Prion diseases are rare, rapidly progressive, and fatal neurodegenerative disorders that remain difficult to diagnose, particularly in their early stages because of nonspecific clinical presentations. However, to our knowledge, there is no publicly available prion-disease-focused dataset designed to capture a broad range of clinically relevant entities from the biomedical literature. We introduce PrionNER, a manually annotated named entity recognition dataset for prion disease clinical information in PubMed abstracts. The current release comprises 317 abstracts, 2,943 sentences, and 6,955 text-bound entity annotations spanning 15 coarse-grained and 31 fine-grained clinically oriented entity types covering diseases, symptoms, diagnostics, findings, anatomy, treatments, and temporal and statistical evidence. Inter-annotator agreement reaches 81.78 exact-match F1, indicating strong annotation consistency. We benchmark supervised BERT baselines, W2NER, and zero-shot extractors on PrionNER. W2NER is the strongest supervised model, and Gemma-4-31B is the strongest zero-shot model, but the benchmark remains challenging, especially for structurally complex mentions and fine-grained clinically adjacent label distinctions. PrionNER provides a clinically grounded benchmark for prion-disease information extraction and supports research on rare-disease biomedical NLP under low-resource, fine-grained, and non-flat extraction conditions. The dataset, annotation guidelines, and evaluation scripts are available at https://github.com/daotuanan/PrionNER/

Named Entity Recognition (NER) is crucial for extracting domain-specific entities from text, particularly in biomedical and chemical fields. Developing high-quality NER models in specialized domains is challenging due to the limited availability of annotated data, with manual annotation being a key method of data construction. However, manual annotation is time-consuming and requires domain expertise, making it difficult in specialized domains. Traditional data augmentation (DA) techniques also rely on annotated data to some extent, further limiting their effectiveness. In this paper, we propose a novel approach to synthetic data generation for NER using large language models (LLMs) to generate sentences based solely on a set of example entities. This method simplifies the augmentation process and is effective even with a limited set of entities.We evaluate our approach using BERT-based models on the BC4CHEMD, BC5CDR, and TDMSci datasets, demonstrating that synthetic data significantly improves model performance and robustness, particularly in low-resource settings. This work provides a scalable solution for enhancing NER in specialized domains, overcoming the limitations of manual annotation and traditional augmentation methods.