CT prediction of surgical outcome in patients with advanced epithelial ovarian carcinoma undergoing neoadjuvant chemotherapy

Objective: A scoring system has been proposed to predict gross residual disease at primary debulking surgery (PDS) for advanced epithelial ovarian cancer. This scoring system has not been assessed in patients undergoing neoadjuvant chemotherapy (NACT). The aim of this study is to assess the reproducibility and prognostic significance of the scoring system when applied to women undergoing NACT followed by interval debulking surgery (IDS). Methods: A retrospective cohort study was conducted of patients with advanced ovarian cancer who underwent NACT and IDS between 2005 and 2014. Change in tumor burden using computed tomography (CT) at diagnosis (T0) and after initiation of NACT but before IDS (T1) was independently assessed by two radiologists blinded to outcomes using two read criteria: a scoring system utilizing clinical and radiologic criteria and RECIST 1.1. Relationship between CT assessments to surgical outcome, progression free survival (PFS) and overall survival (OS) were evaluated. Reader agreement was measured using Fleiss's kappa (ĸ). Results: 76 patients were analyzed. Optimal surgical outcome was achieved in 69 (91%) of patients. Median progression free survival was 13.2 months and overall survival was 32.6 months, respectively. Predictive score change from T0 to T1 of >1 (denoting an improvement in disease burden) was associated with optimal cytoreduction (p = 0.02 and 0.01 for readers 1 and 2, respectively). Neither predictive score nor RECIST 1.1 assessment was predictive of OS or PFS. Reader agreement was substantial for predictive score (κ = 0.77) and moderate for RECIST (κ = 0.51) assessments. Conclusions: A change in score before and after neoadjuvant chemotherapy minimizes reader variability and predicts surgical outcome.