Probability of axillary lymph node metastasis when sentinel lymph node biopsy is negative in women with clinically node negative breast cancer: a Bayesian approach.

Background: Although sentinel lymph node biopsy(SLNB)is highly accurate in predicting axillary nodal status in patients with breast cancer, it has been shown that the procedure is associated with a few false negative results. The risk of leaving metastatic nodes behind in the axillary basin when SLNB is negative should be estimated for an individual patient if SLNB is performed to avoid conventional axillary lymph node dissection(ALND).

Methods: A retrospective analysis of 512 women with T1-3N0M0 breast cancer was conducted to derive a prevalence of nodal metastasis by T category as a pre-test(i.e., before SLNB)probability and to examine potential confounders on the relationship between T category and axillary nodal involvement. Probability of nodal metastasis when SLNB was negative was estimated by means of Bayes' theorem which incorporated the pre-test probability and sensitivity and specificity of SLNB.

Results: Axillary nodal metastasis was observed in 6.1% of T1a-b, 25.1% of T1c, 28.7% of T2, 35.0% of T3 tumors. Point estimates for the probability of nodal involvement when SLNB was negative ranged from 0.3-1.3% for T1a-b, 1.6-6.3% for T1c, 2.0-7.5% for T2, and 2.6-9.7% for T3 tumors with representative sensitivities of 80%, 85%, 90% and 95%, respectively. The risk may be higher when the tumor involves the upper outer quadrant of the breast, while it may be lower for an underweight woman.

Conclusions: The probability of axillary lymph node metastasis when SLNB is negative can be estimated using a Bayesian approach. Presenting the probability to the patient may guide the decision of surgery without conventional ALND.