Background and Aim: There are conflicting results of studies on accuracy of positron emission tomography (PET)/computed tomography (CT) for axillary staging. The aim of this study is to determine the factors affecting the efficacy of 18F-fluorodeoxyglucose (F-18-FDG) PET/CT in detecting axillary metastases in invasive breast cancer. Materials and Methods: Data of 232 patients with invasive breast cancer who underwent F-18-FDG PET/CT examination before surgery between January 2013 and September 2017 were reviewed retrospectively. Histopathological examination of axillary lymph nodes (ALNs) was used as a reference to assess the efficacy of F-18-FDG PET/CT in detecting axillary metastases. Results: While 134 (57.8%) patients had axillary metastases as detected in F-18-FDG PET/CT scans, histopathologically axillary metastases were detected in 164 (70.7%) patients. The sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of F-18-FDG PET/CT in detection of axillary metastasis were 72.56%, 77.94%, 88.8%, 54%, and 74.1%, respectively. The false-negative and false-positive rates were 27.4% and 22%, respectively. In univariate analysis, patients' age, estrogen receptor positivity, higher ALN SUVmax, greater tumor size, and lymph node size determined by F-18-FDG PET/CT were all significantly associated with accuracy of F-18-FDG PET/CT for axillary metastasis. In multivariate analysis, tumor size determined by F-18-FDG PET/CT and ALN SUVmax were independent variables associated with axillary metastasis. The accuracy of F-18-FDG PET/CT for axillary metastasis was higher in patients with a larger tumor (>= 19.5 mm) and a higher ALN SUVmax (>= 3.2). Conclusion: F-18-FDG PET/CT should not be routinely used for axillary staging, especially in patients with small tumors. It cannot eliminiate the necessity of sentinel lymph node biopsy. When it is used, both visual information and optimal cut-off value of axillary node SUVmax should be taken into consideration.