Incremental Diagnostic Value of Computed Tomography Attenuation in Differentiating Malignant Pericardial Effusion: A Retrospective Observational Study

Abstract

Objective: Malignant pericardial effusion (MPE) is associated with poor prognosis and frequently presents as cardiac tamponade. While cytology is the diagnostic gold standard, its sensitivity is limited. Computed tomography (CT) attenuation, measured in Hounsfield Units (HU), may reflect fluid composition and assist in the non-invasive differentiation of MPE.

Method: This retrospective, single-center study included 102 patients who underwent percutaneous pericardiocentesis and thoracic CT. Patients were classified as having malignant or non-malignant effusion based on pathological findings. CT attenuation was measured at three standardized axial levels. Diagnostic performance was assessed using multivariate logistic regression and receiver operating characteristic (ROC) analysis. Two predictive models were compared: Model 1 (clinical and laboratory variables) and Model 2 (Model 1 + CT attenuation).

Results: MPE was diagnosed in 44 patients (43.1%). CT attenuation values were significantly higher in the MPE group (median 24.4 HU vs. 9.3 HU, P < 0.001). On multivariate analysis, male sex, elevated pericardial fluid protein, low glucose, and high lactate dehydrogenase were independent predictors of MPE. CT attenuation also emerged as an independent predictor when added to the model (Model 2) (odds ratio [OR] = 1.076, 95% confidence interval [CI]: 1.026-1.128, P = 0.003). The inclusion of CT attenuation improved the model's diagnostic performance (area under the curve [AUC]: 0.893 for Model 2 vs. 0.860 for Model 1). Model 2 demonstrated superior diagnostic performance (AUC = 0.893), with a CT attenuation cut-off of 16.45 HU yielding a sensitivity of 88.2% and a specificity of 78.3%.

Conclusion: CT attenuation provides incremental diagnostic value in identifying MPE when combined with conventional clinical and biochemical parameters. In settings where rapid diagnosis is critical, its non-invasive and reproducible nature may support early detection of malignant conditions.