OBJECTIVE:To evaluate the additive role of breast MRI in axillary nodal staging following a benign axillary US and compare diagnostic performance of US-first versus MRI-first imaging sequences. METHODS:This retrospective study included 524 patients with stage I-III breast cancer who underwent preoperative breast MRI and axillary surgery. Imaging assessments were categorized as US-first (n = 367) or MRI-first (n = 157). Diagnostic performance metrics were calculated based on the first modality used to evaluate the axilla. Incremental detection by MRI after benign US was assessed. Associations between false-negative results and clinicopathologic factors were analyzed. RESULTS:US-first patients were younger (51.8 vs 55.2 years, P = 0.002), had larger tumors (37.8 mm vs 26.0 mm, P < 0.001), and higher Ki-67 indices (42.6% vs 33.0%, P < 0.001) compared to MRI-first. Sensitivity for nodal metastasis was higher for US-first (71.7%) than MRI-first (50%, P = 0.01), while MRI-first demonstrated greater specificity (85.6% vs 74.6%, P = 0.02) and negative predictive value (87.0% vs 69.2%, P < 0.001). Among 182 axillae assessed as benign on initial US, MRI identified 24 additional abnormal cases (13.2%), of which 15 were true positives, yielding an incremental detection rate of 8.2%. False-negative results were associated with younger age in both groups and with larger tumor size and multifocality in US-first cases. CONCLUSION:MRI provided clinically relevant incremental detection of axillary metastases after benign US and demonstrated higher negative predictive value when performed first. False-negative results were more common in patients with younger age, larger tumors, and multifocal disease, highlighting the need to consider these factors when interpreting axillary imaging.
Interval breast cancers are diagnosed between screening rounds. Because they portend a worse diagnosis than screen-detected cancers, decreasing the interval cancer rate is a key measure of screening program effectiveness. Artificial intelligence (AI) tools for breast cancer detection are rapidly emerging with potential to provide earlier detection or even prediction of interval cancers. This article reviews the importance of interval cancers in screening mammography, current evidence for AI performance related to interval cancers, and considerations for interpreting the findings and implications of the current literature.
Given the increasing frequency of reduction mammoplasty (RM) procedures, it is essential for radiologists to develop expertise in the accurate evaluation and interpretation of postoperative breast imaging findings. This pictorial essay reviews common RM surgical techniques with the resultant expected imaging findings on mammography, ultrasound, and MRI. Findings include symmetric parenchymal redistribution, architectural distortion and scar tissue, benign calcifications, and fat necrosis. Recognizing these patterns is essential to avoid misinterpreting benign postoperative changes as malignancy as well as to identify suspicious findings that require further evaluation. Diagnostic challenges include differentiating post-surgical scarring from malignancy, especially in the lower half of the breast. Cases of breast malignancy after RM are presented to highlight important differentiation between expected postoperative changes and breast cancer.
OBJECTIVE:This study evaluates concordance between abbreviated protocol breast MRIs (abMRI) and standard full-protocol MRI in the post-neoadjuvant chemotherapy (NAC) setting. METHODS:This multi-site single institution retrospective study included breast cancer patients with pre- and post-NAC breast MRI during 2016-2023. The abMRI sequences included localizers, pre-contrast fat sat T1, fat sat first post-contrast phase, reconstructed first post-contrast phase subtraction, and maximum-intensity projection images from the institution's standard protocol. Four breast radiologists interpreted abMRIs and full MRIs after a washout period of at least 1 month. Residual tumor size (RTS) was compared to pathology. Statistical analysis with multivariable logistic regression, Receiver Operating Curve analysis, Kappa analysis, and paired t-tests were performed. RESULTS:Ninety-two of 237 patients (38.8%; mean age, 52.6 years) had pathologic complete response (pCR). There was no significant difference between abMRI and full MRI for predicting pCR for any reader individually (P-value range 0.30-0.91) with overall area under curve (AUC) = 0.70. Intraobserver and interobserver agreement for abMRI and full MRI was moderate to good (κ = 0.52-0.78). There was concordance between abMRI and full MRI for predicting pCR when stratified by menopause status, background parenchymal enhancement, tumor stage, tumor biomarkers, and Ki-67. When compared to pathology, all readers demonstrated no significant difference in RTS with abMRI, and one reader underestimated RTS on full MRI by 7 mm (P = 0.005). CONCLUSION:Our study found no significant difference in diagnostic performance between abMRI and full MRI when compared to pathology for evaluating post-NAC patients for pCR and evaluation of residual disease in the breast. Abbreviated MRI and full MRI show similar effectiveness in predicting complete tumor response and assessing residual disease on post-NAC assessment.
Widespread mammography screening has resulted in a significant decline in breast cancer deaths, but it is resource intensive. The WISDOM (Women Informed to Screen Depending on Measures of Risk) trial attempts to assess screening based on individual risk as an alternative to age-based screening (annual screening starting at age 40 years). In their initial report, the authors concluded that risk-based screening is safe and acceptable. Their results do not support this conclusion. The first primary endpoint to test the noninferiority of risk-based screening used the metric of stage IIB and higher breast cancers. Many of these cancers would be apparent on physical exam, thus eliminating some advantages of early detection and ignoring shortcomings of the risk-based approach. In addition, the study set an extremely weak noninferiority margin, compromising their conclusion concerning noninferiority. Details of mammography and MRI use in WISDOM reflect poor compliance with screening protocols in both arms of the study. Substantial non-compliance as well as out-of-study imaging in both cohorts resulted in two nearly identical protocols, rather than a comparison of different screening strategies. The study's findings of substantial non-compliance to the risk-based protocol make it difficult to justify risk-based screening being acceptable to women. The trial's second primary endpoint, to test whether risk-based screening yielded fewer biopsies than annual screening, failed. The risk-based screening protocol is labor intensive and would be challenging to implement in the typical primary care practice. The vast majority of breast cancers occur in average-risk women and could be missed in a risk-based system, especially in women under age 50 and minority women.
