Reliability of fluorescein-assisted stereotactic brain biopsies in predicting conclusive tissue diagnosis.

Background: The purpose of this study was to assess the reliability of fluorescein sodium in predicting conclusive tissue diagnosis in stereotactic brain biopsies and to characterize features of contrast-enhancing and non-enhancing MRI lesions associated with fluorescence.

Methods: A total of 19 patients were studied, 14 of which had contrast-enhancing and 5 of which had non-enhancing lesions on preoperative T1 post-gadolinium MRI scan. All patients received 3 mg/kg fluorescein sodium during anesthesia induction. Biopsy specimens were photographed under the operating microscope, using the Yellow560 module, prior to histopathological analysis. Two observers blinded to the MRI scans and histopathological results categorized the photographs retrospectively as "fluorescent" or "not fluorescent." Inter-rater agreement was assessed using Cohen's kappa coefficient. Sensitivity, specificity, and positive predictive value of fluorescence reliability were calculated for MRI contrast-enhancing lesions and confirmed location-concordance of tumor pathology based on rater's fluorescence status assessment. Results were correlated finally with final results on permanent sections.

Results: Strength of inter-rater fluorescence status agreement was found to be "substantial" (kappa = 0.771). Sensitivity, specificity, and positive predictive value for "fluorescent" and "not fluorescent" specimen in comparison with MRI contrast-enhancing lesions were 97%, 40%, and 82%, respectively. Sensitivity, specificity, and positive predictive value for confirmed tumor pathology were 100%, 63%, and 91%, respectively. Permanent pathology revealed high-grade glioma n = 5, low-grade glioma n = 3, lymphoma n = 5, pineal tumor n = 2, hamartoma n = 1, and nonspecific hypercellularity n = 3.

Conclusions: Fluorescein-assisted stereotactic brain biopsies demonstrated a high likelihood to manifest fluorescence in contrast-enhancing MRI lesions, while adequately predicting conclusive tumor pathology.