Region-Grounded Report Generation for 3D Medical Imaging: A Fine-Grained Dataset and Graph-Enhanced Framework

2026-04-20

AI Summary

This study addresses the challenges of poor clinical interpretability and insufficient lesion-level analysis in automated report generation for 3D medical imaging (e.g., PET/CT), which stem from high-dimensional data and scarce low-resource language annotations. To this end, the authors construct VietPET-RoI, the first fine-grained 3D region-of-interest (RoI) annotated dataset for a low-resource language, and propose the HiRRA framework. HiRRA leverages graph neural networks to model semantic relationships among lesion regions, emulating radiologists’ RoI-based diagnostic reasoning. The work further introduces novel large language model–driven clinical evaluation metrics—RoI Coverage and RoI Quality Index. Experimental results demonstrate that the proposed method outperforms state-of-the-art models by 19.7% on BLEU and 4.7% on ROUGE-L, while achieving a 45.8% improvement on clinical metrics, substantially enhancing report reliability and mitigating hallucination.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Language and Vision

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Automated medical report generation for 3D PET/CT imaging is fundamentally challenged by the high-dimensional nature of volumetric data and a critical scarcity of annotated datasets, particularly for low-resource languages. Current black-box methods map whole volumes to reports, ignoring the clinical workflow of analyzing localized Regions of Interest (RoIs) to derive diagnostic conclusions. In this paper, we bridge this gap by introducing VietPET-RoI, the first large-scale 3D PET/CT dataset with fine-grained RoI annotation for a low-resource language, comprising 600 PET/CT samples and 1,960 manually annotated RoIs, paired with corresponding clinical reports. Furthermore, to demonstrate the utility of this dataset, we propose HiRRA, a novel framework that mimics the professional radiologist diagnostic workflow by employing graph-based relational modules to capture dependencies between RoI attributes. This approach shifts from global pattern matching toward localized clinical findings. Additionally, we introduce new clinical evaluation metrics, namely RoI Coverage and RoI Quality Index, that measure both RoI localization accuracy and attribute description fidelity using LLM-based extraction. Extensive evaluation demonstrates that our framework achieves SOTA performance, surpassing existing models by 19.7% in BLEU and 4.7% in ROUGE-L, while achieving a remarkable 45.8% improvement in clinical metrics, indicating enhanced clinical reliability and reduced hallucination. Our code and dataset are available on GitHub.

Problem

• 3D medical imaging
• report generation
• Region of Interest (RoI)
• low-resource language
• fine-grained annotation

Innovation

• Region-of-Interest (RoI) grounding
• graph-enhanced framework
• fine-grained 3D medical dataset
• clinical report generation
• low-resource language