OBJECTIVE: To collect data from patients undergoing pelvic exenteration in recent clinical practice. The primary aim was 5-year disease-free survival. Second-ary aims were 5-year overall survival, patterns of recurrence, identification of subgroups at higher risk of recurrence and death, survival associated with lymph node metastasis, and development of a prognostic score. METHODS: This was a retrospective, multicenter, international study conducted in tertiary national gynecologic oncology referral centers. Inclusion criteria included cervical, vaginal, vulvar, or endometrial cancer; anterior or total pelvic exenteration performed between January 2005 and March 2023; curative or palliative intent; and with or without laterally extended endopelvic or pelvic resection. Patients were excluded if they underwent posterior pelvic exenteration only or if preoperative computed tomography (CT), positron emission tomography (PET)–CT, or PET was not performed. A prognostic score was developed that was based on multivariable analysis. RESULTS: Eight hundred sixty-two patients were included. Surgical margins were tumor free in 676 (78.4%). In patients treated with curative intent, total pelvic exenteration, positive surgical margins, and presence of lymphovascular space invasion were independently associated with worse disease-free survival. Performance of lymphadenectomy was associated with better disease-free survival. Total pelvic exenteration, positive surgical margins, and presence of lymphovascular space invasion were factors independently associated with decreased overall survival. Performing pelvic exenteration at time of persistent (instead of recurrent) disease negatively affected overall survival. Prognostic score identified four risk groups with a 5-year disease-free survival of 43.7%, 24.9%, 22.2%, and 8.0% (P,.001). The 5-year overall survival in the four risk groups was 54.3%, 40.4%, 24.0%, and 4.3% (P,.001). The most frequent sites of recurrence were distant in 166 patients (32.1%). The 5-year disease-free survival and cancer-specific survival in patients with para-aortic lymph node metastasis were significantly worse compared with those in patients with pelvic-only metastatic nodes or with negative nodes (P5.002 and P,.001, respectively). CONCLUSION: Independent factors associated with worse disease-free survival and overall survival and subgroups of patients at higher risk of recurrence and death were identified. A multivariable prognostic score was developed that can be used for patient counseling and surveillance strategies and for future prospective studies.
