Review of Prognostic Expression Markers for Clear Cell Renal Cell Carcinoma

Abstract

Context: The number of prognostic markers for clear cell renal cell carcinoma (ccRCC) has been increasing regularly over the last 15 years, without being integrated and compared. Objective: Our goal was to perform a review of prognostic markers for ccRCC to lay the ground for their use in the clinics. Evidence Acquisition: PubMed database was searched to identify RNA and protein markers whose expression level was reported as associated with survival of ccRCC patients. Relevant studies were selected through cross-reading by two readers. Evidence Synthesis: We selected 249 studies reporting an association with prognostic of either single markers or multiple-marker models. Altogether, these studies were based on a total of 341 distinct markers and 13 multiple-marker models. Twenty percent of these markers were involved in four biological pathways altered in ccRCC: cell cycle, angiogenesis, hypoxia, and immune response. The main genes (VHL, PBRM1, BAP1, and SETD2) involved in ccRCC carcinogenesis are not the most relevant for assessing survival. Conclusion: Among single markers, the most validated markers were KI67, BIRC5, TP53, CXCR4, and CA9. Of the multiple-marker models, the most famous model, ClearCode34, has been highly validated on several independent datasets, but its clinical utility has not yet been investigated. Patient Summary: Over the years, the prognosis studies have evolved from single markers to multiple-marker models. Our review highlights the highly validated prognostic markers and multiple-marker models and discusses their clinical utility for better therapeutic care.

Keywords: clear cell renal cell carcinoma (ccRCC); cox models; independent datasets; multivariate analysis; prognostic markers.

Figure 1

(A) Consort diagram showing the selection process of studies included in the literature review. (B) Distribution of the studies investigating one marker, several markers, or multiple-marker models. (C) Venn diagram of the distribution of technologies used to quantify the expression level of the 341 genes. IHC, immunohistochemistry; TMA, tissue microarray; RNA-seq, RNA sequencing; RTQ-PCR, reverse-transcription quantitative polymerase chain reaction. (D) Distribution of the number of studies according to the type of biomaterial over the years: Frozen samples and/or formalin-fixed paraffin-embedded (FFPE) samples. The blue line indicates the number of studies using The Cancer Genome Atlas (TCGA) dataset as training or validation dataset.

Figure 2

(A) Barplot of the number of markers cited in one or more studies. (B) Barplot of the most investigated prognostic markers. In orange are indicated prognostic markers specific to clear cell renal cell carcinoma (ccRCC). (C) Barplot of the number of studies investigating markers involved in the main biological pathways: angiogenesis, immunity, cell cycle, and hypoxia. Pies on the right represent the proportion of prognostic markers in the pathway. (D) Distribution of the studies assessing the prognostic value of genes on chromosome 3p over the years. (E) Barplot of the number of studies integrating clinical covariates. ECOG, Eastern Cooperative Oncology Group; VI, vascular invasion; BMI, body mass index; SSIGN, Stage, Size, Grade, and Necrosis; LVI, lymphovascular invasion; MVD, microvessel density; MSKCC, Memorial Sloan Kettering Cancer Center.