TechCrunch reported that the founder of the AI Medical Automation Platform, Keragon, Conno Christou, after identifying a rare and fast-paced non-Hodgkin lymphoma, began to use Claude to organize laboratory, video, wearable equipment data and symptoms records, to complement the comparative treatment programme and to prepare questions for the doctor.
Pre-operative examination found tumors.
Christou keeps a long record of sleep and physical signs, as well as regular blood tests. After one exercise, he was treated for a swelling of his arm, which was intended for the treatment of a vein embolism, but during a pre-operative examination he found an extensive swelling in the back of his chest. The biopsy later confirmed that he had a rare lymphoma.
The tumours are reported to be very short-lived, but progress is rapid. After the diagnosis, instead of accepting a single recommendation, he contacted several blood and oncologists within two days, collecting a total of 12 opinions, eventually choosing a more radical treatment path and completing six periodic chemotherapy sessions.
Continuous input of data during treatment
In the course of his treatment, he had been wearing Whoop equipment to observe changes in his physical state and to record daily symptoms, side effects and medications by voice. He then handed over the blood results, scanned data, the output of wearable equipment and personal records to Claude for processing.
Christou stated that AI did not have a substitute doctor, but helped him to compile his medical records more quickly and to ask more specific questions. As his disease was rare, he believed that access to models that absorbed a great deal of medical literature was not the same as search engine search.
Image error risk verified again.
At the end of the session, his one PET scan showed a blurry signal, and the attending physician began to discuss second-line treatment and even consider treatment near the CPR area. Christou then continued to review the information and entered three PET scans and one MRI results into Claude.
The model suggests that a patient under 40 years of age who has undergone this type of lymphoma chemotherapy may be re-activated and display a similar signal on the image. Christou consulted three more doctors and finally confirmed that it was a repulsion of the chest, not that the tumor was still active, and that no further treatment was required.
It is also mentioned that the medical community remains cautious about the use of universal chat robots for health advice. Mass General Brigham Data Science and AI Clinical Head Danielle Bitterman had previously indicated that such models were often wrong and had not been sufficiently validated for individualized diagnosis. Christou also acknowledges that AI's role is closer to information collation and supporting questions rather than replacing clinical judgements.