Quantifying Small Lymph Node Metastases in Breast Cancer.
Chad Luedtke, Sujata Patil, Lee Tan, Edi Brogi, Dilip Giri. Memorial Sloan-Kettering Cancer Center, New York, NJ; Memorial Sloan-Kettering Cancer Center, New York, NY
Background: The 2010 AJCC staging of breast cancer axillary nodal metastases classifies isolated tumor cells (ITCs) as clusters of cells not greater than 0.2 mm or nonconfluent clusters not exceeding 200 cells in a single histologic lymph node cross section. Because this distinction can impact patient management and prognosis, we evaluated the diagnostic reproducibility of counting cells and measuring tumor clusters using hematoxylin and eosin (HE) and cytokeratin (CK) stains on glass and digitally scanned slides.
Design: Lymph nodes were selected from a cohort of 83 patients initially node negative that converted to node positive following deeper levels with HE and CK stains. 104 lymph node cross sections taken from these 83 patients were assessed twice by reviewer 1 and once by reviewer 2. Each reviewer evaluated a glass and corresponding digital slide, both stained with HE and CK. Spearman correlations were used to compare these different methods of pathological evaluation in counting cells and measuring tumor cluster size. For each method, the intra and inter observer variability was also calculated.
Results: The comparison of cell counts and cluster size measurements for the various combinations of glass and digital slides with HE and CK stains are listed in Table 1. The intra and inter observer variability are summarized in Table 2.
|Methods Compared||Cell Counts*||Largest Cluster Size*|
|Glass HE - Digital HE||0.87||0.85|
|Glass CK - Digital CK||0.91||0.93|
|Glass HE - Glass CK||0.59||0.76|
|Digital HE - Digital CK||0.52||0.75|
|Intra Observer Variability*||Inter Observer Variability*|
|Assessment Method||Cell Counts||Size^||Cell Counts||Size^|