Radiomics Feature Selection From Thyroid Thermal Images to Improve Thyroid Nodules Interpretations Publisher



Etehadtavakol M¹ ; Siratiamsheh M¹ ; Ng EYK² Show Affiliations
Source: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Published:2023

Abstract

The early detection of malignant nodules and timely diagnosis of thyroid abnormalities play a crucial role in improving medical treatment outcomes and minimizing disease progression. Thermography has emerged as an affordable and non-radiation method for detecting thyroid issues, reducing the risks associated with unnecessary invasive biopsies. By extracting radiomics features from thermal images of the thyroid, valuable information about the underlying tissue characteristics can be obtained, offering numerous advantages in the field of medical imaging. In this study, radiomics features were extracted from thermal images of the thyroid, and unsupervised feature selection techniques including Principal Component Analysis (PCA), Independent Component Analysis (ICA), and variance thresholding were employed to reduce the dimensionality of the feature set. It is important to acknowledge that the field of radiomics analysis in thermography thyroid images is still emerging, and further research is required to validate the clinical usefulness of these features. Nevertheless, radiomics analysis holds significant potential to enhance the assessment of thermography thyroid images and provide valuable insights into thyroid function and pathology. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.