19 Apr 2016
Rapid multi-photon approach said to generate images of similar quality to conventional pathology slides; table-top system in development.
Photonics researchers in The Netherlands have developed an ultrafast laser system that produces path-lab quality images distinguishing brain tumor tissue from surrounding healthy tissue in real time.
Marloes Groot and colleagues from Vrije Universiteit Amsterdam reported in the open-access journal Biomedical Optics Express this week that the label-free approach could provide brain surgeons with high-quality pathological information during operations to remove tumors.
They add that the technique could eventually be used both ex vivo, to identify tumor presence in excised brain tissue, and in vivo to tell surgeons the precise extent of malignancies such as diffuse glial tumors that are particularly difficult to eliminate. Groot also told optics.org that a smaller table-top system suitable for use in an operating room could be ready for clinical trials in less than a year.
The work represents the first time that second- and third-harmonic generation (SHG/THG) techniques have shown the ability to recognize the presence of diffuse infiltrative glioma in fresh, unstained human brain tissue.
“Images and a first diagnosis can be provided in seconds, with the ‘inspection mode’, by moving the sample under the scanning microscope, or in about five minutes if an area has to be inspected with sub-cellular detail,” reports the LaserLab Amsterdam team.
That provides an obvious advantage over conventional histopathology approaches, which take several hours and are unsuitable for use during surgery, where it is extremely difficult for neurosurgeons to determine the boundaries of diffuse gliomas and completely remove the malignant tissue without also damaging surrounding healthy tissue.
And, unlike other real-time optical techniques like Raman spectroscopy and optical coherence tomography (OCT) – both of which are the subject of clinical trials for in vivo screening – the images generated by the multi-photon approach are very similar to those from conventional pathology.
“Our images are really of a quality similar to the pathology images, and they visualize the same information - making it a directly comparable technique,” Groot said. "When I showed these images to the pathologists that we work with, they were amazed."
She also explained that Raman spectra must be compared against a set of reference spectra, and that the differences between spectra of healthy and tumor tissue can be very subtle. “Combined with noise, it must always be a battle to get high reliability numbers,” she added.
Groot and the Amsterdam team used a 200 fs pulse, 1200 nm wavelength laser in combination with a two-photon laser-scanning microscope from LaVision Biotec to generate the images. The laser source was an optical parametric oscillator, pumped by one of Coherent’s Chameleon Ultra II Ti:sapphire sources.
At the moment, the system sits on a 2 meter by 1.5 meter optical table, but Groot indicates that scaling that down to a size more suited to clinical use will be relatively straightforward – partly because the current laser is over-specified for the application.
“Much smaller laser systems are available, making a 1 x 1 meter table-top possible,” she told optics.org. “We are now in the designing process, and a table-top can be ready for a clinical trial in less than a year.”
While the table-top system has been shown to provide fast, histopathological-quality images of excised brain tissue, what neurosurgeons would really benefit from is an optical biopsy probe in the form of a handheld needle.
And although Groot and colleagues have shown that their approach works, the incoming laser pulses can only reach a depth of about 100 µm into the tissue. To reach deeper, the team has already developed a 7.5 mm-long bioptic needle based around two graded-index (GRIN) lenses with aberration correction.
Endomicroscopy studies with the needle on excised brain tissues generated images at a depth of 20-30 µm below the surface tissue, showing both healthy and glioma tissue, with an acquisition time of 16 seconds at a laser power of 50 mW.
Patient-safe irradiation levels are yet to be determined for THG imaging, but are likely to fall in the 10-50 mW range – suggesting that the fiber-delivered needle technique does warrant further development.
Horizon 2020 application
To that end, Groot says she and colleagues have just submitted a grant application for further work on the needle imager to the European Commission’s Horizon 2020 funding scheme, with advisory support from two companies.
However, the researcher believes that there could be room for both the needle imager and the more conventional microscope in the future. “There is a large demand for in situ characterization,” Groot said. “The most crucial element, the micro-endoscopic needle, we have tested on ex vivo tissue, as shown in the [journal] article.”
Ideally, she wants to develop both applications, performing sample tissue analysis in situ with the needle, and complete screening of excised tissue with the ex vivo setup.
“Ultimately, the ex vivo device may replace H&E (hematoxylin and eosin) histopathology, leaving more tissue, and help identifying tissue, for more advanced pathological analysis,” Groot added.
"With our technique it's potentially possible to diagnose not only during an operation but possibly before surgery.”