Predicting Treatment Outcomes for Kidney Cancer
Author Information
Author(s): Chen Meixu, Wang Kai, Kapur Payal, Brugarolas James, Hannan Raquibul, Wang Jing
Primary Institution: Cornell University
Hypothesis
Can a multi-modal ensemble model improve predictions of treatment outcomes for clear cell renal cell carcinoma patients?
Conclusion
The multi-modal ensemble model outperformed single-modality models in predicting survival outcomes for clear cell renal cell carcinoma patients.
Supporting Evidence
- The final multi-modal ensemble model achieved C-indices of 0.820 for overall survival and 0.833 for disease-free survival.
- The model combined clinical data, multi-omics data, and histopathology data to improve predictions.
- Log-rank tests showed improved performance in survival predictions compared to single-modality models.
Takeaway
This study created a smart model that uses different types of patient data to better predict how long kidney cancer patients will live and how likely they are to stay cancer-free.
Methodology
The study analyzed data from 226 patients using a multi-modal ensemble model that combined clinical, multi-omics, and histopathology data to predict overall and disease-free survival.
Limitations
The model needs independent validation before it can be used in clinical practice.
Participant Demographics
Patients were from The Cancer Genome Atlas Kidney Renal Clear Cell Carcinoma dataset.
Want to read the original?
Access the complete publication on the publisher's website