Eyes on diagnosis

Share this on social media:

Topic tags: 

Jessica Rowbury takes a look at how imaging technologies are being used to test, diagnose and study disease

The most common reason people visit the optician is to have their eyesight tested. But, more than just myopia can be evaluated by examining the eyes − a simple ophthalmic exam can reveal signs of systemic diseases such as hypertension and diabetes and, in the future, could even help to diagnose brain conditions such as dementia and Alzheimer’s.

The eye is a unique organ as it allows for a minimally invasive insight into a person’s health. ‘The reason why a lot of vision is being used in eye applications is that it is the only transparent area of the body where you can look at a person’s blood flow without opening up the body,’ said Joost van Kuijk, vice president of marketing and technology at Adimec. ‘So, you can actually look inside the body through this transparent organ, the eye, and do medical diagnostics.’

To diagnose diabetic-related diseases, a typical examination is Fundus Fluorescein Angiography (FFA), which allows leakages, blockages or aneurisms to be detected using a specific contrast dye, fluorescein. The procedure involves first taking images of both eyes in colour, and also in red to make blood vessels appear dark. A cobalt optical filter may also be used to make the translucent retinal nerve fibre layer visible. ‘The fluorescein will then be injected and a large number of images will be taken, initially in a very rapid sequence, to follow the flow of the dye,’ explained Chris van Wijk, product manager of the Eye Care Division at Canon’s Medical Imaging Group. ‘First the filling of the arteries can be observed and then the veins.’ This is then followed by a so-called ‘late phase’, where the effects of recirculation, dilution and elimination of dye can be observed.

To provide a clear view of the vascular structure and enable accurate conclusions to be drawn from the images, there are many features that a camera must have. ‘To take good quality retinal images, through a pupil, and on a curved surface, is far from easy. There can also be interference from reflections of the cornea and the crystalline lens,’ Van Wijik remarked. ‘A retinal camera therefore needs many optical qualities: an even illumination of the retina, a large depth of field, no ghosts of flare, high optical resolution and no optical distortions. Also, the quality of the internal optical filters can make a huge difference in the final image quality.’

Wavefront analysis

For detecting vision defects such as wavefront aberrations, which are optical imperfections of the eye that prevent light from focusing properly on the retina, camera-based systems can be used for wavefront analysis.

Using the Shack-Hartmann technique − one of the diagnostic principles used to detect aberrations − a laser beam is used to illuminate the ocular fundus. After passing through the lens, the reflected light is focused on a CCD sensor outside of the eye by a lens array. The generated pixels are then compared to reference points of an ideal imaging lens, allowing the wavefront defects to be quantified in a wavefront chart.

Alternatively, for the Tscherning principle, a series of light spots are applied to the retina, rather than a single beam of light. The image of the projected pattern is recorded by a camera and also compared with an ideal image to draw conclusions about the eye’s aberrations.

Apart from image quality, it is the speed of a camera that enables these techniques, explained Volker Zipprich-Rasch, head of marketing and product management at the Baumer Vision Competence Centre: ‘With the Shack-Hartmann principle, you have to apply several hundred light spots, one after the other, as fast as you can. So, you project the light beam, take an image, and then scan point-by-point the entire area of the eye,’ he said. ‘It is the camera that is controlling the illumination, so the faster the camera, the faster the entire measurement.’ Baumer’s TX Series cameras are used for wavefront analysis of the eye.

Monitoring surgery

But cameras are used beyond diagnostic applications. Last year, Adimec partnered with Lensar to provide technology to cataract surgeons, which included the Lensar Laser System, a femtosecond laser built for refractive cataract surgery.

Adimec’s Quartz Qs-4A60 camera is part of the laser’s augmented reality 3D imaging, measurement, and guidance system, which is used to produce a three-dimensional model of each individual patient’s eye prior to surgery, in order to ensure precise delivery of laser pulses. The camera rotates and captures anterior segment imaging and biometry at two different angles at up to eight different positions around the optical axis, for a total of 16 potential images. Multiple scans identify the precise location of the relevant anatomical structures from the anterior cornea to the posterior capsule. Using these images, as well as optical ray-tracing and biometric data, the augmented reality system creates the 3D model of the eye.

Image quality is particularly important when surgeons rely on the images to make medical decisions, commented Adimec’s Joost van Kuijk: ‘We have identified, besides of course resolution and speed of the imaging, 28 different parameters of what determines the image quality,’ he said. ‘That ranges from sensitivity to uniformity and so on, so trying to get the best out of the image sensors on the market.’

But, as medical professionals typically spend years assessing patients using older, conventional imaging technologies, camera manufacturers should take this is into account when developing cameras for this market, Van Kuijk added. ‘Sometimes you have to downgrade the possible image quality to be similar to what the surgeons are used to, because they make decisions based on experience, and it takes years to become a good surgeon,’ he pointed out. ‘Even if you introduce a new technology that can do more, you don’t want to change the visual experience of the experienced user, because it will take many years for them to adjust to an even better image to draw conclusions from it.

‘The system doesn’t draw conclusions, it’s the person – the experienced pathologist or surgeon that makes decisions based on the image – so it has to be consistent to what they’re used to and what they’re trained on.’

Both eye examinations and the diagnosis of vision defects or diabetes-related diseases are simple and minimally invasive; even minor eye surgeries, such as cataracts, can be carried out the same day. Therefore, many ophthalmologic procedures are carried out in a standard exam room or an outpatient clinic, which drives a demand for low-cost medical instruments. ‘The main challenge is keeping the cost down, because you don’t have to go to real surgery,’ explained Van Kuijk. ‘It allows for a business model that is closer to the patient, at a doctor’s office for example. But the closer you get, the lower cost point of these systems must be − these offices that monitor patients cannot afford to have machines that are hundreds of thousands of dollars or Euros,’ he continued. ‘Therefore, building a vision system that is in line with the cost of these systems is challenging.’

Looking to CMOS

As with many other market sectors, camera manufacturers are moving towards CMOS sensors for ophthalmology applications in order to achieve higher resolutions and speeds. ‘There is a shift going on right now, because CMOS is getting to be equal or better than CCD quality. With CMOS you have the ability to increase resolution with speed – much higher than what is possible with CCD,’ noted Van Kuijk. ‘Frame speed is important because you want real-time imaging while a surgeon is looking at the images. The latest camera that we have − the Quartz a180 − can take 12-megapixel pictures at 180 frames per second. CCD will never reach those speeds.’

CCD sensors, however, still provide advantages, according to Baumer’s Zipprich-Rasch: ‘CCD sensors still have some benefits, for example when capturing low light and if you go to long exposure times − CCDs have less noise and you can capture less light with longer exposure times.’

But although Baumer’s TX camera series contain CCD sensors, in the future, the company may look to CMOS as the higher speeds and sensitivity will be advantageous for opticians carrying out wavefront analysis, Zipprich-Rasch said. ‘CMOS sensors are more sensitive, so you can capture more photons in a shorter time, but they are noisier − so as long as the application requires short exposure times, it is okay,’ he explained. ‘Wavefront analysis, for example, requires short exposure times, so CMOS sensors may be used in the future for these applications.’

See the eye, see the brain

Diagnosing systemic disease by examining the eye is a well established technique, but one company is working to extend the capability of an eye test to be able to diagnose brain conditions such as Alzheimer’s and Parkinson’s. Phoenix Research provides imaging systems to researchers studying the effects of these brain diseases on the eye, and, using these systems, researchers have demonstrated the detection of Alzheimer’s in mice through a simple eye exam. According to the company’s chairman and CEO, Dr Bert Massie, there has been early clinical data stating that Alzheimer’s also presents in the eye of humans: ‘To put this in perspective, this [eye exams] would deliver a very inexpensive means to diagnose Alzheimer’s as opposed to the current means of expensive brain scans,’ he commented. ‘As a second major advance, again using the Phoenix Research imager, is the ability to observe the existence and growth of Parkinson’s through eye imaging.’

Phoenix Research’s range of instruments, which have been designed for rodent eyes three millimetres in diameter, can be used for retinal microscopy, fluorescent imaging, electrophysiology, and anterior segment imaging.

The platform for this set of instruments is the Micron IV retinal imaging microscope, with other modalities added to the instrument as attachments. Because the system is used for such a range of functionalities as well as in the near infrared, a three-chip CCD camera is used, explained Massie. ‘This choice was driven by the need to provide full resolution into the red region of fluorescence and a single chip sensor with a Bayer pattern would not accomplish this,’ he said. The three CCD chip was modified to sense the near infrared spectral region at 850nm, capturing long wavelength fluorophores and angiograms.

For detecting Alzheimer’s and Parkinson’s in humans, the retinal microscope would be based on the same imaging platform, Massie noted. ‘All of [the technologies] mimic the instruments used in the clinics, assisting in the eventual transfer of these technologies to human use,’ he commented. ‘There is a company already working to commercialise this for human application.’

Observing movement

But this is not the only instance where equipment initially developed for research purposes may, in the future, be used for clinical diagnosis. SensoMotoric Instruments (SMI) produces gaze and eye tracking systems, including Eye Tracking Glasses that are used for visual perception and visual orientation studies. These types of studies provide an insight into how visual orientation works in real-life settings, or to analyse differences between people, for example.

‘There are so many direct relations between the eye and other important parts of the body,’ said Eberhard Schmidt, managing director of SensoMotoric Instruments. ‘For example, the eye is a platform that is linked to the inner ear and the balance system − so if the balance system isn’t working properly, you can learn about any dysfunction by studying the eye movement response to changing g-force vectors. That is just one example, and there are a number of others.’

The tracking glasses contain, in a spectacle type of frame, a camera that looks at each eye and one camera that looks forward to capture the visual view of the person. ‘Then, the eye focusing cameras capture eye images based on an infrared illumination which is located in the rim of the spectacle. The software calculates – using the infrared image acquired from each eye camera − the position of the eyeball and the size of the pupil, for example, and the gaze direction,’ explained Schmidt. ‘In general our systems are working in the infrared light domain, so we’re using digital image sensors that are specifically designed to work well in the infrared light domain.’

The SMI Eye Tracking Glasses are currently only used in specific research areas rather than for clinical diagnosis, but there are plans for this technology to be used in clinical settings, Schmidt added.

‘We are working with a number of research partners and universities on potential clinical applications. So, we expect our systems to be used in diagnostic and treatment settings in the next years.’ 

About the author

Jessica Rowbury is a technical writer for Electro Optics, Imaging & Machine Vision Europe, and Laser Systems Europe.

You can contact her on jess.rowbury@europascience.com or on +44 (0) 1223 275 476.

Find us on Twitter at @ElectroOptics, @IMVEurope, @LaserSystemsMag and @JessRowbury.

 

 

Other tags: