Scientists from the University of British Columbia and BC Cancer created a new AI system which is better at predicting how long cancer patients will live compared to other methods. Their model also uses data that is easier to get.
Scientists are now using a type of artificial intelligence (AI) to predict the survival times of cancer patients. This AI looks at notes from doctors after a patient’s first appointment, searching for patterns in what it finds. The results were published recently, showing that the AI could predict how long a person would live with accuracy greater than 80% for six months, three years, and five years into the future.
Dr. John-Jose Nunez, who works at the UBC Mood Disorders Centre and BC Cancer, said that predicting cancer survival is really important to provide better care for people with cancer. He hopes that using this technique may allow doctors to give patients the best outcome possible by giving them personalised treatment options or referring them to support services right away.
In the past, doctors calculated survival rates for cancer by looking at only a few basic things such as where in the body the cancer is and what type of cells it affects. Even though these numbers are well-known, it can still be difficult for doctors to accurately predict how long a particular patient will live because of all the different factors involved.
Dr. Nunez and a team of researchers from BC Cancer, UBC’s computer science, and psychiatric departments created a new model that can recognize clues within the patient’s first appointment details to give more detailed answers. This works for all kinds of cancers instead of just specific ones like earlier models do.
Dr. Nunez said that to bring all of this information about a patient together, like their age, type of cancer, past health problems, if they’ve used substances before, and family history, the AI reads consultation documents in the same way as humans do. This helps give a more complete picture of what results might be expected.
Researchers tested a new type of technology called “AI” (Artificial Intelligence) by using data from over 47,000 patients located around British Columbia. To be extra safe, all patient information was stored securely and not even the researchers knew who’s records they were looking at. This form of AI can also help protect people’s privacy since humans won’t see their data.
Dr. Nunez said that our province’s cancer survival rate could be predicted by a model built on BC data. This technology could eventually help cancer patients in Canada and other countries all over the world.
Dr. Nunez said that neural NLP models are really useful because they allow us to quickly use local patient data from any area in order to do better cancer care. This can help people who have cancer get better care no matter where they live in the world.
Dr. Nunez is a very special person who will be given money to work on some great projects. The money is coming from the UBC Institute of Mental Health and the BC Cancer Foundation. He’s trying to figure out how to help cancer patients get better at counseling using artificial intelligence. He also wants AI technology to become part of health care so that people can get better care in the future.
Dr. Nunez said that AI can be like a personal assistant to doctors. As healthcare continues to make progress and develop, AI can help by taking in data and providing suggestions for physicians to improve treatments for patients. This will help people have healthier lives and better results from their healthcare.
This study was funded by the BC Cancer Foundation. It was published in February 2023, in JAMA Network Open and referred to as “Predicting the Survival of Patients With Cancer From Their Initial Oncology Consultation Document Using Natural Language Processing” written by John-Jose Nunez, MD, MSc, Bonnie Leung, MN-NP(F), Cheryl Ho, MD, Alan T. Bates, MD Ph.D., and Raymond T. Ng Ph.D. The main purpose of this study was to use natural language processing to figure out how long cancer patients are likely to survive after they go see a doctor for their initial consultation.
1. What happens when a lot of different strategies are put in place to help with people who have drug-resistant tuberculosis in China?
2. How can we improve the way we deal with tuberculosis across China, and what did we learn from it?
3. Why do some places have more issues than others when it comes to helping patients with tuberculosis in South Africa?
1. Closing the health gap: Studying South Africa’s ways for improving healthcare.
2. Is e-learning a good way to help doctors learn about tuberculosis? An experiment from China.
3. Can taking sericin help protect the eyes of rats with diabetes? Research shows it might!
4. Will a health check center be able to quickly and effectively spot eye diseases? Let’s find out!
We need to compare and study the different types of diabetic retinopathy in animals. It’s crucial to use RB1 genetic testing for diagnosing and treating retinoblastoma, as well as for providing advice about genetics. In adult mice, we can reactivate the plasticity of their vision by reducing caloric intake via a pathway called AMPK-SIRT1 that comes from only using one eye.
