Dynamic breast MR imaging: is parametric mapping superior to image subtraction in lesion detection?

Abstract

We compared the detection of malignant lesions in two different methods of parametric-guided analysis to the widely used early subtraction images. The parametric colour-coded overlays were defined by the increase of signal intensity after contrast injection and the course of the time signal intensity curve. Exams of 30 patients with histopathological evidence of 32 invasive breast carcinomas were evaluated. Five hundred five areas of additional enhancing tissue were found in the early subtraction images. The sensitivity was 100 in the subtraction images and 100 and 93.8 in the parametric images, respectively, the specificity 67.5, 84.2 and 88.7, respectively. On average 4.4 ROI-measurements were needed to find the highest increase of signal intensity in the subtraction series compared to 1.9 in the parametric images. Properly selected thresholds in parametric imaging may be helpful in the differentiation of enhancing tissue. Furthermore the parametric image-guided ROI selection significantly speeds up the analysis and makes it safer as a lower rate of ROI-measurements is needed to find the strongest enhancement.

Can Color-Coded Parametric Maps Improve Dynamic Enhancement Pattern Analysis in MR Mammography?

“…In contrast to Hauth et al, the Kurz group reported needing a greater number of ROI measurements to identify the strongest enhancing part of a lesion. They were able to reduce the number of ROI placements needed by using parametric maps [17]. Our own results are in accordance with Hauth et al and verify the diagnostic benefit of curve type assessment using CCPMs compared to ROI measurements in a large patient collective.…”

“…Kurz et al reported a reduced number of false positive findings using parametric maps calculated by the delayed phase curve type after application of a threshold for initial enhancement. However, this group compared CCPM results only with initial enhancement on early subtraction images [17]. Therefore, no systematic comparison of both methods could be drawn.…”

MRI features of intraductal papilloma of the breast: sheep in wolf's clothing?

“…Disagreements on interpretation were resolved in consensus. The MR images were analyzed with a software that allowed the creation of subtraction images in real time, TIC drawn by 3 × 3 pixel size, maximum intensity projection reconstructions and color-coded parametric images (15, 16). The radiologists were aware of the clinical, mammographic, and sonographic findings consistent with clinical practice.…”

“…For dynamic analysis, the time intensity curve (TIC) was drawn from a 3 × 3 pixels sized subjectively selected region of interest (ROI) in the most enhancing region of the papilloma (16–19). The TIC was classified as one of three types (17, 19): Type 1: Initial increase in signal intensity followed by more than 10% increase in signal intensity in the delayed phase (‘persistent’); Type 2: Initial rapid increase rise in signal intensity, followed by less than 10% change in signal intensity in the delayed phase (‘plateau’); Type 3: Initial rapid increase in signal intensity followed by more than 10% reduction in signal intensity in the delayed phase (‘washout’). Three analytic strategies described in the literature were selected for structured assessment of the lesions: the Göttingen score developed by Fischer et al (Table 2) (20, 21), and the diagnostic rules as described by Kinkel et al (Fig.…”