Volatile organic compounds (VOCs) in urine are valuable biomarkers for noninvasive disease diagnosis. Herein, a facile coordination-driven modular assembly strategy is used for developing a library of gas-sensing materials based on porous MXene frameworks (MFs). Taking advantage of modules with diverse composition and tunable structure, our MFs-based library can provide more choices to satisfy gas-sensing demands. Meanwhile, the laser-induced graphene interdigital electrodes array and microchamber are laser-engraved for the assembly of a microchamber-hosted MF (MHMF) e-nose. Our MHMF e-nose possesses high-discriminative pattern recognition for simultaneous sensing and distinguishing of complex VOCs. Furthermore, with the MHMF e-nose being a plug-and-play module, a point-of-care testing (POCT) platform is modularly assembled for wireless and real-time monitoring of urinary volatiles from clinical samples. By virtue of machine learning, our POCT platform achieves noninvasive diagnosis of multiple diseases with a high accuracy of 91.7%, providing a favorable opportunity for early disease diagnosis, disease course monitoring, and relevant research.
Keywords: MXene frameworks; e-nose; machine learning; noninvasive disease diagnosis; urinary volatiles.