Machine Learning and Deep Learning Approaches in Breast Cancer Survival Prediction Using Clinical Data
importance 7/10
ai assisted
Artificial Intelligence in Healthcare oncology stage: research
Problem
Predicting breast cancer survival is needed to help select the best treatment protocols for patients.
Why it matters
Accurate survival prediction can guide treatment decisions for a major cancer type.
What the AI did
Machine learning and deep learning classifiers were trained on 4,902 patient records from a breast cancer registry to predict survivorship.
Results
MLP 88.2%, RF 83.3%, DT 82.5%, SVM 80.5% accuracy; tumour size most important feature
Impact
Demonstrated ML can predict survivorship with up to 88.2% accuracy in tested samples.
Record
| Venue | Folia Biologica |
|---|---|
| Published | 2019 |
| Work type | article |
| Authors | Elham Yousef Kalafi, Nurul Aqilah Mohd Nor, Nur Aishah Mohd Taib, Mogana Darshini Ganggayah, Christopher P. Town, Sarinder Kaur Dhillon |
| Citations | 108 (OpenAlex, at import time) |
| References | 26 |
| Model or method | multilayer perceptron (MLP), random forest, decision tree, SVM |
| Organization | University of Malaya Medical Centre (data source) |
| Primary source | https://doi.org/10.14712/fb2019065050212 |
| Open access | hybrid |
| Keywords | breast cancer, survival prediction, deep learning, multilayer perceptron, random forest |
Importance rating
7
significant — thousands to millions affected
Survival prediction guides treatment selection for a major cancer type; abstract calls the effect 'extreme' but provides no scale numbers.