Contrast-Invariant Self-supervised Segmentation for Quantitative Placental MRI

2025-05-30

AI Summary

This work addresses three key challenges in T2*-weighted multi-echo placental MRI segmentation: weak inter-echo boundary contrast, absence of full-time-point ground-truth annotations, and motion artifact interference. To this end, we propose the first contrast-invariant representation learning framework specifically designed for multi-echo T2* data. Our method integrates masked autoencoder pretraining, pseudo-label-based unsupervised domain adaptation, global-local feature co-alignment, and a semantic matching loss—explicitly enforcing intra-subject representation consistency across echoes. Evaluated on real clinical data, our approach significantly outperforms single-echo baselines and naive multi-echo fusion strategies, demonstrating strong cross-echo generalization. It establishes a robust, reproducible segmentation foundation for quantitative placental T2* mapping, enabling more reliable downstream biomarker analysis.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Segmentation
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate placental segmentation is essential for quantitative analysis of the placenta. However, this task is particularly challenging in T2*-weighted placental imaging due to: (1) weak and inconsistent boundary contrast across individual echoes; (2) the absence of manual ground truth annotations for all echo times; and (3) motion artifacts across echoes caused by fetal and maternal movement. In this work, we propose a contrast-augmented segmentation framework that leverages complementary information across multi-echo T2*-weighted MRI to learn robust, contrast-invariant representations. Our method integrates: (i) masked autoencoding (MAE) for self-supervised pretraining on unlabeled multi-echo slices; (ii) masked pseudo-labeling (MPL) for unsupervised domain adaptation across echo times; and (iii) global-local collaboration to align fine-grained features with global anatomical context. We further introduce a semantic matching loss to encourage representation consistency across echoes of the same subject. Experiments on a clinical multi-echo placental MRI dataset demonstrate that our approach generalizes effectively across echo times and outperforms both single-echo and naive fusion baselines. To our knowledge, this is the first work to systematically exploit multi-echo T2*-weighted MRI for placental segmentation.

Problem

• Weak boundary contrast in T2*-weighted placental MRI
• Lack of manual annotations for all echo times
• Motion artifacts from fetal and maternal movement

Innovation

• Contrast-augmented segmentation for multi-echo MRI
• Masked autoencoding and pseudo-labeling techniques
• Global-local collaboration for feature alignment