Machine learning-assisted biomass colorimetric dipsticks for portable clinical nephropathy diagnosis by using smart mobile terminal

Donglei Fu, Lin Ding, Bowen Zhang, Shuaibo Zhang, Junjie Deng, Ruyi Wei, Hua Shui, Xinghai Liu,
Machine learning-assisted biomass colorimetric dipsticks for portable clinical nephropathy diagnosis by using smart mobile terminal,
Chemical Engineering Journal,
Volume 522,
2025,
168059,
ISSN 1385-8947,
https://doi.org/10.1016/j.cej.2025.168059.
(https://www.sciencedirect.com/science/article/pii/S1385894725089016)
Abstract: Portable and efficient means are urgently needed for diverse problems of nephropathy diagnosis in the era of the “Great health”, that is, the creatinine level identification in human fluids. Given this urgent requirement, we have developed a biomass colorimetric dipstick for creatinine determination in human fluids by using machine learning to enhance the analysis accuracy and reliability. The main active ingredients of this dipstick are safe and cheap 3,5-dinitrobenzoic acid and sodium alginate, using a facile knife coating molding, which has the potential for large-scale industrial production. Besides, the formula of colorimetric ink for printing is optimized based on a completely new reaction mechanism by plenty of control groups. As a result, the pre-fabricated colorimetric dipstick shows a wide linear detection range of 0.5 to 100 mM with the LOD at 0.18 mM for creatinine determination. As for data processing (creatinine identification), machine vision is utilized to classify the collected images and fit the accurate creatinine content through a Random Forest Regression Algorithm, thus distinguishing the condition of nephropathy. Notably, all creatinine determinations can be performed on a smartphone App based on an Android system to achieve human-computer interaction, which opens up a promising way for intelligent nephropathy diagnosis.
Keywords: Nephropathy diagnosis; Creatinine identification; Biomass colorimetric dipstick; Machine vision; Human-computer interaction