Machine Learning based Computer Aided Diagnosis Models for Thyroid Nodule Detection and Classification: A Comprehensive Survey

Abstract

In recent years, thyroid nodules have become a common illness, and their incidence surges with age. Since most of the nodules are benign/perform insignificantly, timely and accurate detection of thyroid nodules are essential to minimize the patient risks and medical expenses. Computer aided diagnosis (CAD) models were presently employed in a diagnoses of thyroid nodule. Besides, several medical imaging modalities like ultrasound, computed tomography (CT), and SPECT are widely used to detect thyroid diseases. Numerous deep learning (DL) and machine learning (ML) models are being utilized in the design of CAD models to detect and classify the thyroid diseases. From this perspective, research focuses on the survey of automated CAD models to detect and classify thyroid nodules by the use of DL techniques. Besides, a set of different methods employed in the detection of thyroid disease along with their objectives are briefed. In addition, the reviewed methods are investigated based on the methodology and imaging modality used. Finally, the reviewed methods are summarized based on different aspects along with a brief experimental results analysis.

Ähnliche Arbeiten