Exclusive: Color CEO Says AI-for-Oncology Copilots Detect and Treat Cancer Earlier

color openAI, ai

The most popular applications of artificial intelligence (AI) today typically center around automating routine tasks.

But that’s not to say that the innovation doesn’t hold extraordinary potential beyond its mastery of the mundane.

Within healthcare, AI’s ability to ingest vast amounts of data and analyze it instantaneously could help usher in a new era of medical innovation and patient care — particularly when applied to historically intractable problems and diseases, such as cancer.

“When it comes to generative AI, a lot of the applications people have been focusing on is around alleviating the administrative burden of healthcare, the processing, the payments, the bookkeeping and the transcription of clinical notes,” Othman Laraki, co-founder and chief executive of Color Health, told PYMNTS’ CEO Karen Webster.


But, as Laraki explained, what his company is after is something entirely different.

“As opposed to automating what people think of as lower-scale labor to save costs, we partnered with OpenAI to focus on areas where you need a lot of medical expertise and depth, but where that expertise is very scarce and that scarcity comes at a high cost, like cancer,” he said. “We decided, instead of going broad, to go very deep in places where we felt there’d be a very big leverage.”

The result?

A new way of accelerating cancer patients’ access to treatment that uses the capabilities of GPT-4o to help doctors transform cancer care.

Enhancing Clinician Expertise With AI

Cancer is the second most common cause of death in the United States and the leading driver of American healthcare costs.

That’s why, Laraki explained, Color Health’s collaboration with OpenAI aims to address two critical areas: risk-adjusted screening and pre-treatment workup. The collaboration focuses on leveraging AI to enhance the expertise and efficiency of clinicians rather than merely automating administrative tasks surrounding their work.

One of the most impactful uses of AI in oncology is improving risk-adjusted screening. Many individuals with high-risk factors, such as genetics, family history or lifestyle choices like smoking, do not receive appropriate screening. As Laraki noted, AI can bridge this gap by ensuring that established risk-adjusted guidelines are applied more consistently and accurately.

“The majority of people who should be getting risk-adjusted screening guidelines don’t today,” he said, adding that early diagnosis is crucial in cancer treatment and can significantly improve survival rates and reduce treatment costs.

By using AI to identify and monitor high-risk individuals, healthcare providers can detect cancers at an earlier, more treatable stage.

But diagnosis is just the start of the healthcare journey, and the period between cancer diagnosis and the initiation of treatment is often fraught with delays, causing unnecessary anxiety and potentially affecting patient outcomes.

“One of the things that blocks being able to initiate treatment, especially as treatments are getting more and more complex, is the workup that happens so that your oncologist can initiate treatment,” Laraki said.

He explained that AI can streamline this process by expediting the pre-treatment workup. By the time a patient meets their oncologist, AI can ensure that all necessary tests and preparations are completed, allowing treatment to commence promptly. This not only improves patient survival rates but also optimizes healthcare resources.

AI and Healthcare: A Revolutionary Partnership

The application of AI in cancer screening and diagnosis represents a significant leap forward in oncology, but it is an evolution — not a pull-the-rug transformation.

That’s because, as Laraki emphasized, integrating AI into healthcare is not about replacing clinicians but augmenting their capabilities. AI can process vast amounts of patient data, extract relevant information, and apply complex guidelines with precision. This allows clinicians to make more informed decisions quickly. AI models are able to act as co-pilots, providing clinicians with comprehensive analyses and recommendations while leaving the final decisions in human hands.

“It is about leveraging AI tooling to amplify the existing expertise today that is very scarce,” Laraki said. “It is always the clinician who is the driver here.”

Still, the integration of AI into cancer care is not just a technological advancement but also a cultural shift. Historically, cancer has been perceived as an unavoidable, costly burden. However, there is a growing recognition that proactive measures, driven by AI and other technologies, can significantly impact outcomes.

“There’s no silver bullet. It is such a vast surface area that it is about providing an integrated set of solutions that cover the different places relevant to cancer,” Laraki explained about AI’s applications across oncology.

He added that many issues, rather than being “science problems,” are actually “immediacy and logistics and integration” problems.

From ensuring follow-ups on positive screenings to coordinating care across multiple specialists, AI can play a pivotal role in streamlining these logistical problems.

For example, Laraki highlighted that the gap in follow-up care for colorectal cancer screening, where a significant percentage of positive cases do not receive timely follow-ups, can be addressed through AI-driven systems that track and remind patients and healthcare providers of necessary actions.

AI can also facilitate better coordination among healthcare providers, reducing delays and improving the overall patient experience. By integrating various stages of cancer care, from education and screening to diagnosis and treatment, AI can help create a more cohesive and efficient healthcare system.