Purpose: To evaluate the diagnostic accuracy of intraoperative squash and imprint cytologic techniques in ocular and orbital lesions.
Design: Observational case series.
Methods: Forty-five consecutive ocular and orbital biopsy specimens subjected to squash and imprint cytologic techniques for rapid intraoperative diagnosis were included in this study. The indications for which the surgeon requested intraoperative diagnosis included: (1) infiltrative lesions where a preoperative tissue diagnosis was not available, (2) discrepancy between preoperative and intraoperative clinical diagnosis, and (3) unusual clinical presentations with an uncertain diagnosis. Fresh unfixed tissue was used for making squash preparations and imprint smears. Alcohol-fixed smears were stained with hematoxylin and eosin, whereas air-dried smears were stained with the Giemsa stain. The residual tissue was subjected to routine processing for histopathologic examination.
Main outcome measure: Correlation of intraoperative cytologic diagnosis with the final histologic diagnosis was the primary outcome measure. The correlation was considered to have complete concordance if the diagnosis matched with respect to nature of the lesion (benign vs. malignant) as well as with the further classification of the benign or the malignant lesion. Partial concordance was considered when there was agreement with respect to the nature of the lesion, but there was a mismatch in the further classification of the benign or the malignant lesion.
Results: Biopsies were obtained from 45 patients whose ages ranged from 1 to 79 years (median, 21 years). The sites of biopsy included eyelids (n = 4), conjunctiva (n = 6), intraocular tissue (n = 3), and orbit (n = 32). Adequate cellularity was obtained in all cases by squash and imprint techniques. A cytologic diagnosis could be made within 5 to 8 minutes after receiving the sample. Of 44 cases in which a cytologic diagnosis could be made, 21 (48%) were malignant and 23 (52%) were benign. Cytologic analysis was inconclusive in 1 case. The cytologic-histologic correlation showed complete concordance in 41 of 45 cases (91%) and partial concordance in 3 cases. Sensitivity and specificity of the squash and imprint technique in detecting malignant lesions was 100%.
Conclusions: Assuming the availability of a skilled cytologist, ophthalmic pathologist, or both, squash and imprint preparations are useful for rapid and reliable intraoperative diagnosis of ocular and orbital lesions.