Evidence-based medicine applied to sentinel lymph node biopsy in patients with breast cancer.

Sentinel lymph node biopsy (SLNB) has not been examined using the principles of evidence-based medicine (EBM). Specifically, likelihood ratios have not been used to assess the validity of SLNB. The Surveillance, Epidemiology, and End Results (SEER) public database of the National Cancer Institute was used to establish the baseline or pretest probability of finding a positive lymph axillary node for each stage of breast cancer. Rates of false negative results of SLNB for all breast cancer stages were determined from the surgical literature. Positive and negative likelihood ratios (LR) were calculated. For each stage of breast cancer, the Bayesian nomogram was used to find the post-test probability of missing a metastatic axillary node when the SLN was negative. The SEER database of 213,292 female patients with breast cancer yielded the following rates of positivity of axillary lymph nodes for each breast tumor size: T1a, 7.8 per cent; T1b, 13.3 per cent; T1c, 28.5 per cent; T2, 50.2 per cent; T3, 70.1 per cent. The combined data from 13 published studies of SLNB (6444 successful SLNBs) demonstrated a false negative rate of 8.5 per cent. The LR of a negative test is 0.086. According to the nomogram, the chances of missing a positive node for stage of cancer are as follows: T1a, 0.7 per cent; T1b, 1.5 per cent; T1c, 3.0 per cent; T2, 7 per cent; T3, 18 per cent. The risk of missing a positive axillary node can accurately be estimated for each stage of breast cancer using the LR, which is much more useful than the simple false negative rate. Surgeons should use this information when deciding whether to perform SLNB and in their informed consent discussions.