Predicting sentinel lymph node metastasis in breast cancer with lymphoscintigraphy.

OBJECTIVE Lymphoscintigraphy is an effective method for detecting sentinel lymph nodes (SLNs). However, the rate and degree of SLN detection is not uniform. We quantified SLNs detected with lymphoscintigraphy, and investigated correlations with factors that may influence detection. We then attempted to predict SLN metastasis from lymph node counts, comparing the predictions to subsequent biopsy results. METHODS We assessed lymph node counts in 100 breast cancer patients in whom a single SLN was detected with a fixed lymphoscintigraphy procedure. We examined correlations between the counts and factors known to influence lymphoscintigraphic SLN detection (age, body mass index, tumor size, and presence or absence of metastasis), and determined reference values (lymph node counts of 10.0, 19.4 and 53.0) which were used to predict SLN metastasis in 100 subsequent patients. The predictions were then compared with the SLN biopsy findings. RESULTS SLN counts correlated strongly with the presence or absence of metastasis, with metastasis-positive lymph nodes showing significantly lower counts than negative nodes (p < 0.001). Prediction of SLN metastasis achieved a 100% positive predictive value at a reference value of 10.0, and a 100% negative predictive value at a reference value of 53.0. At a reference value of 19.4, the sensitivity, specificity, and diagnostic accuracy were 77.8, 73.2, and 74.0%, respectively. CONCLUSIONS The SLN counts detected with lymphoscintigraphy were significantly lower in metastasis-positive lymph nodes than in metastasis-negative lymph nodes. This suggests that prediction of SLN metastasis in breast cancer is possible using lymphoscintigraphy.