OBJECTIVESurgical site infection (SSI) following a neurosurgical operation is a complication that impacts morbidity, mortality, and economics. Currently, machine learning (ML) algorithms are used for outcome prediction in various neurosurgical aspects. The implementation of ML algorithms to learn from medical data may help in obtaining prognostic information on diseases, especially SSIs. The purpose of this study was to compare the performance of various ML models for predicting surgical infection after neurosurgical operations.METHODSA retrospective cohort study was conducted on patients who had undergone neurosurgical operations at tertiary care hospitals between 2010 and 2017. Supervised ML algorithms, which included decision tree, naive Bayes with Laplace correction, k-nearest neighbors, and artificial neural networks, were trained and tested as binary classifiers (infection or no infection). To evaluate the ML models from the testing data set, their sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), as well as their accuracy, receiver operating characteristic curve, and area under the receiver operating characteristic curve (AUC) were analyzed.RESULTSData were available for 1471 patients in the study period. The SSI rate was 4.6%, and the type of SSI was superficial, deep, and organ/space in 1.2%, 0.8%, and 2.6% of cases, respectively. Using the backward stepwise method, the authors determined that the significant predictors of SSI in the multivariable Cox regression analysis were postoperative CSF leakage/subgaleal collection (HR 4.24, p < 0.001) and postoperative fever (HR 1.67, p = 0.04). Compared with other ML algorithms, the naive Bayes had the highest performance with sensitivity at 63%, specificity at 87%, PPV at 29%, NPV at 96%, and AUC at 76%.CONCLUSIONSThe naive Bayes algorithm is highlighted as an accurate ML method for predicting SSI after neurosurgical operations because of its reasonable accuracy. Thus, it can be used to effectively predict SSI in individual neurosurgical patients. Therefore, close monitoring and allocation of treatment strategies can be informed by ML predictions in general practice.

Background:Motor vehicle is a major transportation in Southern Thailand as the result of road traffic injury and death. Consequently, severe disability and mortality in pediatric traumatic brain injury (TBI) were observed from traffic accident, particularly motorcycle accident. To identify the risk of intracranial injury in children, the association of treatment outcome with various factors including mechanisms of injury, clinical characteristics, and intracranial pathology can be assessed.Materials and Methods:This was a retrospective study conducted on children, who were younger than 15 years old with TBI and were enrolled from 2004 to 2015. Several clinically relevant issues were reviewed and statistically analyzed.Results:A total of 948 casualties were enrolled. Compared with falling down, the motorcycle accident was significantly associated with intracranial injury (odds ratio 1.73, 95% confidence interval [CI] 1.08–2.76). Other factors associated with intracranial injury were hemiparesis (odds ratio 5.69, 95% CI 1.44–22.36), positive of basal skull fracture signs (odds ratio 15.66, 95% CI 3.44-71.28), and fixed reaction to light of both pupils (odds ratio 5.74, 95% CI 1.71–19.23). Mortality found in thirty cases (3.2%). Furthermore, the risk of death correlated with motorcycle accident (P = 0.02) and severe head injury (P < 0.001). Neurosurgical intervention was not associated with outcome, but severe head injury, hemorrhagic shock, epidural, and subdural hematoma were impact factors.Conclusion:The findings demonstrate road traffic injury, especially motorcycle accident leading to brain injury and death. Prevention program is a necessary key to decrease mortality and disability in pediatric TBI.

Background Prognosis of low-grade glioma are currently determined by genetic markers that are limited in some countries. This study aimed to use clinical parameters to develop a nomogram to predict survival of patients with diffuse astrocytoma (DA) which is the most common type of low-grade glioma. Materials and Methods Retrospective data of adult patients with DA from three university hospitals in Thailand were analyzed. Collected data included clinical characteristics, neuroimaging findings, treatment, and outcomes. Cox's regression analyses were performed to determine associated factors. Significant associated factors from the Cox regression model were subsequently used to develop a nomogram for survival prediction. Performance of the nomogram was then tested for its accuracy. Results There were 64 patients with DA with a median age of 39.5 (interquartile range [IQR] = 20.2) years. Mean follow-up time of patients was 42 months (standard deviation [SD] = 34.3). After adjusted for three significant factors associated with survival were age ≥60 years (hazard ratio [HR] = 5.8; 95% confidence interval [CI]: 2.09-15.91), motor response score of Glasgow coma scale < 6 (HR = 75.5; 95% CI: 4.15-1,369.4), and biopsy (HR = 0.45; 95% CI: 0.21-0.92). To predict 1-year mortality, sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve our nomogram was 1.0, 0.50, 0.45, 1.0, 0.64, and 0.75, respectively. Conclusions This study provided a nomogram predicting prognosis of DA. The nomogram showed an acceptable performance for predicting 1-year mortality. AbstractKeywords ► diffuse astrocytoma ► nomogram ► survival analysis

Background: Primary spinal cord oligodendroglioma is extremely rare. In an extensive review of this disease, 53 cases were reported. Furthermore, the authors summarize the characteristics of the primary spinal cord oligodendroglioma; chronological presentation , neurological imaging, treatment and the outcome obtained in the present case as well as review the literature. Case Presentation: A 46-year-old male who had progressive neck pain for a year. Magnetic resonance imaging showed an intramedullary mass from level C2 to T4. A radical resection was performed. Histology revealed oligodendroglioma. Thereafter, the patient was treated with adjuvant radiotherapy. A year later, tumor developed recurrence. The patinet died in 3 years and 6 months. Conclusions: The available data of this disease was limited. Base on 11 published papers and the present case, surgical resection is the treatment of choice although recurrence of the tumor tends to occur after partial resection with or without radiotherapy. From the literature, the management of the recurrent disease is still surgery. Moreover, Temozolomide may be an advantage in recurrent situations.

Hemangiomas have rarely been found in the spinal cord. A few cases of spinal capillary hemangioma have been reported since 1987. The authors reported the two cases of capillary hemangioma including the tumor at conus medullaris and the another mimicked von Hippel-Lindau disease. A 15-year-old man was presented with coccydynia and left leg pain. A magnetic resonance imaging (MRI) revealed an intradural extramedullary enhancing mass at conus medullaris. Another case, a 31-year-old man was presented with a history of familial history of brain tumor, retinal hemangioma both eyes, multiple pancreatic cyst and syringobulbia with syringohydromyelia. On MRI, a well-circumscribed intramedullary nodule was detected at C5-6 level and multiple subpial nodule along cervicothoracic spinal cord. All patients underwent surgery, and the histological diagnosis confirmed capillary hemangioma. Although rare and indistinguishable from other tumors, capillary hemangioma should be in the differential diagnosis of the spinal cord tumor.

No abstract

OBJECTIVE The overuse of head CT examinations has been much discussed, especially those for minor traumatic brain injury (TBI). In the disruptive era, machine learning (ML) is one of the prediction tools that has been used and applied in various fields of neurosurgery. The objective of this study was to compare the predictive performance between ML and a nomogram, which is the other prediction tool for intracranial injury following cranial CT in children with TBI. METHODS Data from 964 pediatric patients with TBI were randomly divided into a training data set (75%) for hyperparameter tuning and supervised learning from 14 clinical parameters, while the remaining data (25%) were used for validation purposes. Moreover, a nomogram was developed from the training data set with similar parameters. Therefore, models from various ML algorithms and the nomogram were built and deployed via web-based application. RESULTS A random forest classifier (RFC) algorithm established the best performance for predicting intracranial injury following cranial CT of the brain. The area under the receiver operating characteristic curve for the performance of RFC algorithms was 0.80, with 0.34 sensitivity, 0.95 specificity, 0.73 positive predictive value, 0.80 negative predictive value, and 0.79 accuracy. CONCLUSIONS The ML algorithms, particularly the RFC, indicated relatively excellent predictive performance that would have the ability to support physicians in balancing the overuse of head CT scans and reducing the treatment costs of pediatric TBI in general practice.

Background Early posttraumatic seizure (PTS) is a significant cause of unfavorable outcomes in traumatic brain injury (TBI). This study was aimed to investigate the incidence and determine a predictive model for early PTS. Materials and Methods A prospective cohort study of 484 TBI patients was conducted. All patients were evaluated for seizure activities within 7 days after the injury. Risk factors for early PTS were identified using univariate analysis. The candidate risk factors with p < 0.1 were selected into multivariable logistic regression analysis to identify predictors of early PTS. The fitting model and the power of discrimination with the area under the receiver operating characteristic (AUROC) curve were demonstrated. The nomogram for prediction of early PTS was developed for individuals. Results There were 27 patients (5.6%) with early PTS in this study. The final model illustrated chronic alcohol use (odds ratio [OR]: 4.06, 95% confidence interval [CI]: 1.64–10.07), epidural hematoma (OR: 3.98, 95% CI: 1.70–9.33), and Glasgow Coma Scale score 3–8 (OR: 3.78, 95% CI: 1.53–9.35) as predictors of early PTS. The AUROC curve was 0.77 (95% CI: 0.66–0.87). Conclusions The significant predictors for early PTS were chronic alcohol use, epidural hematoma, and severe TBI. Our nomogram was considered as a reliable source for prediction.