gms | German Medical Science

27. Deutscher Krebskongress

Deutsche Krebsgesellschaft e. V.

22. - 26.03.2006, Berlin

Effectiveness of pretherapeutic gene expression profiling in prediction of tumor response and disease free survival after preoperative 5FU based chemoradiotherapy in rectal cancer (UICC stage II/III)

Meeting Abstract

  • corresponding author presenting/speaker B. Michael Ghadimi - Universitätsklinikum, Göttingen, Deutschland
  • Torsten Liersch - Universitätsklinikum, Göttingen
  • Marian Grade - Universitätsklinikum, Göttingen
  • Sudhir Varma - National Cancer Institute, Bethesda, U.S.A.
  • Richard Simon - National Cancer Institute, Bethesda, U.S.A.
  • Michael Diflippantonio - National Cancer Institute, Bethesda, U.S.A.
  • Lazlo Füzesi - Universitätsklinikum, Göttingen
  • Claus Langer - Universitätsklinikum, Göttingen
  • Clemens Hess - Universitätsklinikum, Göttingen
  • Thomas Ried - National Cancer Institute, Bethesda, U.S.A.
  • Heinz Becker - Universitätsklinikum, Göttingen

27. Deutscher Krebskongress. Berlin, 22.-26.03.2006. Düsseldorf, Köln: German Medical Science; 2006. DocOP598

Die elektronische Version dieses Artikels ist vollständig und ist verfügbar unter:

Veröffentlicht: 20. März 2006

© 2006 Ghadimi et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen ( Er darf vervielfältigt, verbreitet und öffentlich zugänglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.



Introduction: According to the results of the CAO-/ARO-/AIO-94 trial of the German Rectal Cancer Study Group, preoperative 5-fluorouracil (5-FU) based long-term chemoradiotherapy (CT/RT) is recommended for UICC stage II/III rectal cancer. However, tumor response towards neoadjuvant chemoradiation as well as disease-free survival differs significantly in patients. Thus, we have studied firstly whether gene expression profiling might assist in tumor response prediction as well as secondly whether array technology might support the prediction of disease-free survival.

Patients and Methods: Pretherapeutic biopsies from 23 participants of the CAO-/ARO-/AIO-trial, all of whom received preoperative CT/RT, were analyzed for gene expression signatures using 10K cDNA microarrays. Local T-level downsizing, lymph node downcategorizing and histopathomorphologically measured tumor regression grading (TRG) were correlated with DFS and overall survival (OS). Leave-one-out cross-validation (LOOCV) analysis was performed using the gene expression data in order to predict local or distant cancer recurrence. Data were analyzed using Principal Component Analysis (PCA) and Diagonal Linear Discriminant (DLD) classifier analyses.

Results: Tumor response was correctly predicted using a classifier of 54 differentielly expressed genes in 83% of patients (p=0.02). Sensitivity (correct prediction of response) was 78% and specificity (correct prediction of non-response) was 86% with a positive predictive value of 78% and negative predictive value of 86%. After 34.5 months (median) of postoperative follow-up, DFS was 77% and OS 87%. All five patients with metastatic disease belonged to the group of 16 non-responders (p=0.0135). LOOCV resulted in correct prediction of all five patients with recurrence. Based on gene expression profiles, all responsive patients as well as two non-responsive patients were predicted to remain recurrence free.

Conclusion: Our results suggest that gene expression profiling may assist both in the prediction of local tumor response to preoperative CT/RT and to DFS. This provides incentive to now include gene expression profiling in trials for treatment stratification and clinical management.