Preoperative prediction of the size of pure ductal carcinoma in situ using three imaging modalities as compared to histopathological size: does magnetic resonance imaging add value?
MetadataShow full item record
PURPOSE: The purpose of this study was to evaluate whether magnetic resonance imaging (MRI) and ultrasonography add value to traditional mammography in an Asian population with ductal carcinoma in situ (DCIS). METHODS: Data of 244 patients with pure DCIS treated at Severance Hospital between 2013 and 2015 were analyzed retrospectively. Data extracted included age, preoperative diagnosis, tumor size on preoperative imaging studies, and final histopathological tumor type and size, including hormone receptor status. The extent of correlation between imaging and histopathological tumor sizes was evaluated using a variety of methods, including Bland-Altman analysis. RESULTS: The mean patient age was 52.39 years (SD = 10.31). The mean measurements of the tumor on preoperative ultrasonography, mammography, MRI, and histopathology were 1.80 (SD = 1.23) cm, 2.97 (SD = 1.92) cm, 2.53(SD = 1.84) cm, and 1.88 (SD = 1.36) cm, respectively. The mean differences in tumor size between ultrasonography, mammography, and MRI compared with histopathology were -0.09 (SD = 1.39), 1.09 (SD = 1.89), and 0.65 (SD = 1.78), respectively. The correlation between the sizes was significant with r values for ultrasonography, mammography, and MRI of 0.447 (SE = 0.061), 0.375 (SE = 0.042), and 0.409 (SE = 0.043), respectively. Mammography and MRI estimated tumor size significantly better for patients older than 50 years (p = 0.045 and <0.001, respectively). Mammography also provided good estimation for patients with a body mass index under 25 (p = 0.041). CONCLUSION: MRI is better at estimation of histopathological DCIS size compared with mammography. However, ultrasonography had better estimation compared with MRI and mammography, probably owing to the high breast density in this population.
The following license files are associated with this item: