A specialized reasoning large language model for accelerating rare disease diagnosis: a randomized AI physician assistance trial

Source

Source name: arXiv
Source kind: paper
Authors / organization: Haichao Chen, Songchi Zhou, Zhengyun Zhao, Shikai Hu, Xianghong Jin, Hongwei Ji, Li He, Shuli Li, Yiming Qin, Xin Tan, Runfeng Shi, Yih Chung Tham, Jiaye Zhu, Ye Li, Ye Jin, Longhao Cao, Dawei Li, Honghan Wu, Hongqiu Gu, Guanqiao Li, Tudor Groza, Chunying Li, Dian Zeng, Weihong Yu, Gareth Baynam, Saumya Shekhar Jamuar, Min Shen, Shuyang Zhang, Bin Sheng, Sheng Yu, Tien Yin Wong
Venue: arXiv: cs.AI
Published: 2026-06-23
Collected: 2026-06-24
URL: http://arxiv.org/abs/2606.24510v1
PDF: https://arxiv.org/pdf/2606.24510v1
DOI: UNKNOWN

Plain-Language Brief

Researchers have developed a specialized language model called RaDaR (Rare Disease navigatoR) to assist in the diagnosis of rare diseases. The model, which has 32 billion parameters, was trained on both publicly available free-text cases and synthetic cases with reasoning-enhanced training. RaDaR outperformed other open-source models in various benchmarks and validation centers. In a retrospective cohort, the model was able to suggest a final diagnosis before clinical suspicion in 61.06% of cases, potentially reducing diagnostic time by 1.87 months. In a randomized trial, RaDaR improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared to using internet search alone.

What the Document Says

A specialized language model called RaDaR was developed for rare disease diagnosis.
RaDaR has 32 billion parameters.
The model was trained on 49,170 publicly available free-text cases and 104,666 synthetic cases with reasoning-enhanced training.
RaDaR outperformed other open-source models in various benchmarks and validation centers.
In a retrospective cohort, RaDaR suggested a final diagnosis before clinical suspicion in 61.06% of cases, potentially reducing diagnostic time by 1.87 months.
In a randomized trial, RaDaR improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared to using internet search alone.

Key Concepts Mentioned

RaDaR
rare disease diagnosis
large language model
training data
performance evaluation
retrospective cohort
randomized trial

Claims Made by the Document

RaDaR is a specialized language model for rare disease diagnosis.
RaDaR outperformed other open-source models in various benchmarks and validation centers.
In a retrospective cohort, RaDaR suggested a final diagnosis before clinical suspicion in 61.06% of cases.
In a randomized trial, RaDaR improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared to using internet search alone.

Evidence Used

The model was trained and tested on various benchmarks and validation centers.
Results from a retrospective cohort and a randomized trial were presented.
The paper provides details about the training process of RaDaR.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Rare diseases affect millions of individuals worldwide, yet timely diagnosis remains a major public health challenge due to scarcity of specialized clinical expertise. While large language models (LLMs) show promise to support rare disease diagnosis, current models are constrained by insufficient clinical deployability, limited clinically grounded evidence, and scarcity of training data. Here we present RaDaR (Rare Disease navigatoR), an open-source, compact reasoning LLM (32B parameters) for rare disease diagnosis. RaDaR was trained with 49,170 publicly available free-text cases and 104,666 synthetic cases with reasoning-enhanced training. RaDaR showed the strongest performance among evaluated open-source models, including the 671B DeepSeek-R1, across public benchmarks and four external validation centers. In a retrospective cohort, RaDaR prioritized the final diagnosis before documented clinical suspicion in 61.06 percent of cases, corresponding to a potential lead time of 1.87 months and 50.18 percent of the within-center interval. In a randomized physician-assistance trial, RaDaR assistance improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared with internet search alone. Synthetic-data ablations suggested that phenotype-anchored narratives provide useful training signal for long-tail rare diseases, with a monotonic scaling trend within the tested data range. Together, RaDaR and its development and validation framework provide a deployable rare-disease reasoning model and a reproducible development framework for diagnostic AI under data scarcity.