Machine learning-based cytokine microarray digital immunoassay analysis

Abstract

Serial measurement of a large panel of protein biomarkers near the bedside could provide a promising pathway to transform the critical care of acutely ill patients. However, attaining the combination of high sensitivity and multiplexity with a short assay turnaround poses a formidable technological challenge. Here, the authors develop a rapid, accurate, and highly multiplexed microfluidic digital immunoassay by incorporating machine learning-based autonomous image analysis. The assay has achieved 12-plexed biomarker detection in sample volume < 15 μL at concentrations < 5pg/mL while only requiring a 5-min assay incubation, allowing for all processes from sampling to result to be completed within 40 min. The assay procedure applies both a spatial-spectral microfluidic encoding scheme and an image data analysis algorithm based on machine learning with a convolutional neural network (CNN) for pre-equilibrated single-molecule protein digital counting. This unique approach remarkably reduces errors facing the high-capacity multiplexing of digital immunoassay at low protein concentrations. Longitudinal data obtained for a panel of 12 serum cytokines in human patients receiving chimeric antigen receptor-T (CAR-T) cell therapy reveals the powerful biomarker profiling capability. The assay could also be deployed for near-real-time immune status monitoring of critically ill COVID-19 patients developing cytokine storm syndrome.

Ultrasensitive Multiparameter Phenotyping of Rare Cells Using an Integrated Digital‐Molecular‐Counting Microfluidic Well Plate

“…The weak signal intensity could yield counting error with a conventional image processing algorithm, we therefore used a modified version of our machine learning code developed in our previous work. [ 41–43 ] We observed almost no fluorescent background (≈0.1%) in all of Aβ‐free samples whether or not LPS presented in them. The fraction of phagocytosis‐active cells steadily increased from 11.9% to 55.0% with Aβ increasing from 0.25 µм to 1 µм.…”

“…Finally, 40 µL of the QuantaRed (Qred) substrate solution was loaded to the bead channels followed by sealing and isolating the bead microwells with HFE‐7500 fluorinated oil (45 µL) for the digital counting process on a motorized microscope scanning system reported in our previous works. [ 41–43 ] The system scans the image of the bead‐filled microwell arrays on the chip right after the oil sealing step to detect the enzyme–substrate reaction activity. The motorized stage was pre‐programmed to follow the designated path to scan the entire dPP (48 cell microarrays or 48 bead arrays).…”

“…Herein, we propose an ultrasensitive multiparameter cellular phenotyping method employing digital molecular counting immunosensors [ 41–43 ] integrated into a microfluidic platform named the “digital Phenoplate (dPP).” The dPP enables analyzing cellular phenotypes associated with cytokine secretion and other cellular functions (e.g., phagocytic activity and cell death) that can be simultaneously observed by digital protein immunosensing and on‐chip image‐based cytometry. The integrated digital immunosensors preserve analyte concentration‐signal linearity even at a sub‐pg mL −1 level and achieve a limit of detection of 0.25 pg mL −1 for mouse tumor necrosis factor alpha (TNF‐α), which allows for testing fewer than 20 cells per assay.…”

Machine-Learning-Assisted Microfluidic Nanoplasmonic Digital Immunoassay for Cytokine Storm Profiling in COVID-19 Patients

“…To address this issue, we developed a machine-learning-based image processing method using convolutional neural network (CNN) visualization for particle counting (named “CNN method”). 44 − 46 Figure 2 a shows the algorithm architecture of the CNN method. It involves dark-field image data read-in/preprocessing (including noise filtering and contrast enhancement), detection signal/background image segmentation by pretrained CNN, postprocessing, and result output.…”

Microfluidics Technology in SARS-CoV-2 Diagnosis and Beyond: A Systematic Review

“…It also can facilitate acute immune disorder monitoring that guides timely treatment plans due to its quick assay turnaround and analytical power. After analyzing longitudinal blood samples from human patients who had cytokine release syndrome (CRS) after receiving CAR-T therapy, the data showed the evolution of 12 circulating cytokines throughout illness development [ 135 ].…”