Answers are grounded in this paper and cite their sources.
Introduction: Low blood attenuation, measured in Hounsfield units (HU) on non-contrast chest computed tomography (CT), has been proposed as an imaging marker of anemia.This study evaluated the correlation between blood attenuation (HU) on non-contrast chest CT and serum hemoglobin levels, and determined the optimal HU cutoff value, based on a combined assessment of multiple cardiovascular structures, for detecting anemia in adult patients.Material and methods: This casecontrol study included adult patients who underwent non-contrast chest CT and a complete blood count within 24 hours.A region of interest (ROI) was defined in five cardiovascular structures -the ascending aorta, aortic arch, descending aorta, and the left and right ventricles -to measure blood attenuation.Correlation analyses and ROC curves were calculated to assess performance metrics for each structure and for the combined cardiovascular structures.Results: Eighty-eight patients were included: 44 with anemia and 44 controls with normal serum hemoglobin.Mean hemoglobin levels were 9.5 ± 1.7 and 13.7 ± 1.3 g/dL, respectively (p < 0.001).A significant positive correlation was observed between blood attenuation and serum hemoglobin across all cardiovascular structures (Spearman's ρ = 0.697-0.817;p < 0.001).In the ROC curve analysis, a blood attenuation cutoff ≤ 38.0 HU in at least three of five cardiovascular structures showed high diagnostic performance for detecting anemia, with sensitivities of 77.3% to 86.4%, specificities of 93.2% to 100%, and accuracies of 71.6% to 85.2%.Moreover, diagnostic performance increased progressively with anemia severity.Conclusions: A cutoff value ≤ 38 HU based on combined blood attenuation measurements in at least three of five cardiovascular structures on non-contrast chest CT is a simple and reliable method for detecting anemia; the diagnostic performance of this cutoff value increases in direct proportion to anemia severity.
No takes yet. Share an insight, caveat, or question.
Garcia-Armas et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: