Wide-area ultrastructural mapping of brain tumour tissue to assist with diagnosis of surgical and biopsy tissue samples.

Abstract number
European Microscopy Congress 2020
Corresponding Email
[email protected]
LSA.7 - Pathology, immunocytochemistry and biomolecular labelling
Prof Murray Killingsworth (2, 1, 3, 4), Dr Tzipi Cohen Hyams (1, 3, 4)
1. Ingham Institute for Applied Medical Research
2. New South Wales Health Pathology (NSWHP)
3. University of New South Wales (UNSW)
4. Western Sydney University (WSU)

ultrastructural mapping, brain tumour, diagnosis, biopsy, Ependymoma, Glioblastoma multiforme

Abstract text

Background: Transmission electron microscopy (TEM) is still commonly used to characterise surgical brain tumour tissue for diagnosis in Anatomical Pathology. The complexity of these lesions and the interaction of the tumour with adjacent normal tissue generally requires screening of histological tissue by light microscopy to obtain contextual information before an ultrastructural examination can be started. As these two microscopy modalities require different samples and processing it is not possible to directly correlate one with the other. Here we show a new approach to ultrastructural imaging of brain tumour tissue using an automated scan generator system working with a high-resolution field emission scanning electron microscope (FESEM) to generate wide-area ultrastructural images. The images may be zoomed into to obtain high power views, offer similar contextual information to that obtained by light microscopy and have a TEM-like appearance.

Methods: The tumour tissue was fixed routinely with 2.5% glutaraldehyde in 0.1M sodium cacodylate buffer pH 7.4. Processing was routine as for conventional TEM including osmication, staining with uranyl acetate, dehydration and embedding in Spurr low viscosity resin. Sections 250 nm thick were cut on an ultramicrotome (Powertome, RMC Boekeler, USA) and stretched with chloroform vapour before being placed on the polished face of a 5x4 mm piece of silicon wafer. Images are then collected using a backscattered electron detector fitted to a GeminiSEM 300 FESEM (Carl Zeiss, Germany). Image tiling over tissue areas sized approximately 7x5 mm was performed using an Atlas automated scan generator system (Carl Zeiss, Germany). Image contrast reversal was done using PhotoShop (Adobe Systems, USA).

Results: Two examples of brain tumour tissue ultrastructural mapping are shown. A) Ependymoma. This tumour shows characteristic diagnostic features including “zipper-like” tight junctions and small acinar structures lined with microvilli, which can be challenging to locate using TEM. B) Glioblastoma multiforme (GBM). This tumour contained regions of variable cellular composition and diagnostic value. Utilising a large ultrastructural map made it possible to quickly locate the region most likely to contain diagnostic features. An additional advantage was the full tissue area available to view, not obscured by grid bars as can be the case with TEM imaging.

Conclusion: Wide-area ultrastructural imaging is a new approach for generating electron microscopy data from surgical and biopsy brain tumour. Using an automated scan generator system working with a high resolution field emission scanning electron microscope (FESEM) it is possible to generate ultrastructural tumour maps with resolution approaching 2-4 nm. The images may be zoomed into to obtain high power views, offer contextual information equivalent to that from light microscopy and have a TEM-like appearance.