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.

Cross-Neutralizing Activity Against Omicron Could Be Obtained in SARS-CoV-2 Convalescent Patients Who Received Two Doses of mRNA Vaccination.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variant omicron is now under investigation. We evaluated cross-neutralizing activity against omicron in coronavirus disease 2019 (COVID-19) convalescent patients (n = 23) who had received 2 doses of an mRNA vaccination (BNT162b2 or mRNA-1273). Intriguingly, after the second vaccination, the neutralizing antibody titers of subjects against SARS-CoV-2 variants, including omicron, all became seropositive, and significant fold-increases (21.1-52.0) were seen regardless of the disease severity. Our findings thus demonstrate that 2 doses of mRNA vaccination to SARS-CoV-2 convalescent patients can induce cross-neutralizing activity against omicron.

Cross-Neutralizing Breadth and Longevity Against SARS-CoV-2 Variants After Infections.

Background: Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is the virus responsible for the Coronavirus Disease 2019 (COVID-19) pandemic. The emergence of variants of concern (VOCs) has become one of the most pressing issues in public health. To control VOCs, it is important to know which COVID-19 convalescent sera have cross-neutralizing activity against VOCs and how long the sera maintain this protective activity. Methods: Sera of patients infected with SARS-CoV-2 from March 2020 to January 2021 and admitted to Hyogo Prefectural Kakogawa Medical Center were selected. Blood was drawn from patients at 1-3, 3-6, and 6-8 months post onset. Then, a virus neutralization assay against SARS-CoV-2 variants (D614G mutation as conventional strain; B.1.1.7, P.1, and B.1.351 as VOCs) was performed using authentic viruses. Results: We assessed 97 sera from 42 patients. Sera from 28 patients showed neutralizing activity that was sustained for 3-8 months post onset. The neutralizing antibody titer against D614G significantly decreased in sera of 6-8 months post onset compared to those of 1-3 months post onset. However, the neutralizing antibody titers against the three VOCs were not significantly different among 1-3, 3-6, and 6-8 months post onset. Discussion: Our results indicate that neutralizing antibodies that recognize the common epitope for several variants may be maintained for a long time, while neutralizing antibodies having specific epitopes for a variant, produced in large quantities immediately after infection, may decrease quite rapidly.

Cross-Neutralizing Activity Against SARS-CoV-2 Variants in COVID-19 Patients: Comparison of 4 Waves of the Pandemic in Japan.

BACKGROUND: As of March 2021, Japan is facing a fourth wave of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. To prevent further spread of infection, sera cross-neutralizing activity of patients previously infected with conventional SARS-CoV-2 against novel variants is important but has not been firmly established. METHODS: We investigated the neutralizing potency of 81 coronavirus disease 2019 (COVID-19) patients' sera from the first to fourth waves of the pandemic against SARS-CoV-2 D614G, B.1.1.7, P.1, and B.1.351 variants using their authentic viruses. RESULTS: Most sera had neutralizing activity against all variants, showing similar activity against B.1.1.7 and D614G, but lower activity especially against B.1.351. In the fourth wave, sera-neutralizing activity against B.1.1.7 was significantly higher than that against any other variants, including D614G. The sera-neutralizing activity in less severe patients was lower than that of more severe patients for all variants. CONCLUSIONS: The cross-neutralizing activity of convalescent sera was effective against all variants but was potentially weaker for B.1.351. The high neutralizing activity specific to B.1.1.7 in the fourth wave suggests that mutations in the virus might cause conformational change of its spike protein, which affects immune recognition of D614G. Our results indicate that individuals who recover from COVID-19 could be protected from the severity caused by infection with newly emerging variants.

The Neutralizing Antibody Response against Severe Acute Respiratory Syndrome Coronavirus 2 and the Cytokine/Chemokine Release in Patients with Different Levels of Coronavirus Diseases 2019 Severity: Cytokine Storm Still Persists Despite Viral Disappearance in Critical Patients.

Patients with coronavirus disease 2019 (COVID-19) exhibit a wide clinical spectrum ranging from mild respiratory symptoms to critical and fatal diseases, and older individuals are known to be more severely affected. The underlying mechanism of this phenomenon is unknown. A neutralizing antibody against viruses is known to be important to eliminate the virus. In addition, this antibody is induced at high levels in patients with severe COVID-19, followed by a termination of virus replication. Severe COVID-19 patients exhibit high levels of cytokines/chemokines, even after the disappearance of the virus. This indicates that cytokines/chemokines play significant roles in disease severity. These findings also suggest that antiviral therapy (monoclonal antibody and/or convalescent plasma therapy) should be administered early to eliminate the virus, followed by steroid treatment after viral genome disappearance, especially in patients with severe symptoms.

Machine Learning to Predict In-Hospital Morbidity and Mortality after Traumatic Brain Injury.

Recently, successful predictions using machine learning (ML) algorithms have been reported in various fields. However, in traumatic brain injury (TBI) cohorts, few studies have examined modern ML algorithms. To develop a simple ML model for TBI outcome prediction, we conducted a performance comparison of nine algorithms: ridge regression, least absolute shrinkage and selection operator (LASSO) regression, random forest, gradient boosting, extra trees, decision tree, Gaussian naïve Bayes, multi-nomial naïve Bayes, and support vector machine. Fourteen feasible parameters were introduced in the ML models, including age, Glasgow Coma Scale (GCS), systolic blood pressure (SBP), abnormal pupillary response, major extracranial injury, computed tomography (CT) findings, and routinely collected laboratory values (glucose, C-reactive protein [CRP], and fibrin/fibrinogen degradation products [FDP]). Data from 232 patients with TBI were randomly divided into a training sample (80%) for hyperparameter tuning and validation sample (20%). The bootstrap method was used for validation. Random forest demonstrated the best performance for in-hospital poor outcome prediction and ridge regression for in-hospital mortality prediction: the mean statistical measures were 100% sensitivity, 72.3% specificity, 91.7% accuracy, and 0.895 area under the receiver operating characteristic curve (AUC); and 88.4% sensitivity, 88.2% specificity, 88.6% accuracy, and 0.875 AUC, respectively. Based on the feature selection method using the tree-based ensemble algorithm, age, Glasgow Coma Scale, fibrin/fibrinogen degradation products, and glucose were identified as the most important prognostic factors for poor outcome and mortality. Our results indicate the relatively good predictive performance of modern ML for TBI outcome. Further external validation is required for more heterogeneous samples to confirm our results.

[Clinical Features and Treatment Strategy of Vertebral Artery Injury Associated with Cervical Spine Trauma].

OBJECTIVE: Vertebral artery injury(VAI)associated with cervical spine trauma has the potential to cause catastrophic vertebrobasilar stroke. However, there are no well-defined treatment recommendations for VAI. The purpose of this study was to identify an effective treatment strategy for VAI following cervical spine trauma. METHODS: Ninety-seven patients with blunt cervical spine trauma were treated at Hyogo Prefectural Kakogawa Medical Center between January 2013 and September 2017. Of these patients, 49 underwent computed tomographic angiography or magnetic resonance angiography for evaluation of the vertebral artery. Eighteen patients(36.7%)had a diagnosis of VAI. We retrospectively analyzed the clinical features, treatment, and outcomes in these 18 patients. RESULTS: Seven patients(38.9%)had bilateral VAI, 16(88.9%)had cervical dislocation, and 2(11.1%)had transverse process fractures extending into the transverse foramen. Surgical reduction was performed in 14 patients. Five patients with either bilateral or unilateral occlusion underwent parent artery occlusion before reduction. There were no complications after this procedure. Two patients with bilateral VAI had a stroke before treatment. There were no infarctions in the distribution of the vertebrobasilar artery after intervention. The perioperative stroke rate was relatively good, and almost all Glasgow Outcome Scale scores were related to the degree of spinal cord injury. CONCLUSIONS: Aggressive screening for VAI is important in patients with cervical spine trauma in order to ensure adequate treatment. Although the treatment strategy described here could yield good results, it may require modification according to the needs of the individual patient.

A case of non-cardiogenic pulmonary edema provoked by intravenous acetazolamide.

Case: A 61-year-old man was diagnosed with severe chest trauma after a car accident and had had difficulty in weaning from a ventilator because of flail chest and dilated cardiomyopathy. On the 17th day in the intensive care unit, he received i.v. acetazolamide to increase urine output. One hour after the injection, he suddenly developed severe hypoxia. Chest radiography revealed a butterfly shadow. He received other diuretics and a vasodilator, which seemed slowly to resolve the respiratory failure. Five days later, acetazolamide was again given and he experienced the same deterioration. Outcome: We concluded that the episodes were attributed to pulmonary edema provoked by acetazolamide. Conclusion: Acute non-cardiogenic pulmonary edema is an uncommon and lethal adverse effect of acetazolamide. Careful attention may be warranted when administering acetazolamide to critically ill patients.

Combination therapy with early nasopancreatic drainage and minimum surgery for blunt high-grade pancreatic trauma.

CASE: A 22-year-old man suffered blunt, high-grade, pancreatic trauma; however, duct-related complications were avoided by combining early nasopancreatic drainage with minimal surgery. Endoscopic retrograde pancreatography confirmed rupture of the main pancreatic duct and massive retroperitoneal extravasation. A nasopancreatic catheter was placed across the rupture site, laparotomy was carried out, and a grade IV pancreatic head laceration was sutured. Because the nasopancreatic catheter shifted out intraoperatively, another was inserted and left in place for 12 days. OUTCOME: The patient recovered well without any duct-related complications such as pseudocyst or external pancreatic fistula. CONCLUSION: A combination of early therapeutic endoscopic retrograde pancreatography and surgical repair may be useful for treating high-grade pancreatic trauma when pancreatic resection is unnecessary.

Classification of acute pancreatitis based on retroperitoneal extension: application of the concept of interfascial planes.

OBJECTIVE: This study aimed to provide a classification system for acute pancreatitis by applying the principle that the disease spreads along the retroperitoneal interfascial planes. MATERIALS AND METHODS: Medical records and computed tomography (CT) images of 58 patients with acute pancreatitis treated between 2000 and 2005 were reviewed. The retroperitoneum was subdivided into 10 components according to the concept of interfascial planes. Severity of acute pancreatitis was graded according to retroperitoneal extension into these components. Clinical courses and outcomes were compared with the grades. The prognostic value of our classification system was compared with that of Balthazar's CT severity index (CTSI). RESULTS: Retroperitoneal extension of acute fluid collection was classified into five grades: Grade I, fluid confined to the anterior pararenal space or retromesenteric plane (8 patients); Grade II, fluid spreading into the lateroconal or retrorenal plane (16 patients); Grade III, fluid spreading into the combined interfascial plane (8 patients); Grade IV, fluid spreading into the subfascial plane beyond the interfascial planes (15 patients); and Grade V, fluid intruding into the posterior pararenal space (11 patients). Morbidity and mortality were 92.3% and 38.5% in the 26 patients with Grade IV or V disease, and 21.9% and 0% in the 32 patients with Grade I, II, or III disease. Morbidity and mortality were 86.7% and 33.3% in patients with disease classified "severe" according to the CTSI, and 37.5% and 9.4% in patients with disease classified "mild" or "moderate". CONCLUSION: Classification of acute pancreatitis based on CT-determined retroperitoneal extension is a useful indicator of the disease severity and prognosis without the need for contrast-medium enhanced CT.

[Problems in hospital preparation of acute poisoning antagonist (methylene blue injection)].

Since some antagonists or antidotes in cases of acute poisoning are not commercially available in Japan, in many hospitals they are prepared on their premises for clinical use. However, no specific legislation for the procedures of quality assurance and informed consent of these hospital-prepared products as yet exists. Further, the standard procedures for clinical use of the hospital-prepared products have yet to be established. For the treatment of patients with methemoglobinemia, we prepared methylene blue for injectable use in our hospital. In this paper, we describe our procedures ranging from its preparation to clinical use of this product. Methylene blue injection was prepared by using reagent-grade chemicals. The quality of hospital-prepared methylene blue injection was examined in accordance with the United States Pharmacopoeia. The contents of methylene blue injection remained constant at room temperature during storage for 12-month. The sterility testing also gave negative results during the same period. In order to obtain approval for its clinical use by the in-hospital ethical committee, relevant documents such as instructions for the preparation method, product information on safety usage and consent form were created. After these procedures, clinical applications of methylene blue injection were finally initiated.

[Smoke inhalation injury: diagnosis and respiratory management].

Smoke inhalation is a significant comorbid factor following major thermal injury. Smoke exposure is only a trigger for the sequence of events responsible for the development of inhalation injury. Noxious chemicals generated by incomplete combustion injure the exposed bronchoepithelium and stimulate the release of chemical mediators that cause a progressive inflammatory process. Airway inflammation and pulmonary edema impair gas exchange and increase the susceptibility to pulmonary infection. Earlier diagnosis and treatment of inhalation injury is an important element to improve the clinical course of severe burn patients. The American Burn Association, however, recently concluded that there are insufficient data to support a treatment standard for the diagnosis of inhalation injury. At present, the diagnosis of inhalation injury is supported by the combination of history, physical examination, bronchoscopy, and laboratory findings For accurate diagnosis of inhalation injury, helical CT scanning and examination to detect activated leukocytes in bronchoalveolar lavage fluid may be warranted. In the respiratory management of inhalation injury, repeated removal of pseudomembrane by fiberoptic bronchoscopy and the use of adequate PEEP to avoid airway obstruction are essential. High-frequency percussive ventilation can be a suitable mode of ventilation for inhalation injury.

[A case of anaphylaxis after a prairie dog bite].

Several cases of anaphylaxy after hamster bites have recently been reported. We report a case of anaphylaxis after a bite from a prairie dog. To our knowledge, this is the first case to be reported in Japan. The patient was a 26 year old male who was transported to our hospital suffering from dyspnea. The dyspnea occurred within several minutes after a bite from a prairie dog to his left little finger. He was successfully treated by the administration of epinephrine, dopamine and hydrocortisone, and transfusion. Allergies to allergens derived from saliva are speculated to be involved in the mechanism of anaphylaxis in cases caused by hamster bites. We speculate similar mechanisms may also be involved in the case of our patient.

[So-called "ampulla" cardiomyopathy associated with coronary vasospasm compared with acute myocardial infarction showing similar abnormal left ventricular wall motion: two case reports].

So-called "ampulla" cardiomyopathy is characterized by transient abnormal left ventricular wall motion showing hypokinesia around the apical area and hyperkinesia at the basal area, without any detectable coronary lesion. We recently treated a patient with "ampulla" cardiomyopathy (Case 1) and a patient with acute myocardial infarction showing similar abnormal left ventricular wall motion (Case 2). A 75-year-old female (Case 1) presented with "ampulla" cardiomyopathy without coronary lesion. Vasospasm was induced at segment 8 on the left anterior descending (LAD) coronary artery by intracoronary administration of acetylcholine. A 58-year-old male (Case 2) presented with acute myocardial infarction due to occlusion at segment 8 and underwent successful coronary reperfusion therapy by direct percutaneous transluminal coronary angioplasty. Both Case 1 and Case 2 revelaed similar abnormal left ventricular wall motion, with hypokinesia around the apical area and hyperkinesia at the basal area by echocardiography, in the acute phase. Furthermore, these two patients showed elevated ST segment at both anterior and inferior leads by electrocardiography, and markedly reduced uptake of beta-methyl-p-iodophenyl-pentadecanoic acid around the apical area in the acute phase by scintigraphy. Interestingly, the LAD perfused a relatively wide area including the anterior, apical and part of the inferior area of the left ventricle in both patients by coronary angiography. The abnormal wall motion of Case 1 disappeared 4 weeks after onset, but that of Case 2 did not disappear. Although the diagnoses of Case 1 and Case 2 were different, abnormal wall motion of these cases might be due to myocardial ischemia due to distal LAD lesion. "Ampulla" cardiomyopathy might develop from transient myocardial ischemia induced by coronary vasospasm at the distal LAD which perfuses a relatively wide area.