In this study, we compared the suitability and accuracy of imprints, smears, and frozen sections of suspected brain and spinal cord tumors of 150 patients. Eighty-six percent of the imprints, 91% of the smears, and 99% of the frozen sections were suitable for interpretation. Among the suitable preparations, 82% of the imprints, 92% of the smears, and 99% of the frozen sections agreed with our diagnosis on paraffin sections. Although frozen sections were clearly more accurate than imprints and smears, the latter two provided diagnoses in patients with acquired immunodeficiency syndrome where frozen sections were not done to avoid contaminating our cryostat and in a patient with an epidermoid cyst of the middle fossa that could not be adequately frozen sectioned. Our study shows that imprints and smears complement frozen sections in the intraoperative diagnosis of tumors of the central nervous system.

No abstract