High grade and non-high grade ductal carcinomain situon dynamic MR mammography: characteristic findings for signal increase and morphological pattern of enhancement

Publisher Website
Google Scholar
Abstract
The objective of this review is to describe characteristic MR mammographic findings for signal increase and morphological patterns of enhancement in pure ductal carcinoma in situ (DCIS) and to differentiate between high grade and non-high grade lesions. The dynamic MR examination (1.5 T unit, contrast enhanced T1 weighted two dimensional fast field echo, 96 ms repetition time, 5.0 ms echo time, 80° flip angle) of 39 consecutive patients with pure DCIS was evaluated retrospectively. Categories were defined for signal increase (C1=normal, C2=slow, continuous, C3=strong initial and slow further increase, C4=strong initial increase followed by a plateau phenomenon, and C5=strong initial increase followed by a washout phenomenon) and morphological patterns (M0=no pattern observed, M1=linear or linear-branched, M2=segmental dotted or granular, M3=segmental homogeneous, and M4=focal spot-like). Time–intensity curves showing a C4 and C5 signal increase were considered suspicious for malignancy. All cases were correlated with histology. 62% of all tumours had a plateau or washout (C4, C5), 77% showed a strong initial signal increase (C3–C5). On evaluation of time–intensity curves alone MR mammography (MRM) findings were suspicious for malignancy in 62% of all DCIS cases. A segmental enhancement was found in 82% of all enhancing tumors and the M2 pattern in 73%. In a combined analysis of signal increase and morphology, 70% of non-high grade and 92% of high grade DCISs were correctly described as suspicious. The difference between non-high grade and high grade DCIS was not significant (p=0.148), while significant differences were found between G1 and G3 DCISs and between G1 and G2 DCISs (ppp<0.05). Different average size of G1, G2 and G3 DCIS on pathology cannot be excluded as a reason for differences found. Normal MRM seems to exclude high grade DCIS.

This publication has 35 references indexed in Scilit: