Department of Radiation Oncology, Cancer Center, Zhongshan Hospital, Fudan University Shanghai, Shanghai
Z. Shi1,2, X. Zhu2, Y. Tian1, L. Gao1, H. Qiu1, J. Wang1, Y. Zhangcai1, J. Chen1, Y. Wu1, and Y. Chen1; 1Cancer center, Renmin Hospital of Wuhan University, Wuhan, China, 2Department of Oncology, The First Affiliated Hospital of Henan University of Science and Technology, Luoyang, China
Purpose/Objective(s): To explore the molecular events between different pathological responses in neoadjuvant chemoradiotherapy (nCRT) for locally advanced esophageal squamous cell carcinoma (ESCC), and to evaluate the predictive performance of key genes for nCRT outcomes in the real-world. Materials/
Methods: Fresh surgical specimens of ESCC undergoing nCRT from prospective clinical trial were included, and whole transcriptome sequencing (WTS) was performed. Differential expression gene analysis and weighted gene correlation network analysis were conducted to identify key genes between different pathological responses. Gene set enrichment analysis was used to reveal signal pathways in ESCC related to key genes. The function and predictive capabilities of key genes were verified in ESCC cell lines and the GEO dataset. Time-continuous ESCC surgical specimens receiving nCRT from another cancer center were used to verify the roles of key genes. The predictive value of key genes was deeply evaluated by integrating the clinical data of patients and machine learning methods. Results: Fifteen fresh specimens were performed WTS, including 8 pathologic complete responses (pCR) and 7 non-pCR. After analysis and internal verification, 2 key genes, namely HOXD11 and HOXD13, were found up-regulated in ESCC and highly associated with non-pCR after nCRT. These key genes were mainly involved in Epithelial-to-mesenchymal transition (EMT), Androgen receptor (AR), and Receptor tyrosine kinase (RTK) signal pathways of ESCC, and their down-regulation could inhibit the proliferation, invasion, and migration of ESCC cells. ROC analysis showed that the AUC of HOXD11 and HOXD13 for predicting pCR was 0.781 (95% CI 0.572-0.989) in the GSE104958 dataset. Among an independent external ESCC cohort of 105 cases, immunohistochemical analysis indicated that the expression of HOXD11 and HOXD13 in non-pCR patients was significantly higher than that in pCR (89% vs. 50%, 95% vs. 50%) (all P<0.001). The important predictive variables of pCR and progression-free survival (PFS) after nCRT by randomforest and best subset regression were HOXD11 and HOXD13. The nomogram prediction models were established to deeply evaluate their predictive performance based on the different expression ranks of HOXD11 and HOXD13, and the AUC for predicting pCR was 0.882 (95% CI 0.800-0.964), and the time-dependent maximum AUC for predicting PFS was 0.728 (95% CI 0.627-0.829). The calibration curves, decision curve analyses, and clinical impact curves demonstrated that the expression of HOXD11 and HOXD13 had favorable predictive value and clinical practicability. Conclusion: Novel roles for HOXD11 and HOXD13 in the regulation of cell survival were uncovered in ESCC, and the combined detection of HOXD11 and HOXD13 could accurately predict the pathological response and PFS for ESCC patients receiving nCRT. These biomarkers should be integrated into future clinical trials to provide accurate decision-making for the management of ESCC patients.
