Portable, Intelligent Fluorescence Sensing Platform for Dense Convolutional Network-Capable Detection of Indophenol...

Indophenol sulfate (IS) and methylmalonic acid (MMA) are biomarkers of chronic kidney disease (CKD) and diabetes polyneuropathy (DPN), respectively. Portable and accurate monitoring of IS and MMA is very important to ensuring human health. The dense convolutional network (DenseNet) with image recognition has great potential in fluorescence sensing, but developing a platform with high precision and portability to diagnose the disease still faces huge challenges. Herein, we developed a high-sensitivity platform with a fluorescence material, a smartphone, and the DenseNet to monitor IS and MMA. A red-emitting Eu@PFC-13 (1) is prepared, and 1 shows high selectivity and low detection limits (DLs) to detect IS and MMA. The sensing mechanism of 1 toward IS and MMA is investigated by experiments and theoretical calculation. For detecting IS and MMA in serum and urine, 1 is fabricated into an Eu@PFC-13/AG (2) film with DLs of 1.4 and 1.6 μM, respectively. In addition, a portable smartphone platform is designed to monitor IS and MMA with high precision. Moreover, the DenseNet is constructed by Python, which can output the concentration of analytes by identifying fluorescence images and judge whether any is in a dangerous range. This work not only proposes a novel method that integrates a fluorescence material, a smartphone, and deep learning to detect analytes but also opens a new way for the diagnosis of CKD and DPN.