Bright spot analysis for photodynamic diagnosis of brain tumors using confocal microscopy.

BACKGROUND: In a previous study of photodynamic tumor diagnosis using 5-aminolevulinic acid (5-ALA), the authors proposed using fluorescence intensity and bright spot analyses under confocal microscopy for the precise discrimination of tumorous brain tissue (such as glioblastoma, GBM) from normal tissue. However, it remains unclear if bright spot analysis can discriminate infiltrating tumor in the boundary zone and whether this method is suitable for GBM with no 5-ALA fluorescence or for other tumor types. METHODS: Brain tumor tissue resected from 5-ALA-treated patients was sectioned to evaluate bright spots under confocal microscopy with a 544.5 - 619.5 nm band-pass filter, which eliminated the fluorescence induced by 5-ALA. Border regions and adjacent normal tissues were observed for differences in bright spot distribution. Histopathology was also conducted by hematoxylin and eosin (H&E) staining of serial slices from the same samples to confirm the locations of tumorous, infiltrating, and normal regions. Bright spot areas were then calculated for the same regions evaluated by histopathology. This method was applied for GBM with and without 5-ALA-induced fluorescence as well as for lower-grade gliomas and other brain tumor types. RESULTS: The bright spot area was substantially smaller in the GBM body than in normal brain tissues. Bright spot area was also smaller in infiltrating tumors than in normal tissue at the margin. The same bright spot pattern was observed in tumorous tissues with no 5-ALA-induced fluorescence and in non-GBM tumors. The bright spot fluorescence is suggested to arise from lipofuscin based on emission spectra (mainly within 544.5 - 619.5 nm) and optimum excitation wavelength (about 405 nm). CONCLUSIONS: Bright spot analysis is useful for discriminating infiltrating tumor from bordering normal tissue as an alternative or complement to photodynamic diagnosis with 5-ALA. This method is also potentially useful for tumors with no 5-ALA-derived red fluorescence and other nervous system tumors.

Fluorescence intensity and bright spot analyses using a confocal microscope for photodynamic diagnosis of brain tumors.

BACKGROUND: In photodynamic diagnosis using 5-aminolevulinic acid (5-ALA), discrimination between the tumor and normal tissue is very important for a precise resection. However, it is difficult to distinguish between infiltrating tumor and normal regions in the boundary area. In this study, fluorescent intensity and bright spot analyses using a confocal microscope is proposed for the precise discrimination between infiltrating tumor and normal regions. METHODS: From the 5-ALA-resected brain tumor tissue, the red fluorescent and marginal regions were sliced for observation under a confocal microscope. Hematoxylin and eosin (H&E) staining were performed on serial slices of the same tissue. According to the pathological inspection of the H&E slides, the tumor and infiltrating and normal regions on confocal microscopy images were investigated. From the fluorescent intensity of the image pixels, a histogram of pixel number with the same fluorescent intensity was obtained. The fluorescent bright spot sizes and total number were compared between the marginal and normal regions. RESULTS: The fluorescence intensity distribution and average intensity in the tumor were different from those in the normal region. The probability of a difference from the dark enhanced the difference between the tumor and the normal region. The bright spot size and number in the infiltrating tumor were different from those in the normal region. CONCLUSIONS: Fluorescence intensity analysis is useful to distinguish a tumor region, and a bright spot analysis is useful to distinguish between infiltrating tumor and normal regions. These methods will be important for the precise resection or photodynamic therapy of brain tumors.