Can the “pine-needle sign” on computed tomography be used to differentiate pulmonary arteriovenous malformation from its mimics? Analysis based on dynamic contrast-enhanced chest computed tomography in adults

2017 
Abstract Objectives/purpose To determine the diagnostic value of morphological features on computed tomography (CT) in the differentiation of pulmonary arteriovenous malformation (pAVM) and its mimics. Materials and methods We retrospectively examined 59 consecutive patients (109 lesions) with initially suspected or occult pAVM who underwent contrast-enhanced chest CT from January 2006 to June 2016. All lesions were divided into two groups based on their diagnosis: AVM (n = 93) and non-AVM (n = 16). The non-AVM group comprised patients with an anomalous unilateral single pulmonary vein and congenital pulmonary venous atresia. Two board-certified radiologists reviewed the CT images and achieved consensus. Paired abnormal vessels were assessed with respect to their configuration, direction and continuity with the hilum, location, angle between them, and dilation ratio. Results All pAVM lesions had parallel, straight-running, paired abnormal vessels; most of the vessels ran through the identical segment. Conversely, in the non-AVM group, paired abnormal vessels did not run in parallel but detoured to the hilum through different segments from each other (i.e., the configuration, direction and continuity with the hilum and the location of the vessels were all significantly different between the two groups). The angle between the paired abnormal vessels was significantly narrower in the AVM group. The dilation ratio was not significantly different between the two groups. Conclusion Morphological features can be useful in the differentiation of pAVM and its mimics without contrast-enhanced CT to directly visualize the connection between the abnormal and normal pulmonary vessels. Pulmonary AVMs characteristically had a narrow angle between the two vessels, appearing like a pair of pine-needle leaves.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    23
    References
    3
    Citations
    NaN
    KQI
    []