Phasefocus: redefining microscopy


Rebecca Pool

Friday, December 8, 2017 - 13:45
Image: Label-free kinetic cytometry from Phasefocus; a new approach to long term live cell microscopy.
Each has either licensed or adopted technology from a rapidly growing, UK-based company set to disrupt microscopy as we know it today.
Phasefocus span out from the University of Sheffield in 2006, following years of dogged research from physicist and electron microscopist, Professor John Rodenburg.
In the 1990s, Rodenburg, then based at the Cavendish Laboratory of the University of Cambridge, was intent on boosting the resolution of electron microscope with mathematics rather than lenses.
He knew that if he could solve the 'phase problem' and process the entire microdiffraction pattern from the scattered electron beam in a STEM, extremely high spatial resolution would follow.
A decade later he had developed a ptychographic phase retrieval algorithm to do just this and was starting to demonstrate lensless, also known as coherent diffractive, imaging could work in practice.
Instead of using a lens to capture light, Rodenburg's method collected the light scattered by a sample with a sensor.
The scattered light would then be analysed and an image reconstructed using the phase retrieval algorithm (see 'Imaging process').
Crucially, Rodenburg had also patented his process and come 2006, with seed-funding from investor Fusion IP, Phasefocus was born.
Two years later, the researcher and colleagues had completed their first visible light standalone prototype, which was actually networked with an eight-strong fleet of PlayStation PS3s to reduce image construction times from 40 minutes to 20 seconds. 
As Dr Martin Humphry, chief executive officer of Phasefocus explains: "The primary requirement of the company back then was to get the technology working reliably and produce demonstrator equipment to show it worked."
"It was also crucial to identify markets in which the technology would be useful and this was very much a theme for the early days of the company," he adds.
Quantitative phase image of live cells in a 96 well plate [Phasefocus]
Rodenburg had already been working with several X-ray synchrotron groups, demonstrating the value of ptychography in X-ray imaging.
"You can get significant gains on resolution as well as having the added bonus of extracting phase information," explains Humphry. "So given this, Diamond, for example, had always been interested in using ptychography on beamlines built specifically for diffractive imaging."
However, Rodenburg, Humphry and Phasefocus colleagues had also realised that the ptychography approach produced incredible phase contrast in transparent materials that had variations in thickness.
"We realised that contact lenses are by definition transparent but are designed to have this thickness variation," says Humphry. "So we looked at one under our prototype instrument and it showed a really great pattern with very strong contrast."
Armed with results, Rodenburg set off for a British Contact Lens Association annual conference and was immediately approached by numerous contact lens manufacturers; Phasefocus' first commercial application had materialised.
Phasefocus team: driving quantitative phase imaging forward.
By 2011, the company had constructed and sold its first LensProfiler, based what was now known as its 'Virtual Lens' technology, that could provide a wealth of quantitative data on the performance and safety of hydrated soft contact lenses.
As Humphry puts it: "This contact lens measurement instrument did all sorts of things for us; it gave us a target to aim for in an industry where repeatability and accuracy of measurements are critical."
"With this, we were able to refine the technology to a point where it was reliable; the whole experience really gave us a step change in the technology itself," he adds. 
And in the meantime, the company had also been developing label-free microscopy for live cell imaging.
Screening live cells
From day one, the potential to use lensless imaging, such as Phasefocus' ptychography method, for the quantitative phase imaging of live cells had not been lost on Rodenburg, Humphry and colleagues.
So, come 2009, the researchers started to work alongside Professor Peter O'Toole from the University of York, to develop a prototype cell imaging device, called the Virtual Lens 20, or VL20.
By 2010, the first VL20 upright system for label-free live cell imaging based on an Olympus microscope was sold to the University of York.
This provided quantitative, high contrast images of cells for long-term time-lapse studies, and according to Humphry, was 'very similar to the contact lens profiler'.
By the Summer of 2012, Phasefocus has bagged a hefty £3 million in equity funding from UK venture firm, Ombu Group, as well as an additional £220,000 from existing funder, Fusion IP.
LensProfiler sales continued and with more cash in tow, live cell imaging development continued apace.
User feedback from the VL20 had come in fast, so Rodenburg, Humphry and colleagues set to upgrading the VL20 and by 2014, the VL21 inverted live cell imaging system was launched.
"We took these [second-generation] devices to many different laboratories and research institutes, and loaned them out to even more researchers," he says. "This provided us with yet more feedback to build the third version of this system and provided a very clear specification of what people needed with this kind of live cell imaging work."
Launched in April last year, the Livecyte label-free cell analysis system has already been bought by a major pharmaceutical company as well as the Francis Crick Institute, the Universities of Nottingham and York, and more.
Dr Martin Humphry, Phasefocus: "When it comes to electron microscopy, it's always turned out to be more difficult than expected to produce something commercial and reliable." 
According to Humphry, the latest version captures phase images using a red laser diode but also includes brightfield and fluorescence imaging capabilities.
At the same time, a scientific CMOS sensor is used for high sensitivity phase and fluorescence detection and the system can image over multiple wells rather than just a single well.
The system is controlled via a computer interface and, as Humphry highlights, it also introduces automatic, individual cell tracking.
"For the first time, our instrument allows researchers to understand how complex drug interactions behave differently with different individual cells within a population," highlights Humphry. "But building on this, we want to be able to use artificial intelligence for more analytics and ultimately realise personalised medicine."
"For example, in a few years we would like to be able to take cells from a patient's tumour, test different drug treatments and then automatically determine the optimum combination of drugs for that particular patient's tumour," he adds.
But light microscopy applications aside, Rodenburg's original intent with ptychography was to raise the resolution of electron microscopy.
And not surprisingly, Phasefocus researchers have been busy trying to do just this.
Phase images showing the proliferation of NIH-3T3 cells; despite the large size of the 3.2 by 3.2 mm region, individual cells are clearly identified, as shown in the digital zoom of the image (yellow box). [Phasefocus]
In 2010, they were able to demonstrate a five times resolution improvement in a commercial FEI Quanta SEM equipped with a Gatan CDD camera by processing dark-field, high resolution, scattered intensity data.
And come 2012, the researchers had published details of ptychographic electron microscopy using high-angle, dark-field scattering for sub-nanometre resolution imaging in Nature Communications.
More recently, the researchers have developed the Phasefocus πbox, a diffractive image reconstruction engine for ptychographic processing at multiple wavelengths including electron microscopy.
Early adopter sales are underway and the company is currently working with manufacturers and end users in electron microscopy to refine the technology and produce a truly commercial product that could soon be easily integrated to an electron microscope.
"When it comes to electron microscopy, it's always turned out to be more difficult than expected to produce something commercial that is also reliable," says Humphry. "But more recently, for example, fast, sensitive direct electron detectors have been developed that allow you to take a diffraction patterns very quickly in an electron microscope."
"This key, enabling technology means it's now feasible to take very nice ptychography measurements in an electron microscope that just haven't been possible before," he adds. "Now is really a very exciting time for ptychography and electron microscopy."
Looking forward, Humphry is confident that the future is bright for quantitative phase imaging in general.
As he highlights, computational imaging is already having a huge impact on the world of optical microscopy, with several start-ups commercialising technology and industry heavyweights taking note as dynamic live cell imaging gathers momentum.
"In the next year or two, I believe that the benefits of quantitative phase imaging will become well understood and it will be a staple in research labs all over the world," he predicts.
But for Humphry, Phasefocus' technology clearly has the edge. "Relative to Phasefocus' method, many quantitative phase imaging techniques are not so suited to electron, X-ray and EUV microscopy," he says. "So I believe ptychography will be the primary method adopted in these fields."
Extra information: imaging process
Phasefocus's method for coherent diffractive imaging transfers the task of image formation from the lens to a software algorithm.
Crucially, the method creates a quantitative phase image as well as the conventional brightfield image from a single, coherent light source.
A specimen is first illuminated by the large area probe, with its phase and amplitude distribution automatically computed.
The probe is then shifted to a number of overlapping positions on the specimen and at each position, the transmitted or reflected diffraction pattern is recorded on a 2D array detector.
Phasefocus's phase retrieval algorithm then processes these diffraction patterns to create an amplitude image and phase image pair from the specimen.
The amplitude image is similar to a conventional brightfield microscope image, and is a quantitative map of the specimen’s transmittance or reflectance.
Meanwhile, the specimen’s phase function is a quantitative measure of the phase delay introduced as the wavefront travels through, or is reflected by, the specimen.
This phase data may be used to measure thickness, refractive index, dielectric constant, surface topography, the local magnetic field environment, and other parameters of interest.
Website developed by S8080 Digital Media