OpenAI has partnered with healthcare startup Color Health to embed GPT-4 into an AI-powered “copilot” that helps doctors develop personalized cancer treatment plans.
Developed by Color Health, Copilot uses OpenAI's models to research patient data, including personal risk aspects and family history, alongside clinical guidelines.
By identifying missing diagnostic tests and creating tailored screening and pre-treatment plans, the AI assistant helps healthcare providers make evidence-based decisions.
“Color’s vision is to make cancer expertise accessible when it may possibly have the best impact on a patient’s medical decisions,” Saidhman LarakiChairman of the Board of Color health.
The copilot's potential to streamline cancer care is remarkable, as delays in screening, diagnosis and treatment can have serious consequences for patients.
studies show that a one-month delay in treatment can increase mortality by 6 to 13 percent.
Color Health's copilot test has already shown promising leads to reducing this number. Doctors can analyze patient records in only five minutes on average, in comparison with the weeks it may possibly take without the AI assistant.
“I've seen firsthand the complexity of developing individualized cancer prevention plans for my high-risk patients,” says Dr. Keegan Duchicela, a primary care physician at Color. “Guidelines are continuously evolving, and individual risk aspects will not be at all times immediately clear.”
OpenAI and Color Health have been working together since 2023. Their goal is to make use of AI to enhance cancer care and health equity. They call it a ““Clinician-in-the-Loop workflow” that fundamentally supports clinical decision-making without replacing it in any way.
“We see an ideal complement for AI technology and language models as they could be really helpful in each of those dimensions,” said Brad Lightcap, Chief Operating Officer (COO) of OpenAI.
“They can bring relevant information to the surface faster. They may give clinicians more tools to grasp medical records, data, lab results and diagnoses.”
To measure the impact of the copilot, Color Health works with the Helen Diller Family Comprehensive Cancer Center of the University of California, San Francisco (UCSF HDFCCC).
The partnership will involve several evaluation phases, followed by a targeted rollout with the potential of integrating the copilot into clinical workflows for all recent cancer cases at UCSF.
“UCSF is a frontrunner in implementing cutting-edge technology to enhance patient care,” said Dr. Alan Ashworth, president of the UCSF HDFCCC.
“Patients often come to oncologists with incomplete diagnostic workups, and the time it takes to compile these workups and accurately discover completion prevents providers from taking full advantage of their credentials. We are excited about tools that may improve the efficiency and accuracy of pre-visit documentation and avoid costly delays in treatment initiation for cancer patients at UCSF.”
Color Health intends to introduce the copilot slowly, starting with an introductory phase for its own clinicians and applying several levels of quality assurance.
By the second half of 2024, the corporate plans to make use of the Copilot application to offer AI-generated, personalized, physician-supervised treatment plans to over 200,000 patients.
The promise of AI within the fight against disease
The key here is using language models to democratize expertise. It has grow to be much easier to optimize models for various purposes, allowing researchers to construct domain-specific models using medical data.
The same applies to eye health. Chatbot was developed to offer clinicians with information on retinal problems and glaucoma. It met or exceeded expert advice in answering clinical questions, demonstrating again how AI can democratize expertise.
The potential of AI in detecting, diagnosing and treating diseases, including cancer, is otherwise well known.
AI tools beat doctors Identification of complex cancers and accelerating drug discovery and Drugs identified using AI are even near entering clinical trials.
