Diagnostic Accuracy of Intraoperative Brain Smear: A Meta-Analysis of Studies from Resource-Limited Settings

Abstract

BackgroundIntraoperative brain smear is an easy, rapid, and cost-effective technique for immediate diagnosis of brain tumors. Earlier studies have gauged its application on a limited number of samples, but its diagnostic accuracy especially in low-resource settings, where its practice would be extremely helpful is still undetermined. ObjectiveTo investigate the diagnostic accuracy of intraoperative brain smear in resource-limited settings for diagnosis of brain tumors. MethodsA systematic search was conducted on PubMed, Google Scholar, Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Embase for all articles utilizing intraoperative brain smears that were extracted. Studies from low- and middle-income countries (LMIC) with test performance characteristics were selected and subsequent values were summarized using a hierarchical summary receiver operating characteristic (ROC) curve via STATA and pooled using a random-effects model on MetaDiSc 2.0. ResultsTwelve studies consisting of 1124 patients were identified. Six studies included both adult and pediatric population groups, while four investigated adults and two included pediatric patients. The pooled diagnostic odds ratio was calculated to be 212.52 (CI: [104.27 - 433.13]) of Bivariable pooled specificity and sensitivity were 92% (CI: [86%-96%]) and 96% (CI: [93%-98%]), respectively. ConclusionOur study shows that intraoperative brain smear is not only an accurate and sensitive diagnostic modality in resource-rich settings, but it is also equally useful in resource-limited settings, making it an ideal method for rapid diagnosis.