Binary classification of gynecological cancers based on ATR-FTIR spectroscopy and machine learning using urine...

Area of Science:

• Oncology
• Biochemistry
• Medical Diagnostics

Background:

• Early cancer diagnosis remains a significant challenge in modern medicine.
• Effective screening methods are lacking for endometrial and ovarian cancers.
• Existing screening tools like mammography and Pap tests are underutilized in resource-limited settings.

Purpose of the Study:

• To investigate the potential of Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) spectrometry for early cancer detection.
• To develop and validate a machine learning model for classifying cancer patients based on liquid urine analysis.
• To identify specific spectral wavelengths indicative of cancer presence.

Main Methods:

• Analysis of liquid urine samples from 309 patients using ATR-FTIR spectrometry.
• Application of machine learning algorithms to spectral data for classification.
• Identification of discriminant wavelengths associated with cancer.

Main Results:

• Achieved a classification accuracy exceeding 91% in distinguishing cancer patients from healthy individuals.
• Identified specific discriminant wavelengths at 2093 cm⁻¹ and 1774 cm⁻¹.
• These wavelengths show potential correlation with tumor presence and progression.

Conclusions:

• Liquid urine analysis via ATR-FTIR spectrometry, coupled with machine learning, offers a promising, accurate, and cost-effective approach for cancer screening.
• The identified spectral markers warrant further investigation for their role in cancer diagnosis and monitoring.
• This method could potentially improve early detection rates, particularly for gynecologic cancers.