A Novel Blood-Based Screening Test for Colorectal Cancer Using Raman Spectroscopy

Area of Science:

• Oncology
• Biotechnology
• Spectroscopy

Background:

• Colorectal cancer (CRC) screening traditionally uses immunochemical faecal occult blood (iFOBT) tests with low compliance in Australia.
• Label-free vibrational spectroscopy, specifically Raman spectroscopy, offers a novel approach for cancer diagnostics.
• Machine learning (ML) integrated with spectroscopic techniques holds significant potential for oncology screening and diagnostics.

Purpose of the Study:

• To evaluate the efficacy of Raman spectroscopy for detecting colorectal cancer (CRC).
• To assess the potential of Raman spectroscopy as a screening tool for CRC.

Main Methods:

• Plasma samples from 370 participants (117 CRC patients, 253 iFOBT-positive controls) were analyzed.
• The control group included individuals with advanced adenomas, non-advanced adenomas, colitis, or no disease.
• Raman spectra were analyzed using ML algorithms in separate training (n=222) and validation (n=148) cohorts.

Main Results:

• The Raman spectroscopy ML model achieved 84% sensitivity and 93% specificity for CRC detection in the validation cohort.
• In subcategory analysis, 95% of CRC patients would be recommended for colonoscopy.
• Approximately 42% of individuals testing positive via iFOBT could potentially avoid colonoscopy.

Conclusions:

• This exploratory study demonstrates the high accuracy of a Raman spectroscopy-based ML model for CRC detection.
• The developed model shows potential as a tool for colorectal cancer screening, possibly improving upon current methods.