Novel method to classify varicocele using electronic nose

Abstract

The prevalence of varicocele is estimated to be around 15-20% of the male population, and of these, about 35-40% are patients requiring infertility treatment. A varicocele is a dilation of the veins in the spermatic cord diagnosed by physical examination of the male genital area and assisted by scrotal ultrasound. The development of electronic nose technology provides an opportunity to detect disease characteristics of volatile organic compounds produced by biological materials. This study aims to utilize the metabolomic gas produced from the odor of the seminal fluid by using an electronic nose. The identification of the pattern of volatile organic compounds formed was labeled as unilateral varicocele, bilateral varicocele and clinical non varicocele as the basis for classification with supervised machine learning. In this study, the accuracy values were quite good for several algorithms, both in training accuracy and testing accuracy with an average accuracy value above 80%.