Broadband Nanocavity Imaging with Machine Vision for Multiplex miRNA Assays

Sensitive and multiplexed quantification of microRNAs (miRNAs) remains challenging due to their short length, low abundance, and sequence homology, particularly in complex biological matrices. This paper presents an approach that combines a distributed Bragg reflector (DBR)-coupled silver nanoparticle (AgNP) gap nanocavity with deep learning instance segmentation for automated image readout. Applied to A549 lung cancer cell extracts, the assay directly quantifies endogenous miR-191, miR-25, and miR-130a without enrichment or amplification. The nanocavity concentrates fields and favors radiative decay, enhancing collection and suppressing quantum dot (QD) blinking, while Mask R-CNN enables robust, high-throughput counting and classification. The system achieves an attomolar LOD (∼10^-17 mol/L), a linear dynamic range spanning five orders of magnitude, and >99% correct identification across spectrally encoded channels. These results establish an AI-enabled nanophotonic biosensing platform that is sensitive, specific, robust, and scalable for multiplexed miRNA analysis in research and clinically relevant matrices.