CTA-derived Hounsfield units >105 distinguish leaflet fibrosis from thrombosis with 84% sensitivity and 91% specificity in patients with bioprosthetic heart valve dysfunction.
Observational (n=141)
Blinded image analysis
No
Does computed tomography angiography-derived Hounsfield unit measurement differentiate between thrombosis and leaflet fibrosis in patients with bioprosthetic heart valve dysfunction?
CTA-derived Hounsfield units can accurately differentiate between bioprosthetic valve thrombosis and fibrosis, with values >105 HU strongly suggesting fibrosis that is unresponsive to oral anticoagulation.
Effect estimate: AUC 0.94 (95% CI 0.897–0.983)
Absolute Event Rate: 137% vs 87%
p-value: p=<0.001
Abstract Purpose Our hypothesis was that computed tomography angiography (CTA)-derived Hounsfield units (HU) can differentiate between thrombosis and leaflet fibrosis (defined by a lack of response to oral anticoagulation) in patients with bioprosthetic heart valve dysfunction. Materials and methods Valvular leaflet HU were retrospectively assessed in 95 patients (derivation cohort) undergoing CTA 35 days after bioprosthetic heart valve (BHV) implantation showing signs of subclinical leaflet thrombosis (hypoattenuated leaflet thickening, HALT). A second (validation) cohort included 46 patients undergoing CTA for suspected BHV dysfunction 2 years interquartile range IQR 1.5–5.0 after valve replacement. This study included CTA between May 2012 and December 2017. Results In the derivation cohort, the median HU (95 patients) was 87 (IQR 77; 96). In the validation cohort, patients with resolution of findings in a follow-up CTA after newly initiated anticoagulation (“thrombosis” subgroup, 19 patients) similarly demonstrated HU of 87 (IQR 74; 100) ( p = 0.816). In contrast, patients without improvement under oral anticoagulation (“fibrosis” subgroup, 27 patients) exhibited a median of 137 HU (IQR 116, 164; p < 0.001 vs. thrombosis subgroup). In multivariable Cox regression analysis, lower HU were an independent predictor of thrombosis. C-statistics demonstrated an area under the receiver operating characteristic curve of 0.94 ± 0.02 (CI 0.897–0.983, p < 0.001) with a value of 105 HU resulting in a sensitivity of 84% and a specificity of 91% for the differentiation between thrombosis and fibrosis. Conclusion A value of 105 HU on CTA provides good discriminatory power to distinguish between leaflet fibrosis (as defined by a lack of response to oral anticoagulation) and thrombosis after bioprosthetic valve replacement and may help in choosing optimal treatment. Graphical Abstract
Schmitt et al. (Thu,) conducted a observational in Patients with bioprosthetic heart valve dysfunction suspected of leaflet thrombosis or leaflet fibrosis (n=141). Computed tomography angiography-derived Hounsfield units assessment vs. Clinical diagnosis of thrombosis vs. leaflet fibrosis defined by response to oral anticoagulation was evaluated on Differentiation between leaflet thrombosis and leaflet fibrosis defined by response to oral anticoagulation (AUC 0.94, 95% CI 0.897–0.983, p=<0.001). CTA-derived Hounsfield units >105 distinguish leaflet fibrosis from thrombosis with 84% sensitivity and 91% specificity in patients with bioprosthetic heart valve dysfunction.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: