Dynamic assessment of risk in clear cell renal cell carcinoma: The Emory Dynamic Model.

Abstract

e15123 Background: Kidney cancer is the most deadly of the common genitourinary malignancies. No biomarkers exist to guide therapy beyond the initial course of treatment. The purpose of this study was to develop a dynamic prognostic model for kidney cancer. Methods: Candidate biomarkers were identified from published prognostic models for clear cell renal cell carcinoma (ccRCC). Data quality for biomarker studies was assessed by REMARK guidelines. The biomarkers were classified as either static or dynamic with dynamic being defined as measurable in the circulation. Studies of dynamic biomarkers were reviewed to determine their impact on externally validated prognostic models. The most robust of the dynamic biomarkers were selected for inclusion in the Emory Dynamic Model (EDM). Results: Ten of the 20 dynamic biomarkers that were identified are acute phase reactants while the remaining 10 are commonly associated with inflammation. Of these, only C-reactive protein (CRP) has been included in an externally validated prognostic model, the TNM-C model. Serum amyloid A1 (SAA1) and CRP belong to the serum pentraxin (SP) family. Each has been shown to be an independent, adverse prognostic factor for ccRCC with a high degree of concordance. In the EDM, the SPs are viewed as atypical or “phenotypic” tumor markers. Patients are considered to be at high risk (HR) with an aggressive inflammatory phenotype (AIP) if pretreatment SP levels are elevated. When pretreatment SP levels are not elevated, patients are considered to be at low risk (LR) with a non-inflammatory phenotype (NIP). Simple algorithms further refine risk after therapeutic intervention including the perioperative CRP (pCRP) algorithm for localized disease and the Normalized CRP Period (NCP) algorithm for metastatic disease. Patients with persistent SP elevations after treatment are at extremely high risk (EHR) for rapid disease progression and death. Conclusions: The EDM is the first truly dynamic prognostic model for localized and metastatic ccRCC. With prospective validation, the EDM could become a useful tool for risk stratification in clinical trial design. Limitations include the need to extend follow up, refine SP cut-off levels and assess the NCP in localized disease.