Related Experiment Videos
Colon cancer prognosis prediction by gene expression profiling
Alain Barrier1, Antoinette Lemoine, Pierre-Yves Boelle
1Service de Chirurgie Digestive, Hôpital Tenon, Université Pierre et Marie Curie, Assistance Publique, 75020 Paris, France. barrier@stat.Berkeley.edu
Oncogene
|August 11, 2005
Summary
Accurate prognosis predictors for stage II and III colon cancer were developed using gene expression measures from either tumor or non-neoplastic mucosa. These predictors achieved high accuracy, suggesting a new approach for patient outcome assessment.
Area of Science:
- Oncology
- Genomics
- Bioinformatics
Background:
- Accurate prognosis prediction is crucial for managing stage II and III colon cancer.
- Gene expression profiling offers a potential avenue for developing novel prognostic tools.
Purpose of the Study:
- To assess the feasibility of creating a prognosis predictor for stage II and III colon cancer patients using microarray gene expression data.
- To determine if gene expression from tumor (T) or non-neoplastic mucosa (NM) is more effective for prognosis prediction.
Main Methods:
- Microarray gene expression profiling of T and NM mRNA samples from 18 colon cancer patients (9 recurrence, 9 no recurrence).
- Application of the k-nearest neighbour method for prognosis prediction.
- Utilized six-fold and double cross-validation for predictor selection and accuracy estimation.
Main Results:
- A 30-gene tumor-based predictor and a 70-gene non-neoplastic mucosa-based predictor were developed.
- Estimated accuracies for the predictors were 78% (T-based) and 83% (NM-based).
- Both tumor and non-neoplastic mucosa gene expression data proved valuable for prognosis prediction.
Conclusions:
- It is possible to build accurate prognosis predictors for stage II and III colon cancer patients based on gene expression measures.
- Both tumor and non-neoplastic mucosa samples can be effectively utilized for developing these prognostic tools.