GynoSight: a breakthrough in multispectral imaging for cervical cancer diagnosis

Background: Cervical cancer is a significant threat in low-income countries due to limited access to screening, resulting in delayed diagnoses and increased mortality rates. Multispectral imaging enables early detection of cancer by capturing images at multiple wavelengths to identify subtle alterations in cervical tissue. Objective: This work aims to develop a multispectral transvaginal imaging probe that illuminates the tissue with five different wavelengths of LED source namely, 450nm, 545nm, 575nm, 620nm, and white light for early diagnosis of cervical cancer in a low-resource setting. Methods: The probe consists of five pairs of LEDs and an endoscopy camera (Supereyes). The LEDs of different wavelengths are chosen based on the absorption peak of cancer biomarkers. The LED of 450nm matches the absorption peak of protoporphyrin and LEDs of 545nm and 575nm match the absorption peak of oxygenated hemoglobin and deoxygenated hemoglobin. The constant current supply is ensured by using LM317 in the LED driving circuit. Turning ON and OFF LEDs and the camera is controlled by Arduino and Raspberry Pi modules. The intensity ratio R610/R545 will give the extent of malignancy. Results: Multispectral imaging provides a more comprehensive view of tissues and structures. It is seen that intensity ratio R610/R545 tends to increase in case of malignancy. Conclusion: The visualization of the cervix using GynoSight is non-invasive and can provide real-time feedback to the clinician. It reduces the subjectivity often associated with visual examination.