The purpose of this study was to compare the efficacy of short-t inversion-recovery (STIR) MRI and 18 F-FDG PET/CT for the detection of metastasis in mediastinal and hilar lymph nodes in patients with lung cancer. Methods: Ninety-three patients with known or suspected lung cancer with mediastinal and hilar lymph node swelling underwent STIR MRI and 18 F-FDG PET/CT examinations. STIR MRI scans were obtained with a 2% copper sulfate phantom placed along the back of each patient, with the lymph node-to-phantom ratio calculated for quantitative analysis. For qualitative analysis, the results of all STIR MRI scans were evaluated using a 5-point visual scoring system. To evaluate the diagnostic capabilities of STIR MRI and 18 F-FDG PET/CT, we used receiver-operating-characteristic curve analysis to determine the optimal thresholds for the lymph node-to-phantom ratio, visual score, and maximal standardized uptake value. Further, the capability of each to determine N-stage was compared in each patient using the McNemar test. Results: A total of 137 lymph nodes (82 malignant lesions, 55 benign lesions) were analyzed. When optimal threshold values were adopted, the quantitative and qualitative sensitivity, specificity, and accuracy of STIR MRI were not significantly different from those of 18 F-FDG PET/CT. However, 18 F-FDG PET/CT in combination with qualitative STIR MRI analysis had a significantly higher capability to detect nodal involvement on an individual-patient basis (96.9% specificity, 90.3% accuracy) than did 18 F-FDG PET/CT alone (65.6% specificity, 81.7% accuracy). Conclusion: We found that the diagnostic capability of STIR MRI was not significantly different from that of 18 F-FDG PET/CT. However, when those methods were combined, the diagnostic capability for N-staging was significantly improved.