Relationship between CT Number and Haemoglobin Level - A Descriptive Study

Abstract

Background: Haemoglobin, the iron-containing metalloprotein in red blood cells, is the principal determinant of blood density. CT numbers (Hounsfield Units, HU) represent relative X-ray attenuation of tissues. Because iron attenuates X-rays more than soft tissue, it is physiologically plausible that haemoglobin concentration influences CT attenuation in blood-filled structures. Despite this, routine radiological reporting does not account for the effect of haemoglobin on CT numbers. This study aimed to quantify that relationship and assess its utility for incidental anaemia detection.

Methods: A descriptive study was conducted over 18 months at the Department of Radiodiagnosis, Government T.D. Medical College, Alappuzha. Two hundred consecutive patients undergoing unenhanced CT of the thorax and/or abdomen with a contemporaneous haemoglobin value were enrolled by convenience sampling. Regions of interest (ROIs) were placed in the aorta, inferior vena cava (IVC), left ventricle (LV), liver, and spleen. CT numbers were extracted using RadiAnt Viewer and MATLAB. Mean, median, mode, minimum, and maximum pixel values from each ROI were recorded. Pearson correlation and linear regression were performed using SPSS.

Results: Of 200 subjects (105 males, 95 females), mean haemoglobin was 12.3 g/dL in males and 11.6 g/dL in females. Strong positive Pearson correlations were observed between haemoglobin and mean CT numbers of the aorta (r = 0.816), IVC (r = 0.751), and LV (r = 0.737), all significant at p < 0.01. Mean, median, and mode values showed similar correlations; minimum and maximum values did not correlate. CT numbers of the liver (r = 0.19) and spleen (r = 0.14) showed no clinically meaningful correlation.

Conclusion: Haemoglobin levels are positively correlated with CT numbers in vascular blood pools (aorta, IVC, left ventricle) but not in parenchymal organs (liver, spleen). These findings support the feasibility of incidental anaemia detection and haemoglobin estimation from routine unenhanced CT images, with potential applications in clinical management and forensic radiology.