Transcriptomic signatures can aid in post-operative risk stratification of clear cell renal cell carcinoma.

Abstract

383 Background: The ability to risk stratify patient’s post-nephrectomy is critical to select candidates for surveillance and adjuvant therapy. Whether transcriptomic information adds value to current prognostic information is unclear. Methods: Data from confirmed clear cell renal cell carcinoma (ccRCC) used in the TCGA Pan-RCC analysis was downloaded from cBioportal and GDC bioportal. Clinicopathologic variables were used to calculate the integrated staging system, SSIGN (stage, size, grade, and necrosis). The 16 gene recurrence score (RS) signature was generated as described in Rini 2015 using the TCGA data with log TPM normalization. The RS was stratified into high and low risk groups by median. SSIGN was divided into low (0-1) intermediate (2-4) and high ( > = 5) risk groups. RS and SSIGN prognostic significance were evaluated using disease recurrence as an end point, censoring by death. Results: SSIGN and recurrence scores (RS) were calculated for 369 (71.2%) of 518 available TCGA subjects with non-metastatic clear cell renal carcinoma. On multivariable analysis, the RS (continuous variable) was independently associated with disease-free status (HR 1.18 [95%CI (1.02-1.37)], p = 0.031) for each 10 point increase in RS) after adjusting for SSIGN. Categorizing SSIGN into low, intermediate and high-risk groups showed 3.9%, 18.5% and 39.8% 3-year recurrence rates. Stratifying SSIGN risk groups by RS scores we found an RS to further risk stratify the SSIGN intermediate risk group (HR 1.60 [95%CI 1.21-2.12], p < 0.001]). The AUC for at 3 years for SSIGN in the intermediate risk group was 0.68, for RS was 0.74 and the combination was 0.78 (with boot-strapping for optimism adjustment). At 3 years, patients with SSIGN intermediate risk disease (n = 135) with low RS had 3-year recurrence rate of 7.1% compared to 31.0% for those with high RS scores. Conclusions: Transcriptomic recurrence scores can risk stratify intermediate SSIGN clinical risk patients. Patients with intermediate risk disease but a high RS score had poorer outcomes similar to clinically high risk patients. Transcriptomic signatures may add value to existing clinicopathologic variables in ccRCC and clinical implementation is warranted.