ABSTRACT
PURPOSE: To compare "cut-film" (film hard-copy) angiography (CFA) with digital pulmonary angiography in the detection of pulmonary embolism (PE). MATERIALS AND METHODS: Thirty-six adult patients (39 lungs) underwent selective digital pulmonary angiography for suspected PE. Imaging was repeated in one selected projection by using cut film. The standard was consensus interpretation of both CFA and digital angiographic images and clinical course. Three vascular radiologists subsequently reviewed the digital and cut-film images in a blinded fashion and ranked the likelihood of the presence of PE on a five-point scale. The two modalities were compared by means of receiver operating characteristic (ROC) analysis. Image quality (i.e., sharpness, opacification of subsegmental vessels, and exposure) was judged on a three-point scale. The highest-order pulmonary artery branch seen on each study was recorded. RESULTS: ROC curve analyses for all three operators showed similar diagnostic performance for digital pulmonary angiography and CFA, with one operator showing better performance with digital subtraction angiography than with CFA (P = .04). Compared with the final diagnosis, single-plane digital pulmonary angiography had higher sensitivity for the detection of PE than had CFA. The specificity was 100% for both modalities. The mean score in patients with findings positive for PE was higher in the digital pulmonary angiography group than in the CFA group (P < .005). There was no difference in the mean score in patients who did not have a PE. There also was no difference in the smallest detectable subsegmental branch (P = .87) or in the average estimate of image quality. CONCLUSION: Selective digital pulmonary angiography and CFA offer similar diagnostic performance and image quality. Digital pulmonary angiography is a reasonable alternative to CFA in the diagnosis of PE.
