Differentiation of ductal carcinoma in situ versus fibrocystic changes by magnetic resonance imaging: are there pathognomonic imaging features?. (October 2017)

• Record Type: Journal Article
• Authors: Dietzel, Matthias; Kaiser, Clemens G; Wenkel, Evelyn; Clauser, Paola; Uder, Michael; Schulz-Wendtland, Rüdiger; Baltzer, Pascal AT
• Is Part Of: Acta radiologica. Volume 58:Number 10(2017:Oct.)
• Page Start: 1206
• Page End: 1214
• Publication Date: 2017-10
• Subjects: Magnetic resonance imaging -- ductal carcinoma in situ (DCIS) -- fibrocystic changes -- diagnostic accuracy -- breast cancer; Radiology, Medical -- Periodicals; Radiography, Medical -- Periodicals; Radiotherapy -- Periodicals; 616.0757
• DOI: 10.1177/0284185117690420
• Languages: English
• ISSNs: 0284-1851

Abstract

Background: In breast magnetic resonance imaging (MRI), the diagnosis of ductal carcinoma in situ (DCIS) remains controversial; the most challenging cause of false-positive DCIS diagnosis is fibrocystic changes (FC). Purpose: To search for typical and pathognomonic patterns of DCIS and FC using a standard clinical MRI protocol. Material and Methods: Consecutive patients scheduled for breast MRI (standardized protocols @ 1.5T: dynamic-T1-GRE before/after Gd-DTPA [0.1 mmol/kg body weight (BW)]; T1-TSE), with subsequent pathological sampling, were investigated. Sixteen MRI descriptors were prospectively assessed by two experienced radiologists in consensus (blinded to pathology) and explored in patients with DCIS (n = 77) or FC (n = 219). Univariate and multivariate statistics were performed to identify the accuracy of descriptors (alone, combined). Furthermore, pathognomonic descriptor-combinations with an accuracy of 100% were explored (χ 2 statistics; decision trees). Results: Six breast MRI descriptors significantly differentiated DCIS from FC ( P corrected < 0.05; odds ratio < 7.9). Pathognomonic imaging features were present in 33.8% (n = 100) of all cases allowing the identification of 42.9% of FC (n = 94). Conclusion: Pathognomonic patterns of DCIS and FC were frequently observed in a standard clinical MRI protocol. Such imaging patterns could decrease the false-positive rate of breast MRI and hence might help to decrease the number of unnecessary biopsies in this clinically challenging subgroup.