Subdividing BI-RADS category 4 breast lesions observed on magnetic resonance imaging: Is it feasible? Classificar as lesões mamárias da categoria BI-RADS 4 pela ressonância magnética em subdivisões: é viável?

In clinical practice, there are various purposes for which magnetic resonance imaging (MRI) of the breasts is indicated, from the screening of high-risk patients to the staging and treatment planning for patients with breast cancer. This method has better sensitivity than conventional imaging (mammography and ultrasound) for the diagnosis of malignant breast lesions and has greater accuracy in evaluating the size and morphological features of tumors, as well as in detecting multifocal and multicentric lesions. However, despite the high sensitivity of MRI, many studies have reported that its specificity is low and that it produces a large number of false positives, which can lead to unnecessary biopsies and surgical procedures. The Breast Imaging Reporting and Data System (BI-RADS), developed by the American College of Radiology and continually updated since 1992, is a guide with recommendations for the standardization of breast imaging (mammography, ultrasound, and MRI) reports and for the auditing of centers employing such methods (1–3) . Its objective is to standardize the nomenclature used in the reports, which should have a diagnostic conclusion and should propose management, according to the probability of malignancy. However, the cases classified as suspicious (BI-RADS category 4) show wide variation in the risk of malignancy (2–95%), which led to the subdivision of this category, as follows: 4A (low suspicion, risk of 2–10%); 4B (intermediate suspicion, risk of 11–50%); and 4C (high suspicion, risk of 51–95%). In the most recent editions of the BI-RADS, this subdivision was incorporated into the lexicon of mammography and ultrasound, although it has yet to be incorporated into that of MRI, because there is a lack of published studies to support such assessment (4) . Published in this issue of Radiologia Brasileira, the article “Predictive performance of BI-RADS magnetic resonance imaging descriptors in the context of suspicious (category 4) findings” is one of the first in the literature to assess the likelihood of malignancy related to MRI findings in lesions classified as BI-RADS category 4 (5) . In that study, Almeida et al. (5) present consistent methodology and statistical analysis, emphasizing the credibility of their findings. This type of study is essential to defining the criteria to be used for the subdivision of suspicious findings into the categories