Accuracy of unenhanced MR imaging in the detection of axillary lymph node metastasis: study of reproducibility and reliability.

Purpose: To investigate the accuracy, reproducibility, and reliability of unenhanced magnetic resonance (MR) imaging techniques for detecting metastatic axillary lymph nodes in patients with newly diagnosed breast carcinoma.

Materials And Methods: Institutional review board approval and informed consent were obtained. Seventy-four consecutive women with invasive breast carcinoma were recruited to undergo preoperative breast MR imaging. Thirteen patients were excluded, two because they were undergoing preoperative chemotherapy and 11 because of the presence of movement or susceptibility artifacts on images. Thus, 61 patients (mean age, 53 years; range, 33-78 years) were included in this study. Axial T1-weighted MR images without fat saturation and diffusion-weighted (DW) MR images were analyzed by two experienced radiologists, who were blinded to the histopathologic findings. Visual and quantitative analyses of unenhanced MR images were performed. Sensitivity, specificity, and accuracy were calculated. To assess the intraobserver agreement, a second reading was performed. Statistical analysis was conducted on a patient-by-affected side basis.

Results: The sensitivity, specificity, and accuracy were 88%, 82%, and 85%, respectively, for axial T1-weighted MR imaging and 84%, 77%, and 80% for DW imaging. Apparent diffusion coefficients (ADCs) were significantly lower in the malignant group (P<.05 for all four readings), with the average of the four readings ranging from 0.333×10(-3) mm2/sec to 2.843×10(-3) mm2/sec. The mean Lin coefficient comparing the mean ADC reading for each observer was 0.959 (95% confidence interval: 0.935, 0.975), suggesting very high interobserver agreement between the two observers in terms of reproducibility of ADCs. The Bland-Altman plot showed good inter- and intraobserver agreement.

Conclusion: Unenhanced MR imaging techniques showed high accuracy in the preoperative evaluation of axillary status in patients with invasive breast cancer. Results indicate reliable and reproducible assessment with DW imaging, but it is unlikely to be useful in clinical practice.