A comparative analysis of prediction models for complete gross resection in secondary cytoreductive surgery for ovarian cancer

Abstract

Objective: We sought to examine compliance and outcomes using Memorial Sloan Kettering "(MSK) criteria" to predict complete gross resection (CGR) and compare them with the validated Tian and AGO models.

Methods: Patients who underwent SCS for recurrent platinum-sensitive ovarian cancer from 5/2001-6/2014 were identified. The AGO and Tian models were applied to the study population; appropriate statistical tests were used to determine ability to predict CGR.

Results: 214 SCS cases were identified. Since the implementation of MSK criteria, the CGR rate has been 86%. The AGO model had a 49% accuracy rate in predicting CGR, and predicted gross residual disease (RD) in 51%; however, CGR was achieved in 86%. The Tian model had an 88% accuracy rate. Of the 4% scored as Tian high risk for gross RD, 33% achieved a CGR. Comparing models, McNemar's p-value was 0.366 between the Tian and MSK models and <0.001 between AGO and MSK criteria. Median PFS was 21.3 (95%CI, 18.2-24.5), 22.5 (95%CI, 19.4-25.3), and 14.1months (95%CI, 9.7-22.1) for the entire cohort, for those achieving CGR, and for those left with RD, respectively (p=0.013). OS was 82.2 (95%CI, 60.2-123.3), 95.6 (95%CI, 63.6-NE), and 57.5months (95%CI, 27.5-113.9), respectively (p=0.014).

Conclusion: CGR during SCS is associated with extended PFS and OS. We report a high rate of CGR using MSK criteria. There was good concordance between the Tian and MSK models; however, the latter has fewer variables and is more user-friendly. Tian criteria may be applied to intermediate MSK cases for further stratification.

Keywords: AGO DESKTOP criteria; Complete gross resection; Ovarian cancer; Prediction models; Secondary cytoreductive surgery; Tian model.