Keywords:
Breast, MR, Diagnostic procedure
Authors:
P. Jagmohan1, C. Grippo2, C. Sodano3, P. Clauser4, P. Kapetas3, T. H. Helbich3, P. A. T. Baltzer3; 1Singapore/SG, 2Rome/IT, 3Vienna/AT, 4Vienna, Vienna/AT
DOI:
10.26044/ecr2019/C-2750
Methods and materials
Patient population
An anonymized database of consecutively collected histologically verified breast MRI cases (n=70,
24 malignant,
46 benign) was read by two independent off-site fellowship trained readers.
Histopathological analysis was obtained either by image-guided biopsy (core / vacuum assisted biopsy under US/mammographic / MRI guidance) or surgical excision with the histopathological assessment done by board-certified breast pathologists.
Data collection and analysis
The structured reading of predefined lesions included T2w and contrast-enhanced dynamic T1w scans.
Ratings were assigned to each lesion as per the Kaiser score.
The time signal intensity curve type was determined by comparing the initial and delayed enhancement.
The initial enhancement was determined by using three approaches:
-first (1st) ,
second (2nd) and peak(maximum) of either 1st or 2nd post-contrast timepoints.
-last time point was used to determine the delayed enhancement.
-the initial enhancement determined by three approaches were then compared to the last timepoint to determine the curve type.
Visual assessment of curve types was used for this study.
The curve type was based on the BIRADS lexicon.
For the Kaiser score a continuous signal increase was defined as persistent,
a steady signal over time was called a plateau; a signal drop from the initial to the delayed phase was referred to as wash-out.[2] (Fig.
2)
Statistical analysis included ROC-analysis of the Kaiser score ratings based on these variations against the reference standard (benign vs malignant).