Abstract
Colorectal cancer (CRC) remains a leading cause of cancer-related mortality worldwide. Current cancer detection methods are hindered by invasiveness and limited sensitivity. In this study, we present a novel diagnostic approach that uses sensor cells expressing insect-derived olfactory receptors (ORs), which respond to volatile organic compounds (VOCs) in urine samples. The response of these sensor cells to urinary VOCs is detected via luminescence emission, generating time-series luminescence data. We screened a library of 483 ORs to identify CRC-discriminative ORs, using urine samples collected from 75 CRC patients and 75 non-cancer controls. After identifying the ORs that could distinguish between urine from CRC patients and urine from non-cancer controls, we applied machine learning to analyze the time-series luminescence data, enabling the creation of a diagnostic model with 80% sensitivity and an ROC-AUC of 0.84. This study demonstrates the potential of OR-based biosensors as noninvasive liquid biopsy tools for CRC diagnosis.