity 32.9%, specificity 91.1%, positive predictive value 85.2%, and negative predictive value 41.9%. Two patients with neglect had a diagnosis of functional illness, one a seizure, and one a brain tumour. Neglect was present in 7 out of 8 patients with large vessel occlusion (sensitivity 87.5%) and was absent in all patients who did not have large vessel occlusion on angiogram.ConclusionWhen present, neglect is a strong predictor of organic pathology and large vessel occlusion. However, it is not 100% specific and can be seen in functional presentations.ReferenceGargalas S, Weeks R, Khan-Bourne N, Shotbolt P, Simblett S, Ashraf L, Doyle C, Bancroft V, David AS: Incidence and outcome of functional stroke mimics admitted to a hyperacute stroke unit. J Neurol Neurosurg Psychiatry 2017, 88:2–6.


2008 ◽  
Vol 122 (11) ◽  
pp. 1230-1234 ◽  
Author(s):  
S M Ragab ◽  
F A Erfan ◽  
M A Khalifa ◽  
E M Korayem ◽  
H A Tawfik

AbstractObjectives:To conduct a prospective study (1) to evaluate and compare the efficacies of nasopharyngeal endoscopy and computed tomography in the diagnosis of local failure of external beam radiotherapy for nasopharyngeal carcinoma, and (2) to assess whether multiple endoscopic nasopharyngeal biopsies are superior to a single, targeted biopsy, for the same purpose.Methods:Forty-six patients who had been treated with external beam radiotherapy for primary nasopharyngeal carcinoma were enrolled in the study. For every patient recruited, computed tomography, rigid nasopharyngeal endoscopy and nasopharyngeal biopsies were performed 12 weeks after radiotherapy.Results:Twelve weeks after treatment, six patients (13 per cent) had evident disease on histological examination of biopsies. Nasopharyngeal endoscopy showed a sensitivity, specificity, positive predictive value and negative predictive value of 66.6, 95, 66.6 and 95 per cent, respectively. There was statistically significant agreement between the endoscopic findings and the histological findings (Kappa reliability coefficient = 0.617, p < 0.01). Computed tomography showed a sensitivity, specificity, positive predictive value and negative predictive value of 50, 45, 12 and 85.7 per cent, respectively. There was no statistically significant agreement between the computed tomography findings and the histological findings (Kappa reliability coefficient = 0.021, p > 0.05). A targeted, single biopsy performed under endoscopic control demonstrated excellent sensitivity, specificity, positive predictive value and negative predictive value, being 83.3, 100, 100 and 97.5 per cent, respectively. The Kappa test showed a very statistically significant agreement between the histological findings for the single and the multiple endoscopic biopsies (Kappa reliability coefficient = 0.897, p < 0.001).Conclusions:Rigid nasopharyngeal endoscopy should be considered the primary follow-up tool after radiotherapy treatment of nasopharyngeal carcinoma, with computed tomography being reserved for patients with histological or symptomatic indications. Routine postnasal biopsies are not necessary, given the excellent specificity and negative predictive value of rigid nasopharyngeal endoscopy. Single, targeted endoscopic biopsy provides an excellent alternative to the usual multiple biopsies. In addition, it reduces cost, time, morbidity and patient discomfort.


2014 ◽  
Vol 47 (1) ◽  
pp. 9-13 ◽  
Author(s):  
Eduardo Just da Costa e Silva ◽  
Giselia Alves Pontes da Silva

Objective To evaluate the accuracy of computed tomography for local and lymph node staging of Wilms' tumor. Materials and Methods Each case of Wilms' tumor was evaluated for the presence of abdominal lymph nodes by a radiologist. Signs of capsule and adjacent organ invasion were analyzed. Surgical and histopathological results were taken as the gold standard. Results Sensitivity was 100% for both mesenteric and retroperitoneal lymph nodes detection, and specificity was, respectively, 12% and 33%, with positive predictive value of 8% and 11% and negative predictive value of 100%. Signs of capsular invasion presented sensitivity of 87%, specificity of 77%, positive predictive value of 63% and negative predictive value of 93%. Signs of adjacent organ invasion presented sensitivity of 100%, specificity of 78%, positive predictive value of 37% and negative predictive value of 100%. Conclusion Computed tomography tumor showed low specificity and low positive predictive value in the detection of lymph node dissemination. The absence of detectable lymph nodes makes their presence unlikely, and likewise regarding the evaluation of local behavior of tumors.


Stroke ◽  
2020 ◽  
Vol 51 (11) ◽  
pp. 3361-3365 ◽  
Author(s):  
Fareshte Erani ◽  
Nadezhda Zolotova ◽  
Benjamin Vanderschelden ◽  
Nima Khoshab ◽  
Hagop Sarian ◽  
...  

Background and Purpose: Clinical methods have incomplete diagnostic value for early diagnosis of acute stroke and large vessel occlusion (LVO). Electroencephalography is rapidly sensitive to brain ischemia. This study examined the diagnostic utility of electroencephalography for acute stroke/transient ischemic attack (TIA) and for LVO. Methods: Patients (n=100) with suspected acute stroke in an emergency department underwent clinical exam then electroencephalography using a dry-electrode system. Four models classified patients, first as acute stroke/TIA or not, then as acute stroke with LVO or not: (1) clinical data, (2) electroencephalography data, (3) clinical+electroencephalography data using logistic regression, and (4) clinical+electroencephalography data using a deep learning neural network. Each model used a training set of 60 randomly selected patients, then was validated in an independent cohort of 40 new patients. Results: Of 100 patients, 63 had a stroke (43 ischemic/7 hemorrhagic) or TIA (13). For classifying patients as stroke/TIA or not, the clinical data model had area under the curve=62.3, whereas clinical+electroencephalography using deep learning neural network model had area under the curve=87.8. Results were comparable for classifying patients as stroke with LVO or not. Conclusions: Adding electroencephalography data to clinical measures improves diagnosis of acute stroke/TIA and of acute stroke with LVO. Rapid acquisition of dry-lead electroencephalography is feasible in the emergency department and merits prehospital evaluation.


2011 ◽  
Vol 93 (8) ◽  
pp. 639-641 ◽  
Author(s):  
VCY Tang ◽  
A Attwell-Heap

INTRODUCTION The aim of this study was to validate the use of non-contrast computed tomography (CT) with a ureteral stent in situ instead of ureteroscopy for identification of renal tract stones. METHODS All patients who had stents inserted for renal tract stones and underwent non-contrast CT with the stent in situ followed by ureteroscopy between May 2008 and October 2009 at The Canberra Hospital, Australia, were analysed retrospectively. Statistical analysis was performed to compare any differences between CT and ureteroscopy in the identification of stones. RESULTS Overall, 57 patients were included in the study. The difference between CT and ureteroscopy findings was statistically significant. CT identification of stones with a stent in situ had a sensitivity of 86%, a specificity of 46%, a positive predictive value of 63%, a negative predictive value of 76% and an accuracy of 67%. CONCLUSIONS Our study suggests that non-contrast CT is inferior to the ‘gold standard’ of ureteroscopy. It lacks sensitivity, specificity, positive predictive value, negative predictive value and accuracy. Therefore, we cannot recommend using non-contrast CT to replace ureteroscopy.


Sign in / Sign up

Export Citation Format

Share Document