Preoperative MRI Evaluation of Axillary Lymph Nodes in Invasive Ductal Carcinoma: Comparison of Luminal A Versus...

The purpose of this study is to assess whether luminal A versus luminal B molecular subtypes of breast cancer affect the diagnostic utility of preoperative MRI evaluation of the axilla.

This study identified 125 patients who underwent preoperative breast MRI evaluation of tumors classifiable as luminal A or luminal B molecular subtypes between January 2012 and August 2014. The subtypes were classified on the basis of immunohistochemical staining surrogates combining receptor status and the Ki-67 proliferation index. Statistical analysis was performed using chi-square analysis and one-way ANOVA. When a statistically significant difference was found, follow-up analysis involving pairwise comparison using the Bonferroni correction was performed. Multivariate logistic regression analysis was also used to determine whether the molecular subtype was independently predictive of lymph node involvement.

A total of 80 patients had tumors classifiable as the luminal A molecular subtype, whereas 45 patients had tumors classifiable as the luminal B subtype. Pathologically proven axillary lymph node (ALN) disease occurred statistically significantly more frequently in luminal B tumors (18/45 [40.0%]) than in luminal A tumors (11/80 [13.8%]; p < 0.01). In addition, pathologically proven ALN disease was 4.3 times more likely to occur in luminal B tumors after controlling for patient age, tumor size, and tumor grade (p < 0.01). We found no difference in the negative predictive value of the MRI assessment of the axilla when luminal A tumors were compared with luminal B tumors. The positive predictive value of MRI evaluation of ALNs is statistically significantly higher for luminal B tumors than for luminal A tumors (76.2% vs 28.0%, respectively; p = 0.004).

With the use of molecular subtype classification including the Ki-67 proliferation index, suspicious nodes detected in luminal B tumors by MRI are likely to have positive findings and warrant preoperative tissue sampling. Conversely, in view of the high false-positive rate of suspicious nodes detected in luminal A tumors by MRI, proceeding straight to sentinel lymph node biopsy for all cases, except for those with highly suspicious findings, is suggested. These results should, however, be validated in a larger prospective study.