While the buzz around generative AI continues to build, physicians at advanced cancer centers, startups and biomedical companies are seeing the potential for gen AI solutions to advance cancer research and treatment.
During Amazon Web Services' annual summit in Washington, D.C., cancer care pioneers detailed the innovative solutions they’ve built thus far with Amazon Bedrock, AWS’ machine learning platform.
Rowland Illing, chief medical officer and director of international public sector health at AWS, touted that AWS’ Bedrock solution democratizes access to machine learning. “Bedrock is democratizing access to foundation models and putting it in the hands of those who don’t necessarily have to be hardcore, machine learning experts or data scientists,” he said.
The cancer innovation deep dive follows a funding commitment by AWS on Wednesday of $10 million to support research of pediatric rare diseases, including cancer.
Hurone AI is one company using Bedrock to advance cancer care globally.
Hurone is a two-way AI supported platform that connects clinicians and patients once they are diagnosed with cancer. The app can prompt patients to answer questions about the side effects from their treatment, which can help the clinician identify issues or emergencies. It also gives cancer patients more access to their clinician and ensures all interactions are recorded in a centralized medical record.
The platform analyzes data and generates messages so clinicians can support patients remotely and prompts them to do things to manage their care. Hurone also uses generative AI to assess patient eligibility and exclusion criteria for clinical trials.
Kingsley Ndoh, founder and CEO of Hurone AI, said it’s using Bedrock to tailor care to the standards, drug approvals and languages of countries in Africa and South America.
“We are building culturally sensitive AI,” Ndoh said. “A lot of the tools that are being built are really focused on advanced healthcare systems, systems in North America. And so what we're doing in Africa [and] Latin America … we iterate our outputs based on the regional guidelines and not just the general guidelines.”
Patients can also pick the language they want to use with the platform. While a patient can choose to message a clinician in Kinyarwanda, a language widely spoken in Rwanda, a clinician can view the message and respond in English which will be translated for the patient.
Genomics England, a genetic testing company that partners with the National Institutes of Health, is using advanced computing through AWS to improve the diagnosis of rare diseases. Francisco Azuaje, director of bioinformatics at Genomics England, said the company has identified new gene associations linked to intellectual disabilities with the advanced technology.
It is also using gen AI to extract and synthesize thousands of academic articles on genetic research to further its findings.
Praveen Meka, M.D., a physician at Dana-Farber Cancer Institute, said the leading cancer treatment center has been cautious about implementing AI solutions. It has ventured into using a large language model to summarize clinician notes and answer research questions, as well as for medical students to search their curriculum, he said. Dana-Farber is also running a pilot using generative AI to generate plain language descriptions of advanced blood tests for patients to understand their test results.
“[There’s] a theme emerging around complex data … [AI] assists in that bringing together,” Illing said.
Adam Resnick, Ph.D., director of the center for data driven discovery in biomedicine at the Children’s Brain Tumor Network, part of the Children’s Hospital of Philadelphia, said one of the biggest challenges in AI is communicating with relevant stakeholders and aligning their visions.
Other experts on panels agreed that clinicians and the rest of the healthcare system will have to shift their mindsets and be open to adopting AI technologies. They said the mindset shift is critical and will be harder to overcome than issues with the technology.