12 June 2013
Research from Korea proposes a low-cost system for microwave imaging breast cancer detection that could produce microwave images with scan-times of a few milliseconds. This would remove the problem of patient movement and reduce their discomfort whilst keeping the system cost down.
According to the World Health Organisation’s International Agency for Research on Cancer, breast cancer is the most common cancer in women worldwide. However, it can also be highly treatable with over 90% recovery rates possible when it is detected at an early stage.
The most commonly used detection technology is X-ray mammography but this has shortcomings that are driving research into alternatives. X-ray mammography obviously involves use of, and exposure to, X-ray ionising radiation. It also requires the compression of the breast between parallel plates to even out and reduce tissue thickness, which is uncomfortable and can be painful for the patient.
This reduces the required radiation dose and improves image quality - by reducing scattering and holding the breast still to avoid motion blur. However, as well as being inherently important, patient comfort may play a part in their willingness to continue to take part in a voluntary screening programme.
The results can also be hard to interpret in cases where the breast tissue is naturally dense. There has also been studies and discussions on the prevalence of false, positive results with X-ray mammography and the negative consequences of possible over diagnosing.
One alternative being investigated is microwave imaging. Microwaves have inherently less energy than X-rays making them safer to use and there is no need to compress the breast between plates, reducing discomfort. In microwave imaging the breast image is reconstructed from the scattering of the microwaves by contrasting dielectric properties of the tissues within the breast.
The majority of proposed systems are based on frequency-domain analysis and a couple of research groups have conducted clinical trials including those at Bristol University (UK) and Dartmouth College (USA). However, image motion blur due to patient movement is a significant issue here.
This is because these frequency-domain imaging methods take tens of seconds to complete the scans required to construct the breast image, requiring the patient to remain as still as possible with their breast in the system for this period, which can be both difficult and uncomfortable.
A system based on time-domain analysis could significantly reduce scan duration. Recently such a system was reported, however this relied on the use of expensive components including a high-precision pulse generator and a very high-speed oscilloscope.
In this issue the researchers at Ewha Womans University propose a time-domain analysis system with a less than 10 millisecond scan time using much cheaper CMOS components for a greatly reduced system cost.
A cheaper system lends itself to wider use and the use of low-cost mass-producible CMOS technology would not only reduce the costs of the system but also reduce its size; allowing for compact devices which, would be a further aid to their use in screening programmes.
Ewha team member Sollip Kwon explained that “Our system consists of 16 UWB transceivers and a master controller with a single system clock. The transceivers transmit a UWB pulse one at a time, while the rest of them receive the signal.” With a single cycle of the 16 transceivers taking only 1.32 µs, thousands of scans can be taken in a matter of milliseconds.
This then not only improves the image quality by eliminating motion blur but also allows the summation of multiple measurements to produce a better quality image by greatly improving signal-to-noise ratio (SNR). This was a key issue the team faced as the SNR of a single scan is low due to high attenuation within breast tissue at the higher frequencies of the UWB pulse.
Since completing the work reported in their Letter the Ewha team have been working to produce a prototype of their system: “We’ve simulated it using a realistic MRI breast model, and are currently implementing an experimental system employing an array of wideband patch antennas, CMOS chips, master controller digital chip, and a tissue equivalent phantom,” said Kwon.
“There remain several challenges for real-world use. The major obstacle is the lack of well-established diagnostic methodology. As the breast image reconstructed through microwave imaging is a little bit different from an anatomic image acquired from X-ray mammography, interpretation of the image is not straightforward and even confusing in some cases.” However, the team believe that this is essentially a problem of getting to know this new kind of image and that it should be solved as more clinical trials are conducted to build up an extensive database.
The Ewha group have a wider interest in the development of systems for future biomedical applications, and they are currently also working on an implantable neuro-stimulator for epileptic seizure detection and suppression. This would monitor neural signals from a targeted region of the brain via electroencephalogram (EEG), looking for signals indicative of an imminent seizure. The implant would then apply electrical impulses to the brain to suppress the seizure.
This article is based on the Letter: Instantaneous microwave imaging with time - domain measurements for breast cancer detection (new window).
A PDF version (new window) of this feature article is also available.
Browse or search all papers in the latest or past issues of Electronics Letters on the IET Digital Library.