Department of Diagnostic and Interventional Radiology, University Hospital Frankfurt, Frankfurt am Main, Germany.
Department of Biomedical Sciences and Morphological and Functional Imaging, University Hospital Messina, Messina, Italy.
NMR Biomed. 2023 Feb;36(2):e4828. doi: 10.1002/nbm.4828. Epub 2022 Oct 6.
Whole-body magnetic resonance imaging (MRI) has become increasingly popular in oncology. However, the long acquisition time might hamper its widespread application. We sought to assess and compare free-breathing sequences with conventional breath-hold examinations in whole-body MRI using an automated workflow process. This prospective study consisted of 20 volunteers and six patients with a variety of pathologies who had undergone whole-body 1.5-T MRI that included T1-weighted radial and Dixon volumetric interpolated breath-hold examination sequences. Free-breathing sequences were operated by using an automated user interface. Image quality, diagnostic confidence, and image noise were evaluated by two experienced radiologists. Additionally, signal-to-noise ratio was measured. Diagnostic performance for the overall detection of pathologies was assessed using the area under the receiver operating characteristics curve (AUC). Study participants were asked to rate their examination experiences in a satisfaction survey. MR free-breathing scans were rated as at least equivalent to conventional MR scans in more than 92% of cases, showing high overall diagnostic accuracy (95% [95% CI 92-100]) and performance (AUC 0.971, 95% CI 0.942-0.988; p < 0.0001) for the assessment of pathologies at simultaneously reduced examination times (25 ± 2 vs. 32 ± 3 min; p < 0.0001). Interrater agreement was excellent for both free-breathing (ϰ = 0.96 [95% CI 0.88-1.00]) and conventional scans (ϰ = 0.93 [95% CI 0.84-1.00]). Qualitative and quantitative assessment for image quality, image noise, and diagnostic confidence did not differ between the two types of MR image acquisition (all p > 0.05). Scores for patient satisfaction were significantly better for free-breathing compared with breath-hold examinations (p = 0.0145), including significant correlations for the grade of noise (r = 0.79, p < 0.0001), tightness (r = 0.71, p < 0.0001), and physical fatigue (r = 0.52, p = 0.0065). In summary, free-breathing whole-body MRI in tandem with an automated user interface yielded similar diagnostic performance at equivalent image quality and shorter acquisition times compared to conventional breath-hold sequences.
全身磁共振成像（MRI）在肿瘤学领域越来越受欢迎。然而，较长的采集时间可能会阻碍其广泛应用。我们旨在评估和比较使用自动化工作流程的全身 MRI 中的自由呼吸序列和传统屏气检查。这项前瞻性研究包括 20 名志愿者和 6 名患有各种病理的患者，他们接受了包括 T1 加权径向和 Dixon 容积内插屏气检查序列的全身 1.5-T MRI 检查。自由呼吸序列通过使用自动化用户界面进行操作。两位有经验的放射科医生评估了图像质量、诊断信心和图像噪声。此外，还测量了信噪比。使用受试者工作特征曲线下面积（AUC）评估了整体检测病理的诊断性能。研究参与者在满意度调查中对他们的检查体验进行了评分。磁共振自由呼吸扫描在超过 92%的情况下被评为至少与传统磁共振扫描相当，显示出较高的整体诊断准确性（95%[95%置信区间 92-100]）和性能（AUC 0.971，95%置信区间 0.942-0.988；p<0.0001），同时检查时间缩短（25±2 分钟与 32±3 分钟；p<0.0001）。自由呼吸（κ=0.96[95%置信区间 0.88-1.00]）和传统扫描（κ=0.93[95%置信区间 0.84-1.00]）的观察者间一致性均为极好。两种 MRI 图像采集的图像质量、图像噪声和诊断信心的定性和定量评估无差异（均 p>0.05）。与屏气检查相比，自由呼吸的患者满意度评分显著提高（p=0.0145），包括噪声等级（r=0.79，p<0.0001）、紧密度（r=0.71，p<0.0001）和身体疲劳（r=0.52，p=0.0065）的显著相关性。总之，与传统屏气序列相比，结合自动化用户界面的自由呼吸全身 MRI 在具有相似诊断性能的情况下可获得等效的图像质量和更短的采集时间。
