Faculty Bibliography

BACKGROUND:Early prediction of hematoma expansion (HE) following nontraumatic intracerebral hemorrhage (ICH) may inform preemptive therapeutic interventions. We sought to identify how accurately machine learning (ML) radiomics models predict HE compared with expert clinicians using head computed tomography (HCT). METHODS:We used data from 900 study participants with ICH enrolled in the Antihypertensive Treatment of Acute Cerebral Hemorrhage 2 Study. ML models were developed using baseline HCT images, as well as admission clinical data in a training cohort (n = 621), and their performance was evaluated in an independent test cohort (n = 279) to predict HE (defined as HE by 33% or > 6 mL at 24 h). We simultaneously surveyed expert clinicians and asked them to predict HE using the same initial HCT images and clinical data. Area under the receiver operating characteristic curve (AUC) were compared between clinician predictions, ML models using radiomic data only (a random forest classifier and a deep learning imaging model) and ML models using both radiomic and clinical data (three random forest classifier models using different feature combinations). Kappa values comparing interrater reliability among expert clinicians were calculated. The best performing model was compared with clinical predication. RESULTS:The AUC for expert clinician prediction of HE was 0.591, with a kappa of 0.156 for interrater variability, compared with ML models using radiomic data only (a deep learning model using image input, AUC 0.680) and using both radiomic and clinical data (a random forest model, AUC 0.677). The intraclass correlation coefficient for clinical judgment and the best performing ML model was 0.47 (95% confidence interval 0.23-0.75). CONCLUSIONS:We introduced supervised ML algorithms demonstrating that HE prediction may outperform practicing clinicians. Despite overall moderate AUCs, our results set a new relative benchmark for performance in these tasks that even expert clinicians find challenging. These results emphasize the need for continued improvements and further enhanced clinical decision support to optimally manage patients with ICH.

BACKGROUND:Hemorrhagic stroke survivors may have cognitive impairment. We sought to identify preadmission and admission factors associated with cognitive impairment after hemorrhagic stroke. DESIGN/METHODS:Patients with nontraumatic intracerebral or subarachnoid hemorrhage (ICH or SAH) were assessed 3-months post-bleed using the Quality of Life in Neurological Disorders (Neuro-QoL) Cognitive Function short form. Univariate and multivariate analysis were used to evaluate the relationship between poor cognition (Neuro-QoL t-score ≤50) and preadmission and admission factors. RESULTS:Of 101 patients (62 ICH and 39 SAH), 51 (50 %) had poor cognition 3-months post-bleed. On univariate analysis, poor cognition was associated with (p < 0.05): age [66.0 years (52.0-77.0) vs. 54.5 years (40.8-66.3)]; private insurance (37.3 % vs. 74.0 %); BMI > 30 (13.7 % vs. 34.0 %); and admission mRS score > 0 (41.2 % vs. 14.0 %), NIHSS score [8.0 (2.0-17.0) vs. 0.5 (0.0-4.0)], and APACHE II score [16.0 (11.0-19.0) vs. 9.0 (6.0-14.3)]. On multivariate analysis, poor cognition was associated with mRS score > 0 [OR 4.97 (1.30-19.0), p = 0.019], NIHSS score [OR 1.14 (1.02-1.28), p = 0.026], private insurance [OR 0.21 (0.06-0.76), p = 0.017] and BMI > 30 [OR 0.13 (0.03-0.56), p = 0.006]. CONCLUSIONS:Cognitive impairment after hemorrhagic stroke is less common in patients with BMI > 30 and private insurance. Heightened surveillance for non-obese patients without private insurance is suggested. Additional investigation into the relationship between cognition and both BMI and insurance type is needed.

OBJECTIVES/OBJECTIVE:Posterior circulation stroke (PCS) presents diagnostic challenges due to its diverse clinical presentations. Timely detection is crucial, yet a highly sensitive, non-invasive screening tool for PCS is lacking. This study explores gait assessment as a readily accessible diagnostic tool for ruling out PCS in acutely vertiginous patients. MATERIALS AND METHODS/METHODS:In this retrospective case-control study, we examined medical records of 311 acutely vertiginous patients from the Get with the Guidelines Database at an academic hospital in New York City. Of these, 40 were diagnosed with PCS and 271 did not have PCS based on imaging and clinical criteria. We used multivariable logistic regression models and ROC curves to evaluate the association between objective gait abnormality (OGA) and PCS. RESULTS:Objective gait abnormality (OGA) was observed in 38/40 (95 %) posterior circulation stroke (PCS) cases and 57/271 (21 %) controls (adjusted odds ratio 144, 95 %CI 24.4-855, p < 0.0001). In a predictive model, objective gait abnormality (OGA) exhibited excellent discrimination between cases and controls (AUC 0.9599, sensitivity 95.0 %, specificity 75.6 %, positive predictive value 36.5 %, negative predictive value 99.0 %). CONCLUSIONS:Gait assessment emerges as a highly-sensitive screening tool for ruling out posterior circulation stroke (PCS) in acutely vertiginous patients, enabling more efficient triage and patient management. Further prospective research is warranted to validate these findings in larger and more diverse patient populations.

BACKGROUND:Although it is well-known that intracerebral hemorrhage (ICH) is associated with physical and psychological morbidity, there is scant data on factors influencing social engagement after ICH. Understanding the relationship between functionality, psychological outcome and social engagement post-bleed may facilitate identification of patients at high risk for social isolation after ICH. METHODS:Patients ≥18-years-old with non-traumatic ICH from January 2015-March 2023 were identified from the Neurological Emergencies Outcomes at NYU (NEON) registry. Data on discharge functionality were collected from the medical record. 3-months post-bleed, patients/their legally-authorized representatives (LARs) were contacted to complete Neuro-QoL social engagement, anxiety, depression, and sleep inventories. Patients were stratified by ability to participate in social roles and activities (good=T-score>50, poor=T-score≤50) and satisfaction with social roles and activities (high=T-score>50 and low=T-score≤50). Univariate comparisons were performed to evaluate the relationship between post-bleed social engagement and both functionality and psychological outcome using Pearson's chi-square, Fisher's Exact test, and Mann-Whitney U tests. Multivariate logistic regression was subsequently performed using variables that were significant on univariate analysis (p<0.05). RESULTS:The social engagement inventories were completed for 55 patients with ICH; 29 (53 %) by the patient alone, 14 (25 %) by a LAR alone, and 12 (22 %) by both patient and LAR. 15 patients (27 %) had good ability to participate in social roles and activities and 10 patients (18 %) had high satisfaction with social roles and activities. Social engagement was associated with both functionality and psychological outcome on univariate analysis, but on multivariate analysis, it was only related to functionality; post-bleed ability to participate in social roles and activities was associated with discharge home, discharge GCS score, discharge mRS score, and discharge NIHSS score (p<0.05) and post-bleed satisfaction with social roles and activities was related to discharge mRS score and discharge NIHSS score (p<0.05). CONCLUSION/CONCLUSIONS:In patients with nontraumatic ICH, social engagement post-bleed was related to discharge functionality, even when controlling for depression, anxiety, and sleep disturbance.

OBJECTIVES/OBJECTIVE:The objective of this study was to determine factors associated with negative disease-related stigma after hemorrhagic stroke. MATERIALS AND METHODS/METHODS:Patients with non-traumatic hemorrhage (ICH or SAH) admitted between January 2015 and February 2021 were assessed by telephone 3-months after discharge using the Quality of Life in Neurological Disorders (Neuro-QoL) Negative Disease-Related Stigma Short Form inventory. We evaluated the relationship between disease-related stigma (T-score >50) and pre-stroke demographics, admission data, and poor functional outcome (3-month mRS score 3-5 and Barthel Index <100). RESULTS:We included 89 patients (56 ICH and 33 SAH). The median age was 63 (IQR 50-69), 43 % were female, and 67 % graduated college. Admission median GCS score was 15 (IQR 13-15) and APACHE II score was 12 (IQR 9-17). 31 % had disease-related stigma. On univariate analysis, disease-related stigma was associated with female sex, non-completion of college, GCS score, APACHE II score, and 3-month mRS score (all p < 0.05). On multivariate analysis, disease-related stigma was associated with female sex (OR = 3.72, 95 % CI = 1.23-11.25, p = 0.02) and 3-month Barthel Index <100 (OR = 3.46, 95 % CI = 1.13-10.64, p = 0.03) on one model, and female sex (OR = 3.75, 95 % CI = 1.21-11.58, p = 0.02) and 3-month mRS score 3-5 (OR = 4.23, 95 % CI = 1.21-14.75, p = 0.02) on a second model. CONCLUSION/CONCLUSIONS:Functional outcome and female sex are associated with disease-related stigma 3-months after hemorrhagic stroke. Because stigma may negatively affect recovery, there is a need to understand the relationship between these factors to mitigate stroke-related stigma.

OBJECTIVES/OBJECTIVE:Hospital readmissions are associated with poor health outcomes including illness severity and medical complications. The objective of this study was to identify characteristics associated with 30-day post-stroke readmission in an academic urban hospital network. MATERIALS AND METHODS/METHODS:We collected data on patients admitted with stroke from 2017 through 2022 who were readmitted within 30 days of discharge and compared them to a subset of non-readmitted stroke patients. Chart review was used to collect demographics, characteristics of the stroke, co-morbid conditions, in-hospital complications, and post-discharge care. Univariate analyses followed by regression analysis were used to assess characteristics associated with post-stroke readmission. RESULTS:We identified 4743 patients with stroke (18 % hemorrhagic, mean age 70.1 (standard deviation (SD) 17.2), 47.3 % female) discharged from the stroke services, of whom 282 (5.9 %) patients were readmitted within 30 days of index hospitalization. Univariate analyses identified 18 significantly different features between admitted and readmitted patients. Regression analysis revealed characteristics associated with readmission included private insurance (odds ratio (OR) 0.4, confidence interval (CI) 0.3-0.6, p < 0.001), comorbid peripheral vascular disease (PVD) (OR 2.7, CI 1.3-5.5, p = 0.009), malignancy (OR 1.6, CI 1.0-2.6, p = 0.04), seizure (OR 3.4, CI 1.4-8.2, p = 0.007), thrombolytic administration (OR 0.4, CI 0.2-0.7, p = 0.003), undergoing thrombectomy (OR 5.4, CI 2.9-10.1, p < 0.001), and higher discharge modified Rankin Scale score (OR 1.2, CI 1.0-1.3, p = 0.047). CONCLUSIONS:Our data demonstrate that thrombectomy, high discharge Rankin score, comorbid malignancy, seizure or PVD, and lack of thrombolytic administration or private insurance predict readmission.

BACKGROUND AND PURPOSE/UNASSIGNED:Sleep disturbance after hemorrhagic stroke (intracerebral or subarachnoid hemorrhage) can impact rehabilitation, recovery, and quality of life. We sought to explore preclinical and clinical factors associated with sleep disturbance after hemorrhagic stroke assessed via the Quality of Life in Neurological Disorders (Neuro-QoL) short form sleep disturbance inventory. METHODS/UNASSIGNED:We telephonically completed the Neuro-QoL short form sleep disturbance inventory 3-months and 12-months after hemorrhagic stroke for patients >18-years-old hospitalized between January 2015 and February 2021. We examined the relationship between sleep disturbance (T-score >50) and social and neuropsychiatric history, systemic and neurological illness severity, medical complications, and temporality. RESULTS/UNASSIGNED:= .046). CONCLUSION/UNASSIGNED:This exploratory analysis did not demonstrate a sustained relationship between any preclinical or clinical factors and sleep disturbance after hemorrhagic stroke. Larger studies that include comparison to patients with ischemic stroke and healthy individuals and utilize additional techniques to evaluate sleep disturbance are needed.

INTRODUCTION/BACKGROUND:Hematoma expansion (HE) in patients with intracerebral hemorrhage (ICH) is a key predictor of poor prognosis and potentially amenable to treatment. This study aimed to build a classification model to predict HE in patients with ICH using deep learning algorithms without using advanced radiological features. METHODS:Data from the ATACH-2 trial (Antihypertensive Treatment of Acute Cerebral Hemorrhage) was utilized. Variables included in the models were chosen as per literature consensus on salient variables associated with HE. HE was defined as increase in either >33% or 6 mL in hematoma volume in the first 24 h. Multiple machine learning algorithms were employed using iterative feature selection and outcome balancing methods. 70% of patients were used for training and 30% for internal validation. We compared the ML models to a logistic regression model and calculated AUC, accuracy, sensitivity and specificity for the internal validation models respective models. RESULTS:[5.03-18.17] and 25.2% had HE. The best performing model across all feature selection groups and sampling cohorts was using an artificial neural network (ANN) for HE in the testing cohort with AUC 0.702 [95% CI, 0.631-0.774] with 8 hidden layer nodes The traditional logistic regression yielded AUC 0.658 [95% CI, 0.641-0.675]. All other models performed with less accuracy and lower AUC. Initial hematoma volume, time to initial CT head, and initial SBP emerged as most relevant variables across all best performing models. CONCLUSION/CONCLUSIONS:We developed multiple ML algorithms to predict HE with the ANN classifying the best without advanced radiographic features, although the AUC was only modestly better than other models. A larger, more heterogenous dataset is needed to further build and better generalize the models.

BACKGROUND AND PURPOSE/OBJECTIVE:Non-O blood types are known to be associated with thromboembolic complications (TECs) in population-based studies. TECs are known drivers of morbidity and mortality in intracerebral hemorrhage (ICH) patients, yet the relationships of blood type on TECs in this patient population are unknown. We sought to explore the relationships between ABO blood type and TECs in ICH patients. METHODS:Consecutive adult ICH patients enrolled into a prospective observational cohort study with available ABO blood type data were analyzed. Patients with cancer history, prior thromboembolism, and baseline laboratory evidence of coagulopathy were excluded. The primary exposure variable was blood type (non-O versus O). The primary outcome was composite TEC, defined as pulmonary embolism, deep venous thrombosis, ischemic stroke or myocardial infarction, during the hospital stay. Relationships between blood type, TECs and clinical outcomes were separately assessed using logistic regression models after adjusting for sex, ethnicity and ICH score. RESULTS:Of 301 ICH patients included for analysis, 44% were non-O blood type. Non-O blood type was associated with higher admission GCS and lower ICH score on baseline comparisons. We identified TECs in 11.6% of our overall patient cohort. . Although TECs were identified in 9.9% of non-O blood type patients compared to 13.0% in O blood type patients, we did not identify a significant relationship of non-O blood type with TECs (adjusted OR=0.776, 95%CI: 0.348-1.733, p=0.537). The prevalence of specific TECs were also comparable in unadjusted and adjusted analyses between the two cohorts. In additional analyses, we identified that TECs were associated with poor 90-day mRS (adjusted OR=3.452, 95% CI: 1.001-11.903, p=0.050). We did not identify relationships between ABO blood type and poor 90-day mRS (adjusted OR=0.994, 95% CI:0.465-2.128, p=0.988). CONCLUSIONS:We identified that TECs were associated with worse ICH outcomes. However, we did not identify relationships in ABO blood type and TECs. Further work is required to assess best diagnostic and prophylactic and treatment strategies for TECs to improve ICH outcomes.

BACKGROUND AND OBJECTIVES/OBJECTIVE:Mortality after intracerebral hemorrhage (ICH) is common. Neighborhood socioeconomic status (nSES) is an important social determinant of health (SDoH) that can affect clinical outcome. We hypothesize that SDoH, including nSES, contribute to differences in withdrawal of life-sustaining therapies (WLSTs) and mortality in patients with ICH. METHODS:tests. We performed multivariable analysis using backward stepwise logistic regression. RESULTS:≤ 0.01 for both). In multivariable analysis adjusting for age and clinical severity scores, patients who lived in zip codes with high-income levels were more likely to have WLST (adjusted odds ratio [aOR] 1.88; 95% CI 1.29-2.74) and mortality before discharge (aOR 1.5; 95% CI 1.06-2.13). DISCUSSION/CONCLUSIONS:SDoH, including nSES, are associated with WLST after ICH. This has important implications for the care and management of patients with ICH.