ICSI Protocol Advisor: A decision support system for infertility protocol suggestion

Abstract

Intra-Cytoplasmic Sperm Injection (ICSI) is one of the most common infertility treatments in which ovarian stimulation is carried out to extract the eggs from the ovaries. There are three, short, long, and pure treatment protocols of ovarian stimulation that vary by the type of medicine, the dosage of medicine, and the treatment term. Today, physicians choose an appropriate treatment protocol based on the patient's condition, such as age, and hormonal condition. This could be a relatively subjective and inaccurate method, particularly if the physician is not highly experienced. The present study investigates whether a decision support system can propose a more objective treatment protocol based on the patients' data and data mining methods like logistic regression, decision tree, and SVM. Such a system draws upon classification methods to propose proper treatment protocols for ICSI. Moreover, a separate module was developed to calculate the success rate of the proposed protocols. The system was tested with real data of treated patients at a Hospital in Tehran, Iran. The results showed the proposed system can predict the most proper treatment protocol with an accuracy of 81.90%. The proposed system can help inexperienced physicians to feel more confident about their advice.