Radiomics based assessment of Pulmonary nodules detected on high Resolution computed tomography of the Chest: a one year hospital based Observation study

Abstract

BACKGROUND Pulmonary nodules are a frequent incidental finding on thoracic CT scans, observed in approximately 30-50% of adult scans and 0.2% of chest radiographs. While most pulmonary nodules stem from previous infections or benign causes, their clinical significance varies based on size, growth rate, and malignancy potential. Small nodules typically follow a benign course and require no intervention, whereas larger or suspicious nodules necessitate histopathological evaluation to exclude malignancy (Swensen et al., 2000). The diagnostic challenge posed by pulmonary nodules is significant due to overlapping imaging features between benign and malignant lesions, variations in nodule characteristics over time, and limitations of imaging modalities in detecting small or subtle lesions. Conventional imaging relies on visual interpretation of descriptive parameters such as size, shape, and borders, but these subjective assessments are prone to inter- and intra-observer variability, leading to potential misdiagnosis (McWilliams et al., 2013). Radiomics, an emerging field, offers a promising solution by extracting a plethora of quantitative features from medical images. This data-driven approach aims to enhance diagnostic accuracy, prognosis, and therapy response predictions. Radiomics transforms standard medical images into high-dimensional data, capturing details about pixel intensity distributions, texture patterns, and spatial relationships within the imaged tissue that are not discernible by visual inspection alone (Aerts et al., 2014). viii In the context of pulmonary nodules, several studies have demonstrated the potential of radiomics in pulmonary nodule evaluation. For instance, Hawkins et al. (2016) showed that machine learning models incorporating radiomic features achieved sensitivities of up to 90% and specificities of 85% in distinguishing malignant from benign nodules. Additionally, McWilliams et al. (2013) developed a risk prediction model based on nodule characteristics, patient demographics, and smoking history, providing valuable insights into the likelihood of malignancy in early-stage nodules. Radiomics represents a paradigm shift in medical imaging, transforming conventional images into high-dimensional data that reveal detailed tissue characteristics. The extraction and analysis of quantitative imaging features allow for a comprehensive evaluation of pulmonary nodules beyond what is discernible through traditional visual inspection. By examining features related to texture, shape, and intensity, radiomics provides an in-depth understanding of the nodule's biological behavior. This research aims to harness the full potential of radiomics, emphasizing its role in differentiating between benign and malignant nodules with greater precision. The objective of this study is to use IBEX, a radiomics based open-source software for the evaluation and qualitative assessment of pulmonary nodules on high resolution chest tomography over a period of one year OBJECTIVES The primary objective of this research is to explore and validate the use of radiomics in the evaluation and qualitative assessment of pulmonary nodules through the utilization of the Imaging Biomarker Explorer (IBEX), an open-source radiomics software.