CrossSpine：具有解剖学先验知识的多尺度跨序列注意力机制用于自动 Pfirrmann 分级

中文摘要 AI 辅助

腰椎间盘退变的自动分级对于客观量化与腰痛相关的结构变化至关重要。鉴于基线模型在我们的数据上表现不佳，我们提出了一个旨在克服这些限制的框架。首先，我们提出了跨序列注意力脊柱（CrossSpine）框架，这是一种新颖的架构，它采用跨序列注意力机制在多个空间尺度上自适应融合来自不同 MRI 序列的特征。其次，我们贡献了一个精心策划的数据集用于自动 Pfirrmann 分级。最后，我们引入了一种椎间盘感知分类技术，该技术整合了解剖学椎间盘水平信息，使模型能够学习特定水平的退变先验知识。我们的实验证明了这种方法的优越性：与基线相比，CrossSpine 在 Macro F1 分数上实现了超过 125%的相对提升，同时将 Mean AUPRC 提高了 99%，将 Mean AUROC 提高了 36%。

英文摘要

Automated grading of Lumbar Disc Degeneration is essential for the objective quantification of structural changes associated with low back pain. Observing that baseline models underperformed on our data, we propose a framework designed to overcome these limitations. First, we present the Cross-sequence Attention Spine (CrossSpine) framework, a novel architecture that employs a cross-sequence attention mechanism to adaptively fuse features from different MRI sequences at multiple spa- tial scales. Second, we contribute a meticulously curated dataset aimed at automated Pfirrmann grading. Finally, we introduce an IVD-aware classification technique that integrates anatomical disc-level information, enabling the model to learn level-specific degeneration priors. Our experi- ments demonstrate the superiority of this approach: CrossSpine achieved a relative improvement exceeding 125% in the Macro F1 score, while boosting the Mean AUPRC by 99% and the Mean AUROC by 36% com- pared to the baseline.

发表机构

• Business AI Lab, College of Technology, National Economics University（商业人工智能实验室，技术学院，越南国民经济大学）
• A2I Lab, Phenikaa School of Computing, Phenikaa University（A2I实验室，培美卡计算机学院，培美卡大学）
• Medical Imaging & Radiological Technology Department, Faculty of Medical Technology, Phenikaa School of Medicine & Pharmacy, Phenikaa University（医学影像与放射技术系，医学技术学院，培美卡医学院与药学院，培美卡大学）
• Radiology & Functional Exploration Center, Phenikaa University Hospital（放射学与功能探索中心，培美卡大学医院）
• Deparment of Pediatrics, Hospital of University Medicine and Pharmacy, Vietnam National University, Hanoi（儿科系，河内越南国立大学医学与药学院医院）

机构由 AI 辅助整理，请以论文原文为准。