Carnegie Mellon spinout Xlue gains the backing of Sarah Guo’s VC firm Conviction to use AI to help identify which patients are most at risk of developing cancer. (To get the Forbes weekly InnovationRx newsletter in your inbox, subscribe here .)

C henyan Xiong has been working on large language models for a long time, first as a researcher at Microsoft, then as a professor at Carnegie Mellon. So after he spun out a company last year that was building a new clinical model that could flag people with higher cancer risk, he was quickly connected through friends also doing AI research to investors at Conviction, the AI-focused firm founded by Midas lister Sarah Guo.

His startup, Xlue, has now raised a $35 million seed round led by Conviction. The company’s co-founders include top health AI researchers from Georgetown, Stanford and the NIH. “Our model is best in class to tell which patients have an elevated risk of cancer,” Xiong tells Forbes. “This is not diagnostic, it is risk stratification. This is about who the doctors should test early.”

Xlue’s model trains on large-scale, longitudinal health records from the U.S., Taiwan and elsewhere to learn patients’ health trajectories. It then analyzes patients’ electronic health records to predict their likelihood of developing cancer by looking for patterns. “What happens to you and your body is not a next symptom prediction problem, but modeling how your body will change over time,” says Pranav Reddy, a partner at Conviction who led the investment. “We think that’s a different view of the problem.”

A retrospective study of 80,000 patients at Kaohsiung Medical University in Taiwan looked at Xlue’s ability to identify seven cancer types, including breast, colorectal and pancreatic. It found that the model flagged the risk of cancer 12.6% of the time at least 180 days before diagnosis and 19.5% of the time at least 90 days before diagnosis. It beat both a classic tree-based machine learning model and a rule-based system using clinical cancer screening guidelines.

“Xlue is fundamentally shifting the paradigm toward early prediction, prevention and intervention–with the potential to save lives, improve quality of life and reduce the global burden of disease,” KMU president Ming-Lung Yu said by email.

Finding cancers earlier is good for patients, who have a much greater likelihood of being successfully treated for their disease. It’s also good for healthcare costs. While other AI health companies talk about curing cancer or replacing doctors, Xiong says that prevention like this is “an application where we can make a positive impact very quickly.”

While Xlue started with cancer, Xiong figures the model can also be used to predict other diseases. He’s also looking at how to deploy it to help pharmaceutical companies better manage their drug development pipelines. “Clinical trials many times are driven by a hunch,” he says. “It would be helpful to have a better understanding of the risk, the cost and the chances of success.”

While these two use cases are quite different, Xiong notes that the same technology underlies both. The startup is now talking with both hospitals and drug companies about potential partnerships, he says. “Our choice is more about, is the technology ready? And what is the risk if it’s not perfect?” he says.

A housekeeping note: The InnovationRx newsletter will be on hiatus next week and return after Thanksgiving on December 2.

FDA Rejects Summit’s Highly Anticipated Lung Cancer Drug

T he FDA rejected Summit Therapeutics' highly anticipated lung cancer drug. The medication, known as ivonescimab, was developed with Chinese drugmaker Akeso. Hopes had been high for the drug since its positive results and subsequent approval in China, however the phase 3 trial underpinning its regulatory application in the U.S. was mixed. While the drug was found to slow the worsening of disease in patients who took it, it didn’t increase their lifespans. And that inability to show it could extend lung cancer patients’ lives was behind the FDA decision.

The market’s reaction was swift and harsh. Following the weekend decision, Summit stock plummeted 25% (through Tuesday’s close), while Akeso’s shares fell 21%. Bob Duggan , the billionaire co-founder and co-CEO of Summit, saw his fortune drop to $9 billion from $11.3 billion. Summit said in a statement that it would continue to pursue clinical development for the drug for multiple cancer types. “With five positive phase 3 studies between Summit and Akeso, nothing changes with respect to our ever-growing belief in the opportunity for ivonescimab and our determination to advance ivonescimab’s development,” Duggan said in the statement.

Although rare, this is not the first time that the FDA rejected a drug that was developed in China and had already received approval there. Pharma data and consulting firm Evaluate counts five other instances where the FDA said no in similar circumstances (though two of those drugs were subsequently approved and a third is now pending). In 2022, for example, the agency notably rejected a lung cancer drug developed by Eli Lilly and Innovent, a $19 billion (market cap) Chinese firm, that had completed its trials in Asia. The rejections followed a closer look at drugs from China following a 2021 piece that Richard Pazdur, then the head of the FDA’s Oncology Center of Excellence, wrote in the New England Journal of Medicine questioning data generated solely or predominantly in China when seeking approval in the U.S.

While Summit’s phase 3 trial on ivonescimab had expanded on the Chinese data, it didn’t prove strong enough. “There have been a lot of challenges replicating results in China in the U.S.,” says Cody Powers, a principal at consulting firm ZS. “The FDA flipped its position and the whole door was shut.”

The Commoditization Of AI Scribing

When AI scribing first arrived in doctors’ offices and hospitals, it was a huge deal, allowing physicians to save time on burdensome administrative work and focus on patients. No question, ambient listening has changed the way that medicine is practiced. But AI scribing is now a commodity. Just how much the market has changed recently is clear in a new report from VC firm Menlo Ventures . By their reckoning, startups garnered 67% of a then-$600 million scribing market last year. But this year, the tables have turned as incumbents, led by software giant Microsoft and electronic health records behemoth Epic, won 68% of new spending on scribes, according to the Menlo Ventures report. Whether their technology is better or just cheaper, the result is clear: The incumbents now have a 46% share of the market, which has increased to $950 million. AI-native scribes, meanwhile, have suffered a 13% loss in market share.

Isomomorphic’s $5 Billion Fundraise

Alphabet spinoff Isomorphic Labs officially closed its gargantuan $5 billion fundraise, co-led by Greenoaks and Sequoia Capital, in an oversubscribed round just six months after raising $2.1 billion. The firm is now reportedly worth $50 billion, by far the most of any AI drug developer. With the cash horde, Isomorphic is both refining its core foundational models and moving toward clinical development on its first oncology drug, president Max Jaderberg tells Forbes . While he declined to give a timeline for the cancer drug’s progress or to specify how many other therapies were in its internal pipeline, he noted that the company now employs 500 people spread between its London headquarters and outposts in Cambridge, Massachusetts, and Lausanne, Switzerland. Of the $5 billion investment, he says, “I think people recognize the big ambition and the big mission.”

Air ambulance company PHI Health and other providers are suing employers for unpaid out-of-network bills.

The American Cancer Society now recommends that most men over age 50 get screened for prostate cancer in a shift from its previous guidance.

Doctors on RFK Jr.’s influential preventive care panel have significant conflicts of interest, including owning private medical practices that could benefit.

Trump promised to care for veterans. But his dismantling of the VA left a veterans’ hospital in crisis.

Billionaire Max Hodak’s Science Corp., which makes retina chips for blindness, is now worth $3.5 billion.

Noubar Afeyan’s Flagship Pioneering committed some $132 million to build AI-driven drug companies in the U.K. in its first expansion outside of the United States.

The U.S. has lost its measles-elimination status after 26 years in a huge setback for public health.