All-in-one SERS biochip for high-efficiency isolation and multiplexed profiling of exosomes for osteosarcoma diagnosis

Abstract

Early diagnosis of osteosarcoma (OS) is critical for improving patient survival. Although exosomes have emerged as promising noninvasive biomarkers for early cancer detection, their clinical translation remains hindered by technical limitations in high-efficiency isolation and multiplex molecular profiling. To address these challenges, we develop a novel all-in-one surface-enhanced Raman spectroscopy (SERS) biochip that integrates herringbone-embedded microfluidics, tangential flow filtration (TFF), and SERS detection, enabling simultaneous Ca2+-dependent TIM4 affinity capture, EDTA-mediated release, and multiplexed on-device detection of exosomes. Leveraging TIM4-based affinity SERS substrates in conjunction with TFF, the platform achieves a high exosome recovery efficiency of 87.8%, outperforming the gold-standard ultracentrifugation method and avoiding SERS substrate aggregation and microchannel clogging. Multiplex quantitative analysis of three exosomal protein biomarkers using antibody-conjugated SERS probes further enables precise discrimination among different OS cell lines. Using plasma-derived exosomes from OS patients (n = 15) and healthy controls (n = 15), we construct a logistic regression diagnostic model. On an independent test set (40% of samples), the model achieves a diagnostic accuracy of 93.33%, with a sensitivity of 90%, a specificity of 96.67%, and an area under the curve of 0.93. To our knowledge, this is the first study to integrate TIM4-based affinity capture, TFF purification, and SERS-based multiplex profiling with machine learning for noninvasive OS diagnosis. This strategy not only overcomes key technical barriers in exosome-based cancer diagnostics, such as low isolation efficiency and limited multiplexing capability, but also presents significant clinical value for noninvasive OS detection.