What type of study is this?

September 5, 2025JMIR Medical InformaticsOpen Access

Development of a Large-Scale Dataset of Chest Computed Tomography Reports in Japanese and a High-Performance Finding Classification Model: Dataset Development and Validation Study

View Full Paper

Ask AI

Bookmark

Authors

Dataset development enhances natural language processing in radiology, improving language model performance.

A specialized language model, CT-BERT-JPN, significantly improved classification in Japanese radiology reports.
Validation results showed statistically significant enhancements in translation quality, scoring P < .001 across dimensions.
Machine translation combined with expert radiologist revisions ensured high-quality data and model accuracy.
The study addresses the gap in Japanese medical datasets, providing essential resources for future AI research.

Yamagishi et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark