KneePreM: Towards 3D Knee MRI Foundation Models via Large-Scale Unlabeled Pretraining and Label-Efficient Fine-Tuning

2026-09-25

AI Summary

This study addresses the underutilization of massive unlabeled knee MRI data and the high annotation costs for downstream tasks by proposing KneePreM, the first knee-specific 3D self-supervised foundation model. Methodologically, it employs a 3D U-Net-based masked autoencoder for large-scale unsupervised pre-training, integrated with parameter-efficient fine-tuning strategies and an MCP deployment interface to enable end-to-end applications. Experimental results demonstrate that the model exhibits significantly enhanced transferability in few-shot scenarios, surpassing fully supervised baseline performance using only limited annotations. It substantially improves classification ROC AUC and segmentation Dice scores, thereby providing an efficient and generalizable paradigm for medical image analysis.

Abstract

Background: Large volumes of unlabeled knee MRI scans are available across repositories but remain insufficiently leveraged. We developed KneePreM, a knee-specific 3D self-supervised model, and evaluated transfer and label efficiency for classification and segmentation. Methods: A 3D U-Net masked autoencoder was pretrained on 19,011 unlabeled Osteoarthritis Initiative (OAI) MRI series from 4,791 participants. Downstream fine-tuning used full and reduced training sets for fastMRI+ two-label classification (1,172 examinations), Arthroscopic Partial Meniscectomy (APM) eight-target classification (1,716 examinations), SKM-TEA segmentation (155 examinations), and APM segmentation (25 examinations). Baselines were random initialization and SuPreM. Deployment workflow was implemented with a Model Context Protocol interface. Evaluation metrics included balanced accuracy, F1 score, ROC AUC, PR AUC, and Dice score. Statistical analysis used bootstrap confidence intervals and paired bootstrap tests for classification and Wilcoxon signed-rank tests for segmentation. Results: KneePreM achieved higher full-data macro ROC AUC than both baselines for fastMRI+ and APM (all p < .001). For fastMRI+ classification, KneePreM achieved a ROC AUC of 0.722 using 50% of the training data, exceeding both full-data baselines. In APM classification, KneePreM reached a ROC AUC of 0.740 with 70% of the data, matching the full-data random baseline and outperforming SuPreM. For SKM-TEA segmentation, its 70%-data Dice of 0.838 exceeded the full-data random baseline (0.835) and both same-budget comparators. In APM segmentation, its 75%-data Dice of 0.746 exceeded the full-data random baseline (0.731) and both same-budget comparators. Conclusion: KneePreM improves transfer performance and label efficiency across knee MRI classification and segmentation tasks, particularly when labeled training data are limited.

Problem

• Knee MRI
• Unlabeled data
• Label efficiency
• Self-supervised learning
• Foundation model

Innovation

• Self-supervised learning
• Masked autoencoder
• Foundation model
• Label efficiency
• 3D knee MRI