Speakers

A ​​​Shear-Horizontal Surface Acoustic Wave (SH-SAW) biosensor was used to address critical unmet needs in oral cancer screening. Conventional methods like tumor biopsies and brush cytology are often invasive, painful, and expensive, while laboratory techniques such as ELISA and LC-MS are time-consuming and require trained professionals. Instead, the proposed solution utilizes liquid biopsy to detect salivary exosomes because of their high-fidelity cargo from their mother cells. A QCM model was implemented to seek some measures for optimization. This approach is less invasive, more affordable, and allows for dynamic monitoring through easily accessible body fluids.​

​The SH-SAW biosensor operates at 250 MHz and functions by measuring a phase shift caused by velocity reduction when target exosomes are captured on the sensing surface. To enhance detection, AI was first utilized to analyze large-scale proteomic data from online LC-MS archives, such as the CPTAC HNSCC collection, to determine potential biomarkers from a pool of candidates. After evaluating these candidates across independent samples, a final panel of three biomarkers—CD9, CD44, and CD59—was selected. These three biomarkers were then used to inspect and validate the accuracy of the exosomal biosensor for the prediction of oral cancer. Experimental results demonstrated that the SH-SAW biosensor achieved a limit of detection for exosomes of approximately 10⁹ particles/mL in just 30 min with only a 5-µL sample.​

The study concluded that a novel SH-SAW biosensor for liquid biopsy was successfully established and validated. The AI-driven selection process proved that the combination of CD44, CD59, and CD9 could distinguish cancer patients with high accuracy, reaching a ROC-AUC of 0.98, sensitivity of 90%, and specificity of 95%. Among various machine learning algorithms in this study, K-means clustering reached the highest accuracy at 91.7%. Finally, the exosomal biosensor provides an insight to future rapid oral cancer screening.