It's Complicated: Managing Nonmass Enhancement Found at Breast MRI in Patients with Newly Diagnosed Cancer

See also the article by Nguyen and Lotfalla et al in this issue.

Mary S. Newell, MD, is a professor within Emory's Department of Radiology and Imaging Sciences’ breast imaging division. She is currently chair of the committees on Breast Imaging Reporting and Data System as well as Practice Parameters and Technical Standards within the American College of Radiology, president of the Society of Breast Imaging, associate executive director of the American Board of Radiology, a member of the board of directors for the American Board of Medical Specialties, and an associate editor for Radiology.

Here is a dreaded scenario: You are reviewing images in a patient with breast cancer a few days prior to (or even worse, the morning of) the scheduled localization procedure, and you note that the patient underwent a recent MRI examination that shows greater extent of disease than demonstrated at mammography or US. Your task is to delineate for the surgeon how much tissue needs to be removed to achieve negative excision margins while preserving as much benign breast tissue as possible via your localization plan.

As demonstrated by Nguyen and Lotfalla and colleagues (1) in their detailed and well-conceived study in this issue of Radiology: Imaging Cancer, your task is complicated. The authors examined the specific situation in which patients with an index breast cancer presenting as a mass also had accompanying contiguous nonmass enhancement (NME) detected at staging breast MRI. They explored whether there were any imaging or clinical features that could predict whether the NME was also malignant and, therefore, needed to be removed with the index cancer. Based on their results (as well as those of other studies), the answer at this point is no. The associated NME was malignant in the majority of patients (64%, 37 of 58) but not with high enough frequency to warrant automatic inclusion in the localization field without presurgical histologic confirmation. There were no MRI characteristics of either the index mass or the NME that could be relied on to improve the positive predictive value for cancer. Segmental distribution of the NME was more common in malignant than benign cases but did not reach statistical significance. Additionally, use of MRI overestimated the size of cancerous NME by an average of 1.3 cm, with only 52% of patients having an NME span within 2 cm of the actual pathologic tumor span. In practice, this size discrepancy can potentially result in a larger excision specimen than needed, with attendant implications for cosmetic outcomes. The authors found that the presence of mammographic calcifications did not help predict whether the NME was malignant. Calcifications, when present, were noted to be a better predictor of tumor extent than NME, with the median calcification span essentially matching true histologic span. Interestingly and unexpectedly, the authors found that a low Ki-67 index, indicative of low tumor aggressiveness, was positively associated with malignant NME.

Other recent literature shows areas of agreement, as well as some differing results when compared with the current study. Park et al (2) (66 cases of NME) found that while, on average, the pathologic size of invasive cancer was significantly smaller than NME size, the pathologic size of ductal carcinoma in situ was larger than NME size, but this was not statistically significant. Malignant NME was also correlated with the presence of mammographic calcifications, unlike the findings by Nguyen and Lotfalla et al. Kim and colleagues (3) found a 59.1% malignancy rate for index cancer–accompanying NME in 93 cases and noted that malignant NME was significantly associated with clumped or clustered ring internal enhancement as well as with histologically present extensive intraductal component in the index tumor. Jirarayapong et al (4) recently studied all types of ipsilateral-to–index cancer NME (NME not only directly contiguous with the index cancer but also remote from it) and found an overall malignancy rate of 48.5%, with contiguous NME showing a significantly higher malignancy rate (86.7%) compared with noncontiguous NME (and compared with other studies outlined here). All calcified NME lesions contiguous with calcified index cancers were malignant. Overall, NME with segmental distribution (85.7%), internal clumped enhancement (66.7%), or nonpersistent kinetics (77.1%) was significantly more likely to be malignant.

It is clear that while there are some shared features regarding MRI-detected contiguous NME found in patients with new cancers, achieving negative margins while preserving normal breast tissue in these patients remains a challenge because of uncertainties about exactly where tumor ends. The authors rightly point out that close communication with the surgeon is paramount to achieving the best outcomes. Depending on breast size, the plan for oncoplastic reduction, and the philosophy of the surgeon, the approach to localization will necessarily differ. Most conservatively, MRI-guided biopsy of the distal margins of the MRI enhancement with marker placement, and subsequent localization based on pathology results, can be pursued. Even then, there may be a tough decision to be made about what to target during MRI-guided biopsy: furthest remote enhancement (possible false-negative biopsy results) versus remote but still very conspicuous and suspicious tissue (marker may not outline the actual margin, assuming the results are positive). This approach has the disadvantage of substantially extending the time until surgery. In certain conditions, such as if the patient has large breasts, there is no concern about the removal of a substantial volume of benign tissue (eg, patient to undergo oncoplastic reduction), or “guaranteed” negative margins are valued more than ultimate cosmesis, the extent of the NME can be marked under MRI guidance without prior biopsy, either using MRI-compatible markers (subsequent mammography-guided localization will then be needed), nonwire localization devices, or by wire placement under MRI guidance the day of surgery. Another approach would be to use available comparative landmarks or measurements (eg, distances from index cancer clip to the remote extents of the NME) to estimate where to place localization devices. This approach requires close consultation with the surgeon and patient so that the risks and benefits are well understood. As reported in the literature, calcifications, when present, can be a good surrogate landmark for NME. The least conservative approach would be to localize only the index lesion. This approach may be preferable for surgeons who have internalized the concept that NME is often not malignant; however, a prelocalization discussion between radiologist and surgeon (with documentation of the surgeon's preference) is appropriate in this setting.

While the scenario described initially appears complicated and daunting, and at this point, there is no clear consensus on how to manage mass-associated NME, consider how much better off the patient is for our having identified it. Before MRI, the additional unsuspected cancer in the form of contiguous NME was present but just not detected. Those patients would likely have had positive margins at lumpectomy, requiring re-excision or completion mastectomy. Now, we at least know it is there and can address it with intention rather than blindly. I look forward to continued work similar to the study by Nguyen and Lotfalla colleagues, which, with time, will guide us toward an answer regarding the types of NME that are important and actionable.