Assessment of chemical-shift and diffusion-weighted magnetic resonance imaging in differentiating malignant and benign vertebral lesions in oncologic patients. A single institution experience

评估化学位移和弥散加权磁共振成像在鉴别肿瘤患者椎体良恶性病变中的作用:单中心经验

阅读:1

Abstract

BACKGROUND: To analyze the contribution of two non-standard magnetic resonance imaging (MRI) techniques the chemical-shift image (CSI), and diffusion-weighted imaging (DWI) in distinguishing malignant and benign vertebral bone marrow lesions (VBMLs). PATIENTS AND METHODS: Conventional spine MRI protocol, followed by CSI and DWI was performed with a 1.5 T system on 102 oncologic patients between January 2020 and December 2023. From the identified 325 VBMLs, 102 representative lesions (one per patient) were selected. VBMLs were divided into malignant (n = 74) and benign (n = 28) based on histopathology, or imaging follow-up. The quantitative parameters for VBMLs assessment were signal intensity ratio (SIR) derived from CSI and apparent diffusion coefficient (ADC) derived from DWI. RESULTS: The malignant VBMLs had significantly higher SIR values (p < 0.05) and lower ADC values compared to benign VBMLs (p < 0.05). The area under the curve (AUC) was 0.953 (p < 0.001) for SIR, and 0.894 for ADC (p < 0.001) (cut-off at > 0.82, and ≤ 1.57x10(-3) mm(2)/s, respectively). The sensitivity and specificity for SIR were 93.6%, and 88.5%, while for ADC were 88.2% and 92.3% (respectively). The combined use of SIR and ADC improved the diagnostic accuracy to AUC of 0.988 (p < 0.001, cut-off at > 0.19), sensitivity, and specificity of 100.0% and 90.9% (respectively). CONCLUSIONS: Quantitative parameters, SIR and ADC, derived from two non-standard MRI techniques, CSI, and DWI, showed diagnostic strength in differentiating malignant and benign VBMLs. Combining both methods can further enhance the diagnostic performance and accuracy of spine MRI in clinical practice.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。