Radiomics analysis of the tissue content of the esophageal hiatus of normal subjects without esophageal symptoms (establishing control values)

Abstract

Introduction: The opening in the diaphragm through which the esophagus enters from the thorax into the abdomen, i.e., esophageal hiatus is formed by the right crus of the diaphragm. During esophageal peristalsis, the lower end of the esophagus, i.e., lower esophageal sphincter (LES), slides from the abdomen into chest, which we suspect is due to the presence of fat in the hiatus that serves as a lubricant. A recent study from our laboratory found that in patients with achalasia esophagus, the LES and crus of diaphragm are firmly anchored together, which we suspect is due to the loss of fat from the hiatus. Gray scale intensity analysis of the radiological images can be used to determine the nature of tissue, i.e., fat, connective tissue, muscle etc. Goal: To establish gray scale values of the esophageal hiatus from the CT scan images in controls. Methods: 42 patients who had undergone an abdominal CT scan for symptoms unrelated to upper gastrointestinal pathology were selected for this analysis. These patients had no esophageal symptoms and were not taking any medications related to the esophagus and stomach, which included acid inhibition medications. The CT scan was visualized in the Visage (San Diego, CA), a software program that allows rotation of images in X, Y and Z planes to visualize the esophageal hiatus in a 2D image. A software (Segcrus) was built to perform the grayscale intensity analysis of the 3 regions of equal size in the image, 1) subcutaneous fat, 2) muscle, and 3) esophageal hiatus. The 3 regions of interest were demarcated manually in the 2D image. The inner margin of the hiatus was selected as the region of interest for the hiatus image. The software generated a statistical non-parametric method approximation using kernel density estimation (KDE) of the grey scale intensity distributions of the 3 regions of interest. The median values were computed to represent the central tendencies of the distributions, providing an insight into the typical greyscale values of the tissue types. Results: distinct patterns in the region intensity distributions for hiatal region, muscle, and fat were seen. For the hiatus, the values were (Median = 105), interquartile range (IQR = 80.3), and range (Min = 61.5, Max = 137.95), providing a comprehensive description of central tendency, variability, and data spread. For muscle, (Median = 139.92), interquartile range (IQR = 71.1), and range (Min = 119.67, Max = 157.17), and for fat, (Median = 48.24), interquartile range (IQR = 52.9), and range (Min = 23.99, Max = 67.79. Hiatus to fat tissue intensity overlap was not related to BMI (r=0.17, p=0.29), or Age (r=-0.04, p=0.82). There was a notable correlation between the overlap of Hiatal-Muscle tissue intensity and Age (r=0.38, p=0.012). However, such correlation was not observed with BMI (r=-0.3, p=0.06). Additionally, there was no significant correlation between the overlap of Hiatal-Fat tissue intensity and gender (r=-0.26, p=0.09), and similarly, the overlap of Hiatal-Muscle intensity showed no correlation with sex (r=0.09, p=0.58). The results highlight the variability and central tendencies within these tissue types across the normal subjects. Conclusion: These analysis in subjects with no esophageal symptoms may be used as control values to determine if there are difference in the tissue content of the esophageal hiatus between normal and patients with various esophageal motility disorders. R01 DK109376. This is the full abstract presented at the American Physiology Summit 2024 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.