TrajOnco: a multi-agent framework for temporal reasoning over longitudinal EHR for multi-cancer early detection

TL;DR

TrajOnco is a multi-agent LLM framework that performs temporal reasoning over longitudinal EHR data for multi-cancer early detection, achieving competitive risk prediction and interpretability.

Contribution

The paper introduces TrajOnco, a novel multi-agent LLM architecture with long-term memory for scalable, interpretable multi-cancer risk prediction from EHRs.

Findings

Achieved AUROCs of 0.64-0.80 in zero-shot risk prediction across 15 cancers.

Performed comparably to supervised models in lung cancer detection.

Enabled effective temporal reasoning with smaller models like GPT-4.1-mini.

Abstract

Accurate estimation of cancer risk from longitudinal electronic health records (EHRs) could support earlier detection and improved care, but modeling such complex patient trajectories remains challenging. We present TrajOnco, a training-free, multi-agent large language model (LLM) framework designed for scalable multi-cancer early detection. Using a chain-of-agents architecture with long-term memory, TrajOnco performs temporal reasoning over sequential clinical events to generate patient-level summaries, evidence-linked rationales, and predicted risk scores. We evaluated TrajOnco on de-identified Truveta EHR data across 15 cancer types using matched case-control cohorts, predicting risk of cancer diagnosis at 1 year. In zero-shot evaluation, TrajOnco achieved AUROCs of 0.64-0.80, performing comparably to supervised machine learning in a lung cancer benchmark while demonstrating better temporal reasoning than single-agent LLMs. The multi-agent design also enabled effective temporal reasoning with smaller-capacity models such as GPT-4.1-mini. The fidelity of TrajOnco's output was validated through human evaluation. Furthermore, TrajOnco's interpretable reasoning outputs can be aggregated to reveal population-level risk patterns that align with established clinical knowledge. These findings highlight the potential of multi-agent LLMs to execute interpretable temporal reasoning over longitudinal EHRs, advancing both scalable multi-cancer early detection and clinical insight generation.