What is artificial intelligence?
Artificial intelligence (AI) is a technology that can simulate human intelligence characteristics including learning, predicting, analyzing, and deriving solutions using data. AI was first defined in 1956, but the use of the technology was limited through the early 2000s due to poor computer performance and other economic factors.
In the 2010s, increased computer performance, reduced costs for computer storage, and faster internet led to the Big Data Era. This was most apparent with the rise of social media, smartphones, and other physical objects that connect and exchange data with each other (known as the Internet of Things).
Many of us now seamlessly interact with AI daily, perhaps without even knowing it. Artificial intelligence helps us shop more efficiently, secures living spaces, enables interaction with friends and families, and affects everyday life in various other ways.
How is artificial intelligence being applied to cancer diagnosis and treatment?
Around the same time smartphones and internet connectivity were becoming universal in the early 2010s, a transition was also occurring in medical research. The cost of Next Generation Sequencing decreased significantly. Coupled with powerful new molecular biology techniques, new scientific discoveries were made regarding disease mechanisms. This led to more precise disease definitions and treatment plans.
Since then, medicine and oncology, in particular, have become highly data-driven. Many computational tools are now used to help organize information and make decisions regarding patient care.
Determining care for cancer patients is very complex. Doctors must consider a number of factors, including clinical presentation, patient history, histopathology, medical imaging, genomics, and ongoing clinical trials. As medicine becomes more personalized, care decisions require even more analysis and consideration.
This is where artificial intelligence can help.
AI enhances current computational tools in a way that allows more data to be used while decreasing the burden on clinicians so they can focus on the more humanistic components of healthcare. Artificial intelligence is currently used in cancer diagnosis and treatment in three primary ways:
- Classification: These tasks are the most common and include identifying a tumor in a clinical image like an MRI or CT scan. More advanced uses include identifying tumor location and borders for radiation therapy, alerting staff if a dangerous combination of drugs or procedures has been inadvertently requested, and subtyping a tumor based on complex genetics.
- Translation: This area of AI converts information from one data type to another, such as conversion from an audio recording to text. This enables greater documentation from clinical visits without requiring additional effort from patients, doctors, or nursing/support staff.
- Data generation: This is a powerful tool that can be used to reduce image acquisition time, as well as create private and secure versions of patient data that are safer to share.
You’ve made possible artificial intelligence research to help kids with cancer
Because of support from generous people like you, The Morgan Adams Foundation has been funding artificial intelligence research in pediatric cancer since 2017.
In 2020, researchers in the Morgan Adams Foundation Pediatric Brain Tumor Research Program showed that it is possible to predict the diagnosis of rare suprasellar brain tumors known as craniopharyngioma using clinical images like MRI and CT scans. Their optimized AI model performed as well as human neuroradiologists, with about 85% accuracy. This work was published and has been presented at major international and national scientific conferences.
There are currently four artificial intelligence research projects underway. Researchers in Dr. Todd Hankinson’s laboratory are:
- Expanding the AI classifier of suprasellar tumors from covering 5% of pediatric brain tumors to more than 40%
- Developing translational AI to predict the risks of neurosurgical intervention. This will improve the quality and safety of patient care
- Exploring how clinicians perceive the information provided by AI systems to improve the human-centered design
- Developing a new web-based platform to facilitate centralized and secure patient data handling, AI model development, visual analytics design, and user studies.
These four projects will lay a foundation for further development and utilization of AI tools for clinical neuro-oncology. The overarching goal of this AI research is to improve patient care by developing a tool that is satisfying for clinicians to use and improves decision-making.
Thanks to your support, doctors and researchers in the Morgan Adams Foundation Pediatric Brain Tumor Research Program have been recognized as leaders in the application of AI to improve pediatric brain tumor care, which has led to additional funding from Children’s Hospital Colorado and the National Institutes of Health.
Our team is collaborating with their counterparts at the Colorado School of Public Health, St. Jude Children’s Research Hospital, and the University of Birmingham (England).
Thanks to Senior Researcher and doctoral student Eric Prince for his help with this update and for putting together this graphic illustrating the artificial intelligence work you’re making possible to help kids with cancer!
