From: Predicting multi-class responses to preoperative chemoradiotherapy in rectal cancer patients
1. Collecting 77 rectal cancer samples with clinical features |
2. Gene expression profiling with Affymetrix ST1.0 array |
3. Feature selection by pval, norm and rank |
4. Designing most accurate prediction models for MI or TO |
5. Testing the prediction models for MI or TO |
6. Multi-class prediction model |
7. Internal validation of prediction model |
8. External validation of the prediction model |