Quantitative analysis of background parenchymal enhancement in whole breast on MRI: Influence of menstrual cycle and comparison with a qualitative analysis

Publication date: Available online 11 April 2018
Source:European Journal of Radiology
Author(s): Yongsik Jung, Seong Kyun Jeong, Doo Kyoung Kang, Yeorae Moon, Tae Hee Kim
ObjectiveWe quantitatively analyzed background parenchymal enhancement (BPE) in whole breast according to menstrual cycle and compared it with a qualitative analysis method.Materials and MethodsA data set of breast magnetic resonance imaging (MRI) from 273 breast cancer patients was used. For quantitative analysis, we used semiautomated in-house software with MATLAB. From each voxel of whole breast, the software calculated BPE using following equation: [(signal intensity [SI] at 1 min 30 sec after contrast injection – baseline SI)/baseline SI] × 100%.ResultsIn total, 53 patients had minimal, 108 mild, 87 moderate, and 25 marked BPE. On quantitative analysis, mean BPE values were 33.1% in the minimal, 42.1% in the mild, 59.1% in the moderate, and 81.9% in the marked BPE group showing significant difference (p = 0.009 for minimal vs. mild, p < 0.001 for other comparisons). Spearman's correlation test showed that there was strong significant correlation between qualitative and quantitative BPE (r = 0.63, p < 0.001).The mean BPE value was 48.7% for patients in the first week of the menstrual cycle, 43.5% in the second week, 49% in the third week, and 49.4% for those in the fourth week. The difference between the second and fourth weeks was significant (p = 0.005). Median, 90th percentile, and 10th percentile values were also significantly different between the second and fourth weeks but not different in other comparisons (first vs. second, first vs. third, first vs. fourth, second vs. third, or third vs. fourth).ConclusionQuantitative analysis of BPE correlated well with the qualitative BPE grade. Quantitative BPE values were lowest in the second week and highest in the fourth week.



from Imaging via alkiviadis.1961 on Inoreader https://ift.tt/2H2Dl4A