Abstract:Brain tumor segmentation is crucial for accurate diagnosisand treatment planning, but the small size and irregular shapeof tumors pose significant challenges. Existing methods of-ten fail to effectively incorporate medical domain knowledgesuch as tumor grade, which correlates with tumor aggres-siveness and morphology, providing critical insights for moreaccurate detection of tumor subregions during <a class="link-external link-http" href="http://segmentation.We" rel="external noopener nofollow">this http URL</a> propose an Automated and Editable Prompt Learning(AEPL) framework that integrates tumor grade into the seg-mentation process by combining multi-task learning andprompt learning with automatic and editable prompt gen-eration. Specifically, AEPL employs an encoder to extractimage features for both tumor-grade prediction and segmen-tation mask generation. The predicted tumor grades serveas auto-generated prompts, guiding the decoder to produceprecise segmentation masks. This eliminates the need formanual prompts while allowing clinicians to manually editthe auto-generated prompts to fine-tune the segmentation,enhancing both flexibility and precision. The proposed AEPLachieves state-of-the-art performance on the BraTS 2018dataset, demonstrating its effectiveness and clinical <a class="link-external link-http" href="http://potential.The" rel="external noopener nofollow">this http URL</a> source code can be accessed online.

What problem does this paper attempt to address?

This paper attempts to solve several key problems in brain tumor segmentation: 1. **Challenges of small size and irregular shape**: Brain tumors usually have a small size and an irregular shape, which makes automated tumor segmentation very difficult. Existing methods often perform poorly when dealing with these complexities, resulting in inaccurate segmentation results. 2. **Lack of integration of medical domain knowledge**: Existing segmentation methods usually fail to effectively integrate professional knowledge in the medical field, such as tumor grade. And the tumor grade is closely related to the invasiveness and morphology of the tumor and is crucial for more accurate detection of tumor sub - regions. The lack of this domain - specific knowledge will lead to inaccurate demarcation of tumor boundaries, especially in complex cases. 3. **Requirement for manual prompts**: Existing prompt - based learning methods usually require human intervention to provide input prompts, which may be limited by time and expertise in the clinical environment, thus limiting their practical applications. To solve these problems, the authors propose a framework named AEPL (Automated and Editable Prompt Learning), which achieves brain tumor segmentation by combining multi - task learning and prompt learning. Specifically, the main contributions of AEPL include: 1. **Automated prompt learning strategy**: By seamlessly integrating tumor grade prediction into the segmentation process, the need for manual prompts is eliminated. 2. **Editable prompt paradigm**: A new editable prompt system is designed, allowing clinicians to interactively modify the generated prompts, thereby improving the accuracy of the segmentation results. 3. **Superior performance**: Extensive experiments on the BraTS 2018 dataset show that AEPL outperforms several state - of - the - art methods, demonstrating its effectiveness and clinical potential. Through these improvements, AEPL not only improves the accuracy of brain tumor segmentation but also enhances the flexibility and clinical applicability of the model.

AEPL: Automated and Editable Prompt Learning for Brain Tumor Segmentation

DeepRecS: From RECIST Diameters to Precise Liver Tumor Segmentation

Unsupervised Brain Tumor Segmentation with Image-based Prompts

A continuous learning approach to brain tumor segmentation: integrating multi-scale spatial distillation and pseudo-labeling strategies

PE-MED: Prompt Enhancement for Interactive Medical Image Segmentation

Automated post-operative brain tumour segmentation: A deep learning model based on transfer learning from pre-operative images

PFPs: Prompt-guided Flexible Pathological Segmentation for Diverse Potential Outcomes Using Large Vision and Language Models

Medical Visual Prompting (MVP): A Unified Framework for Versatile and High-Quality Medical Image Segmentation

Improving Brain Tumor Segmentation with Anatomical Prior-Informed Pre-Training.

A Deep-Learning Model with Learnable Group Convolution and Deep Supervision for Brain Tumor Segmentation

Brain Tumor Segmentation Based on Deep Learning, Attention Mechanisms, and Energy-Based Uncertainty Prediction

Knowledge-Guided Prompt Learning for Lifespan Brain MR Image Segmentation

A Prior Knowledge Based Tumor and Tumoral Subregion Segmentation Tool for Pediatric Brain Tumors

Attention-Enhanced Hybrid Feature Aggregation Network for 3D Brain Tumor Segmentation

DPAFNet: A Residual Dual-Path Attention-Fusion Convolutional Neural Network for Multimodal Brain Tumor Segmentation

MPEDA-Net: A lightweight brain tumor segmentation network using multi-perspective extraction and dense attention

Brain Tumor Segmentation via Multi-Modalities Interactive Feature Learning

Point-supervised Brain Tumor Segmentation with Box-prompted Medical Segment Anything Model

Interactive Image Selection and Training for Brain Tumor Segmentation Network

CAT: Coordinating Anatomical-Textual Prompts for Multi-Organ and Tumor Segmentation

A deep learning-based framework for segmenting invisible clinical target volumes with estimated uncertainties for post-operative prostate cancer radiotherapy