US surveillance of regional lymph node recurrence after breast cancer surgery.

PURPOSE: To determine the diagnostic indexes of lymph node ultrasonography (US) of the axillary and supraclavicular regions for detecting lymph node recurrence (LNR) after breast cancer surgery and assess the effect of lymph node evaluation on prognosis during bilateral breast US. MATERIALS AND METHODS: Institutional review board approved this retrospective study and waived informed consent. Between January 2003 and December 2004, 3982 lymph node US examinations, including bilateral axillary and supraclavicular areas, were performed in 1817 women (mean age, 49.9 years; range, 22-86 years) after breast cancer surgery, nine of whom had palpable lesions. Final diagnosis was based on cytopathologic results, clinical follow-up, and imaging studies for at least 12 months after breast US. Diagnostic indexes of US for detecting LNR were assessed. The frequency of distant metastases between patients with ipsilateral LNR and those without was compared. Three-year mortality rates of patients with ipsilateral LNR only and those with distant metastases were evaluated. RESULTS: Of 1817 patients, 54 had suspicious LNR at US (28 at first, 20 at second, five at third, and one at fourth US examination). Thirty-nine of 1817 patients (2.1%), including nine with palpable lesions, had LNR, 11 of whom had ipsilateral LNR only. At first lymph node US, LNR was detected in 17 patients; at second, in 10; at third, in two; and at fourth, in one. Nine had false-negative results. The respective sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of lymph node US for detecting LNR per woman was 76.9% (30 of 39), 98.7% (1754 of 1778), 98.2% (1784 of 1817), 55.6% (30 of 54), and 99.5% (1754 of 1763); those of first lymph node US were 85.0% (17 of 20), 99.4% (1786 of 1797), 99.2% (1803 of 1817), 60.7% (17 of 28), and 99.8% (1786 of 1789); and those of total US examinations were 78.0% (32 of 41), 99.4% (3917 of 3941), 99.2% (3949 of 3982), 57.1% (32 of 56), and 99.8% (3917 of 3926). Distant metastases were found more frequently in patients with ipsilateral LNR (62%) than in those without (2.3%) (P < .0001). The 3-year mortality rate of patients with ipsilateral LNR only was significantly lower than that in patients with distant metastases (P = .03). CONCLUSION: Ipsilateral LNR is a predictor of distant metastasis, and lymph node evaluation during breast US is useful for early detection of LNR in asymptomatic patients.ope