New Model Predicts Cancer Survival Rates with 30% Greater Accuracy
Machine-learning techniques have revolutionized the medical field by enabling the prediction of survival rates and life expectancies for patients with various diseases. Now, researchers from The University of Texas at Arlington have developed a new model that predicts cancer survival rates with 30% greater accuracy than previous models.
The traditional approach to modeling the probability of a cure in cancer patients relied on generalized linear models, which could not capture non-linear or complex relationships between important variables and the cure probability. However, the new model developed by Professor Suvra Pal and doctoral student Wisdom Aselisewine combines a previously tested promotion time cure model with a machine-learning algorithm called a support vector machine. This integration allows for the identification of non-linear relationships between variables and the cure probability.
Supported by a grant from the National Institute of General Medical Sciences, this new model, called the SVM-integrated PCM model (PCM-SVM), provides a simple interpretation of covariates to predict which patients will be uncured at the end of their initial treatment and require additional interventions.
To validate the effectiveness of the new model, the researchers used real survival data for patients with leukemia, a type of blood cancer often treated with a bone marrow transplant. By comparing the predictions of the PCM-SVM model with those of the previous technique, they found that the PCM-SVM model was 30% more accurate in predicting which patients would be cured by the treatments.
These results have significant implications for patient care. With the improved predictive accuracy of the PCM-SVM model, patients with high cure rates can avoid unnecessary high-intensity treatments and the associated risks. On the other hand, patients with low cure rates can be recommended timely treatment to prevent the disease from progressing to an advanced stage with limited therapeutic options.
Professor Pal emphasizes the importance of the proposed model in defining optimal treatment strategies. By identifying patients with a high likelihood of being cured, the model can help protect them from unnecessary treatments. Furthermore, it can ensure that patients with low cure rates receive timely interventions to prevent disease progression.
Overall, the new PCM-SVM model offers a significant advancement in predicting cancer survival rates. Its ability to capture non-linear relationships between variables and the cure probability makes it a valuable tool for healthcare professionals in making informed treatment decisions. As machine-learning techniques continue to evolve, they are likely to play an even more prominent role in improving patient outcomes in the future.
References:
1. University of Texas at Arlington. New Model Predicts Cancer Survival Rates with 30% Greater Accuracy. Mirage News, 21 Jan. 2022, https://www.miragenews.com/new-model-predicts-cancer-survival-rates-with-726697/.
2. University of Texas at Arlington. New Model Predicts Cancer Survival Rates with 30% Greater Accuracy. National Institute of General Medical Sciences, https://www.nigms.nih.gov/.
