A 3D SAM-Based Progressive Prompting Framework for Multi-Task Segmentation of Radiotherapy-induced Normal Tissue Injuries in Limited-Data Settings

2026-04-15 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Computer Vision & Pattern Recognition · Depth: Expert, quick

Summary

Researchers have curated a new head-and-neck radiotherapy-induced normal tissue injury dataset to address the challenge of automatic segmentation of these lesions. This dataset covers three specific manifestations: osteoradionecrosis (ORN), cerebral edema (CE), and cerebral radiation necrosis (CRN). Alongside this, they propose a 3D SAM-based progressive prompting framework designed for multi-task segmentation, particularly effective in limited-data environments. The framework integrates text prompts for task adaptation, dose-guided box prompts for initial localization, and click prompts for iterative refinement. A small-target focus loss is also incorporated to enhance local prediction and boundary delineation for small and sparse lesions. Experiments on ORN, CE, and CRN demonstrate that this method achieves reliable segmentation performance across diverse injury types, outperforming existing methods.

Key takeaway

For Computer Vision Engineers developing medical image segmentation tools, this research indicates that integrating progressive prompting and a small-target focus loss within a 3D SAM framework can significantly improve accuracy for diverse and sparse radiotherapy-induced injuries. You should consider adopting this multi-prompting strategy to enhance lesion boundary delineation and overall segmentation performance in your models, especially when working with limited annotated data.

Key insights

A 3D SAM-based framework improves radiotherapy injury segmentation using progressive prompting and a small-target focus loss.

Principles

• Progressive prompting enhances segmentation accuracy.
• Small-target focus loss improves sparse lesion delineation.

Method

The framework uses text prompts for task adaptation, dose-guided box prompts for coarse localization, and click prompts for iterative refinement, combined with a small-target focus loss.

In practice

• Apply 3D SAM for medical image segmentation.
• Utilize dose data for initial lesion localization.

Topics

• Radiotherapy Injury Segmentation
• 3D SAM Framework
• Progressive Prompting
• Limited-Data Segmentation
• Osteoradionecrosis

Best for: Computer Vision Engineer, AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• A 3D SAM-Based Progressive Prompting Framework for Multi-Task Segmentation of Radiotherapy-induced Normal Tissue Injuries in Limited-Data Settings — A progressive prompting framework enhances 3D SAM for accurate, multi-task segmentation of small, sparse radiotherapy injuries with limited data.
• Parameter-Efficient Adaptation of SAM 3 for Automated ITV Generation from 4DCT Images — Parameter-efficient fine-tuning of SAM 3 with temporal coherence effectively automates ITV generation from 4DCT images using minimal data.
• Cross-Source Supervision for Bone Infection Segmentation in Dual-Modality PET-CT — Decoupled dual-source learning with PET-CT fusion improves bone infection segmentation by accommodating diverse clinical annotation philosophies.
• Evidential learning driven Breast Tumor Segmentation with Stage-divided Vision-Language Interaction — TextBCS enhances breast tumor segmentation in MRI by combining text prompts with evidential learning to address low contrast and blurred boundaries.
• Geometry-Aware Superpixel Graph Transformer with Metadata for Skin Lesion Classification — Explicitly modeling skin lesions as geometry-aware superpixel graphs with integrated metadata improves classification accuracy.
• End-user validation of BRIGHT with custom-developed graphical user interface applied to cervical cancer brachytherapy — BRIGHT with its new GUI offers an excellent, user-friendly semi-automated planning tool for cervical cancer brachytherapy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.