A combination of health wearables, biomarkers, and an artificial intelligence model successfully overturned a scheduled radiotherapy procedure for a cancer patient in 2026, revealing a benign post-chemotherapy tissue reaction instead of active disease.
As reported by mezha.
net on June 27, 2026, 35-year-old startup founder Conno Christou utilized biometric data and the Claude AI assistant to question a post-treatment PET scan that initially suggested a need for radiation near his heart and lungs.
Christou discovered that the false-positive rate for his specific lymphoma at the end of PET scans can reach approximately 60 percent, prompting him to seek further evaluation of his medical imaging.
"Happy in my misfortune; it was revealed only because I came with a different purpose," said Christou.
Christou gathered 12 medical opinions after receiving conflicting treatment strategies from two leading oncologists regarding his initial chemotherapy regimen, with 11 specialists favoring a more aggressive six-month plan.
"As founders, we steer the helm; you hear many voices. It isn’t necessary to follow the first piece of advice," he said.
Throughout his subsequent treatment, Christou tracked his health indicators using a Whoop bracelet to monitor immune system declines, documented his side effects via voice transcription, and fed all medical records into Claude.
"This has a bigger impact than anything else," Christou noted.
He maintained focus on sleep, nutrition, and psychology, never asking why the illness occurred because there was no useful answer available.
"This did not replace doctors, – but it helped me ask the right questions," he said.
The AI model processed three PET scans and an MRI to estimate a 90 percent probability of a thymic rebound phenomenon, which is a reactivation of the thymus gland in individuals under 40 after chemotherapy that mimics active disease.
"Sixty percent," Christou said, referring to the false-positive rate.
The AI finding was later confirmed by a fourth doctor, which led to the cancellation of the unnecessary radiation treatment.
Following the experience, Christou founded Keragon, an AI-supported company designed to automate administrative processes within medical practices.
"This is happening today," Christou said.
Public chatbots often make errors and have not been fully evaluated for personalized diagnostics according to Danielle Bitterman of Mass General Brigham, who emphasized that such tools cannot replace professional doctors.
