Topology-Informed Prompt-Conditioned Universal Segmentation of Uterine Structures from Ultrasound and MRI

2026-10-05

AI Summary

This study addresses the challenge of multi-modal uterine segmentation across ultrasound and MRI, where significant inter-domain discrepancies and inconsistent label spaces hinder the development of unified models. To this end, this work proposes TPUS, a universal segmentation framework that pioneers the integration of topological constraints with prompt conditioning. Specifically, TPUS employs a graph-structured multi-dataset backbone network, a dynamic convolution adaptation module incorporating task-aware prompts, and a topology-aware loss function to achieve precise cross-modal multi-structure segmentation. The proposed approach effectively mitigates negative transfer among heterogeneous tasks. Experimental results demonstrate that TPUS achieves Dice coefficients of 0.898 and 0.693 on ultrasound and MRI test sets, respectively, significantly outperforming existing baseline methods.

Abstract

Multi-structure segmentation of the uterus is important for computer-assisted screening, diagnosis, and treatment planning of uterine diseases, where ultrasound and MRI provide complementary clinical information. However, developing a unified model across these modalities is challenging due to their substantially different image appearances, anatomical contexts, spatial resolutions, and label spaces. Moreover, existing datasets often define different segmentation targets, making joint learning challenging and potentially leading to negative transfer across heterogeneous tasks. To this end, we propose a Topology-informed Prompt-conditioned Universal Segmentation (TPUS) framework for segmenting multiple uterine structures across ultrasound and MRI. TPUS introduces a graph-based multi-dataset backbone comprising modality-specific stems and a modality-shared graph-based encoder-decoder to support modality-sensitive input adaptation, structural feature reasoning, and joint representation learning across heterogeneous uterine segmentation tasks. In addition, TPUS uses task-aware class prompts to condition the segmentation process for different datasets and label spaces, a dynamic convolutional adaptation module to generate task-specific output responses, and a topology-informed loss to encourage anatomically consistent predictions. Experiments on a uterine ultrasound dataset and a T2-weighted uterine myoma MRI dataset demonstrate that TPUS achieves Dice scores of 0.898 and 0.693 on the two held-out test sets, respectively, outperforming several generic and universal segmentation baselines. Source code can be accessed at https://github.com/YonghengSun1997/TPUS.

Problem

• multi-structure segmentation
• cross-modality
• universal model
• negative transfer
• uterine imaging

Innovation

• Universal Segmentation
• Topology-Informed Loss
• Prompt-Conditioned
• Graph-Based Encoder-Decoder
• Dynamic Convolutional Adaptation