Objective The aim of this study was to assess the safety and feasibility of macroscopically complete (R0) secondary debulking surgery (SDS) in a single-center cohort of patients with recurrent ovarian cancer. The performances of existing prediction models (DESKTOP score, Tian model) for R0 SDS were evaluated in this cohort.

Methods Patient, disease, and treatment characteristics of 102 patients undergoing SDS for recurrent ovarian cancer at the University Hospitals Leuven between 1997 and 2014 were collected.

Results R0 SDS was achieved in 73% of patients and associated with improved progression-free survival (P = 0.0002) and overall survival (P = 0.0003) compared with non-R0 resection. Variables associated with R0 SDS were site of relapse (P = 0.046) and absence of ascites (P = 0.045). The DESKTOP score and Tian model showed positive predictive values for R0 SDS of 80% and 73%, respectively. However, a false-negative rate for R0 resection of 61% and 70% was observed in our study. Progression-free survival and overall survival did not significantly differ between DESKTOP score–positive and –negative patients with R0 SDS.

Conclusions We confirmed a high positive predictive value in the selection of candidates for R0 SDS with the DESKTOP score and the Tian model. However, because 61% and 70% of the patients with a negative score were debulked to R0, we suggest that other selection criteria based on anatomic and metabolic imaging such as whole-body diffusion-weighted magnetic resonance imaging should be evaluated when selecting patients for SDS.