Project Lead(s): Siva Rama Krishna Vanjari, Sumohana Channappayya
Thyroid diseases are among common endocrine disorders worldwide.
Despite being relatively easy to diagnose and treat, it is estimated that about 42 million people in India continue to suffer from thyroid diseases that are untreated, due to a lack of proper diagnostics.
Thyroid Stimulating Hormone (TSH) is an important biomarker that can be measured to determine thyroid diseases and a low-cost, point-of-care diagnostic device to measure TSH could be used for triage and be of particular use in rural India.
The object of the project was to develop a low-cost, semi-quantitative, triaging diagnostic platform for the detection of TSH, using a fabric-woven Lateral Flow Immunoassay (LFIA) platform in conjunction with a mobile-enabled image-processing solution.
The key concept in this approach was the extraction of TSH concentration information from the fully qualitative LFIA.
The intensity of the test line can be used as an indication of the concentration TSH and this intensity can be extracted accurately from a mobile image, provided the image processing techniques take into account the variations in the illumination and exposure conditions.
An attempt to correlate intensity with concentration, while removing any artifacts caused by different, ambient conditions, was conducted using novel image-processing algorithms.
In the proof of concept stage, three such algorithms were evaluated: the gamut mapping algorithm, the colour flow algorithm and the colour cat algorithm. A support vector machine (SVM) classifier was used to quantify TSH concentration into one of three levels.
The team was able to successfully demonstrate that it is possible to perform colourimetric analysis of the output from the fabric-based TSH detection solution.
However, it was not possible to convert this into a triaging tool, due to the poor sensitivity of the fabric-based immunoassay platform. This was because, when a LFIA device is woven on a fabric, the warp region is coated with the antibodies and the weft region is not coated with the antibodies. As the weft region covers half of the warp region, sensitivity is cut in half.
More importantly, the silk fibers used for weaving the device are not porous and this lack of porosity significantly limits the number of active antibodies.
The study has also developed a microfluidic platform, using proprietary technology, in conjunction with image-processing algorithms, and this microfluidics chip for TSH detection had high sensitivity and specificity.