Learning from Anatomy: Supervised Anatomical Pretraining (SAP) for Improved Metastatic Bone Disease Segmentation in Whole-Body MRI

2025-06-24

AI Summary

Segmenting metastatic bone disease (MBD) in whole-body MRI is challenging due to highly variable lesion morphology, ill-defined boundaries, and severe class imbalance. To address these issues—particularly in low-data regimes—we propose supervised anatomical pretraining: leveraging high-quality skeletal annotations from healthy subjects to encode anatomical priors, training an MRI skeletal segmentation model to learn robust bone morphology representations, and thereby injecting domain-specific inductive bias. Evaluated on 44 patient scans, our method achieves Dice score 0.64, surface Dice 0.76, and F2 score 0.44—outperforming both conventional supervised baselines and self-supervised approaches. It attains 100% sensitivity (28/32) for lesions >1 mL. This work introduces the first supervised anatomical pretraining framework for MBD segmentation, establishing a novel paradigm for few-shot medical image segmentation.

Technology Category

Computer Vision: SegmentationMachine Learning: Multi-instance/Multi-view LearningData Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

Web Mining and Content Analysis: Large pretrained models with web dataSearch and Retrieval-Augmented AI: Web evaluation methodologies and metricsUser Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

The segmentation of metastatic bone disease (MBD) in whole-body MRI (WB-MRI) is a challenging problem. Due to varying appearances and anatomical locations of lesions, ambiguous boundaries, and severe class imbalance, obtaining reliable segmentations requires large, well-annotated datasets capturing lesion variability. Generating such datasets requires substantial time and expertise, and is prone to error. While self-supervised learning (SSL) can leverage large unlabeled datasets, learned generic representations often fail to capture the nuanced features needed for accurate lesion detection. In this work, we propose a Supervised Anatomical Pretraining (SAP) method that learns from a limited dataset of anatomical labels. First, an MRI-based skeletal segmentation model is developed and trained on WB-MRI scans from healthy individuals for high-quality skeletal delineation. Then, we compare its downstream efficacy in segmenting MBD on a cohort of 44 patients with metastatic prostate cancer, against both a baseline random initialization and a state-of-the-art SSL method. SAP significantly outperforms both the baseline and SSL-pretrained models, achieving a normalized surface Dice of 0.76 and a Dice coefficient of 0.64. The method achieved a lesion detection F2 score of 0.44, improving on 0.24 (baseline) and 0.31 (SSL). When considering only clinically relevant lesions larger than 1~ml, SAP achieves a detection sensitivity of 100% in 28 out of 32 patients. Learning bone morphology from anatomy yields an effective and domain-relevant inductive bias that can be leveraged for the downstream segmentation task of bone lesions. All code and models are made publicly available.

Problem

• Segmenting metastatic bone disease in whole-body MRI accurately
• Overcoming class imbalance and ambiguous lesion boundaries in MBD
• Reducing reliance on large annotated datasets for lesion detection

Innovation

• Supervised Anatomical Pretraining for bone segmentation
• MRI-based skeletal model from healthy individuals
• Outperforms baseline and SSL in lesion detection