Near‐Infrared Spectroscopy Mapping for Uterine Cancer and Fibroid Detection
ABSTRACTEndometrial cancer and uterine leiomyomas (fibroid) are common uterine pathologies that require early diagnosis to improve a patient's symptoms and increase the success rate of interventional procedures. In this work, we report on near‐infrared spectroscopy (NIRS) spectral features of uterine cancer and fibroids from 69 surgical specimens obtained from 24 patients following hysterectomies. Normal uterus, cancer, and fibroid tissue were identified by NIR spectral contrast parameters based on the differences in spectrum morphology. Using the significant optical features and spectral principal components, a classification model was able to classify uterus tissue with a prediction accuracy higher than 70%, identifying cancer specimens with 70% sensitivity and 93% specificity, and fibroid samples with 86% sensitivity and 83% specificity. These results demonstrated NIRS mapping has promise as a complementary method for gynecologic imaging.
