Development and Validation of a Raman Spectroscopic Classification Model for Cervical Intraepithelial Neoplasia (CIN)

Area of Science:

• Biomedical Optics
• Spectroscopy
• Gynecologic Oncology

Background:

• Cervical cancer mortality is reducible with early precancer detection.
• Current diagnostic methods for cervical precancer face limitations in subjectivity, cost, and time.
• Optical spectroscopic techniques like Raman spectroscopy offer label-free, non-destructive biochemical analysis.

Purpose of the Study:

• To demonstrate the clinical utility of Raman spectroscopy for identifying cervical precancer.
• To validate Raman spectroscopy's diagnostic capability using a large sample set and an independent test set.

Main Methods:

• Collected liquid-based cervical cytology samples (n=662) comprising negative, cervical intraepithelial neoplasia (CIN)1, and CIN2+ cases.
• Acquired Raman spectra from single-cell nuclei.
• Applied partial least squares discriminant analysis (PLSDA) for classification and validated the model on an independent blinded test set (n=69).

Main Results:

• Achieved a classification accuracy of 91.3% on the independent test set.
• Successfully classified negative, CIN1, and CIN2+ cervical cytology samples.
• Only six samples in the blinded test set were misclassified.

Conclusions:

• Raman spectroscopy shows significant clinical utility for cervical precancer detection.
• The method provides accurate classification of cervical intraepithelial neoplasia stages.
• This technique offers a promising alternative to current subjective and time-consuming diagnostic methods.