Radiomics biomarkers are associated with survival in patients with Oesophageal Adenocarcinoma

Abstract Objective: Radiomics studies have shown promising results defining image surrogates that predict treatment response and survival in oesophageal squamous cell carcinoma. There are limited studies examining radiogenomics profiles in oesophageal adenocarcinoma (OAC) and this warrants further exploration. Background: Although advances have been made in the treatment of OAC, the five-year overall survival remains poor. Hence, there is a need for biomarkers to improve therapeutic approaches. Methods: Image analysis was performed on pre-treatment PET/CT scans from 80 OAC patients. In the discovery cohort (n=50), CT texture analysis was performed using a filtration-histogram method to evaluate the prognostic value of image features. Whole-genome sequencing, RNA sequencing (RNAseq) and immunohistochemistry characterised the molecular pathways associated with the CT markers. A validation cohort (n=30) was used to confirm radiomics biomarkers. Results: Analysis showed 3 CT markers were significantly associated with both disease-specific and disease-free survival. Combining these markers provided further prognostic significance. Furthermore, we identified a correlation between CT markers and major pathological response rate. Both RNAseq and immunohistochemistry confirmed that lower CD8 T-cell expression in the tumour correlated with the poor survival group predicted using CT markers. RNAseq gene set enrichment analysis identified pathways associated with these radiomics biomarkers. Finally, radiomics analysis in a validation cohort confirmed CT marker (SD, standard deviation) as a biomarker of survival and identified additional image features associated with patient survival. Conclusions: This study demonstrates that radiogenomics has the potential to identify OAC subgroups with prognostic significance and support clinical decision making.