Machine Learning to Predict Three Types of Outcomes After Traumatic Brain Injury Using Data at Admission: A Multi-Center Study for Development and Validation.

The difficulty of accurately identifying patients who would benefit from promising treatments makes it challenging to prove the efficacy of novel treatments for traumatic brain injury (TBI). Although machine learning is being increasingly applied to this task, existing binary outcome prediction models are insufficient for the effective stratification of TBI patients. The aim of this study was to develop an accurate 3-class outcome prediction model to enable appropriate patient stratification. To this end, retrospective balanced data of 1200 blunt TBI patients admitted to six Japanese hospitals from January 2018 onwards (200 consecutive cases at each institution) were used for model training and validation. We incorporated 21 predictors obtained in the emergency department, including age, sex, six clinical findings, four laboratory parameters, eight computed tomography findings, and an emergency craniotomy. We developed two machine learning models (XGBoost and dense neural network) and logistic regression models to predict 3-class outcomes based on the Glasgow Outcome Scale-Extended (GOSE) at discharge. The prediction models were developed using a training dataset with n = 1000, and their prediction performances were evaluated over two validation rounds on a validation dataset (n = 80) and a test dataset (n = 120) using the bootstrap method. Of the 1200 patients in aggregate, the median patient age was 71 years, 199 (16.7%) exhibited severe TBI, and emergency craniotomy was performed on 104 patients (8.7%). The median length of stay was 13.0 days. The 3-class outcomes were good recovery/moderate disability for 709 patients (59.1%), severe disability/vegetative state in 416 patients (34.7%), and death in 75 patients (6.2%). XGBoost model performed well with 69.5% sensitivity, 82.5% accuracy, and an area under the receiver operating characteristic curve of 0.901 in the final validation. In terms of the receiver operating characteristic curve analysis, the XGBoost outperformed the neural network-based and logistic regression models slightly. In particular, XGBoost outperformed the logistic regression model significantly in predicting severe disability/vegetative state. Although each model predicted favorable outcomes accurately, they tended to miss the mortality prediction. The proposed machine learning model was demonstrated to be capable of accurate prediction of in-hospital outcomes following TBI, even with the three GOSE-based categories. As a result, it is expected to be more impactful in the development of appropriate patient stratification methods in future TBI studies than conventional binary prognostic models. Further, outcomes were predicted based on only clinical data obtained from the emergency department. However, developing a robust model with consistent performance in diverse scenarios remains challenging, and further efforts are needed to improve generalization performance.

The impact of hormonal dynamics and serum sodium fluctuations on symptomatic vasospasm after subarachnoid hemorrhage.

BACKGROUND: Symptomatic vasospasm (SVS) is a major cause of morbidity and mortality in aneurysmal subarachnoid hemorrhage (SAH), and serum sodium frequently decreases before SVS. Serum sodium changes might be regulated by sodium metabolism-related hormones. This multi-institutional prospective cohort study therefore investigated the measurement of sodium metabolism-related hormones to elucidate the pathophysiology of serum sodium changes in SAH. METHODS: SAH patients were treated with clipping or coiling from September 2017 to August 2020 at five hospitals. The laboratory data of 133 SAH patients were collected over 14 days and correlations between changes in serum sodium, sodium metabolism-related hormones (plasma adrenocorticotropic hormone (ACTH), serum cortisol, plasma arginine vasopressin (AVP)), and SVS were determined. Serum sodium concentrations were measured every day and serum sodium levels >135 mEq/L were maintained until day 14. RESULTS: Of the 133 patients, 18 developed SVS within 14 days of subarachnoid hemorrhage onset (SVS group) and 115 did not suffer from SVS (non-SVS group). Circulating AVP, ACTH, and cortisol concentrations were significantly higher on day 1 in the SVS group compared with the non-SVS group. Fluctuations in serum sodium in the SVS group were significantly higher than those in the non-SVS group. There were antiparallel fluctuations in serum sodium and potassium from days 2 to 14. CONCLUSIONS: Elevated levels of ACTH/cortisol and AVP on day 1 may be predictive markers for the occurrence of SVS. Multiple logistic regression analysis showed that serum sodium fluctuations were associated with SVS occurrence. Serum sodium fluctuations were associated with stress-related hormonal dynamics. (249 words).

Fibrous granulation mimicking cranial intraosseous tumor.

A cranial intraosseous lesion is a rare disease with a limited number of subtypes. We report a case of a cranial intraosseous fibrous granulation that mimicked an intraosseous tumor. A 57-year-old man was incidentally found to have a cranial intraosseous lesion on brain computed tomography. Total resection was performed to establish a pathological diagnosis and to achieve cosmesis, and the pathological diagnosis was fibrosis and fibrous granulation in the medullary cavity. Fibrous granulation tissue occurs in the calvarium due to bone defects secondary to acquired factors, including trauma. Since its pathological diagnosis is established through surgery, surgery should be carefully considered based on the patient's chief complaint, location of the lesion, and suspicion of malignancy based on imaging findings.

[Locked-in Syndrome due to Primary Brainstem Injury:A Case Report].

Most cases of the primary brainstem injuries(PBSI)are fatal, and disturbance of consciousness is often prolonged even if lifesaving is obtained. The mechanisms of PBSI are as follows: diffuse axonal injury from acceleration/deceleration, shear strain at the midbrain, direct injury of neurovascular structures by tentorial margin, and lower brainstem injury by hyperextension of the cervical vertebrae. Though we can use both CT and MRI to diagnose, MRI is more helpful than CT in detecting, localizing, and characterizing PBSI. When the location of PBSI is limited in the ventral side of pons, it may occasionally result in locked in syndrome(LIS). Generally it is difficult to diagnose LIS with severe trauma due to the rarity of this syndrome caused by head injury. Here, we report a case of an elderly man with traumatic brainstem hemorrhage, who transiently presented LIS and finally improved.

Decreased serum sodium levels predict symptomatic vasospasm in patients with subarachnoid hemorrhage.

BACKGROUND: Symptomatic vasospasm is a major cause of morbidity and mortality in subarachnoid hemorrhage patients. Hyponatremia and dehydration due to natriuresis after subarachnoid hemorrhage are related to symptomatic vasospasm. Therefore, most institutions are currently targeting euvolemia and eunatremia in subarachnoid hemorrhage patients to avoid complications. We retrospectively investigated the predictors of symptomatic vasospasm with respect to water and sodium homeostasis, while maintaining euvolemia and eunatremia after subarachnoid hemorrhage. METHODS: We monitored changes in serum sodium levels, serum osmolarity, daily sodium intake, daily urine volume, and daily water balance for 14days after subarachnoid hemorrhage. Outcomes were assessed using the modified Rankin scale at 1month after subarachnoid hemorrhage. RESULTS: Among 97 patients, 27 (27.8%) had symptomatic vasospasm. Patients with symptomatic vasospasm were older than those without symptomatic vasospasm; the occurrence of symptomatic vasospasm affected outcomes. Serum sodium levels were sequentially significantly decreased, but within the normal range from 1day before the occurrence of symptomatic vasospasm. Serum osmolarity of the spasm group was lower than that of the non-spasm group. CONCLUSIONS: Symptomatic vasospasm occurs more often in older patients and affects outcomes. A decrease in serum sodium levels occurs a day before symptomatic vasospasm. This observation may help predict symptomatic vasospasm.

[A Case of Spontaneous Cerebrospinal Fluid Leak Associated with Cervical Spondylosis].

Spontaneous cerebrospinal fluid leak and intracranial hypotension associated with cervical spondylosis have rarely been observed, and only a few cases are reported. A 69-year-old woman, previously treated for rectal and thyroid cancer, complained of a non-postural persistent headache. The patient regularly practiced aerobic exercise, but a month earlier she had started experiencing headache and neck pain while exercising. Computed tomography(CT)showed bilateral chronic subdural hematomas, and magnetic resonance imaging(MRI)revealed diffuse dural enhancement and tonsillar herniation. We drained the subdural hematomas and replaced the ventricular reservoir to safely access the cerebrospinal fluid space. After surgery, the persistent headache disappeared for several days, but a postural headache emerged. CT myelogram showed extradural accumulation of the contrast medium at the C2-5 level with cervical spondylosis. The patient was treated with conservative therapy of bed rest and intravenous fluid hydration for two weeks, and the headache improved. CT myelogram after treatment showed no extradural accumulation of the contrast medium. Spontaneous cerebrospinal fluid leak associated with cervical spondylosis could be induced by the repeated minor mechanical stress caused by physical exercise. Therefore, the possibility that non-postural persistent headache may be caused by spontaneous cerebrospinal fluid leak should not be underestimated.

Varicella-zoster vasculitis presenting with cerebellar hemorrhage.

BACKGROUND: Varicella zoster virus (VZV) is known as one of the rare, but important, causes of both ischemic and hemorrhagic stroke. Most previously reported VZV-related hemorrhagic stroke and cerebral vasculitis are associated with anterior circulation because VZV spreads from trigeminal ganglia to the anterior circulation of Willis. The present study presents a patient with cerebellar hemorrhage, who was diagnosed with VZV encephalitis and vasculitis of the posterior inferior cerebellar artery. CASE REPORT: A 75-year-old man with stupor was admitted to our hospital. Computed tomography revealed right intracerebellar hemorrhage, and magnetic resonance imaging revealed multiple high-intense signals throughout the brainstem and temporal lobe on fluid attenuation inversion recovery, suggestive of encephalitis. Cerebral angiography revealed stenosis of left posterior inferior cerebellar artery. Based on cerebrospinal fluid analysis, including anti-VZV IgG antibody and VZV DNA polymerase chain reaction, the patient was diagnosed with VZV encephalitis, vasculitis, and cerebellar hemorrhage. CONCLUSIONS: Both cerebral vasculitis and hemorrhagic stroke due to VZV can occur in the vertebrobasilar system. VZV may enter the central nervous system not only from trigeminal ganglia but also from other pathways. We should be aware that a VZV infection could cause cerebral vasculitis and hemorrhagic stroke in the vertebrobasilar system and in anterior circulation.