The Selection for Cytoreductive Nephrectomy (SCREEN) Score: Improving Surgical Risk Stratification by Integrating...

Careful patient selection is critical when considering cytoreductive nephrectomy (CN) for metastatic renal cell carcinoma (mRCC) but few studies have investigated the prognostic value of radiologic features that measure tumor burden.

To develop a prognostic model to improve CN selection with integration of common radiologic features with known prognostic factors associated with mortality in the first year following surgery.

Data were analyzed for consecutive patients with mRCC treated with upfront CN at five institutions from 2006 to 2017. Univariable and multivariable models were used to evaluate radiographic features and known risk factors for associations with overall survival. Relevant factors were used to create the SCREEN model and compared to the International mRCC Database Consortium (IMDC) model for predictive accuracy and clinical usefulness.

A total of 914 patients with mRCC were treated with upfront CN during the study period. Seven independently predictive variables were used in the SCREEN score: three or more metastatic sites, total metastatic tumor burden ≥5 cm, bone metastasis, systemic symptoms, low serum hemoglobin, low serum albumin, and neutrophil/lymphocyte ratio ≥4. Predictive accuracy measured as the area under the receiver operating characteristic curves was 0.76 for the SCREEN score and 0.55 for the IMDC model. Decision curve analysis showed that the SCREEN model was useful beyond the IMDC classifier for threshold first-year mortality probabilities between 15% and 70%.

The SCREEN score had higher predictive accuracy for first-year mortality compared to the IMDC scheme in a multi-institutional cohort and may be used to improve CN selection.

This study provides a model to improve selection of patients with metastatic kidney cancer who may benefit from surgical removal of the primary kidney tumor. We found that radiographic measurements of the tumor burden predicted the risk of death in the first year after surgery. The model can be used to improve decision-making by these patients and their physicians.