An integrated microneedle device for skin interstitial fluid multiplexed detection and deep learning-based analysis of acute myocardial infarction biomarkers.
Acute myocardial infarction (AMI) remains difficult to diagnose rapidly outside hospital settings because current evaluation still relies mainly on electrocardiography and blood-based biomarker testing. Here, we developed a coin-sized, Rapid Electrochemical Skin interstitial fluid microneedle device for multiplexed Cardiac-biomarker detection Utilizing deep-learning-based Evaluation (RESCUE). The RESCUE system integrates a mesoporous gold-coated microneedle electrode patch (MNE) with a miniaturized three-channel microelectrochemical workstation (MEW) and Bluetooth data link to enable fully portable operation. Antibodies immobilized on amino-functionalized multiwalled carbon nanotubes enabled simultaneous detection of C-reactive protein (CRP), cardiac troponin I (cTnI), and myoglobin (Myo). Electrochemical measurements showed log-linear responses, with limits of detection of 4.09 ng/mL, 0.078 ng/mL, and 1.642 pg/mL for C-reactive protein, cardiac troponin I, and myoglobin, respectively. In simulated skin and artificial interstitial fluid, the device showed good selectivity and average recoveries above 98%. Biocompatibility studies demonstrated preserved fibroblast viability, rapid closure of microneedle-induced micropores, and no detectable histological or serum biochemical toxicity in mice. In murine model of AMI induced by left anterior descending coronary artery ligation, RESCUE tracks ISF trajectories of all three biomarkers. A seven-feature one-dimensional convolutional neural network (CNN1D) trained on interstitial-fluid and blood biomarker features classified infarct size with 80% accuracy. These results support the feasibility of integrated interstitial-fluid sensing for preclinical AMI stratification.
Authors
Liu Liu, Chen Chen, Li Li, Li Li, Li Li, He He, Huang Huang, Chang Chang, Liu Liu, Xiong Xiong, Zhang Zhang, Chen Chen, Chen Chen, Zhang Zhang, Xiao Xiao, Mou Mou, Guo Guo
View on Pubmed