Intraoperative frozen section (FS) is an important tool in the management of neoplastic and non-neoplastic central nervous system (CNS) lesions. Although the final pathological diagnosis is based on the findings in paraffin tissue sessions, interpreting FS of CNS lesion is to assist the neurosurgeon in making the accurate judgment regarding the nature of the lesion and tissue adequacy. This study was to evaluate the agreement between FS and paraffin sections (PS) results; and to highlight the possible causes of error in discrepancy between FS and PS. Materials & Methods:This is a retrospective study of 85 cases of FS. H&E stained sections of both FS and PS of all these cases were reviewed with their histopathological reports. FS and PS results were compared. Possible causes of errors were reviewed and recorded.Results:Out of 85 cases; 76 cases (89.4%) showed no discrepancy, 2 cases (2.4%) minor discrepancy and 7 cases (8.2%) discrepancy. Sensitivity and specificity of FS were 90.1% and 85.7% respectively. Positive predictive value was 97.0% and negative predictive value was 63.1%. The false positive rate was 14.2% and false negative rate was 36.8%. The overall accuracy of FS was 89.4%. Conclusion:Our results showed high accuracy and specificity of FS. Therefore, FS is still useful, an effective tool for neurosurgeon although there are some challenges for histopathologists in reporting FS. The commonest discrepancy was seen between reactive gliosis and low-grade glioma. Over-diagnosis or under-diagnosis of FS can be reduced by closed cooperation among pathologist, neurosurgeon and radiologist.