EPSRC (EP/F040202/1)

This EPSRC-funded project (01/04/08-60/06/12) aimed to provide new label-free imaging tools for minimally invasive diagnosis of diseases including cancer.  When irradiating tissue with light at an appropriate wavelength, many molecules absorb this “excitation” energy and emit new radiation called “fluorescence”.  If these fluorescent molecules are naturally occurring in biological tissue, their emission is called autofluorescence.  By analysing such autofluorescence signals, it is possible to detect the presence of particular kinds of molecule and also to learn about their local environment – e.g. the structure of surrounding tissue and whether they are bound to other molecules.  Autofluorescence may therefore provide a means to detect the early onset of diseases that cause changes in the concentration, distribution and interaction of biological molecules.  Because autofluorescence measurements do not require the addition of any chemicals, this approach is label-free and can be non-invasive, making it attractive for diagnostic applications.  Biological tissue is usually heterogeneous, however, often containing several kinds of fluorescent molecule in unknown quantities, and strongly scatters optical radiation, making quantitative fluorescence measurements difficult.  It is therefore desirable to analyse tissue autofluorescence in a way that avoids intensity artefacts and to acquire images so that variations in the autofluorescence signal can be correlated with the observed structures in the tissue.  This is analogous to conventional histopathology, where diagnoses are made following biopsy using images of sections of biological tissue that have been stained with dyes to indicate the distributions of different types of molecule.  In this project, we developed special endoscopes to provide microscope-like images of biological tissue with the autofluorescence signal providing the molecular contrast.

To quantify the autofluorescence signals from different types of molecule, we will exploit the fact that different molecular species radiate fluorescence at different rates and so it is possible to distinguish them by observing the fluorescence decay times (lifetimes) following excitation by a short pulse of light.  In this project we developed single-point fibre-optic endoscopic probes to measure the autofluorescence lifetimes and undertook clinical trails to see if we coudl distinguish normal and diseased tissues. If we measure the fluorescence lifetime for each pixel in the field of view, we can obtain a fluorescence lifetime image that can provide a map of different types of tissue. By combining fluorescence lifetime imaging (FLIM) with a special endoscope that provides microscope images with depth resolution, we may ultimately be able to perform “optical biopsy” in situ, analysing optically sectioned images with fluorescence lifetime providing the molecular contrast.

We have investigated FLIM of biological tissue since 1998, demonstrating some of the first label-free FLIM of ex vivo disease in tissue (e.g. cancer, osteoarthritis and atherosclerosis) and have developed a range of sophisticated laboratory-based FLIM instrumentation including proof-of-principle FLIM endoscopy. We are working to correlate endoscopic FLIM with histopathology and with existing FLIM instrumentation at Imperial, particularly using instruments capable of resolving excitation and emission spectra as well as fluorescence lifetime.  This will help elucidate the molecular origins of the autofluorescence contrast we observe between normal and diseased tissues. In this project we undertook extensive spectrally and lifetime resolved ex vivo studies of a range of tissue types (colon cancer, Barrett’s oesophagus, adenomas and inflammatory bowel disease). In a study of 65 specimens from 30 patients we observed statistically significant lifetime contrast between normal and dysplastic or inflammatory tissues.

In order to progress to in vivo FLIM, we worked with world-leading confocal "endomicroscope" technology developed by Mauna Kea Technologies (MKT) that provides optically sectioned imaging with subcellular resolution via flexible fibre-optic bundles that relay fluorescence images from inside the body to external detectors.  Currently the commercial products are limited to intensity imaging at a single excitation wavelength (usually 488 nm) but we developed a FLIM version of this endoscope incorporating a tunable excitation laser in order to excite more biological molecules.  This instrument has been implemented as a trolley-based platform that can be deployed in operating theatres or in medical research laboratories.  We have developed new software tools for montaging of fluorescence lifetime iamges and it now provides real time FLIM with subcellular resolution. Unfortunately it is difficult to apply this instrument to tissue autofluorescence because the fibre-optic bundles fluoresce at similar wavelengths to tissue and provide unwanted background noise. We are therefore developing it for in vivo imaging of exogenous fluorophores including its application to cell signalling mechanisms in disease models and to image photsensitisers in patients.

To realise subcellular endomicroscopy of tissue autofluorescence, we are instead developing a new concept for ultracompact endoscopy (patent application filed) that utilises multiphoton excitation to minimise background (fibre-optic bundle) fluorescence and, uniquely, eliminates unwanted nonlinear optical effects by distributing the excitation power across 100’s of cores in a fibre bundle and uses adaptive optics to control the phase of the distal beams such that they can be scanned and focussed with no optical or mechanical components.  This permits 3-D imaging via ultrathin (~100’s microns) endoscopes - potentially providing unprecedented access to internal tissues.  We are now working towards a clinical demonstrator.

We were particularly interested to investigate in vivo tissue autofluorescence lifetime contrast in order to compare it to our ex vivo data and to explore its diagnostic potential.  For this we developed a compact single-point autofluorescence endoscopic (SAFE) fibre-optic probe to measure the spectral and lifetime properties of tissue autofluorescence in vivo. This instrument is self-contained and trolley mounted for easy deployment and was designed to be sufficiently thin to pass through the biopsy channel of a conventional GI endoscope so that it could be used during regular clinical examinations.

Using this instrument we have undertaken two clinical trials. The first study was of skin cancer in 25 patients (undertaken with colleagues at Lund University Hospital) with which we demonstrated significant autofluorescence  lifetime contrast for blue excitation, in agreement with ex vivo studies.  The second study used the SAFE probe to measure GI tissue in 17 patients in vivo (at Charing Cross Hospital), with initial data indicating blue excitation of autofluorescence provides contrast broadly consistent with ex vivo studies

We are also developing a clinically viable wide-field FLIM endoscope to provide a larger viewing area and shorter excitation wavelengths, albeit without optical sectioning. To this end we ha ve developed a trolly-mou nted wide-field FLIM pla tform that we have applied to freshly excised ENT tissue at Northwick Park Hospital.  This will be translated to an in vivo clini cal trial using a newly designed wide-field flexible FLIM endoscope once we have the regulatory approval.