Acute Ischemic Stroke Cohort Research Articles

Development of a Natural Language Processing (NLP) model to automatically extract clinical data from electronic health records: results from an Italian comprehensive stroke center

IntroductionData collection often relies on time-consuming manual inputs, with a vast amount of information embedded in unstructured texts such as patients’ medical records and clinical notes. Our study aims to develop a pipeline that combines active learning (AL) and NLP techniques to enhance data extraction in an acute ischemic stroke cohort. Materials and methodsConsecutive acute ischemic stroke patients who received reperfusion therapies at IRCCS Humanitas Research Hospital were included. The Italian NLP Bidirectional Encoder Representations from Transformers (BERT) model was trained with AL to automatically extract clinical variables from electronic health text. Simulated active learning performances were evaluated on a set of labels representing patients’ comorbidities, comparing Bayesian Uncertainty Sampling by Disagreement (BALD) and random text selection. Prognostic models predicting patients’ functional outcomes using Gradient Boosting were trained on manually labelled and semi-automatically extracted data and their performance was compared. ResultsThe active learning process initially showed null performance until around 20% of texts were labelled, possibly due to root layers freezing in the BERT model, yet overall, active learning improves model learning efficiency across most comorbidities. Prognostic modelling showed no significant difference in performance between models trained on manually labelled versus semi-automatically extracted data, indicating effective prediction capabilities in both settings. ConclusionsWe developed an efficient language model to automate the extraction of clinical data from Italian unstructured health texts in a cohort of ischemic stroke patients. In a preliminary analysis, we demonstrated its potential applicability for enhancing prediction model accuracy.

Role of Circadian Rhythm Changes on Functional Dependence Despite Successful Repercussion in Patients with Endovascular Treatment.

Increasing evidence of circadian biology may influence the physiopathologic mechanism, progression, and recovery of stroke. However, few data have shown about circadian rhythm on futile recanalization (FR) in patients treated with endovascular treatment (EVT). From 2017 to 2021, an observational cohort of acute ischemic stroke (AIS) patients with large vessel occlusion (LVO) underwent EVT was conducted. FR was defined as the failure to achieve functional independence in patients at 90 days after EVT, although the occluded vessels reached a recanalization. The effect of circadian rhythm on FR was investigated using the logistic regression model. Of 783 patients, there were 149 patients who had stroke onset between 23:00-6:59, 318 patients between 7:00-14:59, and 316 patients between 15:00-22:59. Patients suffered a stroke during 15:00-22:59 had shorter OTP (p =0.001) time, shorter OTR (p<0.001) time, higher rate of intravenous thrombolysis (p =0.001) than groups of other time intervals. The rate of FR post-EVT in patients who had a stroke between 15:00-22:59 was significantly higher than in those with stroke onset between 23:00-6:59 (p =0.017). After adjusting for confounding factors, the time of stroke occurring during 15:00-22:59 (adjusted OR [aOR], 1.652; 95%CI, 1.024-2.666, p =0.04) was an independent predictor of FR. Circadian rhythm can directly or indirectly affect the occurrence, development, and prognosis of AIS. More studies may be needed in the future to validate the results of our study and to explore the potential mechanisms behind the effects of circadian rhythms on FR.

Cerebral collaterals are associated with pre-treatment brain-blood barrier permeability in acute ischemic stroke patients.

To investigate the relationship between collaterals and blood-brain barrier (BBB) permeability on pre-treatment MRI in a cohort of acute ischemic stroke (AIS) patients treated with thrombectomy. We conducted a retrospective analysis of the HIBISCUS-STROKE cohort, a single-center observational study that enrolled patients treated with thrombectomy from 2016 to 2022. Dynamic-susceptibility MRIs were post-processed to generate K2 maps with arrival-time correction, which were co-registered with apparent diffusion coefficient (ADC) maps. The 90th percentile of K2 was extracted from the infarct core-defined by an ADC ≤ 620 × 10-6 mm2/s with manual adjustments-and expressed as a percentage change compared to the contralateral white matter. Collaterals were assessed using pre-thrombectomy digital subtraction arteriography with an ASITN/SIR score < 3 defining poor collaterals. Out of 249 enrolled, 101 (40.6%) were included (median age: 72.0 years, 52.5% of males, median NIHSS score at admission: 15.0). Patients with poor collaterals (n = 44) had worse NIHSS scores (median: 16.0 vs 13.0, p = 0.04), larger infarct core volumes (median: 43.7 mL vs 9.5 mL, p < 0.0001), and higher increases in K2 (median: 346.3% vs 152.7%, p = 0.003). They were less likely to achieve successful recanalization (21/44 vs 51/57, p < 0.0001) and experienced more frequent hemorrhagic transformation (16/44 vs 9/57, p = 0.03). On multiple variable analysis, poor collaterals were associated with larger infarct cores (odds ratio (OR) = 1.12, 95% confidence interval (CI): [1.07, 1.17], p < 0.0001) and higher increases in K2 (OR = 6.63, 95% CI: [2.19, 20.08], p = 0.001). Poor collaterals are associated with larger infarct cores and increased BBB permeability at admission MRI. Poor collaterals are associated with a larger infarct core and increased BBB permeability at admission MRI of AIS patients treated with thrombectomy. These findings may have translational interests for extending thrombolytic treatment eligibility and developing neuroprotective strategies. In AIS, collaterals and BBB disruption have been both linked to hemorrhagic transformation. Poor collaterals were associated with larger ischemic cores and increased BBB permeability on pre-treatment MRI. These findings could contribute to hemorrhagic transformation risk stratification, thereby refining clinical decision-making for reperfusion therapies.

Service delivery in acute ischemic stroke patients: Does sex matter?

Women with acute ischemic stroke (AIS) are older and have greater preexisting handicap than men. Given that these factors do not fully explain their poorer long-term outcomes, we sought to investigate potential sex differences in the delivery of acute stroke care in a large cohort of consecutive AIS patients. We analyzed all patients from ASTRAL (Acute Stroke Registry and Analysis of Lausanne) from March 2003 to December 2019. Multivariable analyses were performed on acute time metrics, revascularization therapies, ancillary examinations for stroke workup, subacute symptomatic carotid artery revascularization, frequency of change in goals of care (palliative care), and length of hospital stay. Of the 5347 analyzed patients, 45% were biologically female and the median age was 74.6 years. After multiple adjustments, female sex was significantly associated with higher onset-to-door (adjusted hazard ratio [aHR] = 1.09, 95% confidence interval [CI] = 1.04-1.14) and door-to-endovascular-puncture intervals (aHR = 1.15, 95% CI = 1.05-1.25). Women underwent numerically fewer diagnostic examinations (adjusted odds ratio [aOR] = 0.94, 95% CI = 0.85-1.04) and fewer subacute carotid revascularizations (aOR = 0.69, 95% CI = 0.33-1.18), and had longer hospital stays (aHR = 1.03, 95% CI = 0.99-1.07), but these differences were not statistically significant. We found no differences in the rates of acute revascularization treatments, or in the frequency of change of goals of treatments. This retrospective analysis of a large, consecutive AIS cohort suggests that female sex is associated with unfavorable pre- and in-hospital time metrics, such as a longer onset-to-door and door-to-endovascular-puncture intervals. Such indicators of less effective stroke care delivery may contribute to the poorer long-term functional outcomes in female patients and require further attention.

Abstract TP127: Pretreatment CT Collateral Parameters Predict Good Outcomes in Successfully Recanalized Middle Cerebral Artery Distal Medium Vessel Occlusions

Background/Purpose: Distal medium vessel occlusions (DMVOs) account for a large percentage of vessel occlusions resulting in acute ischemic stroke (AIS) with disabling symptoms. We aim to assess whether pretreatment CTP collateral status (CS) parameters can serve as imaging biomarkers for good clinical outcomes prediction in successfully recanalized middle cerebral artery (MCA) DMVOs. Methods: We performed a retrospective analysis of consecutive patients with AIS secondary to primary MCA-DMVOs who were successfully recanalized by mechanical thrombectomy (MT) defined as modified thrombolysis in cerebral infarction (mTICI) 2b, 2c, or 3. We evaluated the association between CBV index and HIR independently with good clinical outcomes (modified Rankin score 0-2) using Spearman rank correlation, logistic regression, and ROC analyses. Results: From 8/22/2018 to 10/18/2022, 60 consecutive patients met our inclusion criteria (mean age 71.2 +- 13.9 years old [mean+-SD], 35 female). CBV index (r = -0.693, p &lt; 0.001) and HIR (0.687, p &lt; 0.001) strongly correlated with 90 day mRS. A CBV index &gt;= 0.7 ((OR 2.27 [6.94 - 21.23], p = 0.001)) and lower likelihood of prior stroke (0.13 [0.33 - 0.86]), p = 0.024) were independently associated with good outcomes.ROC analysis demonstrated good performance of CBV Index in predicting good 90 day mRS (AUC 0.73, p = 0.003) with a threshold of 0.7 for optimal sensitivity (71% [52.0-85.8%]) and specificity (76% [54.9 - 90.6%]). HIR also demonstrated strong performance in predicting good 90 day mRS (AUC 0.77, p = 0.001) with a threshold of 0.3 for optimal sensitivity (64.5% [45.4-80.8%]) and specificity (76.0% [54.9 - 90.6%]). Conclusion: A CBV index ≥ 0.7 may be independently associated with good clinical outcomes in our cohort of AIS caused by MCA-DMVOs that were successfully treated with MT. An HIR &lt; 0.3 is also associated with good clinical outcomes. This is the first study of which we are aware to identify a CBV index threshold for MCA-DMVOs.

Pretreatment CTP Collateral Parameters Predict Good Outcomes in Successfully Recanalized Middle Cerebral Artery Distal Medium Vessel Occlusions.

Distal medium vessel occlusions (DMVOs) account for alarge percentage of vessel occlusions resulting in acute ischemic stroke (AIS) with disabling symptoms. We aim to assess whether pretreatment quantitative CTP collateral status (CS) parameters can serve as imaging biomarkers for good clinical outcomes prediction in successfully recanalized middle cerebral artery (MCA) DMVOs. We performed aretrospective analysis of consecutive patients with AIS secondary to primary MCA-DMVOs who were successfully recanalized by mechanical thrombectomy (MT) defined as modified thrombolysis in cerebral infarction (mTICI) 2b, 2c, or 3. We evaluated the association between the CBV index and HIR independently with good clinical outcomes (modified Rankin score0-2) using Spearman rank correlation, logistic regression, and ROC analyses. From 22 August 2018 to 18 October 2022 8/22/2018 to 10/18/2022, 60consecutive patients met our inclusion criteria (mean age 71.2 ± 13.9 years old [mean ± SD], 35female). The CBV index (r = -0.693, p < 0.001) and HIR (0.687, p < 0.001) strongly correlated with 90-day mRS. ACBV index ≥ 0.7 (odds ratio, OR, 2.27, range 6.94-21.23 [OR] 2.27 [6.94-21.23], p = 0.001)) and lower likelihood of prior stroke (0.13 [0.33-0.86]), p = 0.024)) were independently associated with good outcomes. The ROC analysis demonstrated good performance of the CBV index in predicting good 90-day mRS (AUC 0.73, p = 0.003) with athreshold of 0.7for optimal sensitivity (71% [52.0-85.8%]) and specificity (76% [54.9-90.6%]). The HIR also demonstrated adequate performance in predicting good 90-day mRS (AUC 0.77, p = 0.001) with athreshold of 0.3for optimal sensitivity (64.5% [45.4-80.8%]) and specificity (76.0% [54.9-90.6%]). ACBV index ≥ 0.7may be independently associated with good clinical outcomes in our cohort of AIS caused by MCA-DMVOs that were successfully treated with MT. Furthermore, a HIR < 0.3 is also associated with good clinical outcomes. This is the first study of which we are aware to identify aCBV index threshold for MCA-DMVOs.

Racial Impact on Inpatient Stroke Quality of Care in Two Community Hospitals.

This analysis was conducted as a part of a quality improvement project aiming at identifying racial disparity in inpatient stroke quality of care. The Get With The Guidelines (GWTG) database was used to identify all patients discharged with any stroke diagnosis between January and December 2021. An additional chart review was conducted to ensure the accuracy of racial/ethnic categorization. The sample was dichotomized into white vs. non-white groups and compared with univariate analysis. The study sample comprised 1408 encounters (1347 patients) with Mean age of 71 ± 15 years, 51% women, 82% white patients, 15% non-white patients, 72% acute ischemic stroke (AIS); 15% transient ischemic attack (TIA), 9% intracerebral hemorrhage (ICH), 3% subarachnoid hemorrhage (SAH), and 1% stroke not otherwise specified. Non-white patients were younger and had fewer concomitant diagnoses, a lower proportion of TIA, and a higher proportion of ICH (p = 0.004). In the AIS cohort, compared to white patients, non-white patients had less frequent ambulance (p = 0.009), arrived at the hospital later than white patients (7.7 h longer; p < 0.001), had more severe strokes, and had less frequent IV thrombolysis utilization (7% vs. 13%; p = 0.042). Similarly, in the TIA cohort, non-white patients' utilization of EMS was lower than that of white patients, and their hospital arrival was delayed. In the ICH cohort, non-white patients were younger and had a lower frequency of atrial fibrillation and a non-significant trend toward higher disease severity. The SAH cohort had only eight non-white patients, six of whom were transferred to a higher level of hospital care within a few hours of arrival. Importantly, the hospital-based quality metrics, such as door-to-CT time, door-to-needle time, and the Joint Commission stroke quality metrics, were similar between the two groups. There is a racial disparity in the pre-hospital phase of the stroke chain of survival of non-white patients, impacting IV thrombolysis utilization. The younger age and worse lipid profile and hemoglobin A1c of non-white patients suggest the need for better preventative care starting at a young age.

Blood-Brain Barrier Permeability and Kinetics of Inflammatory Markers in Acute Stroke Patients Treated With Thrombectomy.

The aim of this study was to investigate the relationship between baseline blood-brain barrier (BBB) permeability and the kinetics of circulating inflammatory markers in a cohort of acute ischemic stroke (AIS) patients treated with mechanical thrombectomy. The CoHort of Patients to Identify Biological and Imaging markerS of CardiovascUlar Outcomes in Stroke includes AIS patients treated with mechanical thrombectomy after admission MRI and undergoing a sequential assessment of circulating inflammatory markers. Baseline dynamic susceptibility perfusion MRI was postprocessed with arrival time correction to provide K2 maps reflecting BBB permeability. After coregistration of apparent diffusion coefficient and K2 maps, the 90th percentile of K2 value was extracted within baseline ischemic core and expressed as a percentage change compared with contralateral normal-appearing white matter. Population was dichotomized according to the median K2 value. Univariable and multiple variable logistic regression analyses were performed to investigate factors associated with increased pretreatment BBB permeability in the whole population and in patients with symptom onset <6 hours. In the whole population (n = 105 patients, median K2 = 1.59), patients with an increased BBB permeability had higher serum levels of matrix metalloproteinase (MMP)-9 at H48 (p = 0.02), a higher C-reactive protein (CRP) serum level at H48 (p = 0.01), poorer collateral status (p = 0.01), and a larger baseline ischemic core (p < 0.001). They were more likely to have hemorrhagic transformation (p = 0.008), larger final lesion volume (p = 0.02), and worst neurologic outcome at 3 months (p = 0.04). The multiple variable logistic regression indicated that an increased BBB permeability was associated only with ischemic core volume (odds ratio [OR] 1.04, 95% CI 1.01-1.06, p < 0.0001). Restricting analysis to patients with symptom onset <6 hours (n = 72, median K2 = 1.27), participants with an increased BBB permeability had higher serum levels of MMP-9 at H0 (p = 0.005), H6 (p = 0.004), H24 (p = 0.02), and H48 (p = 0.01), higher CRP levels at H48 (p = 0.02), and a larger baseline ischemic core (p < 0.0001). The multiple variable logistic analysis showed that increased BBB permeability was independently associated with higher H0 MMP-9 levels (OR 1.33, 95% CI 1.12-1.65, p = 0.01) and a larger ischemic core (OR 1.27, 95% CI 1.08-1.59, p = 0.04). In AIS patients, increased BBB permeability is associated with a larger ischemic core. In the subgroup of patients with symptom onset <6 hours, increased BBB permeability is independently associated with higher H0 MMP-9 levels and a larger ischemic core.

Clinical score for early diagnosis and treatment of stroke-like episodes in MELAS syndrome.

Stroke-like episodes (SLEs) in patients with MELAS syndrome are often initially misdiagnosed as acute ischemic stroke (AIS), resulting in treatment delay. We aimed to determine clinical features that may distinguish SLEs from AISs and explore the benefit of early L-arginine treatment on patient outcomes. We looked retrospectively for MELAS patients admitted between January 2005 and January 2022 and compared them to an AIS cohort with similar lesion topography. MELAS patients who received L-arginine within 40days of their first SLE were defined as the early treatment group and the remaining as late or no treatment group. Twenty-three SLEs in 10 MELAS patients and 21 AISs were included. SLE patients had significantly different features: they were younger, more commonly reported hearing loss, lower body mass index, had more commonly a combination of headache and/or seizures at presentation, serum lactate was higher, and hemiparesis was less common. An SLE Early Clinical Score (SLEECS) was constructed by designating one point to each above features. SLEECS ≥ 4 had 80% sensitivity and 100% specificity for SLE diagnosis. Compared to late or no treatment, early treatment group patients (n = 5) had less recurrent SLEs (total 2 vs. 11), less seizures (14% vs. 25%, p = 0.048), lower degree of disability at first and last follow-up (modified ranking scale, mRS 2 ± 0.7 vs. 4.2 ± 1, p = 0.005; 2 ± 0.7 vs. 5.8 ± 0.5, p < 0.001, respectively), and a lower mortality (0% vs. 80% p = 0.048). The SLEECS model may aid in the early diagnosis and treatment of SLEs and lead to improved clinical outcomes.

β2-Microglobulin is a Novel and Reliable Biomarker for Predicting Ischemic Stroke Recurrence: A Prospective Cohort Study.

Immune and inflammatory mechanisms play key roles in the development and outcome of acute ischemic stroke (AIS). β2-Microglobulin (β2M) is the light chain of major histocompatibility complex-1 (MHC-1), which can directly and quickly reflect the immune and inflammatory state of the body. Previous studies have shown a close relationship between β2M and AIS, but its relationship with the recurrence of AIS has not been reported. This study attempted to explore the relationship between β2M and the recurrence of AIS. A single-center AIS cohort involving 135 patients was followed for approximately 26–46 months. Clinical and laboratory data from the patients were collected when hospitalized. The endpoint was the occurrence of recurrent AIS after patients were discharged. Propensity score matching was used to match cohort groups. Cox regression analysis was used to predict risk factors for recurrent AIS, and receiver operating characteristic curve (ROC) analysis was used to calculate the optimal cutoff value for discriminating recurrence in patients with AIS. The rate of recurrence was 29.6% [95% CI, 21.8%–37.3%] in the follow-up group. Patients with higher levels of serum β2M had a higher risk of AIS recurrence than patients with lower levels of β2M (adjusted hazard ratio, 3.214 [95% CI, 1.557–6.633]; adjusted hazard ratio after matching, 5.831, [95% CI, 2.052–16.572]). A β2M value of 2.31 mg/L was calculated by ROC analysis as the optimal cutoff value for AIS recurrence (area under the curve 0.770, [95% CI, 0.687–0.853]). As a quick responder to the body’s immune and inflammatory states, β2M may be a novel and reliable biomarker in predicting AIS recurrence.

Abstract WMP64: Multimodal Bayesian Modeling Of Stroke Severity

Introduction: Understanding the relations between lesions and outcomes is a particularly promising avenue to support tailored stroke care. We here employed a novel Bayesian framework integrating lesion location and functional lesion connectivity, i.e., lesion network mapping data, aiming to augment the prediction of stroke severity and the interpretability of lesion-symptom associations in a multi-center cohort of acute ischemic stroke (AIS) patients. Methods: Analyses relied on 1,077 AIS patients of the MRI-GENIE study [age: 64.2(14.7), 38%women]. NIHSS-based stroke severity was modeled via Bayesian linear regression. Structural location and functional connectivity information was represented by ten unique patterns each. We compared the out-of-sample predictive performances of structural vs. structural-functional models in Bayesian model comparisons. In ancillary analyses, we extracted functional connectivity pattern relevant for stroke severity and repeated analyses after stratifying for lesion size (50% of patients with larger lesions). Models were generally adjusted for age, sex, comorbidities and total DWI lesion volume. Results: Model comparison indicated a benefit of the combined structural-functional stroke severity prediction model compared to the structural one (model weights: 82% vs. 18%). This benefit of the combined model became even more apparent in case of patients with large lesions (97% vs. 3%). Positive lesion connectivity to the bilateral ventral attention (VA) and somatomotor networks, as well as temporal default mode network (DMN) predicted higher severity, while positive connectivity to more medial parts of the DMN and the visual network predicted lower severity. Additionally, negative lesion connectivity to the bilateral DMN, somatomotor, dorsal attention, and control networks predicted higher, and negative connectivity to the right control and bilateral VA networks predicted lower severity. Conclusions: Model comparisons indicated a benefit of integrating structural and functional information for the prediction of stroke severity, especially in case of larger lesions. Lesion connectivity data can enhance our understanding of why specific lesion locations can lead to symptoms post-stroke.

Abstract TP24: Indicators Of High Morbidity And Poor Functional Outcome In Acute Ischemic Stroke Patients With Concurrent Covid-19

Objectives: Evidence suggests an association of increased cerebrovascular accidents frequency in patients diagnosed with the novel coronavirus disease, COVID-19. Coagulopathy resulting from the 2019 novel coronavirus (SARS-CoV-2) infection is suspected. This study aims at evaluating thrombotic markers in relation to stroke severity and functional outcomes in a patient cohort of acute ischemic stroke (AIS) with concurrent COVID-19. Methods: We performed a retrospective observational cohort study of 28 patients who tested positive for SARS-CoV-2 via polymerase chain reaction and concomitant AIS confirmed by brain imaging. We analyzed data regarding initial stroke presentation, markers of coagulopathy, and 90-day functional outcomes. Results: The patient cohort displayed high rate of comorbidities with 78.6% having at least 1 vascular risk factor. NIHSS had a median of 16 at initial presentation and median stroke volume of 52 mL. Median NIHSS at discharge or prior to death was 19, and median 90-day mRS was 4. Highest fibrinogen level recorded showed a median of 759.54 mg/dL (IQR 653.75-940.75), D-dimer and lactate dehydrogenase (LDH) showed a median highest recorded value 24,106 ng/mL (IQR 6105.00-80165.00) and 442 ng/mL (IQR 277.00-545.50), respectively. LDH (p=0.0008), D-dimer (p=0.001), and maximum fibrinogen levels (p=0.049) near the time of stroke significantly predicted final NIHSS and functional outcome 90-days after discharge. Conclusions: Adult patients with acute ischemic stroke and concurrent COVID-19 disease exhibited abnormally high markers of coagulopathy, and LDH, D-Dimer, and fibrinogen levels were predictors of morbidity and neurological disability at 90-days in this patient population. Further research is necessary to establish a definitive pattern and assess the ability to use these markers as prognostic elements of 90-day functional outcome.

Modified Prehospital Acute Stroke Severity (mPASS) Scale to Predict Emergent Large Arterial Occlusion.

Introduction To date, identifying emergent large vessel occlusion (ELVO) patients in the prehospital stage is important but still challenging. In this present study, we aimed to design a modified prehospital acute stroke severity (mPASS) scale to identify ELVO patients and compared the scale to the PASS scale which has been published. Methods We retrospectively evaluated a consecutive cohort of acute ischemic stroke (AIS) in our stroke unit who visited the emergercy department. These patients underwent CT angiography (CTA), MR angiography (MRA), or digital subtraction angiography (DSA) at admission. The mPASS scale was calculated based on the National Institutes of Health Stroke Scale (NIHSS) items retrospectively, including the level of consciousness commands, gaze, arm weakness, and aphasia/dysarthria. Receiver operating characteristic (ROC) analysis was used to obtain the area under the curve (AUC) of the mPASS scale, NIHSS, and PASS scale. U-statistics was used to compare the AUC of the mPASS scale to the NIHSS and PASS scale. Results A total of 382 AIS patients were enrolled. The AUC and specificity of the mPASS scale (0.92, 84.4) were all higher than those of the PASS scale. Cortical symptoms such as gaze palsy and consciousness disorder were more specific indicators for ELVO than motor deficits. Conclusions The mPASS scale had a better discrimination for identifying ELVO than the PASS scale in our retrospective cohort. It might predict ELVO in an effective and simple way for paramedics in the prehospital triage stage or emergency stage. Moreover, cortical symptoms might have relatively high specificities to predict ELVO on their own.

Abstract P641: Risk and Predictors of New Cancer Diagnoses After Acute Ischemic Stroke

Introduction: Acute ischemic stroke (AIS) may be the first sign of occult cancer. We aimed to better define the incidence of cancer in the year after AIS and to identify clinical factors associated with new cancer diagnoses. Methods: This was a retrospective cohort study using data from the Cornell Acute Stroke Academic Registry (CAESAR) on patients hospitalized at our center with AIS from 2011-2015. Patients with history of cancer were excluded. Through automated electronic data capture and manual abstraction of inpatient and outpatient medical records, we collected data on patients’ demographics, comorbidities, presentation, radiographic characteristics, stroke subtype, and clinical outcomes. Patients were followed for 1 year after the index AIS for a new diagnosis of pathologically-confirmed cancer. Cox hazards regression adjusting for the competing risk of death was used to evaluate associations between clinical factors and incident cancer. Factors significantly associated in multivariable analysis were entered into a risk stratification score, and this score’s discriminatory ability was evaluated by Harrell’s C-statistic. Results: After excluding 253 patients with history of cancer, this analysis included 963 patients with AIS. During a mean follow-up of 222 days, 16 patients (1.7%; 95% CI, 1.0-2.7%) were diagnosed with cancer. The most common cancers were lung (n=7) and leukemia (n=4) and the median time to cancer diagnosis was 13 days (IQR, 7-194 days). Among patients with cryptogenic stroke, the 1-year cancer incidence rate was 1.7% (95% CI, 0.6-3.7%). Clinical factors associated with incident cancer in multivariable analysis were venous thromboembolism during the AIS hospitalization (HR, 12.5; 95% CI, 3.3-47.0), unexplained weight loss within 6 months (HR 11.7; 95% CI, 3.3-42.0), and three-territory acute infarcts (HR, 4.1, 95% CI, 1.3-13.4). These factors were used to create a clinical score that had a C-statistic of 0.7 (95% CI, 0.5-0.8). Conclusions: In a large urban cohort of AIS, the estimated 1-year incidence of first-ever cancer was 1.7%. Unexplained weight loss, concomitant venous thromboembolism, and three-territory acute infarction pattern may serve as clues to occult cancer with AIS.

Temporal Trends, Predictors, and Outcomes of Acute Ischemic Stroke in Acute Myocardial Infarction in the United States.

BackgroundThere are limited contemporary data prevalence and outcomes of acute ischemic stroke (AIS) complicating acute myocardial infarction (AMI).Methods and ResultsAdult (>18 years) AMI admissions using the National Inpatient Sample database (2000–2017) were evaluated for in‐hospital AIS. Outcomes of interest included in‐hospital mortality, hospitalization costs, length of stay, discharge disposition, and use of tracheostomy and percutaneous endoscopic gastrostomy. The discharge destination was used to classify survivors into good and poor outcomes. Of a total 11 622 528 AMI admissions, 183 896 (1.6%) had concomitant AIS. As compared with 2000, in 2017, AIS rates increased slightly among ST‐segment–elevation AMI (adjusted odds ratio, 1.10 [95% CI, 1.04–1.15]) and decreased in non–ST‐segment–elevation AMI (adjusted odds ratio, 0.47 [95% CI, 0.46–0.49]) admissions (P<0.001). Compared with those without, the AIS cohort was on average older, female, of non‐White race, with greater comorbidities, and higher rates of arrhythmias. The AMI‐AIS admissions received less frequent coronary angiography (46.9% versus 63.8%) and percutaneous coronary intervention (22.7% versus 41.8%) (P<0.001). The AIS cohort had higher in‐hospital mortality (16.4% versus 6.0%; adjusted odds ratio, 1.75 [95% CI, 1.72–1.78]; P<0.001), longer hospital length of stay, higher hospitalization costs, greater use of tracheostomy and percutaneous endoscopic gastrostomy, and less frequent discharges to home (all P<0.001). Among AMI‐AIS survivors (N=153 318), 57.3% had a poor functional outcome at discharge with relatively stable temporal trends.ConclusionsAIS is associated with significantly higher in‐hospital mortality and poor functional outcomes in AMI admissions.

Comparative analysis, applications, and interpretation of electronic health record-based stroke phenotyping methods

BackgroundAccurate identification of acute ischemic stroke (AIS) patient cohorts is essential for a wide range of clinical investigations. Automated phenotyping methods that leverage electronic health records (EHRs) represent a fundamentally new approach cohort identification without current laborious and ungeneralizable generation of phenotyping algorithms. We systematically compared and evaluated the ability of machine learning algorithms and case-control combinations to phenotype acute ischemic stroke patients using data from an EHR.Materials and methodsUsing structured patient data from the EHR at a tertiary-care hospital system, we built and evaluated machine learning models to identify patients with AIS based on 75 different case-control and classifier combinations. We then estimated the prevalence of AIS patients across the EHR. Finally, we externally validated the ability of the models to detect AIS patients without AIS diagnosis codes using the UK Biobank.ResultsAcross all models, we found that the mean AUROC for detecting AIS was 0.963 ± 0.0520 and average precision score 0.790 ± 0.196 with minimal feature processing. Classifiers trained with cases with AIS diagnosis codes and controls with no cerebrovascular disease codes had the best average F1 score (0.832 ± 0.0383). In the external validation, we found that the top probabilities from a model-predicted AIS cohort were significantly enriched for AIS patients without AIS diagnosis codes (60–150 fold over expected).ConclusionsOur findings support machine learning algorithms as a generalizable way to accurately identify AIS patients without using process-intensive manual feature curation. When a set of AIS patients is unavailable, diagnosis codes may be used to train classifier models.

Abstract 14739: Temporal Trends, Predictors and Outcomes of Acute Ischemic Stroke in Acute Myocardial Infarction in the United States

Introduction: There are limited contemporary data on prevalence and outcomes of acute ischemic stroke (AIS) complicating acute myocardial infarction (AMI). Methods: Adult (&gt;18 years) AMI admissions using the National Inpatient Sample database (2000-2017) were evaluated for in-hospital AIS. Outcomes of interest included in-hospital mortality, hospitalization costs, length of stay, and use of tracheostomy and percutaneous endoscopic gastrostomy (PEG). The discharge destination was used to classify AIS survivors into good and poor functional outcomes with poor (moderate to severe disability) defined as discharge to extended care facility including short-term hospital/rehabilitation facility, intermediate care or long-term care facilities. Results: Of a total 11,622,528 AMI admissions, 183,896 (1.6%) had concomitant AIS. Over the 18-year period, AIS rates were stable in STEMI admissions and decreased in NSTEMI (p&lt;0.001). Compared to those without, the AIS cohort was on average older, female, of non-White race, with greater comorbidities, and higher rates of arrhythmias. The AMI-AIS admissions received less frequent coronary angiography (46.9% vs. 63.8%) and percutaneous coronary angiography (22.7% vs 41.8%) (p&lt;0.001). Female sex, non-White race, higher comorbidity, ST-segment-elevation AMI presentation, atrial fibrillation/flutter, use of mechanical circulatory support and invasive mechanical ventilation were identified as individual predictors of AIS after AMI. The AIS cohort had higher in-hospital mortality (16.4% vs. 6.0%; adjusted OR 1.75 [95% CI 1.72-1.78]; p&lt;0.001) with a steady decrease in the adjusted in-hospital mortality in over the 18-year period (21% in 2000 vs 17% in 2017). The AIS cohort had longer hospital length of stay, higher hospitalization costs, greater use of tracheostomy and PEG, and less frequent discharges to home (all p&lt;0.001). In AMI-AIS survivors (N=153,318), 57.3% had a poor functional outcome with temporal trends showing a slight increase in recent years (57% in 2000 vs 62% in 2017). Conclusions: AIS is associated with significantly higher in-hospital mortality and poor functional outcomes in AMI admissions. Despite improvement in in-hospital mortality, poor functional outcomes remain high.

Effect of erroneous body-weight estimation on outcome of thrombolyzed stroke patients

Intravenously administered tissue plasminogen activator (IV-tPA), dose determined by patients' body-weight, remains the only approved drug treatment for acute ischemic stroke (AIS). Since a shorter onset-to-treatment time results in better functional outcome, treatment is often initiated according to the estimated or last-known body-weight of the patient. This approach may result in underdosing or overdosing of tPA. In this multicenter retrospective study, we evaluated the extent of error in tPA dosing in our AIS cohort and its impact on functional outcome and symptomatic intracranial hemorrhage (SICH). Consecutive AIS patients, receiving IV-tPA, dose determined by the estimated body-weight, at three tertiary centers between January and December 2017 were included. Collected data included information about demographics, cardiovascular risk factors, stroke subtype and National Institute of Health Stroke Scale (NIHSS) score. Estimated and measured body-weights were recorded. Modified Rankin scale (mRS) of 2 or more defined unfavorable outcome. The study included 150 patients. Median age was 64 -years (IQR 55-75) with male preponderance (67%) and median NIHSS score of 9 points (IQR 6-17). Mean measured weight of our study population was 58 (SD 13) kg. Median difference between actual and estimated body-weight was 3kg (IQR 1.5-6). Difference was more than 10% in 35 (23.3%) patients. Good functional outcome (mRS 0-1) was achieved by 74 (49.3%) patients and 10 (6.8%) developed SICH. NIHSS (OR 1.288; 95% CI 1.157-1.435, p < 0.001) and large artery atherosclerosis (OR 5.878; 95% CI 1.929-17.910, p = 0.002) were independent predictors of unfavorable functional outcome. Our finding of the statistically insignificant 2.5-fold increase in poor outcomes among patients where the estimated and actual weight differed by more than 10% should be interpreted with caution due to the limited sample size. Significant difference occurs between estimated and actual body-weight in a considerable proportion of thrombolysed AIS patients. However, this discrepancy does not affect functional outcome or the risk of SICH.

Abstract TMP24: Soluble St2 Links Innate Inflammation to Secondary Injury After Stroke and Subarachnoid Hemorrhage

Background: We recently found that soluble ST2 (sST2) is a plasma marker that predicts neurological outcome after ischemic stroke and aneurysmal subarachnoid hemorrhage. We sought to extend these findings by studying sST2 associations with markers of secondary brain injury and peripheral innate immunity. Methods: We studied 241 acute ischemic stroke (AIS) patients with sST2 level measured on admission (7.4 ± 3.7 hrs after stroke onset) and at least one follow-up brain imaging study, obtained at a mean of 2.2 ± 1.4 days after stroke onset. A subgroup (n=147) had a second sST2 measurement at a mean of 2.1 ± 1.7 days after stroke onset. A separate cohort of 110 subarachnoid hemorrhage (SAH) patients were studied who had plasma sST2 measured at approximately 3.5 ± 1.2 days after ictus. Markers of secondary brain injury included edema after AIS (midline shift), and delayed cerebral ischemia (DCI) or epileptiform discharges on EEG after SAH. The relationships between sST2 level and outcome measures were assessed. Primary blood mononuclear cells from SAH patients and elective aneurysm controls were analyzed by multiparameter flow cytometry to phenotype the peripheral immune response. Results: In the AIS cohort, baseline plasma sST2 level was associated with the presence of midline shift (46.1 vs. 41.8 ng/mL, p=0.017), which further increased at 48 hr after stroke onset (50.6 vs. 38.4 ng/mL, p=0.006). In the SAH cohort, median sST2 in patients who developed new epileptiform abnormalities was higher compared to patients who did not (114.8 ng/ml vs. 74.7 ng/ml, p=0.024). Higher median sST2 concentration was also observed in those patients with DCI (90.8 vs 53.7ng/mL, p=0.003). In patients with high sST2, flow cytometry identified a peripheral monocyte population with decreased expression of CD14 (4.27x105 ± 2950 A.U. vs. 5.64x105 ± 1290 A.U., p&lt;0.001), and increased expression of CD16 (39,960 ± 272 A.U. vs. 34,869 ± 183 A.U., p&lt;0.001). Conclusion: Plasma sST2 independently predicts edema after stroke, and DCI and epileptiform abnormalities after subarachnoid hemorrhage. Elevated sST2 is also associated with changes in peripheral innate immune cells. These data suggest that sST2 links the innate immune response to secondary brain injury.

Abstract TP476: White Matter Hyperintensity Burden is Associated With Poor Collateral Flow and Worse Outcomes in an Acute Ischemic Stroke Cohort

Introduction: White matter hyperintensity volume (WMHv), a quantitative neuroimaging biomarker of cerebral small vessel disease (CSVD), is associated worse outcomes after ischemic stroke. In this study, we hypothesized that worse outcomes in CSVD patients were due to poor collateral flow during acute ischemia. Methods: 47 patients with acute ischemic stroke (AIS) were prospectively enrolled in this study. Serial MRIs were performed at 3 hours and 30 days after stroke onset. 3-hour FLAIR images were used to determine WMHv, after manually delineating lesions with MIPAV. An index of collateral flow (delayed perfusion to the penumbra) was determined by subtracting core volume (volume of tissue with ADC&lt;600) from the volume of brain tissue with Tmax&gt;2. Patient’s NIHSS was scored at 3 hours and 30 days after stroke onset and the difference was calculated (ΔNIHSS). Log-transformed WMHv was correlated to ΔNIHSS and the collateral flow index, using Pearson correlation. Results: Mean age = 63.9 years (SD 13.5); 37% female; median 3-hour NIHSS = 13 (IQR 6.5-20); median change in NIHSS between 3h and 30d = 4 (IQR: 0-7); median core volume = 13cm3 (IQR 4.3-35.6); median WMHv = 1.257cm3 (IQR 641-3595). WMHv was associated with reduced improvement in ΔNIHSS (R=-0.42, ρ=0.005). Furthermore, WMHv demonstrated a trend for association with poor collateral flow (R=-0.28, ρ=0.062). In this dataset, we will explore the relationship between WMHv and other tissue-based metrics of collateral flow, including the hypoperfusion intensity ratio (HIR) and the cerebral blood volume ratio (rCBV). Conclusions: Our study confirms that patients with CSVD have worse outcomes after AIS. The data also raise the possibility that these worse outcomes in CSVD patients may be mediated by compromised collateral flow in the setting of acute ischemia.