AI Beats Doctors at Writing Cancer Reports, Study Finds

Artificial intelligence just proved it can outperform doctors at one critical task: writing detailed cancer diagnosis reports.

A team of seven researchers at Northwestern's Feinberg School of Medicine tested six AI language models against human-written pathology summaries. Every single AI model scored higher on objective quality measures.

The models tested included Llama 3.0, Llama 3.1, Llama 3.2, Mistral, Gemma, and DeepSeek-R1. Their reports were published in early April 2026.

Dr. Mohamed Abazeed, one of the study's co-authors, said the research targets a real problem in healthcare. Physicians are drowning in information while patient consultation windows keep shrinking.

"Human attention spans are becoming increasingly limited, and more people are overwhelmed with information and tend to disregard some details," Abazeed explained.

The challenge is especially acute in oncology. Cancer patients often have years of treatment history that doctors must review in just 30 minutes before meeting with them.

That's where AI shines. Yirong Liu, the study's lead author and a fifth-year resident in radiation oncology, said the models caught details that even experienced clinicians might miss.

"The model can pick up the change of genetic alteration along the treatment course, which is quite amazing," Liu said. "It would take us, as clinicians, a lot of effort to pick up those."

Why This Inspires

This breakthrough isn't about replacing doctors. It's about giving them superpowers.

Troy Teo, a radiation oncology instructor who co-authored the study, sees AI as a tool that frees physicians to do what they do best: connect with patients. The technology handles the paperwork mountain while doctors focus on healing relationships.

Liu emphasized that no algorithm can replace the human touch in medicine. Patients need to feel comfortable and trust the person guiding their treatment, something only face-to-face human interaction can provide.

The research team is now tackling the next challenge: how to safely integrate these AI tools into real clinical settings. They're carefully considering legal questions like who takes responsibility when an AI makes a mistake.

Teo acknowledges that skepticism about AI in healthcare is healthy and necessary. That's exactly why rigorous testing matters so much.

The beauty of this approach is that it doesn't eliminate the doctor. It eliminates the exhaustion and information overload that prevent doctors from being fully present with their patients when it matters most.