Usage of Classification and Regression Tree (CART) in Cancer Thyroid Nodule Prediction

Abstract

The thyroid gland develops solid or fluid-filled lumps. At the base of the neck, somewhere above the breastbone, is a small gland. Nodules on the thyroid gland may be solid or fluid-filled. According to most current estimates, about 99% of thyroid nodules are benign, causing no symptoms. Only a tiny percentage of cases display cancerous thyroid nodules. This chapter is intended to build a Classification and Regression Tree model that will identify the most concerning thyroid nodules as well as recommend treatment. It has a high predictive value, is repeatable, and has superior clinical utility when compared to alternative methods. Thyroid fine-needle aspiration cytology procedures were investigated while under the supervision of an ultrasound machine. The Bethesda score ranges from 1 to 6 classes based on the diagnostic category. The objective of the work was to develop a figurative value for good prognosis cancer prediction. The Bethesda score has been created by combining qualitative and quantitative indicators like gender, shape, echogenicity, margins, composition, and the nodule for echogenic foci. According to the proposed CART model, the series that leads to the majority of Bethesda 4 and 5 nodules is larger than large, extreme hypoechoic nodule. In addition, strong or predominantly robust nodules, hypoechoic or extremely hypoechoic nodules, and hypoechoic or hypoechoic nodules are all present in all Fine needle aspiration candidate countries.