Non-Mass Like Enhancement (NMLE) in Breast Magnetic Resonance Imaging: A Proposal of Pathological-Radiological Correlation
Sirisa Kandel, Prasanna Kumar, Qian Liu, Song Liu, Thaer Khoury. Roswell Park Cancer Institute, Buffalo, NY
Background: Radiological correlation with the pathological findings in breast core needle biopsy (CNB) is essential in breast disease management. Unlike mammography and breast sonogram, there is no data to suggest correlation between NMLE seen by Magnetic resonance imaging (MRI) and the pathology results. The purpose of this study is to test this correlation.
Design: The radiological and pathological findings of non-palpable mammographically occult 87 consecutive NMLE lesions targeted by CNB were reviewed. MRI imaging was performed utilizing a 1.5Tesla magnet and appropriate breast coil. Pre and post contrast dynamic fat subtracted images were obtained. The radiologic findings were classified according to the established lexicon by American College of Radiology. The distribution was classified into linear/ductal, regional, or segmental and enhancement pattern was classified into focus, clumped, homogenous or heterogeneous. The histologic findings were recorded including carcinoma, fibroadenoma, radial scar, papilloma, stromal fibrosis, inflammation, pseudoangiomatous stromal hyperplasia (PASH), sclerosing adenosis, UDH/ADH/DCIS, apocrine metaplasia, duct ectasia, columnar cell lesion and fat necrosis. One or more histologic entity could be present in one biopsy and one or more radiologic feature could be present in the same MRI image. These lesions were histologically graded from 0 to 3, based on the degree of involvement. 2 X 2 tables Fisher's exact test is used to test the radiological/pathological correlation.
Results: Invasive carcinoma was seen in 4 cases (4.6%). MRI was performed for screening in 3 patients, for staging in 61, and for high risk in 23. All carcinoma cases were seen in the staging group. The MRI features with correlated pathology features are presented in table 1.
Conclusions: Our results suggest it is possible to test the MRI finding of NMLE correlation with a specific pathologic finding. The study is slightly limited, especially for invasive carcinoma; due to the limited number of cases in each category.
|NMLE MRI feature||Histologic correlation||n (%)||p value*|
|Linear and Clumped||Papilloma||4/9 (44.4%)||0.019|
|Focus||Fat necrosis||4/4 (100%)||0.021|
|Clumped||UDH or ADH or DCIS||18/30 (60%)||0.013|
|Focus and Clumped||Apocrine metaplasia||4/9 (44.4%)||0.044|