The Research On The Related Factors Of Graves’ Disease

Graves’ disease is one of the most dangerous diseases. So far, there are manymethods to treat the disease. The most popular method is to calculate the dose of¹³¹I byrules of thumb. However, in the clinical practice, subjective factors still affect thediagnosis of doctors. Besides, quantitative relation between the dosage of administrationand affecting factors is still not clear. How to determine the dose of¹³¹I accurately stillconfuses researches. Since the ability of doctors varies from one to another, and treatingenvironment, pressure and other factors also disturb the doctors’ decision, doctors areeasy to overestimate or underestimate the dose of¹³¹I. However, excess dose could leadto hypothyroidism, and underdosage could affect the curative effect. All in all, we needa tool to give doctors more accurate advice on the dose to treat patients well, thus toreduce risk to patients.In this study,116records of Grave patients with¹³¹I treatment from NanchongCentre Hospital are adopted. Through multi-factors variance analysis, gender and age isunrelated to the dose. Then, Pearson product-moment correlation is employed to studythe relation between mss of goiter,24hours maximum absorption rate of¹³¹I,effectivehalf-life and each gram of thyroid dose. The results show that there is positive relationbetween each gram of thyroid dose and mss of goiter,24hours maximum absorption rateof¹³¹I. There is no relation between the each gram of thyroid dose and effective half-life.The obvious relation is found between rigidity of goiter, volume of goiter, course ofdisease, ATD treatment and each gram of thyroid dose.Then, by adoption the five factors we creatively employed Support Vector Machine（SVM） and BP neural network to classify the subclass of the dose range. There are4classes of the dose:70-82.5μCi/g, B82.5-95μCi/g,95-107.5μCi/g and107.5-120μCi/g.After tested by experimental data, the classification accuracy of the SVM classifier withvolume of goiter, course of disease, rigidity of goiter,24hours maximum absorptionrate of¹³¹I, mss of goiter as classifying features is94.18%. The result shows that SVMcould be used as an effective way to classify subclass of the dose. The proposed methodcould be used to improve the dosing accuracy.Thus, it improves the recovery rate of patients while reduces the possibility of hypothyroidism.