Abstract No. 218 Diagnostic accuracy of computed tomography angiogram chest in detecting pulmonary arteriovenous malformation: a comparison with invasive catheter pulmonary angiogram

Abstract

To determine the diagnostic accuracy of computed tomography angiography (CTA) of the chest in the detection of pulmonary arteriovenous malformations (pAVM) compared to invasive pulmonary catheter angiogram of the chest both in HHT and non HHT patients. This retrospective study was approved by the institutional review board. A total of 341 pulmonary angiograms were performed for either embolization (treatment) of pAVM or for diagnosis of pAVM at our institution from January 2003 to June 2020. Two angiograms were excluded as they did not have prior CTA for pAVM. The remaining 339 pulmonary angiograms had an immediate prior CTA of the chest within a mean 7 ± 30 (SD) days prior to pulmonary angiogram. CTA was acquired after intravenous administration of contrast. The kV and mA were selected based on patient weight and ranges between 80 to 140. The mean fluoroscopy time was 33 ± 19 SD with range 2.4 to 120 minutes. Positive CTA of the chest was defined as the presence of pulmonary arteriovenous malformation with at least one well-defined feeding artery and a draining vein. A 2 × 2 contingency table was constructed and sensitivity, specificity, PPV, NPV, and accuracy were calculated. One hundred and eighty-five patients were included with mean age of 46 ± 19 SD, range 3 to 87 years with 68% of female. A total of 337 CTA of the chest were positive and 319 catheter pulmonary angiograms were positive for pulmonary AVM. There were 318 true positive, 19 false positives, 1 true negative, and 1 false negative. CTA angiogram chest demonstrated a sensitivity of 99.7%, a specificity of 5.0%, a positive predictive value of 94.4%, and a negative predictive value of 50.0%. Overall, the diagnostic accuracy of CTA angiogram chest was 94.10%. False positive CTA causes were catheter angiogram did not detect micro shunt (6/339), pulmonary vein to vein shunt (5/339), CTA wrong call for reperfusion (3/339), calcified granuloma (2/339), and tortuous or dilated pulmonary vessel (2/339). In the false negative CTA was not able to detect micro shunt (1/339). The limitation of our study is low specificity as those patients that have negative CTA angiogram chest do not proceed for catheter pulmonary angiogram unless they have other comorbidities and require CTA of the chest. Our result demonstrates that CTA of the chest is highly sensitive in detecting pulmonary arteriovenous malformation. This again highlights the importance of screening of pulmonary AVM in HHT patients.