AI-Enhanced Dual-Mode Vertical Flow Assay for Multiplexed Cardiac Biomarker Detection. Image Credit: Ozcan Lab @ UCLA
UCLA researchers developed an AI-powered portable diagnostic platform for rapid multiplexed detection of key cardiac biomarkers.
Cardiovascular diseases remain the leading cause of death worldwide, accounting for nearly 20 million deaths each year. Rapid diagnosis and risk assessment of cardiac injury are therefore essential for improving patient outcomes. In clinical practice, physicians often rely on cardiac biomarker measurements; for example, troponin I (cTnI) and CK-MB are commonly used to diagnose myocardial infarction (heart attack), and NT-proBNP is a gold-standard biomarker for heart failure. Because heart attack and heart failure are often interrelated, clinicians increasingly rely on multiplexed biomarker testing for a more precise risk stratification and earlier intervention. However, such multiplexed testing is currently performed using large, expensive centralized laboratory analyzers, limiting its availability in decentralized and time-critical settings.
To address this challenge, UCLA researchers led by Chancellor’s Professor Aydogan Ozcan, in collaboration with Professor Dino Di Carlo, Professor Omai B. Garner and Professor Jeffrey J. Hsu, developed a dual-mode multiplexed vertical flow assay (xVFA) platform capable of detecting multiple cardiac biomarkers in a compact point-of-care format. Dr. Gyeo-Re Han, a Postdoctoral Researcher in the UCLA Electrical & Computer Engineering Department, served as the first author of the study.
Recently published in Light: Science & Applications, a journal of Springer Nature, this work demonstrates that combining colorimetric and chemiluminescent biosensing within a single paper-based test, together with neural network-based signal analysis, enables accurate multiplexed detection of three key cardiac markers (cTnI, CK-MB, and NT-proBNP) across clinically relevant concentration ranges, improving point-of-care diagnostics.
This new platform benefits from two complementary optical modalities, namely colorimetry and chemiluminescence, integrated within a single paper-based cartridge. Colorimetric sensing provides reliable signal detection at higher biomarker concentrations, while chemiluminescence enables highly sensitive detection of extremely low biomarker levels. By combining these two sensing modes in a single assay, the system can accurately quantify cardiac biomarkers across a much broader dynamic range than conventional rapid tests.
“Our goal was to bridge the gap between centralized laboratory diagnostics and point-of-care testing”, said Professor Ozcan, corresponding author of the study. “By combining dual-mode optical sensing with AI, we can achieve sensitive and multiplexed biomarker detection using a compact and accessible diagnostic platform, added Ozcan.
The dual-mode xVFA system operates through a streamlined workflow by loading 50 µL of serum sample into a multiplexed paper-based cartridge, enabling easy operation by minimally trained medical personnel. The cartridge contains multiple testing spots coated with antibodies specific to the three target biomarkers, providing parallel detection within a single test. After the operation, the system generates both colorimetric and chemiluminescent optical signals, which are captured by a portable optical reader. These signals are then rapidly analyzed using a neural network-based algorithm that interprets the multiplexed sensor responses and accurately predicts biomarker concentrations. The sensor provides quantitative results for three cardiac biomarkers – cTnI, CK-MB, and NT-proBNP – in just 23 minutes, all in a single test, to support rapid clinical decision-making.
“Combining complementary sensing modalities allows us to significantly expand the analytical capability of rapid diagnostic tests”, said Dr. Gyeo-Re Han. “By using neural network models for signal interpretation and analysis, we minimize inter-patient and inter-sensor variabilities and achieve high quantification performance comparable to standard laboratory-based analyzers”, added Professor Dino Di Carlo.
UCLA researchers rigorously validated the system using patient serum samples, demonstrating strong agreement with conventional laboratory-based measurements. In particular, neural network models trained and blindly tested in this study yielded robust quantification performance for the three cardiac biomarkers. “Multiplexed biomarker testing provides clinicians with more comprehensive insight into a patient’s cardiovascular condition,” noted Professor Omai B. Garner. “Technologies like this could help bring advanced diagnostic capabilities closer to patients, enabling faster and more informed clinical decisions in emergency departments, clinics, and decentralized healthcare settings,” concluded Professor Jeffrey J. Hsu.
The researchers envision further expansion of the dual-mode xVFA platform to support additional disease biomarkers and broader diagnostic applications. By integrating multiplexed biosensing, portable instrumentation, and AI-based analysis, this technology represents an important step toward more accessible and intelligent diagnostic systems for global healthcare.
See the article:
Gyeo-Re Han, Merve Eryilmaz, Artem Goncharov, Yuzhu Li, Shun Ye, Aoi Tomoeda, Emily Ngo, Margherita Scussat, Xiao Wang, Zixiang Ji, Max Zhang, Jeffrey J. Hsu, Omai B. Garner, Dino Di Carlo, and Aydogan Ozcan “Deep learning-enhanced dual-mode multiplexed optical sensor for point-of-care diagnostics of cardiovascular diseases”, Light: Science & Applications (2026)
https://www.nature.com/articles/s41377-026-02275-9