Data extraction and quality assessment
We extracted detailed information from the full texts of included studies (supplementary p7). Outcomes of interest was the diagnostic performance of FSE of SLN, including sensitivity, true positives (TP) and false negatives (FN). The reference standard was DPE based on the paraffin sections of SLN. Sensitivity was assumed as the primary indicator of the diagnostic performance and defined as the ratio of cases having both positive FSE and DPE within all DPE-positive patients. TP were defined as cases having both positive FSE and DPE, and FN as cases having negative FSE and positive DPE. Isolated tumor cells (ITC) were defined as tumor diameter <0.2 mm, micrometastasis (MIM) as between 0.2 and 2 mm, and macrometastasis (MAM) as >2 mm. We took these parameters directly from the original papers or calculated them with available information. We excluded cases who had no SLN detected or did not underwent FSE. Whenever a calculation was impossible, we contacted the authors for necessary information.
In previous publications the term FN usually referred to cases presenting positive non-SLN without a positive SLN. Of note, in this study we redefined these cases as true negative given that their SLNs were tumor-free. We assessed the qualities of the included studies using Quality Assessment of Studies of Diagnostic Accuracy-2 (QUADAS2) tool.