In a classic example of how automation technology applied right can save lives, the use of MATLABS coding driven analytical algorithms in healthcare and diagnostic devices by Philips, made its MicroDose SI, a mammography diagnostics, more accurate and safer than ever.
The bad news first.
By 2020, India is likely to have over 17.3 lakh new cases of cancer, and almost half of these cases will succumb to the disease. The top three culprits are cancers of the breast, lung and cervix, according to the Indian Council of Medical Research (ICMR), the premier Indian medical body. The worse news is that, of these cases, only 12.5 per cent of patients come for treatment in early enough stages of the disease, while there still is a chance of cure.
Breast cancer tops the list for women, with over 1.5 lakh new cases during 2016, taking up more than 10% of all cases. So, in India, on an average 1 in 28 women is likely to develop breast cancer during her lifetime.
There are very sincere government and private healthcare agencies spreading awareness of getting tested in time, since early detection can actually increase the chances of a cure for the disease. Mammography is one of the most important tests that women need to do regularly, to this end. Aided by latest technologies, the process can be extremely accurate, especially if it uses cutting edge algorithms to ensure high detection and prediction rates, upping the chances of surviving the disease manifold.
However, the diagnostics also pose a challenge; reports say that the risk of radiation is one of the biggest threats.
In order to make the process of mammography safer and smarter, the market needed low dose mammography application devices that could accurately and safely detect even smaller cancers from breast tissue.
Philips Healthcare’s MicroDose SI is a digital low dose spectral mammography which delivers a much higher efficiency of imaging and diagnosis, at much lower doses. The device is based on a Photon counting technology, which provides a high DQE due to energy weighing and scatter rejection. Philips MicroDose is 40% lower dose than others in the market but the detect rate is higher, allowing for much better decisions and more accurate therapy. Since breast tissues are sensitive, the low dosage system is a huge advantage, especially in western countries where the tests are much more frequent, and hence low does systems are a great need.
Scattered radiation is one of the main sources of inaccuracies in X-ray imaging. The MicroDose SI features a scanning multislit collimating system that rejects nearly 97% of scattered radiation, without blocking the direct x-rays. The traditional grids that used to block scatter radiation also block some useful imaging, and hence create inaccuracies. To meet these inaccuracies, most devices compensate with higher radiation. The MicroDose SI eliminates that risk totally. The advanced benefit is the non-invasive spectral imaging that allows for a much more accurate Spectral Breast density measurement.
The added offering is a 3D imaging that helps detect anomalies in overlapping tissues. In 2D, often small cancers get missed. In 3D there are different slices of the data, so very little will be missed out, and the radiologist can study it slide by slide. This will make detection rate much higher.
The breast is constructed of two kinds of tissues – adipose that is fat, and fibro-granular tissue. If there is a cancer, it is located in the latter area, and thus the density of this part of the breast defines the level of risk. An imaging technique that allows technicians and radiologists to define the breast density automatically increases the chances of cancer detection, and predictability of how prone the breast is to develop cancer in the future.
The analysis of the tissue and the detection of the breast density in MicroDose are driven by the MATLAB (Matrix Laboratory) statistical algorithm developed by MathWorks that eliminates manual decisions and thus wrong diagnosis due to human error. On the surface, there is a number that the algorithm can derive for the breast for each test, which will give the radiologist the accurate density of the breast, layer by layer, to correctly gauge its present risk of cancer, as well as predict if it may happen.
MATLAB is a multi-paradigm numerical computing environment, fourth-generation proprietary programming language developed by MathWorks. It allows matrix manipulations, plotting of functions and data, thus helping in implementation of algorithms, and creation of user interfaces. It also assists in interfacing with programs written in other languages, including C, C++, C#, Java, Fortran and Python.
It supports developing applications with graphical user interface (GUI) features, with tightly integrated graph-plotting features. It is this ability that can be harnessed and leveraged to provide accurate imaging data for support in healthcare, to help medical device engineers, radiologists and analysts with tools for analysing and visualising medical images.
The technology also helped Philips develop an advanced imaging algorithm for MicroDose.
To add BDM in the basic mammography data, there was the MATLAB code available, but Philips needed to integrate it in the production environment, in order to make the results more automated and predictive. Since doing it manually would be a huge challenge, they needed a coder support from MathWorks, to arrive at accurate (at least up to 6 decimal spaces), data for diagnostics of breast cancer.
The biggest challenge faced was that the MATLAB code is bring regularly updated, and if it were to be used in a prediction environment, this constant updation would throw everything off gear. MathWorks mitigated this challenge by offering an updater tool – the Coder – that crunched everything to a few hours. The benefit of using MathWorks’ MATLAB coder was the rapid conversion and updation of the code, which is an extremely complex algorithm, and would have taken recoding of more than a thousand lines of code. The coder made it happen in two hours flat. The development process for the MATLAB code to be used in a production environment was thus reduced to two days.
Then, there was a mismatch of the data with some boundary conditions – which are essential to the accuracy of the data in order to be useful as an efficient diagnostic tool. Going with C++ language would have completely excluded the boundary conditions, making the accuracy of the data questionable. MathWorks resolved the issue quickly and efficiently.
This priceless performance makes the MicroDose SI a great choice for mobile cancer detection and mammography environments as well. MicroDose SI’s humidity- and temperature-tolerant detector technology and system stability are well-suited for a mobile environment. The system can tolerate temperatures from -10°C to +50°C while in transport. This enables the mobile units to be easily usable throughout the year in countries like India as well, without any extra expenditure or arrangements for temperature and climatic effects management.
The MATLAB tool, which was extremely highly interactive and accurate, has been a huge support for Phillip’s MicroDose mammography system, to help detect breast tissue anomalies and cancerous growths for women world over – preventing or even curing cancers – the deadly killer.