基于人工智能的多模态MRI在脊髓型颈椎病诊疗决策中的研究进展OA
Research progress on AI-based multimodal MRI in the diagnosis and treatment decision-making of cervical spondylotic myelopathy
以深度学习(deep learning,DL)为首的人工智能(artificial intelligence,AI)技术,由于在复杂图像特征提取及非线性数据分析上表现出显著优势,已成为神经影像领域的研究热点.在脊髓型颈椎病(cervical spondylotic myelopathy,CSM)诊疗中,AI不再局限于自动化影像分割,而是贯穿量化评估、多模态特征分析与功能预后预测等关键环节.其既能实现病灶的早期精准识别与客观分级,弥补传统MRI人工视觉评估主观性强的缺陷,也有助于预测患者术后神经功能的恢复趋势,从而辅助临床医生制定个体化的手术方案与康复策略.尽管已有部分研究探讨了AI在颈椎疾病中的应用,但现有综述多局限于单一常规磁共振成像(magnetic resonance imaging,MRI)序列的形态学分析或单一诊断环节,缺乏对多模态MRI与AI深度融合在CSM中应用价值的系统总结,如扩散张量成像、扩散基谱成像、静息态功能成像等,尤其对AI如何揭示脊髓微观病理演变及"脑-脊髓轴"功能重塑的关注严重不足.为填补这一空白,本文对AI结合多模态MRI技术在CSM早期筛查、辅助诊断、预后预测方面的近期研究成果进行综述,旨在从病理生理学表征和全链条诊疗决策的视角探讨该领域面临的挑战与未来前景,以期为基于AI的多模态MRI技术在CSM精准诊疗中的临床转化实践提供新思路与学术借鉴.
Artificial intelligence(AI)technologies,spearheaded by deep learning(DL),have emerged as a prominent research focus in neuroimaging due to their remarkable capabilities in extracting complex image features and analyzing nonlinear data.In the diagnosis and management of cervical spondylotic myelopathy(CSM),AI is no longer confined to automated image segmentation,but rather encompasses key stages including quantitative assessment,multimodal feature analysis,and functional prognosis prediction.It facilitates not only early and precise lesion identification with objective grading,overcoming the lack of objectivity in conventional MRI manual visual evaluation,but also assists in forecasting postoperative neurological recovery trends,thereby supporting clinicians in devising individualized surgical plans and rehabilitation strategies.Although some studies have explored the application of AI in cervical spine diseases,existing reviews are mostly limited to morphological analysis of a single conventional magnetic resonance imaging(MRI)sequence or a single diagnostic stage,lacking a systematic summary regarding the application value of the deep integration of multimodal MRI and AI in CSM,such as diffusion tensor imaging,diffusion basis spectrum imaging,and resting-state functional MRI.In particular,insufficient attention has been paid to how AI reveals the micro-pathological evolution of the spinal cord and the functional remodeling of the"brain-spinal cord axis".To fill this gap,this review synthesizes recent advances in AI-integrated multimodal MRI for early screening,auxiliary diagnosis,and prognosis prediction in CSM,aiming to explore current challenges and future prospects within this domain from the perspectives of pathophysiological representation and full-chain diagnostic and treatment decision-making,with the hope of providing new insights and academic references for the clinical translation practice of AI-based multimodal MRI technology in the precision diagnosis and treatment of CSM.
戴大城;谢芳芳;崔家禾;汪思宇;蔡君豪;姚斐;王桂茂
上海中医药大学附属市中医医院推拿科,上海 200071上海中医药大学针灸推拿学院,上海 201203上海中医药大学附属市中医医院推拿科,上海 200071上海中医药大学附属市中医医院推拿科,上海 200071上海中医药大学附属市中医医院推拿科,上海 200071上海中医药大学附属市中医医院推拿科,上海 200071||上海中医药大学针灸推拿学院,上海 201203上海中医药大学附属市中医医院推拿科,上海 200071
医药卫生
脊髓型颈椎病人工智能磁共振成像深度学习诊疗决策预后预测
cervical spondylotic myelopathyartificial intelligencemagnetic resonance imagingdeep learningdiagnosis and treatment decision-makingprognosis prediction
《磁共振成像》 2026 (8)
226-234,9
National Natural Science Foundation of China(No.82575252)Shanghai University of Traditional Chinese Medicine(No.02.ZY05.K231201N). 国家自然科学基金项目(编号:82575252)上海中医药大学项目(编号:02.ZY05.K231201N)
评论