MeniMV: A Multi-view Benchmark for Meniscus Injury Severity Grading

2025-12-20

AI Summary

Current MRI-based automated analysis methods struggle with fine-grained localization and severity grading of anterior/posterior horn meniscal tears, typically relying on case-level labels or binary classification—thus lacking clinically essential anatomical localization and quantitative four-level (Grade 0–3) assessment. To address this, we introduce the first multi-view MRI benchmark dataset specifically designed for meniscal horn tear analysis, comprising 3,000 multi-center cases and 6,000 registered sagittal–coronal image pairs, with a novel dual-view collaborative annotation protocol. We propose a CNN-Transformer hybrid multi-view fusion model incorporating cross-plane feature alignment and dual-branch joint training. Extensive experiments establish multiple strong baselines, identify key challenges in ordinal grading, and deliver—by far—the largest, clinically interpretable, and fully reproducible ground-truth dataset and evaluation framework for meniscal pathology grading.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Precise grading of meniscal horn tears is critical in knee injury diagnosis but remains underexplored in automated MRI analysis. Existing methods often rely on coarse study-level labels or binary classification, lacking localization and severity information. In this paper, we introduce MeniMV, a multi-view benchmark dataset specifically designed for horn-specific meniscus injury grading. MeniMV comprises 3,000 annotated knee MRI exams from 750 patients across three medical centers, providing 6,000 co-registered sagittal and coronal images. Each exam is meticulously annotated with four-tier (grade 0-3) severity labels for both anterior and posterior meniscal horns, verified by chief orthopedic physicians. Notably, MeniMV offers more than double the pathology-labeled data volume of prior datasets while uniquely capturing the dual-view diagnostic context essential in clinical practice. To demonstrate the utility of MeniMV, we benchmark multiple state-of-the-art CNN and Transformer-based models. Our extensive experiments establish strong baselines and highlight challenges in severity grading, providing a valuable foundation for future research in automated musculoskeletal imaging.

Problem

• Automated MRI grading of meniscus horn tear severity
• Addressing lack of localized multi-view injury annotations
• Providing benchmark for precise clinical diagnostic models

Innovation

• Multi-view benchmark dataset for meniscus grading
• Four-tier severity labels for anterior and posterior horns
• Benchmarking CNN and Transformer models on MeniMV