Histology across the Spectrum

There is a requirement within the laboratory to take primary tissue through imaging to mathematical modeling so to aid in the discovery and assessment of tissue based biomarkers. This requires the visualisation of multiple markers to study amount and molecular relationship or localisation and examination of rare cells amongst a population. When imaging via immunohistochemistry (IHC) or fluorescence labelling the limiting factor for multi-parameter characterization of the tissue is the camera and filters. A colour camera for IHC imaging has a tonal range that severely restricts faithful image capture, filters used for wavelength discrimination in fluorescence imaging are too broad and not conducive for the discernment of multiple labels due to cross-talk between signals.

To provide a platform for multiple labeling of tissue so more complex questions can be asked measurement of intensity as a function of wavelength is required, hence rather than a camera and/or filters an imaging and analysis system based around classifying tissue spectrophotometrically is essential.

Within the laboratory a PerkinElmer Vectra has been purchased which can visualise whole tissue, tissue micro-arrays, smears and cytospins with multiple markers (IHC, fluorescence or standard labels such as H&E). The system allows for imaging cross the spectrum from 420-720nm, hence every pixel of the digitized image contains 30 separate intensity measures across the spectrum. Using control slides (non-labelled, single labelled) the individual spectra of the labels can be characterised and then used to extract or reject specific spectra from the raw data.

The above multiplex image shows follicular lymphoma tissue stained for CD3 (Vector Red), CD8 (MenaPath Green) and FoxP3 (DAB) with a haematoxylin counterstain. A spectral library was generated by sampling the spectrum in single colour stained control slides. This spectral library was then used to unmix the multiplex image into the single label images shown above. Christian Slater, of the Lymphoma Translational Research Group, has shown that analysis of percentage positivity of cells in the inForm software correlates well between multiplex slides and slides stained for only one marker with haematoxylin. Christian Slater, Lymphoma Translational Research Group.

 

An advantage of the system is that when imaging under fluorescence there is an inherent naturally derived auto-fluorescence (NADH, flavins and porphyrins) and from processing the tissue (e.g. aldehyde fixatives) which are always an issue when imaging. Using a spectrophotometric detection the auto-fluorescence can be removed or separated and then utilised as a general tissue marker.

With the assistance of the Histology department and the Tissue Bank standardisation of preparation, processing and labelling has allowed for variation which can result from tissue handling and processing to be reduced to a minimum by robotizing tissue preparation. This has allowed for us to mathematically model tissue with statistical confidence, any subtle changes between patients or tissues is due to the biology and not the processing.