A Multimodality Imaging Approach to Defining Risk in Patients With Acute Pulmonary Embolism

•Risk assessment in PE uses clinical features and qualitative assessment of RV function. •Risk scores should reflect the consequences of obstruction to pulmonary blood flow. •Imaging may identify characteristics amenable to new reperfusion strategies. •These include clot burden, RV function, and biomarkers of cardiac injury or overload. •Combining quantitative echo, CT, and clinical variables may be useful in PE. Background Morbidity and mortality for acute pulmonary embolism (PE) remain high. Therapies such as catheter-directed thrombolysis may improve outcomes, but these are generally reserved for higher-risk patients. Imaging may help guide the use of the newer therapies, but current guidelines focus more on clinical factors. Our goal was to create a risk model that incorporated quantitative echocardiographic and computed tomography (CT) measures of right ventricular (RV) size and function, thrombus burden, and serum biomarkers of cardiac overload or injury. Methods This was a retrospective study of 150 patients evaluated by a PE response team. Echocardiography was performed within 48 hours of diagnosis. Computed tomography measures included RV/left ventricular (LV) ratio and thrombus load (Qanadli score). Echocardiography was used to obtain various quantitative measures of RV function. We compared characteristics of those who met the primary endpoint (7-day mortality and clinical deterioration) to those who did not. Receiver operating curve analysis was used to assess the performance of different combinations of clinically relevant features and the association with adverse outcomes. Results Fifty-two percent of patients were female, with age 62 ± 17 years, systolic blood pressure 123 ± 25 mm Hg, heart rate 98 ± 19, troponin 3.2 ± 35 ng/dL, and b-type natriuretic peptide (BNP) 467 ± 653. Fourteen (9.3%) were treated with systemic thrombolytics, 27 (18%) underwent catheter-directed thrombolytics, 23 (15%) were intubated or required vasopressors, and 14 (9.3%) died. Patients who met the primary endpoint (44%) versus those who did not (56%) had lower RV S' (6.6 vs 11.9 cm/sec; P < .001) and RV free wall strain (−10.9% vs −13.6%; P = .005), higher RV/LV ratio on CT, and higher serum BNP and troponin levels. Receiver operating curve analysis demonstrated an area under the curve of 0.89 for a model that included RV S’, RV free wall strain and tricuspid annular plane systolic excursion/RV systolic pressure ratio from echo, thrombus load and RV/LV ratio from CT, and troponin and BNP levels. Conclusion A combination of clinical, echo, and CT findings that reflect the hemodynamic effects of the embolism identified patients with adverse events related to acute PE. Optimized scoring systems that focus on reversible abnormalities attributable to PE may allow more appropriate triaging of intermediate- to high-risk patients with PE for early interventional strategy. Morbidity and mortality for acute pulmonary embolism (PE) remain high. Therapies such as catheter-directed thrombolysis may improve outcomes, but these are generally reserved for higher-risk patients. Imaging may help guide the use of the newer therapies, but current guidelines focus more on clinical factors. Our goal was to create a risk model that incorporated quantitative echocardiographic and computed tomography (CT) measures of right ventricular (RV) size and function, thrombus burden, and serum biomarkers of cardiac overload or injury. This was a retrospective study of 150 patients evaluated by a PE response team. Echocardiography was performed within 48 hours of diagnosis. Computed tomography measures included RV/left ventricular (LV) ratio and thrombus load (Qanadli score). Echocardiography was used to obtain various quantitative measures of RV function. We compared characteristics of those who met the primary endpoint (7-day mortality and clinical deterioration) to those who did not. Receiver operating curve analysis was used to assess the performance of different combinations of clinically relevant features and the association with adverse outcomes. Fifty-two percent of patients were female, with age 62 ± 17 years, systolic blood pressure 123 ± 25 mm Hg, heart rate 98 ± 19, troponin 3.2 ± 35 ng/dL, and b-type natriuretic peptide (BNP) 467 ± 653. Fourteen (9.3%) were treated with systemic thrombolytics, 27 (18%) underwent catheter-directed thrombolytics, 23 (15%) were intubated or required vasopressors, and 14 (9.3%) died. Patients who met the primary endpoint (44%) versus those who did not (56%) had lower RV S' (6.6 vs 11.9 cm/sec; P < .001) and RV free wall strain (−10.9% vs −13.6%; P = .005), higher RV/LV ratio on CT, and higher serum BNP and troponin levels. Receiver operating curve analysis demonstrated an area under the curve of 0.89 for a model that included RV S’, RV free wall strain and tricuspid annular plane systolic excursion/RV systolic pressure ratio from echo, thrombus load and RV/LV ratio from CT, and troponin and BNP levels. A combination of clinical, echo, and CT findings that reflect the hemodynamic effects of the embolism identified patients with adverse events related to acute PE. Optimized scoring systems that focus on reversible abnormalities attributable to PE may allow more appropriate triaging of intermediate- to high-risk patients with PE for early interventional strategy.