[en] Objective: To compare the efficacies of MRI, X-ray mammography (XMG) and Ultrasound (US) in detecting and diagnosing breast ductal carcinoma in situ (DCIS). Methods: Two hundred and forty one consecutive patients with pathology-confirmed DCIS were retrospectively recruited from January 2011 to December 2017 in PLA General Hospital. The imaging examination modalities included MRI and/or XMG and/or US. The breast imaging reporting and data system (BI-RADS) categorizations by MRI, XMG and US were compared and their sensitivities of detecting DCIS were calculated. The causes of underestimation on MRI were interpreted with the information of XMG and US. Chi-square test was used to compare the differences. Results: The diagnostic sensitivity of XMG, US and MRI was 65.9% (29/44), 71.6% (101/141) and 91.2% (145/159), respectively, with statistical significant differences (X² = 24.034, P < 0.001). Breast density and lesion type would influence the sensitivity of XMG. And the sensitivity of US was decreased because of non-mass lesion. Of the 14 cases under-evaluated as BI-RADS category 1 to 3 on MRI, 5 were corrected by XMG and/or US to BI-RADS category 4. The cause of underestimation on MRI was the coexistence of DCIS with adenoma or other benign lesion. Conclusion: The retrospective comparison of MRI, XMG and US in this study showed that MRI had significant higher sensitivity in detecting breast DCIS, while the false negative rates of XMG and US were un-negligible. (authors)