Kathryn E. Fenton, Nikolay L. Martirosyan, Mohammed G. Abdelwahab, Stephen W. Coons, Mark C. Preul and Adrienne C. Scheck
For patients with glioblastoma multiforme, median survival time is approximately 14 months. Longer progression-free and overall survival times correlate with gross-total resection of tumor. The ability to identify tumor cells intraoperatively could result in an increased percentage of tumor resected and thus increased patient survival times. Available labeling methods rely on metabolic activity of tumor cells; thus, they are more robust in high-grade tumors, and their utility in low-grade tumors and metastatic tumors is not clear. The authors demonstrate intraoperative identification of tumor cells by using labeled tumor-specific antibodies.
GL261 mouse glioma cells exhibit high expression of a membrane-bound protein called second tyrosinase-related protein (TRP-2). The authors used these cells to establish an intracranial, immunocompetent model of malignant glioma. Antibodies to TRP-2 were labeled by using Alexa Fluor 488 fluorescent dye and injected into the tail vein of albino C57BL/6 mice. After 24 hours, a craniotomy was performed and the tissue was examined in vivo by using an Optiscan 5.1 handheld portable confocal fiber-optic microscope. Tissue was examined ex vivo by using a Pascal 5 scanning confocal microscope.
Labeled tumor cells were visible in vivo and ex vivo under the respective microscopes.
Fluorescently labeled tumor-specific antibodies are capable of binding and identifying tumor cells in vivo, accurately and specifically. The development of labeled markers for the identification of brain tumors will facilitate the use of intraoperative fluorescence microscopy as a tool for increasing the extent of resection of a broad variety of intracranial tumors.
Nikolay L. Martirosyan, Daniel D. Cavalcanti, Jennifer M. Eschbacher, Peter M. Delaney, Adrienne C. Scheck, Mohammed G. Abdelwahab, Peter Nakaji, Robert F. Spetzler and Mark C. Preul
Infiltrative tumor resection is based on regional (macroscopic) imaging identification of tumorous tissue and the attempt to delineate invasive tumor margins in macroscopically normal-appearing tissue, while preserving normal brain tissue. The authors tested miniaturized confocal fiberoptic endomicroscopy by using a near-infrared (NIR) imaging system with indocyanine green (ICG) as an in vivo tool to identify infiltrating glioblastoma cells and tumor margins.
Thirty mice underwent craniectomy and imaging in vivo 14 days after implantation with GL261-luc cells. A 0.4 mg/kg injection of ICG was administered intravenously. The NIR images of normal brain, obvious tumor, and peritumoral zones were collected using the handheld confocal endomicroscope probe. Histological samples were acquired from matching imaged areas for correlation of tissue images.
In vivo NIR wavelength confocal endomicroscopy with ICG detects fluorescence of tumor cells. The NIR and ICG macroscopic imaging performed using a surgical microscope correlated generally to tumor and peritumor regions, but NIR confocal endomicroscopy performed using ICG revealed individual tumor cells and satellites within peritumoral tissue; a definitive tumor border; and striking fluorescent microvascular, cellular, and subcellular structures (for example, mitoses, nuclei) in various tumor regions correlating with standard clinical histological features and known tissue architecture.
Macroscopic fluorescence was effective for gross tumor detection, but NIR confocal endomicroscopy performed using ICG enhanced sensitivity of tumor detection, providing real-time true microscopic histological information precisely related to the site of imaging. This first-time use of such NIR technology to detect cancer suggests that combined macroscopic and microscopic in vivo ICG imaging could allow interactive identification of microscopic tumor cell infiltration into the brain, substantially improving intraoperative decisions.
Nikolay L. Martirosyan, Joseph Georges, Jennifer M. Eschbacher, Daniel D. Cavalcanti, Ali M. Elhadi, Mohammed G. Abdelwahab, Adrienne C. Scheck, Peter Nakaji, Robert F. Spetzler and Mark C. Preul
The authors sought to assess the feasibility of a handheld visible-wavelength confocal endomicroscope imaging system (Optiscan 5.1, Optiscan Pty., Ltd.) using a variety of rapid-acting fluorophores to provide histological information on gliomas, tumor margins, and normal brain in animal models.
Mice (n = 25) implanted with GL261 cells were used to image fluorescein sodium (FNa), 5-aminolevulinic acid (5-ALA), acridine orange (AO), acriflavine (AF), and cresyl violet (CV). A U251 glioma xenograft model in rats (n = 5) was used to image sulforhodamine 101 (SR101). A swine (n = 3) model with AO was used to identify confocal features of normal brain. Images of normal brain, obvious tumor, and peritumoral zones were collected using the handheld confocal endomicroscope. Histological samples were acquired through biopsies from matched imaging areas. Samples were visualized with a benchtop confocal microscope. Histopathological features in corresponding confocal images and photomicrographs of H & E–stained tissues were reviewed.
Fluorescence induced by FNa, 5-ALA, AO, AF, CV, and SR101 and detected with the confocal endomicroscope allowed interpretation of histological features. Confocal endomicroscopy revealed satellite tumor cells within peritumoral tissue, a definitive tumor border, and striking fluorescent cellular and subcellular structures. Fluorescence in various tumor regions correlated with standard histology and known tissue architecture. Characteristic features of different areas of normal brain were identified as well.
Confocal endomicroscopy provided rapid histological information precisely related to the site of microscopic imaging with imaging characteristics of cells related to the unique labeling features of the fluorophores. Although experimental with further clinical trial validation required, these data suggest that intraoperative confocal imaging can help to distinguish normal brain from tumor and tumor margin and may have application in improving intraoperative decisions during resection of brain tumors.