Systematic histopathological analysis of different 5-aminolevulinic acid-induced fluorescence levels in newly diagnosed glioblastomas.

OBJECTIVE Glioblastoma (GBM) is characterized by distinct intratumoral histopathological heterogeneity with regard to variable tumor morphology, cell proliferation, and microvascularity. Maximum resection of a GBM results in an improved prognosis and thus represents the aim of surgery in the majority of cases. Fluorescence-guided surgery using 5-aminolevulinic acid (5-ALA) is currently widely applied for improved intraoperative tumor visualization in patients with a GBM. Three intratumoral fluorescence levels (i.e., strong, vague, or no fluorescence) can usually be distinguished during surgery. So far, however, their exact histopathological correlates and their surgical relevance have not been clarified sufficiently. Thus, the aim of this study was to systematically analyze tissue samples from newly diagnosed GBMs with different fluorescence levels according to relevant histopathological parameters. METHODS This prospective study recruited patients who underwent 5-ALA fluorescence-guided resection of a newly diagnosed radiologically suspected GBM. Each patient received 5-ALA approximately 3 hours before surgery, and a modified neurosurgical microscope was applied for intraoperative visualization of 5-ALA-induced fluorescence. During surgery, tissue samples with strong, vague, or no fluorescence were collected. For each sample, the presence of tumor tissue, quality of tissue (compact, infiltrative, or no tumor), histopathological criteria of malignancy (cell density, nuclear pleomorphism, mitotic activity, and presence of microvascular proliferation/necrosis), proliferation rate (MIB-1 labeling index [LI]), and microvessel density (using CD34 staining) were investigated. RESULTS Altogether, 77 patients with a newly diagnosed, histopathologically confirmed GBM were included, and 131 samples with strong fluorescence, 69 samples with vague fluorescence, and 67 samples with no fluorescence were collected. Tumor tissue was detected in all 131 (100%) of the samples with strong fluorescence and in 65 (94%) of the 69 samples with vague fluorescence. However, mostly infiltrative tumor tissue was still found in 33 (49%) of 67 samples despite their lack of fluorescence. Strong fluorescence corresponded to compact tumors in 109 (83%) of 131 samples, whereas vague fluorescence was consistent with infiltrative tumors in 44 (64%) of 69 samples. In terms of the histopathological criteria of malignancy, a significant positive correlation of all analyzed parameters comprising cell density, nuclear pleomorphism, mitotic activity, microvascular proliferation, and necrosis with the 3 fluorescence levels was observed (p < 0.001). Furthermore, the proliferation rate significantly and positively correlated with strong (MIB-1 LI 28.3%), vague (MIB-1 LI 16.7%), and no (MIB-1 LI 8.8%) fluorescence (p < 0.001). Last, a significantly higher microvessel density was detected in samples with strong fluorescence (CD34 125.5 vessels/0.25 mm2) than in those with vague (CD34 82.8 vessels/0.25 mm2) or no (CD34 68.6 vessels/0.25 mm2) fluorescence (p < 0.001). CONCLUSIONS Strong and vague 5-ALA-induced fluorescence enables visualization of intratumoral areas with specific histopathological features and thus supports neurosurgeons in improving the extent of resection in patients with a newly diagnosed GBM. Despite the lack of fluorescence, tumor tissue was still observed in approximately half of the cases. To overcome this current limitation, the promising approach of complementary spectroscopic measurement of fluorescence should be investigated further.