Molecular Subtypes and Gene Expression Signatures as Prognostic Features in Fully Resected Clear Cell Renal Cell Carcinoma: A Tailored Approach to Adjuvant Trials

Abstract

Background

Trials with adjuvant vascular endothelial growth factor receptor tyrosine kinase inhibitors (VEGFR-TKIs) failed to demonstrate meaningful benefit in clinically high-risk, fully resected clear cell renal cell carcinoma (ccRCC). We evaluated whether the ccrcc1-4 molecular subtypes and gene expression signatures (GES) are associated with outcomes in this setting.

Materials and Methods

We determined molecular subtypes as well as angiogenesis- and immune-related GES through RNA sequencing of 75 fresh frozen (FF) and 62 formalin-fixed, paraffin-embedded (FFPE) tumor samples. We studied disease-free (DFS) and overall survival (OS) and determined correlations among GES and Leibovich score.

Results

Angiogenesis-related GES and molecular subtypes were associated with longer DFS and OS across both cohorts, whereas immune-related GES were not. In the FF cohort, molecular subtypes (ccrcc2 & 3 vs. ccrcc1 & 4) were associated with DFS and OS, on bivariable analysis with Leibovich score (HR 0.62, 95%CI 0.39-0.98, P = .04 and HR 0.35, 95%CI 0.19-0.64, P < .001). In the FFPE cohort, molecular subtypes (ccrcc2 & 3 vs. ccrcc1&4) were also associated with DFS (HR 0.53, 95%CI 0.31-0.93, P = .03), but not OS (HR 0.59, 95%CI 0.31-1.13, P = .11) on bivariable analysis with Leibovich score. Leibovich score was significantly inversely correlated with all angiogenesis-related GES (all P < .01), but not correlated with immune-related GES.

Conclusions

Molecular subtypes and angiogenesis-related GES are prognostic for DFS and OS in fully resected, localized ccRCC. Favorable ccrcc2 & 3 molecular subtypes with high angiogenesis-related GES, which respond best to VEGFR-TKIs, are at lower risk of relapse but were probably underrepresented in the adjuvant VEGFR-TKI trials since they inversely correlate with Leibovich score. Conversely, immune-related GES are not correlated with Leibovich score and clinically high-risk tumors can display both high and low immune-related GES. Therefore, molecular characterization could guide patient selection for adjuvant treatment.

Introduction

Renal cell carcinoma (RCC) represents a significant and increasing cancer burden, accounting for approximately 3% of malignancies worldwide.^1,2 Up to one third of cases treated with surgery will eventually develop metastases. Therefore, in recent years, several trials with adjuvant vascular endothelial growth factor receptor tyrosine kinase inhibitors (VEGFR-TKI) were finalized and trials with adjuvant immune checkpoint inhibitors (ICI) are currently ongoing. Surprisingly, these clinical trials with VEGFR-TKIs have failed to demonstrate a significant benefit in terms of disease-free survival (DFS) and overall survival (OS) (Table 1).3, 4, 5, 6, 7 Only the S-TRAC trial, which randomized ccRCC patients to one year adjuvant sunitinib or placebo, showed a marginal DFS advantage for the sunitinib arm compared to placebo (HR 0.76, 95% CI 0.59-0.98, P = .03), which did not translate into an OS benefit (HR 1.01, 95% CI 0.72-1.44, P = .94), despite considerable treatment-related toxicity.³ In this regard, the use of adjuvant therapy might be associated with a DFS benefit but not an OS benefit, delaying but not preventing an eventual disease recurrence. Thus, although DFS provides an early, clinically relevant endpoint and is not affected by subsequent therapy lines, OS remains a key clinical endpoint for adjuvant trials.

Considering that VEGFR-TKIs are usually very efficient in the treatment of RCC, this raises the question whether patient selection for these adjuvant trials was adequate, and how patient selection could potentially be improved. Patients were included in these adjuvant trials based on clinical risk stratification tools such as the Leibovich score and the University of California Los Angeles Integrated Staging System (UISS). These scores predict patients’ risk of recurrence and are currently recommended as prognostic models by the European Association of Urology guidelines.⁸ The Leibovich score is based on pT-stage, tumor dimensions, lymph node status, Fuhrman nuclear grade and the presence of tumor necrosis, while the UISS categories take into account pT-stage, Fuhrman nuclear grade and eastern cooperative oncology group performance status.9, 10, 11

Recently, several predictive biomarkers for VEGFR-TKI and ICI treatment have emerged from trials in the metastatic setting. We have previously described four molecular subtypes, named ccrcc1 through 4, with distinct differences in mRNA expression, methylation status, mutation profile, copy number aberrations and immune infiltrate.¹² In the metastatic setting, these subtypes are correlated with OS, outcomes on first-line VEGFR-TKI and metastasectomy with curative intent.12, 13, 14 Briefly, ccrcc2 & 3 tumors have a favorable prognosis, with a significant upregulation of angiogenesis-related genes in the ccrcc2 tumors and ccrcc3 tumors exhibiting an expression profile similar to that of normal kidney tissue. Ccrcc2 tumors are highly sensitive to VEGFR-TKIs. Conversely, ccrcc1 & 4 tumors have an intermediate and poor prognosis, respectively. While they both show relative upregulation of Myc-targets, the ccrcc4 subtype has a distinct immune inflamed phenotype with high expression of immunological signatures and is enriched for sarcomatoid features.¹⁵ Similar molecular subtypes have been described by other groups in different settings and are overlapping.16, 17, 18, 19 The IMmotion 150 trial, comparing atezolizumab/bevacizumab vs. atezolizumab vs. sunitinib described a 6-gene angiogenesis-related gene expression signature (GES) (IMmotion Angio) associated with improved outcomes to sunitinib, a 5-gene effector T-cell (IMmotion Teff) GES associated with improved outcomes to atezolizumab/bevacizumab and a 6-gene myeloid-inflammation GES (IMmotion Myeloid) associated with poorer outcomes to atezolizumab or atezolizumab/bevacizumab.²⁰ Similarly, the JAVELIN Renal 101 trial, comparing avelumab/axitinib vs. sunitinib equally described a 26-gene angiogenesis-related GES (JAVELIN Angio) associated with improved outcomes to sunitinib and a 26-gene immune-related GES (JAVELIN Immuno) associated with improved outcomes to avelumab/axitinib.²¹

We hypothesized that these molecular subtypes as well as GES could be associated with post-surgery outcomes in the localized setting and potentially aid patient selection for adjuvant trials. Additionally, we hypothesized that adjuvant trials with VEGFR-TKIs have hitherto failed to show significant benefit due to equivocal patient selection.