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:With the advancement of neuronavigation technologies, frameless stereotactic brain biopsy has been developed. Previous studies proved that frameless stereotactic brain biopsy was as effective and safe as frame-based stereotactic brain biopsy. The authors aimed to find the factors associated with diagnostic yield and complication rate of frameless intracranial biopsy.Materials and Methods:Frameless stereotactic brain biopsy procedures, between March 2009 and April 2017, were retrospectively reviewed from medical records including imaging studies. Using logistic regression analysis, various factors were analyzed for association with diagnostic yield and postoperative complications.Results:Eighty-nine frameless stereotactic brain biopsy procedures were performed on 85 patients. The most common pathology was primary central nervous system lymphoma (43.8%), followed by low-grade glioma (15.7%), and high-grade glioma (15%), respectively. The diagnostic yield was 87.6%. Postoperative intracerebral hematoma occurred in 19% of cases; however, it was symptomatic in only one case. The size of the lesion was associated with both diagnostic yield and postoperative intracerebral hematoma complication. Lesions, larger than 3 cm in diameter, were associated with a higher rate of positive biopsy result (P = 0.01). Lesion 3 cm or smaller than 3 cm in diameter, and intraoperative bleeding associated with a higher percentage of postoperative intracerebral hematoma complications (P = 0.01).Conclusions:For frameless stereotactic brain biopsy, the size of the lesion is the essential factor determining diagnostic yield and postoperative intracerebral hematoma complication.

Objective: The authors aimed to evaluate the necessity of in-hospital neurological observation for mild traumatic brain injury (TBI) patients, who did not have any evidence of intracranial injury from initial computed tomography (CT) brain scans. Material and Methods: We retrospectively reviewed mild TBI patients with initial negative CT brain scans, receiving treatment at Songklanagarind hospital between January and December, 2018. All patients were observed in the emergency department short stay observation unit for 24 hours after injury. Patients’ medical records, initial and official CT brain scan interpretation were collected and analyzed. Results: This study included 493 cases. No patient deteriorated from intracranial injury, while one patient deteriorated from hypoglycemia, associated with his underlying adrenal insufficiency. However, one patient was admitted to the in patient ward, due to a missed diagnosis of acute subdural hematoma from his initial CT interpretation. The incidence of missed intracranial injury from initial CT brain scan interpretation was 1.6%. The need for neurosurgical intervention (in-patient ward admission, anticonvulsant and repeat brain imaging) was 0.2% (1/493). No patient required surgical intervention.Conclusion: Mild TBI patients, with initial negative CT brain scans, have very low risk for deterioration or need of neurosurgical intervention. Patient's underlying major comorbidity may be considered as an indication for in-hospital observation.

No abstract