MRNA expression pattern study of esophageal basaloid squamous cell carcinoma.

Abstract

e16062 Background: Esophageal squamous cell carcinoma (SCC) is one of the most common cancers and the leading cause of cancer-related death in China, and esophageal basaloid squamous cell carcinoma (BSC) is a rare variant (2.2%-11.3%) of SCC with highly aggressive behavior due to its propensity for vessel invasion and distant metastasis. Most reports stated that BSCs had poorer prognosis than conventional esophageal SCC. However, because of the histological morphologic similarity and lacking of diagnostic approaches, esophageal BSC can easily result in misdiagnosis as SCC. Methods: To deepen our understanding of esophageal BSC and seek for diagnostic biomarkers, we analyze the difference of mRNA expression between esophageal BSC and conventional SCC. Nanostring nCounter PanCancer Progression gene expression panel was used on different tumor components from 36 patients diagnosed with esophageal BSC component (Ba) coexisting with conventional SCC component (Co), who underwent esophagectomy at our Hospital (Cancer Hospital, Chinese Academy of Medical Sciences, China). Immunohistochemistry (IHC) was used for verification of candidate biomarkers sFRP1 concluded from mRNA expression analysis. The expression of sFRP1 was evaluated by H-score = % cells staining (0-100%) * intensity (1 to 3). Results: Among 770 RNAs analyzed in the study, 22.9% (176/770) showed downregulation (FDR-adjusted p < 0.05) in Ba than Co, while 5.3% (41/770) showed upregulation. A set of significantly upregulated RNAs (FDR-adjusted p < 0.01, log2(fold change) > 2) were observed in Ba, including sFRP1, FGFR1, ITGA9, VIT, VWA2 and PROM1, which were associated with angiogenesis response, basement membrane, and epithelial-mesenchymal transition (EMT). Besides, significant downregulation (FDR-adjusted p < 0.01, log2(fold change) < -2) of KRT14 was observed in Ba, which was well known as a biomarker for squamous differentiation. Clustering analysis with different pathway scores according to mRNA expression results of Ba yielded two major clusters: Cluster1 with higher pathway scores with resembled to Co, and Cluster2 with lower pathway scores. Survival analysis revealed that patients in Cluster1 seemed to have a longer RFS and a longer OS than those in Cluster2 (Figure 4C&D), although with no statistically significance (RFS: p = 0.322, median RFS = 23.7 vs 10.8 months; OS: p = 0.217, median OS = 51.6 vs 26.5 months), which was probably due to the limited number of patients evolved (10 patients in Cluster1 and 26 in Cluster2). The most upregulated mRNA, sFRP1, was validated by IHC. Receiver operating characteristic (ROC) curve revealed that sFRP1 IHC could be a remarkable diagnosis biomarker to distinguish BSC from conventional SCC (area under ROC = 0.935, p < 0.001 95%CI = 0.871-1.000). Conclusions: Our results revealed BSC as a subtype with distinct RNA expression pattern compared with conventional SCC, and provided a remarkable biomarker, sFRP1, for diagnosis.