Killer potential: time to raise the profile of QPI

Early last year, physicists led by Gabriel Popescu from the University of Illinois at Urbana-Champaign (UIUC) revealed stunning footage, several hours long, of the minuscule microtubules so very critical to cell division and nervous system health.

Only months later, the UIUC professor and his collaborators had delivered 3D images of live cow embryos, a breakthrough for clinical scientists striving to select healthy embryos for in vitro fertilization.

Critically, in each case the biological structures were captured label-free, providing crisp, clear, high-resolution imagery - and a wealth of information on structure, growth dynamics and more. Welcome to the wonderful world of quantitative phase imaging (QPI).

The term QPI covers myriad methods that measure how much light is delayed through a specimen at each point in the field of view. This optical path length - or phase information - relates to a sample's refractive index and thickness, enabling detailed studies on cells and tissue.

Popescu first realized the potential of QPI in the late 1990s. Working on his PhD at the College of Optics and Photonics (CREOL), University of Central Florida, he was combining light-scattering techniques with interferometry and phase measurements to analyze inhomogeneous media, including tissue, as well as blood coagulation.

By 2006, and now at the Massachusetts Institute of Technology's G.R.Harrison Spectroscopy Lab, he had developed some key QPI techniques. Fourier phase, Hilbert phase, and diffraction phase microscopies were all designed to extract quantitative phase images from dynamic biological phenomena. A year later Popescu was honing his methods to explore nanoscale fluctuations in red blood cells, vibrational changes in malaria-infected cells and more, while heading up the Quantitative Light Imaging Laboratory at UIUC.


Gabriel Popescu (center) and collaborators produced 3D images of live cattle embryos that could help determine embryo viability. Credit: L. Brian Stauffer.

Crucially, and like many in the QPI field, usability was at the forefront of Popescu's mind. In his earlier years at UIUC, he and colleagues developed a "spatial light interference microscopy" (SLIM) module that could be attached to the camera port of a conventional microscope for diffraction tomography. The module converted interference patterns, recorded by a CCD device, to quantitative phase images, giving sub-nanometer path-length sensitivity, and, crucially, generating high-resolution images of unstained cells.

But as Popescu points out: "We had realized that if we continued to use SLIM as it was, we would be the only ones that could use it, as it was occupying an entire optical table and required PhD students to run it."

He adds, "I wanted it to be accessible to biology users, without this engineering factor, and have it operate at the push of a button, like any of the major brand microscopes."

So, in 2009, Popescu set up the company Phi Optics to commercialize SLIM. By 2014 its first prototype, the CellVISTA SLIM Basic, was launched. The automated "Pro" version followed in 2015, and could be combined with fluorescence imaging.

"The Pro system allows the user to program different channels of fluorescence while using SLIM at the same time," says Popescu. "This is so important to biologists."

Popescu and his team are not alone in their quest to push the usability of QPI methods for biologists. Belgium-based Ovizio Imaging Systems, along with Phasics in France, Tomocube in South Korea, the Swiss pair Lyncee Tec and Nanolive, and UK-based Phase Focus, are all now delivering market-ready products for label-free cell imaging.

Sweden-based Phase Holographic Imaging is another. Launched by chief executive Peter Egelberg back in 2004, it aims to provide long-term, label-free quantitative analysis of living cell dynamics. After years of persistentresearch and development, the company today supplies automated, affordable time-lapse cytometry.

Right from the start, Egelberg and his team wanted to make the instruments simple, reliable and easy to use. The company's initial prototype, HoloMonitor M1, was delivered in 2004, followed in 2008 by a miniaturized QPI module attached to a Nikon microscope - the HoloMonitor M2.

"We placed six M2 units at Lund University but soon realized that we needed to redesign the software and the instrument itself," highlights Egelberg. "The setup was still too complex and way too expensive, and we were never going to make any money out of it."The company launched the HoloMonitor M3 in 2011, having set about designing an even smaller, cheaper instrument that, this time, could operate inside a cell incubator for long periods of time - a huge draw for biologists.

"Reducing field-rate failures took some time," recalls Egelberg. "For example, we had to switch to optically-coated components that would deter the formation of micro-organisms. But we increased the mechanical and electronic quality of the instrument, and delivered HoloMonitor M4 in 2014."

Four years on, and Phase Holographic Imaging has sold around 100 of those instruments worldwide. Recent examples include a HoloMonitor Wound Healing Assay that provides data on cells migrating around a wound area, including individual cell tracking information.

Clearly success has ensued but as Egelberg points out: "When you are not working with these instruments on a day-to-day basis, you just don't realize how much can go wrong. We're a small company that uses distributors so we really need to try and keep everything super-simple and reliable, otherwise these sales representatives will simply sell something that they are more confident selling."

Killer applications
Physicist YongKeun ‘Paul' Park is a professor at the Korea Advanced Institute of Science and Technology (KAIST), and also co-founder and CTO of Tomocube. He has a slightly different take on QPI. Tomocube has focused on 3D imaging, launching its HT-1 and HT-2 products with fluorescence imaging. They generate high-resolution 3D refractive index tomograms, with example videos of unlabeled single human sperm and cancer cells. Park's team has also combined its platform with a deep convolutional neural network, ‘HoloConvNet', designed to classify holographic images of unlabeled living cells.


Cells in 3D: the central cell is undergoing cell suicide, known as apoptosis. Credit: Phase Holographic Imaging.

"The vision of our company has not just been to provide instrumentation," highlights Park. "I believe [that] in a decade, microscopy platforms will be digitized, automated and also powered by artificial intelligence... so we have been developing that artificial intelligence power."

Indeed, after training with QPI images, HoloConvNet can accurately identify single cells including bacteria such as listeria and E. coli, as well as anthrax spores, outperforming methods such as optical fingerprinting
and surface-enhanced Raman scattering.

Beyond combating terrorism and deadly bacteria, Park also reckons his company's technology will be used in biofuel applications, to quantify lipids within bacteria and algae. He adds: "[Our technology] is also accessible to medical doctors and biomedical scientists for easy, rapid, and accurate point-of-care diagnosis of pathogens."

Probing deeper
Popescu and his team at Phi Optics have also branched out from their SLIM technology mainstay, launching "gradient light interference microscopy" (GLIM) in late 2017. The GLIM module can be implemented to existing inverted microscopes, and extracts 3D information from unlabeled specimens whether thick or thin.

As Popescu points out: "Light scatters in specimens that are hundreds of microns thick, so essentially we have been trying to image these thicker specimens through a cloud; indeed, the majority of other instruments in this space are used for thin samples."

GLIM probes deeper into thick samples by controlling the path length over which light travels through the specimen, allowing users to generate images from multiple depths that are then composited into a single 3D image.

Popescu is convinced this label-free imaging modality will find applications in in-vitro fertilization, and has already released detailed images of bovine embryos monitored over several days. And he has high hopes that the method will also be adopted by neuroscientists.

"Brain slices 300-500 microns thick need to be modeled for connectivity over several hours," he points out. "Before GLIM, these were very difficult to image but right now we are looking at this; the method is really opening up new, complementary applications for us."

But as applications proliferate, without a doubt, cancer research remains as important as ever for QPI companies. Popescu: "Look at the way you can get incredible resolution and beautiful images with single cells, as well as quantify cell growth, which is so difficult to measure otherwise."



Captured using Phase Holographic Imaging kit, the top sequence shows a highly unusual event; a cancer cell dividing into three daughter cells. The daughter cells then also divided in three. This new information led to the development of improved child cancer treatment. The bottom sequence shows a cancer cell "budding" a daughter cell; another highly abnormal process. Credit: Phase Holographic Imaging.

Importantly, in the last year or so, many in the field have realized that the phase maps produced by QPI methods can present a wealth of information on cancer diagnosis, prognosis and biopsies. As a result, many groups are now working on QPI-based pathology, with Popescu's lab also combining artificial intelligence.

"[QPI methods] analyze the same thin biopsies used in traditional pathology, and our lab has imaged many thousands of these biopsies on several different types of cancer," he says. "Importantly, we have found that there actually is a lot of information in the collagen and the stromal tissue surrounding the tumor that hasn't been accessed in the past by traditional stained pathology.

"For Egelberg, cancer biology is also of paramount importance. He noted: "We have other fields such as stem cells and immunology but cancer research is totally dominating."

Egelberg firmly believes that QPI methods will prove instrumental to providing the final cancer cure. "When cancer develops, the controlling mechanisms that have evolved to make cells collaborate have been displaced. The cells start to multiply uncontrollably as their ancestors did a billion years ago, and as micro-organisms still do today," he points out. "Cancer researchers don't really understand why this takes place and need to move from studying at a cellular level to a cell population level."

Right now, Egelberg's company is developing software to automatically extract more detailed information from the time-lapse movies of cells captured by the HoloMonitor M4. In addition to current applications, he would like to provide easy-to-obtain, label-free quantitative cell culture information on viability, division rate, mitosis duration and more.

Importantly, Egelberg reckons that if researchers could automatically map thousands of cells throughout cancer treatment, identify and extract the surviving cancerous cells and then study these cells still continuing to divide, a cure for cancer could be clearer.

"Our goal is really to make a contribution to cancer research, and we can do this with commercial success," he emphasizes. "If we are commercially successful, that means we've created a system useful for cancer researchers, and hopefully this will lead to a cure for cancer. Personally, this is why I am involved in QPI."

Like others in his field, Egelberg is certain that software will be critical to future commercial success. He started his company with the QPI technology at the forefront of his mind, but points out: "We assumed that everyone would understand all of this but that just isn't the case. So what we are really doing now is adapting the software so the technology can be more easily used by cell biologists."

Where next for the wonderful world of QPI? Belgium's Ovizio, regarded as a frontrunner, has established a platform for label-free cervical cytology and is set to launch a rapid, low-cost, screening test for cancer.

Over in the UK, Phasefocus has built on the 2016 launch of its "Livecyte" label-free cell analysis platform by expanding distribution chains, and recently penetrated China. CEO Martin Humphry is looking forward to the QPI sector gathering momentum and continuing to enter more conservative markets such as pathology and pharmaceuticals.

"Academic sales are leading to more research publications that highlight specific applications," he points out. "This will lead to the technology becoming accepted as a must-have in the more conservative pharma sector, for example, that generally wants a more push-button solution."


Label-free kinetic cytometry equipment from the UK company Phasefocus. Credit: Phasefocus.

Indeed, as Popescu also highlights, Phi Optics' strategy has been to build awareness and sell commercial instruments to the research market while building the necessary performance data for those more conservative sectors.

"The clinical market demands huge amounts of data and validations and this can take years to realize," says Popescu. "But just last year, and for the first time, a whole slide imaging system from Philips used in pathology was approved for clinical diagnostics." He predicts that this milestone will open up pathology: "And I think there will be a special place for QPI in there," Popescu added.

Tomocube's Park also sees more medical doctors adopting the technology. "For the last ten years, this field has been driven by engineers, but to find the killer applications we need the medics to join us, and come up with ways to use our products and images to facilitate diagnosis," he told Show Daily.Without a doubt, today's over-arching challenge for QPI companies is to raise the market profile of the technology in general. While academics have wholeheartedly embraced the technique, industry peers have yet to raise an eyebrow.

"The interest is massive in academia," says Popescu. "Look at the number of papers we have received for this conference alone. Yet 99% of potential biology users in industry don't even know that QPI exists; so we need to go to shows and make demonstrations worldwide to raise awareness and increase market adoption."

Egelberg agrees. In his words, most of the cell biologists that his company targets "don't have a clue" what QPI is. "From very early on, we have been reducing our manufacturing costs and to reach volume sales, we're targeting these end users directly," he says. "But this field is now getting more attention and this is very beneficial for all of us."

He concludes: "If you are the only company in a field, you have a hard time attracting customers and investors. To reach success, we need many companies working on the technology, or it simply isn't going to happen."

Rebecca Pool is a UK-based freelance writer. A version of this article appeared in the Photonics West Show Daily in February.

Related SPIE content:

Real-time quantitative phase imaging in biomedicine

Computational phase imaging for light microscopes

Quantitative phase imaging: taking the pulse of living cells