07 Oct 2020
Jürgen Popp describes current key technologies now transferring into clinics.
That was the message from Jürgen Popp of Jena's Leibniz Institute of Photonic Technology (IPHT), in a plenary address to the SPIE Photonex+Vacuum Expo Digital Forum on the topic of translational biophotonics.
"Infections are now an urgent global issue, and optical biophotonics technologies can both provide a precise diagnosis and indicate where to start with targeted therapy." commented Popp. "Raman spectroscopy is one of the simplest technologies to achieve this."
Raman's ability to identify individual spectral fingerprints for organic molecules, in combination with suitable deep learning algorithms to process the data generated, can allow clinicians to image the white blood cells (WBCs) and other immune cells responding to an infection, yielding valuable information about the nature of a disease.
"It need not require a large number of cells to assess the disease state of a patient," said Popp. "We have carried out clinical trials using our HemoSpec platform, an automated system in which in vitro samples are analysed for WBCs using Raman spectroscopy to determine their response to pathogens. In combination with suitable biomarkers this has been able to identify infected cells with an accuracy of around 92 percent, and do so in under one hour from just 1500 cells."
A growing challenge, however, is antibiotic resistance. "If trends continue we will enter a post-antibiotic era, where no functioning antibiotics may be available," cautioned Pepp. "The solution must be to use antibiotics in a targetted way, not in a broad spectrum."
Raman plays a key role here too, by potentially speeding up the process of testing different antibiotics against samples from infected patients to determine the most suitable therapy.
A project named RamanBioAssay at IPHT has developed a chip structure in which different antibiotics at various concentrations can be placed within holes on the chip, and bacterial samples from a patient introduced to all of them. Changes in the morphology and Raman spectra of the cells will then indicate the effectiveness of the different antibiotics.
"RamanBioAssay is an automatic device, reducing the time taken to determine the best antibiotic treatment from up to 24 hours using gold standard microdilution techniques, to under 2 hours," said Popp. "We hope this technology will be translated into clinical routines soon."
An ageing population
A second global challenge identified by Popp is the rising number of medical issues arising from a longer lifespan, including inflamatory bowel disease (IBD) as well as cancers, age-related macular degeneration (AMD), and dementia.
For IBD, a multimodal technique involving coherent anti-Stokes Raman scattering (CARS), two-photon-excited fluorescence (TPEF), and second-harmonic generation (SHG) spectroscopy can provide the molecular information usually lacking when conventional optical examination using endoscopes is carried out.
"CARS can be used for identification of lipids, TPEF for enzymes, and SHG for collagen," commented Popp. "Deep learning algorithms of the kind called generative adversarial networks, applied to the multimodal spectral data produced by these techniques, can allow real-time disease assessment without the need for pathological staining of excised samples."
In cancer treatment, Popp envisages an optical fiber-based approach in which multimodal spectral data from a scan of a suspected tumor site reveals both the nature of the disease and the tumor margins. This is a key step towards in vivo digital histopathology, in which the images created in this way are as clinical useful as traditional pathology samples but obtained much more rapidly.
Time, money and patience
Finally Popp discussed the eye as a window to the brain, and the potential for identifying the early effects of neurodegenerative conditions on the retina and the eye.
"Our goal is to use a multimodal system including OCT, Raman, and a scanning laser ophthalmoscope to assess the effects of both AMD and Alzheimer's, and diagnose them at an early stage," said Popp. "Raman spectra from retina tissues are dominated by lipids and proteins, and overlaying Raman images with data from OCT and white light images allows the chemical and morphological information to be revealed at the same time."
A platform called Multimodal Optical Diagnostics of Ocular and Neurodegenerative Disease (MOON) has been developed on this principle under the EU Horizon 2020 program, using a 785-nanometer laser at a patient-friendly power of 1 milliWatt to perform the multimodal analysis, and do so with an integration time of 30 seconds. This information is complementary to that from conventional examination systems, but measured directly under low-light conditions, a step towards in vivo clinical applications for the technology.
"Medical photonics is a key technology for precision medicine. There is a clear need for medical imaging that is fast, label-free, and chemically sensitive as a means to characterize cells and tissues, and methods to interpret the data in combination with deep learning algorithms," concluded Popp.
"This is a co-operative effort. Translation takes time, money and patience, and without our collaborators we would not be able to make the progress that we have. But I am hopeful that these light-driven technologies will soon make it into the clinics."