Smartphone-based Colorimetric Diagnosis : DEVELOPMENT OF A METHOD FOR AUTOMATIC COMPENSATION OF IMPACT OF LIGHT SETTING

Abstract: During the last years many mobile health applications have emerged on the market. Most of these collect and compiles physical data that can be followed over time. Now the next generation of health care applications are on their way. With an increasing capacity and high quality sensors, smartphones have the potential to be used as diagnostic tools. Calmark Sweden AB is a company that has developed a smartphone based diagnostic platform for analysis of colorimetric assays integrated on a disposable plastic chip. Their first product, the hilda Neo system is a Point of care test (POCT) for semi quantitative measurement of the biomarker Lac- tate dehydrogenase (LDH). The system consists of a disposable colorimetric LDH test with inte- grated chemical assay, a separate light-box for controlled light conditions and a smartphone appli- cation for image acquisition and test analysis. The purpose of this Master Thesis project was to develop and evaluate a method for smartphone based semi quantitative colorimetric analysis of the hilda Neo LDH test that would work without the light-box in different light settings. The method was to be implementable as an iPhone applica- tion and should be able do correctly determine LDH activity in the four LDH ranges; 0-300, 300- 600, 600-900 and >900 units per litre (U/L). Also, the computed LDH levels among cards run with the same sample were not to have a standard deviation higher than 50 U/L. Two methods based on continuous measurements of the colour stimuli given from the assay site were developed. In both methods, measurements were made by using the iPhone camera for taking an image series following the colour development of the assay over time. The image series was then processed in MATLAB and the LDH level was computed in two different ways. None of the two proposed methods did reach the stated objectives. Neither of the methods gave the correct LDH interval in all evaluation cards and the computed LDH levels had a larger standard deviation then aimed for. However the results indicate that the variation in light settings is not the only factor for the unreached objectives. It is believed that with further studies of the colour proper- ties of the hilda Neo assay and with the continuing development of smartphone technology, it is possible to find a method for smartphone-based colorimetric analysis without having to control the light setting.