Conventional procedures for the intraoperative assessment of breast cancer sentinel lymph nodes (SLNs) are frozen section (FS) and touch imprint cytology (TIC). The one-step nucleic acid amplification (OSNA) assay is a novel molecular technique. The aim of this study was to evaluate the optimal approach by comparing OSNA assay, FS, and TIC. Five hundred and fifty-two consecutive patients were enroled from five study centers in China. The SLNs were cut into alternating 2 mm blocks. The odd blocks were tested by the OSNA assay intraoperatively, and the even ones were assessed by postoperative histology (four 4- to 6-μm-thick sections were taken every 200 μm per block). In addition, intraoperative histological assessments were carried out on the even blocks of 211 patients by FS and all blocks of 552 patients by TIC. Overall performance of the assay compared to postoperative histology was: accuracy 91.4%; sensitivity 83.7%; and specificity 92.9%. The sensitivity of the assay was higher than FS (211 patients, 77.6% vs 69.7%; not significant, P = 0.286) and was significantly higher than TIC (552 patients, 83.6% vs 76.2%; P = 0.044). When assessing nodes with micrometastases, the sensitivity of the assay was higher than FS (17 nodes, 47.1% vs 23.5%; not significant, P = 0.289) and was significantly higher than TIC (48 nodes, 62.5% vs 35.4%; P = 0.007). The study indicated that the OSNA assay is an accurate and rapid intraoperative assay for assessing breast SLNs and it can replace FS and TIC for application in general medical practice. The trial was registered as: OSNA assay China Registration Study. Clinical trial registration number: China Breast Cancer Clinical Study Group 001c.