Purpose: A previous retrospective study demonstrated that positron emission tomography with [18F]fluorodeoxyglucose (FDG-PET) was more sensitive than computed tomography for lymph node staging in patients with cervical cancer; the findings on FDG-PET were strongly associated with progression-free survival. Therefore, a prospective cohort study was initiated to evaluate FDG-PET lymph node staging in a larger patient population. Patients and Methods: The study was conducted between July 2000 and March 2009. All 560 patients with cervical cancer underwent pretreatment FDG-PET lymph node staging. Treatment included surgery alone, surgery and postoperative radiation therapy, and definitive radiation or combination radiation and chemotherapy. PET findings were correlated with the risk of disease progression and with survival. Results: Overall, 47% of patients had lymph node involvement by FDG-PET at diagnosis. The frequency of lymph node metastasis increased with clinical stage and was similar to that in historical surgical series. Within a stage, patients with PET-positive lymph nodes had significantly worse disease-specific survival than those with PET-negative lymph nodes (P < .001). Disease-specific survival was stratified into distinct groups based on the most distant level of PET-detected nodal disease (none, pelvic, para-aortic, or supraclavicular; P < .001). The hazard ratios for disease recurrence increased incrementally based on the most distant level of nodal disease: pelvic 2.40 (95% CI, 1.63 to 3.52), para-aortic 5.88 (95% CI, 3.80 to 9.09), and supraclavicular 30.27 (95% CI 16.56 to 55.34). Conclusion: Nodal involvement detected by FDG-PET in cervical cancer relates to clinical stage, is comparable to historical data, and stratifies patient recurrence and survival outcomes.