An AI managed to synthesize decades of messy medical records for cancer patients and reached the same conclusions as a panel of expert doctors.
April 29, 2026
Original Paper
Agentic clinical reasoning over longitudinal myeloma records: a retrospective evaluation against expert consensus
arXiv · 2604.24473
The Takeaway
Cancer patients often have thousands of pages of history spread across different hospitals and years of treatment. Human doctors struggle to process all this fragmented data, leading to delays and errors in care. This new AI system achieved nearly 80% agreement with top oncologists when analyzing longitudinal records for multiple myeloma. It performed significantly better than previous AI models because it can reason through the timeline of a patient's life. This means that in the near future, AI could act as a super assistant that ensures no critical detail of a patient's history is ever missed. It could make the management of complex chronic diseases much safer and more efficient.
From the abstract
Multiple myeloma is managed through sequential lines of therapy over years to decades, with each decision depending on cumulative disease history distributed across dozens to hundreds of heterogeneous clinical documents. Whether LLM-based systems can synthesise this evidence at a level approaching expert agreement has not been established. A retrospective evaluation was conducted on longitudinal clinical records of 811 myeloma patients treated at a tertiary centre (2001-2026), covering 44,962 do