Article Text

Download PDFPDF
OC-060 A clinically applicable three gene signature is independently highly prognostic in oesophageal and junctional adenocarcinoma
  1. C J Peters1,
  2. J R E Rees2,
  3. J S Hardwick3,
  4. S L Vowler4,
  5. C J Ong1,
  6. C Zhang3,
  7. V Save5,
  8. M O'Donovan6,
  9. D Rassl7,
  10. C Caldas8,
  11. D Alderson9,
  12. R H Hardwick10,
  13. R C Fitzgerald1,1,
  14. On behalf of the OCCAMS Study Group
  1. 1Medical Research Council's Cancer Cell Unit, Hutchison/MRC Research Centre, Cambridge, UK
  2. 2Department of Surgery, Bristol Royal Infirmary, Bristol, UK
  3. 3Molecular Profiling, Merck Research Laboratories, North Wales, Pennsylvania, USA
  4. 4Bioinformatics Core, Cambridge Research Institute, Cambridge, UK
  5. 5Department of Pathology, Edinburgh Royal Infirmary, Edinburgh, Scotland, UK
  6. 6Department of Pathology, Addenbrookes Hospital, Cambridge, UK
  7. 7Department of Pathology, Papworth Hospital, Papworth, UK
  8. 8Cancer Genomics, Cambridge Research Institute, Cambridge, UK
  9. 9Academic Deprtment of Surgery, Queen Elizabeth Hospital, Birmingham, UK
  10. 10Department of Surgery, Addenbrookes Hospital, Cambridge, UK


Introduction The incidence of oesophageal and junctional adenocarcinoma has increased sixfold in the last 30 years and 5-year survival remains <14%. Most patients present with advanced disease and current staging is limited in its ability to predict survival. We aimed to generate and externally validate a molecular prognostic signature for oesophageal adenocarcinoma.

Methods Gene expression profiling was performed and the resulting 42 000 gene signatures correlated with clinical features for 91 snap frozen oesophageal and junctional resection specimens. External validation of selected targets was carried out on 371 independent cases using immunohistochemistry to maximise clinical applicability.

Results Gene expression profiles were obtained from 75/91 of the samples (82%). 119 genes were significantly associated with survival and 270 genes with the number of involved lymph nodes. Ten of these genes were taken forward to external validation at the protein level. Three genes were prognostic in the external validation data set (TRIM44, SIRT2 and PAPSS2). Patients with dysregulation of none of these three genes had a significantly better outcome (5-year survival 48%) than those with dysregulation of one gene (5-year survival 29%), who in turn did better than those with two or more genes dysregulated (5-year survival 15%) (p=0.004). The three gene signature was independently prognostic in a multivariable model together with the existing clinical TNM staging system (p=0.003).

Conclusion We have generated and externally validated an immunohistochemical prognostic gene signature which is independently highly prognostic in an external validation data set. In addition these three genes are possible therapeutic targets raising the possibility of personalised therapy based on a patient's molecular prognostic signature. Our three gene signature is suitable for application in routine clinical practice and this study may provide a framework for similar future projects.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.