Surgical planning of breast cancer relies on adequate assessment of the extent of disease including the size of the primary tumor and the presence or absence of additional disease. Although breast MRI is used for surgical planning, recent studies showed that it led to no improvement in outcome with higher rates of mastectomy [1,2]. Critics of breast MRI state that the exam leads to overdiagnosis of clinically irrelevant disease. We evaluated a cohort of women with a known cancer who underwent MRI guided needle localization (MRNL) for a lesion that was not amenable to MRI guided biopsy. The purpose of this study is to examine the IER of synchronous cancers detected on breast MRI, as a possible marker of biological significance synchronous cancers.

This is an institutional review board approved retrospective study of patients with biopsy-proven breast cancer who underwent pre-operative MRNL for a synchronous cancer found on MRI. Mean percentage increase of the signal between the pre-contrast and second postcontrast DCE-MRI series (initial enhancement ratio [IER]) was calculated retrospectively using a commercially available software (DynaCAD, Invivo). The highest percentage increase in signal intensity on the initial contrast-enhanced series for a given lesion based on placement of multiple pixels was identified and recorded as IER. IER was then analyzed with respect to biological significance, defined as invasive cancers, for the index and synchronous cancers. Correlation was made with whether the cancer was invasive or in situ, the morphologic features and size of the cancers, and the grade of the tumor. Sensitivity and specificity were calculated using a cut-off of 135% for IER [6]. Spearman rank correlations were used to characterize the association of pathology, grade, and IER.

All patients that underwent MRNL from 2008-2014 were reviewed and 26 patients with synchronous cancers were identified. A total of 55 cancers were included, 3 patients had 2 synchronous cancers each. Of the 26 index cancers, 22 cancers were invasive and 22 cancers were invasive or high grade DCIS. Of the 29 synchronous cancers, 16 were invasive and 20 were invasive or high grade DCIS. The index cancers were larger tumors: 23 were masses (> 5mm), 3 non mass enhancement (NME), and no foci (<5mm). The synchronous cancers included 9 foci (<5mm), 5 NME, and 15 masses (>5mm). The mean initial enhancement ratio for 38 invasive cancers was 175% (range 5-340%), and for 17 non-invasive cancers was 102% (26-170%). Spearman analysis demonstrated a significant positive correlation between IER and invasive index cancers (p=0.012) but not synchronous invasive cancers (p=0.66). A trend towards significance was noted for index invasive cancers (p=0.09) which may reflect the small sample size or the smaller size of synchronous cancers. There was no significant correlation between synchronous invasive cancers and tumor grade (p=0.857). Using a cut-off of 135%, greater than the BIRAD recommended cut-off for early enhancement of >100%, the sensitivity is 81.6% with a confidence interval (CI) of 65.7-92.3%. The specificity is 76.5% with a CI of 50.1-93.2%. The positive predictive value is 88.6% with a CI of 73.3-96.8% and the negative predictive value is 65.0% with a CI of 40.8-84.6%. The data demonstrates a significant positive correlation IER and invasive index cancers. There is not a significant correlation between IER and the synchronous cancers, however morphology such as tumor size were not considered. When the cancers are combined to form a larger cohort there is a moderate to high sensitivity and specificity when using IER at a threshold of 135% to identify biologically significant cancers. These tumors would likely be detected on AB-MR exam. We found there is a significant correlation between invasive cancers and IER that can aid in the detection of biologically significant cancers on AB-MR exam, demonstrating the utility of an AB-MR exam for diagnostic purposes. Further we demonstrated a potential threshold that could be used to identify biologically significant synchronous cancer and improve surgical outcome. Further evaluation including morphology and a larger cohort would strengthen this conclusion.