Conno Christou is a 35-year-old founder who closely monitors his health: he tracks sleep with a Whoop band and an Oura ring, has nearly 100 biomarkers checked annually, and follows longevity protocols influenced by researchers such as Peter Attia and Rhonda Patrick. His most recent annual check in 2025 was entirely green.
After a workout his arm swelled. Initially he paid it little mind; a week later a doctor found two blood clots and scheduled surgery. Preoperative tests changed the plan: an examining physician returned and reported “an 11-by-11-by-8 centimeter mass behind your sternum.” A biopsy confirmed an aggressive, fast-growing non-Hodgkin lymphoma, a rare diagnosis driven by a random genetic mutation and unrelated to lifestyle. The condition affects roughly one in 420,000 people.
Christou estimates the tumor had only existed for about three months; in three more weeks it would have been close to stage four. He calls himself “lucky in my unluckiness,” because the mass was found incidentally while addressing something else.
Conflicting expert advice and how he decided
His first oncologist, a renowned specialist, recommended the less intensive of two chemotherapy regimens and an infusion was booked for three days later. The night before he sought a second opinion. That doctor immediately recommended the more aggressive option: continuous in-hospital infusion on three-week cycles for six months. The lighter regimen carried roughly a 60% success rate for his presentation; the aggressive plan increased that to about 85%. Two top doctors, opposite recommendations.
Christou did not accept either opinion without further work. Over the next two days he gathered 12 expert opinions via his professional network and international contacts; 11 out of 12 recommended the more aggressive approach, which he then chose. For him the choice was logical rather than a display of bravery — he is data-driven and the stakes felt existential.
Data collection, wearables and journaling
Across six months of treatment he approached chemotherapy like building a company: a series of sprints within a marathon, each cycle finite and monitored. He wore his Whoop continuously, which he found remarkably predictive of days when his immune system would be most depleted. He kept a symptom journal using voice transcription, logging every shift, side effect, medication and counter-medication.
He concentrated on three variables: sleep, nutrition and, above all, psychology. He emphasized psychology’s outsized effect on outcomes and said he never asked “why me,” finding that question unhelpful.
The role of Claude in decision-making
Christou fed all of his data — blood results, scans, wearable output and journal entries — into Claude, a large language model. While experts such as Danielle Bitterman, clinical lead for data science and AI at Mass General Brigham, warn that general-purpose chatbots are frequently wrong and not comprehensively evaluated for personalized diagnoses, Christou says the AI did not replace doctors but “helped me ask the right questions.”
He argues that for a rare condition that an individual oncologist might see only once a year, access to a model that has absorbed the medical literature is not the same as a Google search.
End-of-treatment imaging, AI insight, and avoided radiotherapy
At the end of treatment a PET scan—used to detect active disease—returned an ambiguous result. His oncologist began discussing a second-line therapy, possibly radiation near his heart and lungs. Christou researched and read that for this lymphoma the false-positive rate on end-of-treatment PET scans is around 60%.
He uploaded all three PET scans and his MRI into Claude. The model flagged a known but easily overlooked phenomenon: thymus rebound, in which the thymus gland can reactivate after chemotherapy in patients under 40 and appear on imaging as active disease. Taking into account his age and the specific scan features, Claude put the probability of that explanation at roughly 90%.
Christou sought three more opinions; a fourth doctor confirmed thymus rebound. There was no active disease and radiation was unnecessary. He was declared clear.
Reflections on the healthcare system and personal changes
The experience gave Christou insight into systemic issues: nurses and doctors weighed down by non-care tasks, and a tendency for undifferentiated protocols — he received the same chemotherapy protocol as an 80-year-old woman, with side effects managed through additional drugs that themselves caused further problems. He expects that future observers will look back on this era of treatment with discomfort.
The episode also reshaped his relationship to time and work. He had founded Keragon — an AI-powered platform to automate administrative operations for medical practices — before his diagnosis. After treatment he now takes Sundays off and tries to be more present with friends, his dog, and in conversations he once might have seen as distractions. He recalls advice from a VC friend: “Be happy now,” a message he says he replayed during treatment and now values.
Christou offers to speak with others facing similar situations to compare notes and experiences. His closing observation: the kinds of AI assistance patients are using are not a future promise but a present reality. “It’s not happening in 10 years,” he said. “It’s happening today.”
Numbers and timeline — brief
- Age: 35
- Last checkup: 2025 (green across the board)
- Tumor size: 11×11×8 cm
- Approximate incidence of his lymphoma: 1 in 420,000
- Reported success rates: lighter treatment ~60%, aggressive treatment ~85%
- Expert opinions gathered: 12 (11–1 in favor of aggressive treatment)
- False-positive rate on end-of-treatment PET scans for this lymphoma: ~60%
- Claude’s estimated probability for thymus rebound explanation: ~90%
