Prediction of Colon Cancer Stages and Survival Period with Machine Learning Approach
Problem
Predicting colon cancer TNM stage and five-year disease-free survival from histopathology parameters is difficult.
Why it matters
Accurate staging and survival prediction inform colon cancer treatment decisions.
What the AI did
Various machine learning algorithms, with Random Forest performing best, predicted tumor stage and DFS using 4021-patient registry data including a Tumor Aggression Score.
Results
Random Forest achieved an F-measure of 0.89 with Tumor Aggression Score, evaluated by five-fold cross-validation.
Impact
Retrospective internal validation only; clinical deployment or external validation not stated in source.
Record
| Venue | Cancers |
|---|---|
| Published | 2019 |
| Work type | article |
| Authors | Pushpanjali Gupta, Sum-Fu Chiang, Prasan Kumar Sahoo, Suvendu Kumar Mohapatra, Jeng‐Fu You, Djeane Debora Onthoni, Hsin‐Yuan Hung, Jy‐Ming Chiang et al. |
| Citations | 100 (OpenAlex, at import time) |
| References | 24 |
| Model or method | Random Forest (among various ML algorithms) |
| Organization | Chang Gung Memorial Hospital, Linkou, Taiwan (data source) |
| Primary source | https://doi.org/10.3390/cancers11122007 |
| Open access | gold |
| Keywords | colon cancer, TNM staging, Random Forest, survival prediction, machine learning |
Importance rating
significant — thousands to millions affected
Colon cancer staging and survival prediction directly inform treatment decisions for a major cancer; 4021-patient registry used, but broader impact scale not stated.