MedLatentDx: Latent Multi-Agent Communication for Cross-Hospital Rare-Disease Diagnosis

2026-06-11

AI Summary

This work addresses the challenge of reliable rare disease diagnosis in single hospitals, where limited case volumes hinder performance, while cross-institutional collaboration is constrained by clinical text privacy regulations and risks of sensitive information leakage in existing approaches. To overcome these limitations, the authors propose a privacy-preserving multi-agent diagnostic framework in which each hospital retains clinical records locally and transmits only compact implicit key-value (KV) blocks to a central agent for collaborative inference. The method supports both homogeneous and heterogeneous large model deployments, leveraging implicit KV cache compression, hidden-state distillation within shared backbone architectures, and cross-model-family alignment in latent space to enable efficient cooperation. Evaluated on the newly introduced CrossRare-Bench benchmark, the approach significantly improves cross-institutional rare disease diagnostic accuracy while substantially reducing clinical content reconstructability and privacy leakage risks.

Abstract

Rare diseases affect over $300$ million patients across more than $7{,}000$ conditions, yet no single hospital encounters enough cases of any one condition for reliable diagnosis. Cross-hospital collaboration could help by allowing a diagnosing institution to use distributed, case-specific diagnostic evidence, but privacy regulations restrict the transmission of identifiable clinical text across institutional boundaries. This setting raises two challenges: existing medical agent systems often rely on textual evidence exchange, while raw latent states such as hidden states and KV caches may still reveal prompt-derived clinical content. We introduce MedLatentDx, a latent multi-agent communication framework in which hospital agents keep private clinical records and retrieved cases local, and send compact latent KV blocks to a host agent for rare-disease diagnosis. MedLatentDx supports two deployment settings: same-backbone hospital agents use latent KV distillation, while hospitals with different LLM backbones use cross-family latent alignment. On CrossRare-Bench, a self-built large-scale rare-disease benchmark with hospital-level partitions, MedLatentDx improves cross-hospital diagnostic performance while reducing reconstructable clinical content relative to raw-latent communication baselines.

Problem

• rare-disease diagnosis
• cross-hospital collaboration
• privacy-preserving communication
• medical multi-agent systems
• clinical data sharing

Innovation

• latent multi-agent communication
• cross-hospital diagnosis
• rare disease
• KV distillation
• privacy-preserving AI