Initial Head CT Research Articles

Abstract Number ‐ 215: Distinct presentation of RCVS and PRES co‐existence and overlap of their pathophysiology

Introduction Reversible cerebral vasoconstriction syndrome (RCVS), is characterized by reversible segmental and multifocal vasoconstrictionof cerebral arteries leading to infarcts and cytotoxic edema. Posterior reversible encephalopathy syndrome (PRES) is characterized by vasogenic edema predominantly in the parieto‐occipital regions associated with acute onset encephalopathy, seizures, headache, and visual disturbances.Whether PRES and RCVS are independent syndromes and sometimes overlapped or part of a continuum process, these theories are still debated. We report a case of a patient with distinct overlapping of RCVS and PRES. Methods Case Report Results 52‐year‐old woman with past medical Hx of Hypertension, Diabetes, Migraines and cocaine use, initially presented with hypertensive emergency andsubacute headaches with worsening confusion. Initial CT head was negative for bleed but showed evolving subcortical hypodensities in the right parietal and left frontal lobes. CTA head and neck showed multifocal intracranial irregularities of bilateral ACAs and MCAs with paucity of left anterior M2 branch, left V3/V4 steno‐occlusion. This vasculopathy and clinical presentation is very consistent with RCVS.MRI of brain showed multiple acute / subacute punctate infarcts in the left corona radiata, bilateral basal ganglia, right greater than left thalami, right dorsal pons, right mesial temporal lobe, bilateral occipital regions. Vasogenic edema was noted in the bilateral border zone regions, right parietal being most prominent with some contrast enhancement. The MRI showed vasogenic edema and cytotoxic edema with multiple infarcts which is consistent with PRES and RCVS vasculopathy. CSF analysis was negative for pleocytosis and showed mildly elevated protein. The patient was started on Verapamil and Aspirin and Atorvastatin for secondary stroke prevention. With improvement in mental status was discharged to Rehab. Conclusions The pathophysiological mechanisms of PRES and RCVS are still unknown. Theories whether RCVS and PRES are independent syndromes and sometimes overlap or part of spectrum, are still debated. However, some common characteristics make conceivable a common origin linked with impaired cerebral autoregulation, endothelial dysfunction, and BBB breakdown. Increased clinical awareness of their co‐existence may help in quickerdiagnosis and treatment.

Abstract Number: LBA8 Thin Cut Non‐Contrast Computed Tomography for Periprocedural Planning: Thrombus Burden Assessment Prior to Mechanical Thrombectomy

Introduction In patients presenting with acute ischemic stroke (AIS), non‐contrast CT (NCCT) and CT angiography (CTA) are used to determine large vessel occlusions (LVO). Using thin‐cut NCCT, hyperdense signs represent thrombus. NCCT provides information on thrombus characteristics such as length and morphology that is not evidenced by CTA. The aim of this study is to determine if hyperdense signs identified on thin‐cut NCCT are a valid tool in periprocedural planning for endovascular thrombectomy (EVT). Methods At our comprehensive stroke center, patients presenting with AIS who had thin‐cut NCCT defined as 0.625mm per slice followed by EVT and subsequent High Resolution Photographs (HRPs) of post‐EVT thrombus were identified. High‐resolution photographs were taken by operating physician with adjacent ruler for accurate measurement. Thin‐cut NCCT and associated fully extracted thrombus HRPs were reviewed by two board certified vascular neurologists to determine agreement. Cohen’s K was used to determine kappa inter‐rater agreement. Wilcoxon signed‐rank test was used to determine if significant difference existed between thrombus length and independently measured hyperdense sign. Social science statistics software was used for data analysis. Results From January 2019 to December 2021, out of 87 cases where thrombus was extracted after EVT, 57 met inclusion criteria and had associated initial CT head. Mean age was 68.22 (95% CI 64.26, 72.18), and 49% were female (n = 28). Mean thrombus length was 12.49mm (95% CI 10.19, 14.80). Mean hyperdense sign length measured by Interpreter 1 was 12.86mm (95% CI 10.54, 15.18). Mean hyperdense sign length measured by Interpreter 2 was 12.61mm (95% CI 10.40, 14.83). Kappa score was 1. There was no significant difference in thrombus length and Interpreter 1 hyperdense sign (z = 0.99; p‐value 0.317) or Interpreter 2 hyperdense sign (z = ‐0.92; p‐value = 0.36). Conclusions Our study suggests that in acute ischemic stroke patients, hyperdense sign on thin‐cut NCCT may be a valid reliable marker of thrombus length and morphology that assists in peri‐procedural planning for mechanical thrombectomy. Larger, prospective studies are needed to validate our results.

Abstract WMP72: Acute Ischemic Stroke In Asians In A Large Integrated Health Care Delivery System

Background: Stroke is a major problem in Asia where more than half of the world population resides. Asian Americans are among the fastest growing racial-ethnic groups in the US, yet the incidence of stroke subtypes, stroke risk factors and outcomes have been largely unstudied in Asian Americans. We assessed acute stroke alert presentations in Asians compared to whites in a large integrated health care delivery system. Methods: In 2016, Kaiser Permanente Northern California launched a standardized telestroke program for 21 certified stroke centers to include immediate video evaluation by a teleneurologist. Non-cancelled stroke alerts included those patients who were potential candidates for further workup for acute ischemic stroke treatment. We evaluated all non-cancelled stroke alerts evaluated by Telestroke in 2019. Assessment included demographics, mode of arrival, initial NIHSS, neuroimaging results, IV thrombolytic, door-to-needle (DTN) time, large vessel occlusion (LVO), thrombectomy referral, and discharge disposition. Results: We identified 427 Asian and 1262 White patients with potential acute ischemic stroke. Compared with whites, Asians were younger, just as likely to arrive via EMS and arrived quicker from their last known well time, but presented with more severe NIHSS, were more likely to have bleeding on the initial head CT, and less likely to receive IV thrombolytic within 30 minutes from ED arrival (Table). Asians had less detected LVO, but those with LVO were transferred out for thrombectomy faster compared with whites. There were less stroke mimics in Asians presenting with stroke alerts, and no difference in inpatient mortality compared with whites. Conclusions: Several differences existed between Asian and white patients evaluated by Telestroke as non-cancelled stroke alerts. Further research is needed to understand these differences and assess outcomes for Asians as well as major Asian subgroups presenting with acute ischemic stroke.

Abstract WMP55: Up-front Ct Angiography Of The Head And Neck Imaging For Patients With Low NIH Stroke Scale: Is There Utility?

Background: Current guidelines encourage early vascular imaging to assess for large vessel occlusions (LVO) in patients presenting with NIHSS ≥ 6 to establish candidacy for mechanical thrombectomy (MT). In patients with low NIHSS, the benefit of early vascular imaging is uncertain. Cortical signs such as aphasia, neglect or visual field deficits on the exam have been used to identify patients with LVO. Still, the utility of these clinical signs in low NIHSS is unknown. Methods: This is a retrospective analysis of all acute ischemic stroke patients who presented directly to our comprehensive stroke center (CSC) from December 1, 2020-May 31, 2021. All patients had CT angiography (CTA) or MR angiography (MRA) of the head and neck with timing at the discretion of the treating neurologist. Results: Of the 167 patients [median age 70 IQR (58-78), 50% female, 64% African-American], 85 (51%) presented with NIHSS of 0-5. Compared to patients with NIHSS ≥ 6, patients with NIHSS of 0-5 were younger (median age 66 vs. 70, p= 0.039) and had less congestive heart failure (0.6% vs. 4.8%, p=.014). Upfront CTA with initial non-contrast CT head was done less in patients with NIHSS 0-5 (65% vs. 85%, p=.002). Patients with NIHSS 0-5 had fewer LVO (13% vs. 44%, p<.001), less IV tPA administration (7.1% vs. 22%, p=.006), less MT (2.4% vs. 27%, p<.001), and more posterior circulation strokes (42% vs. 13%, p<.001). Of the 11 patients with NIHSS 0-5 and an LVO, 4 (36%) had no cortical signs (aphasia, extinction, or visual field deficit) on the exam. Conclusions: The absence of cortical signs on exam in ischemic stroke patients presenting with NIHSS 0-5 does not exclude LVO. Our cohort with NIHSS 0-5 were less likely to have upfront CTA despite having LVO in 13% of cases. Early upfront vascular imaging should be considered to assess for LVO in all suspected acute ischemic stroke patients regardless of the presence of cortical signs.

A quantitative model to differentiate nonaneurysmal perimesencephalic subarachnoid hemorrhage from aneurysmal etiology.

Nonaneurysmal perimesencephalic subarachnoid hemorrhage (pmSAH) is considered to have a lower-risk pattern than other types of subarachnoid hemorrhage (SAH). However, a minority of patients with pmSAH may harbor a causative posterior circulation aneurysm. To exclude this possibility, many institutions pursue exhaustive imaging. In this study the authors aimed to develop a novel predictive model based on initial noncontrast head CT (NCHCT) features to differentiate pmSAH from aneurysmal causes. The authors retrospectively reviewed patients admitted to an academic center for treatment of a suspected aneurysmal SAH (aSAH) during the period from 2016 to 2021. Patients with a final diagnosis of pmSAH or posterior circulation aSAH were included. Using NCHCT, the thickness (continuous variable) and location of blood in basal cisterns and sylvian fissures (categorical variables) were compared between groups. A scoring system was created using features that were significantly different between groups. Receiver operating characteristic curve analysis was used to measure the accuracy of this model in predicting aneurysmal etiology. A separate patient cohort was used for external validation of this model. Of 420 SAH cases, 48 patients with pmSAH and 37 with posterior circulation aSAH were identified. Blood thickness measurements in the crural and ambient cisterns and interhemispheric and sylvian fissures and degree of extension into the sylvian fissure were all significantly different between groups (all p < 0.001). The authors developed a 10-point scoring model to predict aneurysmal causes with high accuracy (area under the curve [AUC] 0.99; 95% CI 0.98-1.00; OR per point increase 10; 95% CI 2.18-46.4). External validation resulted in persistently high accuracy (AUC 0.97; 95% CI 0.92-1.00) of this model. A risk stratification score using initial blood clot burden may accurately differentiate between aneurysmal and nonaneurysmal pmSAH. Larger prospective studies are encouraged to further validate this quantitative tool.

Delayed Subarachnoid Hemorrhage in A Patient on Warfarin with Blunt Head Trauma

Closed head trauma in anticoagulated patients is an increasing common presentation to Emergency Department (ED). Within this population, there is a subset of patients who present with delayed intracranial haemorrhage, typically resulting in subdural or parenchymal bleeds, with negative initial non-contrast head CT.

314: EXTRACORPOREAL LIVER ASSIST DEVICE FOR ACUTE LIVER INJURY DUE TO LIKELY PARTIAL UREA CYCLE DISORDER

Introduction: Extracorporeal liver assist device (ELAD) is used to bridge patients in acute liver failure to transplant or in severe acute liver injury until spontaneous recovery. ELAD employs an albumin circuit to dialyze albumin-bound toxins from the blood stream. Urea is an albumin bound molecule formed in the liver as a result of nitrogen breakdown. Urea Cycle Disorders (UCDs) are a rare group of genetic enzymatic deficiencies that typically manifest early in life but partial enzymatic deficiencies can present later in life in times of metabolic stress. Description: A 38-year-old female presented with progressive encephalopathy 2 months after being diagnosed with alcoholic hepatitis at which time she had stopped drinking alcohol. She had no prior liver disease. Initial labs showed an elevated ammonia level of 87 mcmol/L, despite decrease in total bilirubin from 33 mg/dL to 3.5mg/dL, and INR of 1.9 to 1.4 since the peak of her alcoholic hepatitis. She had no imaging evidence of cirrhosis and had normal aminotransferases, platelets, and renal function. Initial head CT, infectious evaluation, and work-up for alternate etiologies of acute liver injury were unremarkable. She was intubated for progressive hepatic encephalopathy on hospital day 9. Her ammonia level significantly rose to 203 mcmol/L and a repeat head CT showed diffuse cerebral edema. Due to a rising ammonia level to 247 mcmol/L and CT evidence of progressive cerebral edema despite mannitol, therapeutic hyperventilation, and sedation. She was started on MARS (TM) therapy on hospital day 11. After one 7-hour session, her ammonia level decreased to 101 mcmol/L, mental status improved and head CT showed improved cerebral edema. A urinary orotic acid level was elevated at 1.8 mmol per mol of creatinine. Discussion: A partial UCD was suspected due to atypical time course of encephalopathy and hyperammonemia in the context of improving liver synthetic function and absence of cirrhosis. Genetic testing is pending however, a partial UCD is likely given the elevated urinary orotic acid. A single case report exists of confirmed acquired glutamine synthetase deficiency after orthotopic liver transplant that also improved after a single MARS session. Our report adds to the scarce literature on this topic.

384: ACTINOMYCES BRAIN ABSCESS AS SEQUELAE FROM PULMONARY ARTERIOVENOUS MALFORMATION: A CASE REPORT

Introduction: Most bacterial brain abscesses are a result of direct spread of bacteria, only a third are through hematogenous dissemination. A less common cause of hematogenous spread is via paradoxical embolism in the presence of a pulmonary arteriovenous malformation (PAVM). Currently PAVMs are not routinely screened for. Description: A 51-year-old male with past medical history of obstructive sleep apnea, and transient ischemic attack presented with left sided weakness. Initial CT head was concerning for subacute stroke. Patient’s physical exam was significant for 2/5 strength in left upper extremity, oral mucosa and skin findings were negative for telangiectasias. No significant family history reported. Initial white count negative. Worsening left sided numbness led to a repeat CT head which showed a 1.5 cm right posterior frontal/anterior mass lesion with surrounding edema. Patient underwent biopsy and drainage of brain abscess on hospital day 4, intraoperative pus was encountered. Patient was treated with ceftriaxone, metronidazole, and vancomycin. Cultures grew Actinomyces meyeri. On hospital day 8 a CT angiogram showed a PAVM in the medial middle lobe. Patient underwent right pulmonary artery angiogram on hospital day 12 which showed that the PAVM was fed by a branch of the right middle lobe pulmonary artery. After embolization follow-up angiogram showed cessation of flow through the PAVM. An MRI showed significantly less brain swelling around the cavity, based off the patient’s clinical and laboratory improvements patient was deemed stable for discharge to rehab facility on IV antibiotics. Discussion: A PAVM typically consists of afferent flow from a branch of the pulmonary artery to a branch of the pulmonary vein. Serious outcomes including strokes, seizures, and brain abscesses can arise from the right-to-left shunting of blood and paradoxical emboli. Most PAVMs are congenital associated with hereditary hemorrhagic telangiectasia, however, sporadic ones occur. Infections caused by Actinomyces are thought to induce hyperplastic changes in bronchial arteries which might lead to PAVMs. A PAVM should be considered in patients with a bacterial brain abscess and no identifiable site of infection. Treatment for PAVM include endovascular embolization vs thoracic surgery for removal.

ODP299 Cabergoline Administration Markedly Improves Cognitive Impairment Associated with Macroprolactinoma

Abstract Introduction Prolactinomas classically present with galactorrhea, infertility, or hypogonadal symptoms. However cognitive impairment is rare presentation. We describe a man with a macroprolactinoma presenting with significant cognitive dysfunction which improved remarkably with cabergoline treatment. Case History A 21-year-old man with no significant past medical history presented to the hospital for evaluation of seizure-like activity. His father noted him limping earlier in the day and witnessed him fall and hit his head, while his brother reported involuntary convulsions and decreased awareness earlier that week. The patient himself had been complaining of blurry vision for 3 months for which he was prescribed glasses. Initial head CT demonstrated a 3.4×4.3×3.4cm cystic mass in the sella, 3.9×3.2×2.6 cm mass in the right basal ganglia and 2.7×2.3×3cm mass in the right temporal lobe. His prolactin was elevated at 16326 ng/mL. He was empirically started on dexamethasone due to concern for cerebral edema and was admitted. Surgery was recommended given the severity of his presentation, but the patient declined and he was discharged on cabergoline 1.5mg/week, dexamethasone taper, hydrocortisone 15mg/day, and levetiracetam 1000mg twice daily. On initial outpatient endocrinology evaluation 3 weeks later, he required the use of a wheel chair. Examination was notably for markedly flat affect, slurred speech, decreased attention span requiring frequent re-direction during the interview but he was otherwise oriented to person, time, and place. Neurological examination was noted for 4/5 muscle weakness throughout, facial asymmetry, decreased facial sensation, and an antalgic gait. He was recommended for expedited neurosurgical evaluation but he again declined. Laboratory drawn at this time showed a prolactin of 1946 ng/mL and total testosterone of 45 ng/dL, but he had normal thyroid and growth hormone axis (adrenal axis was not assessed as he was on dexamethasone taper). Carbergoline was titrated with further improvement in his prolactin levels. Four months after initial presentation, he reported enrolling in college with resolution of any cognitive and neurological deficits. Fourteen months after presentation, serum prolactin was 34 ng/mL on 4mg cabergoline per week and he is active in college and regularly plays basketball. Discussion Prolactinomas are incidentally detected or found as a result of workup for galactorrhea, infertility, or low libido. Typically, goals for treatment are to relieve compressive effects along the optic chiasm or to restore gonadal function. Acute cognitive decline should also be considered when there is a macroprolactinoma, and serial monitoring for improvement may be considered a clinical marker along with MRI imaging and prolactin levels. In this case, the patient had marked cognitive dysfunction on initial presentation but after 3 months of treatment had almost complete recovery as evidenced by his re-enrolling in college classes successfully. Presentation: No date and time listed

Evaluation of Sequential Head Computed Tomography in Traumatic Brain Injuries.

Background: The grading of the severity of head trauma plays a vital role in acute patient management and planning a case-appropriate follow-up protocol. Few studies have been published regarding the Rotterdam scoring. In this study, we have established a correlation between the Rotterdam scores, need for sequential CTs, and the cumulative radiation dose. This correlation has helped develop a preliminary protocol that can be followed for patients hence bringing about better planned and efficient patient care.Materials and methods: From August 2014 to December 2020, 88 cases of traumatic head injury on whom a minimum of one sequential CT was performed, with no surgical intervention, were included and studied. Sequential head CTs of each patient were evaluated by skilled radiologists with a minimum experience of five years, all of whom were blinded to the findings of the initial and previous head CT findings. The serial head CTs were evaluated for the Rotterdam CT score (RCTS). Results: Among the patients with extradural hemorrhage (EDH), only 28.6% (8) progressed over successive CTs and 75.5% (34) of patients with subdural hemorrhage (SDH) showed significant progression over sequential CTs. Maximum number of serial CTs were obtained for cases presenting with a score of 3 (34 cases) with about three of them requiring up to a total of three CTs. However, no significant change in findings was noted on serial CTs. On the contrary, significant disease progression was noted in patients with baseline scores of 4 (76.9%) and 5 (100%), with statistical significance obtained on further analysis (P = 0.001).Conclusions: We are of the opinion that there is no additional role of sequential CT for the cases with Rotterdam score of 1 or 2 in the initial CT unless there is clinical evidence of deterioration. Rotterdam score 3 needs sequential CT after 24 hours and Rotterdam scores 4 and 5 need sequential CT after 12 hours if surgical intervention is delayed. The Rotterdam score may help predict any further need for a second CT, hence decreasing the unwanted radiation exposure.

Traumatic Intracranial Hemorrhage on CT After Ground-Level Fall in Adult Patients Receiving Antithrombotic Therapy: A Retrospective Case-Control Study.

BACKGROUND. Antithrombotic medications may increase the risk and severity of traumatic intracranial hemorrhage (tICH) after minor head trauma. OBJECTIVE. The purpose of this study was to determine the frequency, distribution, and clinical course of tICH in patients receiving antithrombotic therapy who present with good neurologic status after a ground-level fall. METHODS. This retrospective study included 1630 patients (693 women and 937 men; mean age, 80.2 ± 12.7 [SD] years) who underwent head CT after presenting to the emergency department after a ground-level fall between January 1 and December 31, 2020; all patients had a Glasgow Coma Scale score of 14 or higher and no focal neurologic deficit. Patients with tICH were identified on the basis of clinical reports. In patients with tICH, images from initial head CT examinations were reviewed for characteristics of tICH, images from follow-up head CT examinations (performed within 24 hours) were reviewed for hematoma expansion, and clinical outcomes were extracted from medical records. Patients receiving antithrombotic therapy and control patients (those not receiving antithrombotic therapy) were compared. RESULTS. The antithrombotic therapy group included 954 patients (608 receiving anticoagulant therapy; 226, antiplatelet therapy; and 120, both therapies); the control group included 676 patients. A total of 63 patients (3.9%; 95% CI, 2.9-4.8%) had tICH. The antithrombotic therapy and control groups were not significantly different in terms of the frequency of tICH (4.4% vs 3.1%, p = .24), midline shift (10.0% vs 7.1%, p = .76) or regional mass effect (33.3% vs 14.3%, p = .19) on initial CT. Hematoma expansion on follow-up CT occurred in 11 of 42 patients (26.2%) in the antithrombotic group and one of 21 patients (4.8%) in the control group (p = .04). Two patients required neurosurgical intervention, and three deaths related to tICH occurred within 30 days; all five of these patients were receiving antithrombotic therapy. CONCLUSION. Antithrombotic therapy use was not associated with an increased frequency of tICH, although it was associated with an increased frequency of hematoma expansion at follow-up. CLINICAL IMPACT. In patients with good neurologic status after a ground-level fall, the findings suggest application of a similar strategy for selecting patients for initial head CT, regardless of antithrombotic therapy use; if initial head CT shows tICH, early follow-up head CT should be systematically performed in those receiving antithrombotic therapy, though it possibly should be deferred in other patients.

Abstract WP5: Bundling Of Head Computed Tomography Scan And Head And Neck Computed Tomography Angiogram At Initial Telestroke Consultation Is Associated With Faster Thrombectomy Transfer And Treatment Times

Introduction: Telestroke has expanded access to tPA and thrombectomy for acute ischemic stroke and improved neurologic outcomes. A key step in the evaluation of acute stroke is non-contrast head CT (HCT) and head and neck CT angiogram (CTA). These imaging studies can either be bundled at presentation, or a delayed CTA may be performed after a tPA decision based on HCT alone. Bundled imaging may facilitate faster transfer and treatment times, while delaying tPA. We investigated the effect of bundled imaging on thrombectomy transfer and treatment times in our telestroke network. Methods: We reviewed patients in our telestroke network from 2019-2021. We included patients who had received HCT and CTA at a partner hospital prior to transfer to our comprehensive stroke center (CSC) for thrombectomy. We collected baseline clinical variables, and timing of imaging, transfer and treatment relative to time of initial HCT. Patients were divided into a bundled imaging group (simultaneous HCT and CTA) or a delayed CTA group. Wilcoxon Rank sum test was used to compare between the two groups for HCT to tPA needle time, HCT to CSC arrival time, and HCT to Groin Puncture Time Results: 273 patients were assessed via telestroke in the study period. Of these 93 were included. 43 patients received bundled CTA, and 50 patients received a delayed CTA. 20 patients received tPA in the bundled group vs 25 in the delayed CTA group. Median CSC transfer time was lower with a bundled imaging strategy (110 min vs 160 min; p&lt;0.001), as was median time to groin puncture (174 min vs 242 min; p&lt;0.001). Further, there was no difference in tPA needle time (38 min vs 37 min; p=0.86). Conclusions: Bundled HCT and CTA at telestroke sites was associated with significantly faster transfer and groin puncture times for patients with large vessel occlusion. Further, obtaining the CTA prior to a tPA decision did not significantly prolong HCT-to-needle times.

Natural language processing of head CT reports to identify intracranial mass effect: CTIME algorithm.

The Mortality Probability Model (MPM) is used in research and quality improvement to adjust for severity of illness and can also inform triage decisions. However, a limitation for its automated use or application is that it includes the variable "intracranial mass effect" (IME), which requires human engagement with the electronic health record (EHR). We developed and tested a natural language processing (NLP) algorithm to identify IME from CT head reports. We obtained initial CT head reports from adult patients who were admitted to the ICU from our ED between 10/2013 and 9/2016. Each head CT head report was labeled yes/no IME by at least two of five independent labelers. The reports were then randomly divided 80/20 into training and test sets. All reports were preprocessed to remove linguistic and style variability, and a dictionary was created to map similar common terms. We tested three vectorization strategies: Term Frequency-Inverse Document frequency (TF-IDF), Word2Vec, and Universal Sentence Encoder to convert the report text to a numerical vector. This vector served as the input to a classification-tree-based ensemble machine learning algorithm (XGBoost). After training, model performance was assessed in the test set using the area under the receiver operating characteristic curve (AUROC). We also divided the continuous range of scores into positive/inconclusive/negative categories for IME. Of the 1202 CT reports in the training set, 308 (25.6%) reports were manually labeled as "yes" for IME. Of the 355 reports in the test set, 108 (30.4%) were labeled as "yes" for IME. The TF-IDF vectorization strategy as an input for the XGBoost model had the best AUROC:-- 0.9625 (95% CI 0.9443-0.9807). TF-IDF score categories were defined and had the following likelihood ratios: "positive" (TF-IDF score>0.5) LR=24.59; "inconclusive" (TF-IDF 0.05-0.5) LR=0.99; and "negative" (TF-IDF<0.05) LR=0.05. 82% of reports were classified as either "positive" or "negative". In the test set, only 4 of 199 (2.0%) reports with a "negative" classification were false negatives and only 8 of 93 (8.6%) reports classified as "positive" were false positives. NLP can accurately identify IME from free-text reports of head CTs in approximately 80% of records, adequate to allow automatic calculation of MPM based on EHR data for many applications.

Role of Routine Repeat Head CT for Pediatric Patients under 2 Years Old with Mild-to-moderate Traumatic Brain Injury.

Repeat head computed tomography (RHCT) is common and routine for pediatric traumatic brain injury (TBI) patients. In mild (Glasgow Coma Scale; GCS 13–15) to moderate (GCS 9–12) TBI, recent studies have shown that RHCT without clinical deterioration does not alter management. However, the effectiveness of routine RHCT for pediatric TBI patients under 2 years has not been investigated. This study aims to investigate whether routine RHCT changes management in mild-to-moderate TBI patients under 2 years. We performed a retrospective review at the emergency department of the National Center for Child Health and Development between January 2015 and December 2019. Mild-to-moderate TBI patients under 2 years with an acute intracranial injury on initial head CT scan and receiving follow-up CT scans were included. Mechanism, severity of TBI, indication for RHCT, and their findings were listed. Study outcome was intervention based on the findings of RHCT. Intervention was defined as intubation, ICP monitor placement, or neurosurgery. We identified 50 patients who met inclusion criteria and most patients (48/50) had mild TBI. The most common mechanism was ‘fall’ (68%). Almost all RHCT was routine and the overall incidence of radiographic progression on RHCT was 12%. RHCT without clinical deterioration did not lead to intervention, although one patient with moderate TBI required intervention due to radiographic progression with clinical symptoms. Our study showed that routine RHCT without clinical deterioration for mild TBI patients under 2 years may not alter clinical management. We suggest that RHCT be considered when there is clinical deterioration such as decrease in GCS.

Abstract 1122‐000021: Delayed Vasospasm and Cerebral Ischemia Following Mild Traumatic Brain Injury

Introduction : Post‐traumatic vasospasm (PTV) is a significant cause of morbidity and mortality following traumatic brain injury (TBI). 1 Delayed PTV is thought to occur due to inflammation from SAH. 1 The risk of symptomatic PTV is associated with the severity of TBI. 1,2,3 Treatment of PTV traditionally involves agents used in aneurysmal vasospasm such as Nimodipine. Intra‐arterial and intravenous verapamil and milrinone have been utilized to treat PTV. 4 We present a rare case of delayed cerebral ischemia caused by PTV without SAH treated with intra‐arterial milrinone and oral verapamil. Methods : This is a case report of a case of a 16‐day delay of cerebral ischemia secondary to PTV. Results : A 19‐year‐old female without significant medical history presented to the emergency room as the restrained driver in a motor vehicle collision involving a car versus a tree. The patient’s Glasgow Coma Scale score was 13 and the initial head CT did not demonstrate SAH with a Rotterdam score of 0. There were multiple fractures and soft tissue contusions noted on imaging which required surgical correction and splinting. The patient’s hospital stay was complicated by a large retroperitoneal hematoma requiring blood transfusions and surgical evacuation. 16 days after admission, the patient experienced sudden‐onset right upper extremity paralysis and weakness of the right leg. CT head demonstrated loss of gray‐white differentiation in the left middle cerebral artery (MCA) territory and CT angiography demonstrated 70% stenosis of the left supra‐clinoid internal carotid artery (ICA) and proximal left MCA. Cerebral angiography demonstrated 60% stenosis left supra‐clinoid ICA and 60% stenosis at the origin of the left MCA consistent with PTV. Left MCA stenosis improved to 20% post 10 mg intra‐arterial milrinone in the left ICA (Figure 1). The patient also received verapamil orally. Transcranial doppler demonstrated elevated peak systolic velocities at 298 cm/s in the left MCA and 276 cm/s in the left ICA. Six days later this improved to 150 cm/s in the left MCA and 151 cm/s in the left ICA. The patient continued to be unable to move her right side against gravity with dense expressive aphasia and dysarthria at discharge to inpatient rehabilitation. At follow‐up 4 weeks later, her right hemiparesis had improved significantly to being able to stand with assistance and her expressive aphasia had improved from one word to occasionally forming several word sentences. Conclusions : PTV is a potentially devastating complication of TBI. As our case demonstrates, the presence of mild TBI and absence of SAH may be falsely reassuring. 2 Additionally, our case report demonstrates that intra‐arterial milrinone causes radiographic improvement in PTV. Further studies are needed for the best screening and diagnostic exams for PTV and therapeutic interventions.

Abstract 1122‐000201: Is Neutrophilia a Risk Factor for Fast Stroke Progression in Large Vessel Occlusion?

Introduction : Rapid stroke progressors with large vessel occlusion (LVO) have a worse prognosis than their time‐matched cohorts receiving IV thrombolytics and/or mechanical thrombectomy. Our objective was to evaluate the association of neutrophilia with rapid stroke progression. Methods : Initial white blood cell (WBC) and absolute neutrophil counts (ANC) were collected for subjects presenting with acute ischemic stroke secondary to LVO who received IV thrombolytics and/or mechanical thrombectomy within 4.5 and 6 hours, respectively. Rapid stroke progression was determined by Alberta Stroke Program Early CT Score (ASPECTS) on initial CT head. Baseline and discharge NIHSS, age, and follow up mRS were also compared to presenting WBC and ANC. Spearman’s rho was used for correlation. Social Science Statistics was used for data analysis. Results : From October 2020 to April 2021, the association between neurophilia and stroke progression was evaluated in 19 subjects receiving tenecteplase (n = 16; 6 females; age, 63.25 95% CI [54.9207, 71.5793]) and alteplase (n = 4; 2 females; mean age 59, 95% CI [38.13, 79.87]) for LVO causing disabling neurological deficits. Mechanical thrombectomy was attempted in all subjects. The association between higher ANC and lower ASPECTS score reached statistical significance (rs = ‐0.49255, p = 0.04457). There was no significant association of white blood cell (WBC) and ASPECTS score. WBC and ANC were not associated with baseline or discharge NIHSS, age, or follow up mRS. Conclusions : Rapid stroke progression as measured by presenting ASPECTS score may be associated with neutrophilia. Larger prospective clinical trials are needed to validate our results.

POCUS OBTAINED OPTIC NERVE SHEATH DIAMETER IN PATIENTS PRESENTING WITH INTRACEREBRAL HEMORRHAGE

TOPIC: Critical Care TYPE: Original Investigations PURPOSE: Intracerebral hemorrhage (ICH) results in spontaneous bleeding into the brain. This accounts for 10-15% of all strokes in the United States and has poor outcomes due to death or disability. An ICH can result in increased intracranial pressure due to blood occupying extra space. Optic nerve ultrasonography has emerged as a quick noninvasive procedure to detect increased intracranial pressure. A question that has not been answered is whether there is a specific correlation between the size of hematoma in the brain and the optic nerve sheath diameter (ONSD). The optic nerve is wrapped by a sheath derived from the meninges. Cerebrospinal fluid (CSF) can transfer freely between the intracranial and intraorbital spaces. METHODS: 23 patients were recruited into this study. Each of these patient’s received the diagnosis of ICH and admitted to the neuro intensive care unit (NICU). Ultrasound of the optic nerve was performed using a Sonosite machine and the high frequency probe. The exam was performed upon admission to the NICU. Both eyes were scanned in two planes, longitudinal and transverse. Once the ONSD measurements were recorded, the volume of the ICH was noted on initial head CT. RESULTS: The average ICH score was 1.35 and the average APACHE IV score was 27.35. In addition, the average GCS on admission was 13. The most common cause of ICH was hypertension. The most common location for the ICH was lobar hemorrhage. 8 out of 23 patients had midline shift on their initial head CT. 4 out of 23 patients underwent some sort of surgical intervention. The smallest average combined ONSD was 4.43cm. The longest average combined ONSD was 8.03cm. The minimum hemorrhage volume was 0.20cm3. The maximum was 60.0cm3. A moderate, positive correlation was detected between average optic nerve size and hemorrhage volume (correlation=0.5262; p-value=0.0099). A weak positive correlation was detected between APACHE IV and average ONSD (correlation=0.2052; p-value=0.3476). A weak-moderate positive correlation was detected between ICH score and average ONSD. (correlation=0.3622; p-value=0.0895). Those who experienced a midline shift had significantly higher averaged ONSD compared to those who did not. CONCLUSIONS: This study determined there was a moderate positive correlation between the ONSD and volume of hematoma. From a pathophysiological standpoint, conditions like ICH lead to cerebral parenchyma swelling which is transmitted to the optic nerves through the meninges. Additionally, our study revealed a significant difference in average ONSD and midline shift observed on CT scan. In settings where no invasive monitoring is available, ONSD enlargement may be a useful technique for monitoring midline shift. CLINICAL IMPLICATIONS: There are already established prognostic indicators for ICH including an ICH volume of 60cm3or greater on initial CT along with a GCS of 9 or less which predict a 30 day mortality of 19%. ONSD has a role in prognostication. Growth of the hematoma in the first 24 hours is an independent predictor of mortality and poor outcome. Serial exams of a patients ONSD can be applied to indirectly monitor any expanding hematoma. Given the positive correlation between ONSD and hematoma volume, I would be interested in ONSD and the degree of disability scores such as the modified rankin scale. DISCLOSURES: No relevant relationships by Braden Anderson, source=Web Response No relevant relationships by Shekhar Ghamande, source=Web Response No relevant relationships by Sarita Kambhampati, source=Web Response No relevant relationships by Matthew Lohse, source=Web Response No relevant relationships by Alireza Nathani, source=Web Response

A RARE CAUSE OF PARAPLEGIA: DISTRIBUTIVE SHOCK CAUSING BORDER ZONE STROKES

TOPIC: Critical Care TYPE: Medical Student/Resident Case Reports INTRODUCTION: Simultaneous bilateral cerebrovascular infarction is relatively uncommon and is an entity seen rarely in watershed cerebral infarctions which occur at the border between cerebral vascular territories in the setting of severe hypotension. CASE PRESENTATION: 84-year-old male with history of CAD, cerebellar stroke, peripheral arterial disease was admitted to the ICU with distributive shock and intubated for toxic metabolic encephalopathy. Initial CT head on admission was unremarkable. On day 2 of hospitalization his neuro assessment was notable for bilateral lower extremity paresis with preservation of sensation, proprioception and temperature but had sluggish reflexes and loss of motor strength. Repeat CT head did not show a bleed and subsequently, MRI brain with MRA head and neck revealed bilateral acute anterior cerebral artery strokes. Given the location and duration of symptoms he was not a TPA candidate and was treated with medical management with aspirin and statin. Echocardiogram and carotid ultrasounds were done that were unremarkable for embolic cause of stroke. DISCUSSION: The anterior cerebral artery is a major vessel supplying blood to the interhemispheric region. ACA territory infarction accounts for only 0.3% to 4.4% of cerebral infarctions. Bilateral ACA infarction is even rarer. Twenty-seven cases of ACA territory infarction were reported among 1490 cases of cerebral infarction in the Lausanne Stroke Registry; among them, only two cases had bilateral ACA territory infarction. Bilateral ACA infarction can occur due to vasospasm in the setting of subarachnoid hemorrhage caused by a rupture of aneurysms of the anterior communicating arteries or distal ACAs, thrombosis or embolism in case of anomaly of circle of Willis or bilateral anaplastic ACAs. In our patient the area of infarction was in a watershed zone bilaterally, unfortunately affected by hypotension which is rare in the absence of any of the above-mentioned anomalies as predisposing factors. The true incidence of bilateral ACA infarction is unknown, with few cases reported in the literature. CONCLUSIONS: This case report highlights the possibility of developing simultaneous bilateral cerebral infarctions in the setting of profound systemic hypotension and shock, especially if the patient has baseline chronic small vessel changes. REFERENCE #1: Orlandi G, Moretti P, Fioretti C, Puglioli M, Collavoli P, Murri L. Bilateral medial frontal infarction in a case of azygous anterior cerebral artery stenosis. Ital J Neurol Sci. 1998;19:106–108. doi: 10.1007/BF02427567. [PubMed] [CrossRef] [Google Scholar] REFERENCE #2: Gacs G, Fox AF, Barnett HJ, Vinuela F. Occurrence and mechanisms of occlusion of the anterior cerebral artery. Stroke. 1983;14:952–959. [PubMed] [Google Scholar] REFERENCE #3: Bogousslavsky J, Regli F. Anterior cerebral artery territory infarction in the Lausanne Stroke Registry. Clinical and etiologic patterns. Arch Neurol. 1990;47:144–150. [PubMed] [Google Scholar] DISCLOSURES: No relevant relationships by Imama Ahmad, source=Web Response No relevant relationships by Christine Blaski, source=Web Response No relevant relationships by Sneha Lakshman, source=Web Response

Predictors of neurosurgical intervention in complicated mild traumatic brain injury patients: a retrospective cohort study

ABSTRACT Objectives To determine the predicting demographic, clinical and radiological factors for neurosurgical intervention in complicated mild traumatic brain injury (mTBI) patients. Methods Design: retrospective multicenter cohort study. Participants: patients aged ≥16 presenting to all level-I trauma centers in Quebec between 09/2016 and 12/2017 with mTBI(GCS 13–15) and complication on initial head CT (intracranial hemorrhage/skull fracture). Procedure: Consecutive medical records were reviewed and separated into two groups: no neurosurgical intervention and neurosurgical intervention (NSI). Main outcome: neurosurgical intervention. Analysis: multiple logistic regression model. Results Four hundred and seventy-eight patients were included and 40 underwent NSI. One patient had radiological deterioration but no clinical deterioration prior to surgery. Subdural hemorrhage ≥4 mm width (OR:3.755 [95% CI:1.290–10.928]) and midline shift (OR:7.507 [95% CI: 3.317–16.989]) increased the risk of NSI. Subarachnoid hemorrhage was associated with a lower risk of NSI (OR:0.312 [95% CI: 0.136–0.713]). All other intracranial hemorrhages were not associated with NSI. Conclusion Radiological deterioration was not associated with the incidence of NSI. Subdural hemorrhage and midline shift should be predicting factors for neurosurgery. Some patients with isolated findings such as subarachnoid hemorrhage could be safely managed in their original center without being transferred to a level-I trauma center.

Hyperdense middle cerebral artery sign and response to combination of mechanical Thrombectomy plus intravenous thrombolysis in acute stroke patients

Background and purposeCombining intra-arterial mechanical thrombectomy (IAMT) and intravenous thrombolysis (IVT) has shown to have an excellent recanalization rate and better clinical outcome in acute ischemic stroke (AIS) patients. Hyperdense middle cerebral artery sign (HMCAS) on pretreatment non-contrast head CT scan of AIS patients is one of the early ischemic radiological findings in middle cerebral artery territory AIS. We aimed to evaluate whether the presence of HMCAS predicts the outcome of AIS patients receiving combination therapy with IAMT and IVT. MethodsWe retrospectively reviewed medical records and cerebrovascular images of the patients treated with IAMT and IVT for AIS in our center. Patients with occlusion in the terminal internal carotid artery or middle cerebral artery on pretreatment CT angiogram of the head were included. Clinical outcome was compared between subjects with HMCAS and those without. Modified Rankin Score (mRS) and symptomatic intracranial hemorrhage (sICH) were used as measures of efficacy and safety, respectively. ResultsOf 93 patients, 46 (49%) had HMCAS on their initial head CT scan. Both groups had comparable baseline characteristics and stroke severity. After adjusting for age, NIHSS score, time from symptom onset to starting IVT, and history of diabetes mellitus in multivariate logistic regression analysis, there was no difference in terms of a poor outcome (mRS >2) (OR = 0.5 [CI 0.2–1.4], p = 0.188) or rate of sICH (OR = 3.3 [CI 0.6–19.0], p = 0.190) between the two groups. ConclusionsHMCAS is not a predictor of poor outcome in AIS patients receiving combination therapy with IAMT and IVT and does not affect treatment safety.