Researchers have developed a computer algorithm that might be able to accurately predict cancer outcomes in patients.
Cancer is a group of complex diseases, characterized by the abnormal growth of cells. One of the major challenges of caring for cancer patients is the inability to accurately predict long-term health outcomes. Specifically, it is difficult for doctors to know how long patients are expected to live and how patients may respond to chemotherapy. Patients suffer because of these uncertainties since they are unsure about their futures. Currently, doctors predict patient outcomes by assessing the patient’s initial symptoms, determining the cell type that the cancer originated from, and by the size and location of the tumor. However, these approaches are not entirely effective, and they only give a snapshot of the tumor’s progression. They provide no information about how the tumor responds to therapies and whether patients are improving or deteriorating over time. Therefore, it is important that researchers seek out newer effective strategies to be able to accurately predict health outcomes in cancer patients.
Researchers from Stanford University School of Medicine recently published a study in Cell, where they demonstrated the effectiveness of CIRI (Continuous Individualized Risk Index) in predicting cancer outcomes. CIRI is a computer algorithm developed by the authors of the study. It is based on the same techniques that statisticians have used for decades to predict sports matches and elections. In short, the algorithm works by integrating a large amount of data, including tumors responsiveness to treatment and the amount of cancer DNA in the patient’s blood, to generate a single, continuous risk assessment.
The researchers collected data from 2,500 patients with diffuse large B-cell lymphoma, the most common type of blood cancer in the United States. The data was then fed into the computer algorithm so that it could recognize and identify patterns/combinations that it would then use to determine patient outcomes. They found that compared to the standard methods, CIRI did markedly better. However, it is was not entirely perfect in its predictive power and it needs further improvement. They also tested CIRI in predicting outcomes for other cancers, including leukemia and breast cancer. Again, they found that while the predictive power was not perfect, and the level of accuracy differed between cancers, CIRI did much better at predicting health outcomes compared to standard techniques.
These findings could have large implications on how doctors manage patients with cancer. Specifically, the technology could allow patients to have a better idea of what to expect over the course of their disease. While CIRI cannot change health outcomes, it can at least allow individuals to plan their futures appropriately. The researchers believe that the technology can also be used to continually monitor patients over their disease progression, allowing doctors to determine relatively quickly whether the chemotherapy regimen a patient receives is improving or worsening their health. With this information in hand, doctors will be able to make minor adjustments to ensure that they are constantly receiving the best care.
Written by Haisam Shah, BSc
Reference: Kurtz, D. M., Esfahani, M. S., Scherer, F., Soo, J., Jin, M. C., Liu, C. L., … & Westin, J. R. (2019). Dynamic Risk Profiling Using Serial Tumor Biomarkers for Personalized Outcome Prediction. Cell.
Image by Gerd Altmann from Pixabay
