ATHENS — Conno Christou utilized an artificial intelligence model to interpret ambiguous post-treatment PET scan results in 2026, which helped him avoid potentially unnecessary radiotherapy. This occurred after he was diagnosed with an aggressive form of non-Hodgkin's lymphoma.

Christou sought medical attention after his arm swelled following a workout. A doctor discovered two blood clots and pre-operative examinations then revealed an 11-by-11-by-8 centimeter mass behind his sternum. A biopsy confirmed the diagnosis of non-Hodgkin's lymphoma.

He was 35 years old at the time of his diagnosis and was building his second company. His lymphoma was determined to have been caused by a random genetic mutation, with no connection to lifestyle, diet, or stress. The tumor had been present for approximately three months. Christou's medical checkup in 2025 had shown normal results across all metrics.

Christou's first oncologist recommended a lighter chemotherapy regimen. However, Christou sought a second medical opinion the night before his scheduled infusion, leading to a recommendation for a more aggressive chemotherapy regimen involving continuous in-hospital infusion cycles every three weeks for six months. The lighter treatment had an estimated 60% success rate for his condition, while the aggressive regimen increased the success rate to approximately 85%. In total, Christou gathered 12 medical opinions over two days, with 11 doctors favoring the harder chemotherapy path. He subsequently underwent six months of this chemotherapy treatment.

During his treatment, Christou engaged in practices such as wearing a Whoop band and maintaining a symptom journal using voice transcription. He focused on sleep, nutrition, and psychology as key variables. Christou input blood results, scan data, wearable output, and journal entries into Claude, an AI model.

His final PET scan at the end of treatment yielded ambiguous results, prompting his oncologist to discuss a potential second line of therapy, including radiotherapy near his heart and lungs. Christou had learned that the false-positive rate on end-of-treatment PET scans for his specific type of lymphoma is around 60%. He then input the three PET scans and an MRI into Claude.

Claude indicated that the thymus gland could reactivate after chemotherapy in patients under 40 recovering from this type of lymphoma, potentially appearing as active disease on imaging. The AI model estimated the probability of thymus rebound at approximately 90% based on Christou's age and scan characteristics. Following this, Christou sought three additional medical opinions regarding the ambiguous scan. A fourth doctor subsequently confirmed the presence of thymus rebound and verified the absence of active disease, making radiotherapy unnecessary.