Affiliation:
1. Departmentof Ultrasound, Shanghai General Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China
Abstract
Objective: To compare the diagnostic value of contrast-enhanced ultrasound (CEUS)+conventional ultrasound vs MRI for malignant non-mass breast lesions (NMLs). Methods: A total of 109 NMLs detected by conventional ultrasound and examined by both CEUS and MRI were retrospectively analysed. The characteristics of NMLs in CEUS and MRI were noted, and agreement between the two modalities was analysed. Sensitivity, specificity, positive-predictive value (PPV), negative-predictive value (NPV), and area under the curve (AUC) of the two methods for diagnosing malignant NMLs were calculated in the overall sample and subgroups of different sizes(<10 mm, 10–20 mm, >20 mm). Results: A total of 66 NMLs detected by conventional ultrasound showed non-mass enhancement in MRI. Agreement between ultrasound and MRI was 60.6%. Probability of malignancy was higher when there was agreement between the two modalities. In the overall group, the sensitivity, specificity, PPV, and NPV of the two methods were 91.3%, 71.4%, 60%, 93.4% and 100%, 50.4%, 59.7%, 100%, respectively. The diagnostic performance of CEUS+conventional ultrasound was better than that of MRI (AUC: 0.825 vs 0.762, p = 0.043). The specificity of both methods decreased as lesion size increased, but sensitivity did not change. There was no significant difference between the AUCs of the two methods in the size subgroups (p > 0.05). Conclusion: The diagnostic performance of CEUS+conventional ultrasound may be better than that of MRI for NMLs detected by conventional ultrasound. However, the specificity of both methods decrease significantly as lesion size increases. Advances in knowledge: This is the first study to compare the diagnostic performance of CEUS+conventional ultrasound vs that of MRI for malignant NMLs detected by conventional ultrasound. While CEUS+conventional ultrasound appears to be superior to MRI, subgroup analysis suggests that diagnostic performance is poorer for larger NMLs.
Publisher
Oxford University Press (OUP)
Subject
Radiology, Nuclear Medicine and imaging,General Medicine