MRI quantitation of abdominal skeletal muscle correlates with CT-based analysis: implications for sarcopenia measurement
MRI quantitation of abdominal skeletal muscle correlates with CT-based analysis: implications for sarcopenia measurement
复制标题
DOI：
10.1139/apnm-2018-0473
复制
发表时间：
2019-08-01
影响因子：
3.4
通讯作者：
Master, Viraj A.
中科院分区：
文献类型：
--
作者：
Khan, Amir Ishaq;Reiter, David A.;Master, Viraj A.
肌肉减少症与多种疾病的不良预后相关，包括恶性肿瘤。使用计算机断层扫描 (CT) 进行的腹部骨骼肌区域 (SMA) 分割已被证明是识别肌肉减少症的准确替代方法。虽然 SMA 的磁共振成像 (MRI) 分割已在尸体四肢中得到验证，但很少有研究使用 MRI 在腰椎中 L3 水平验证腹部 SMA 分割。我们的目标是评估 L3 中期 SMA 的 CT 和 MRI 分割分析的再现性和一致性。这项回顾性分析包括 10 名肾细胞癌 (RCC) 患者和腹部/骨盆 CT 患者的随机样本，用于评估使用 CT 进行的 SMA 测量的观察者内部变异性。使用 9 名 RCC 患者和 CT 和 T2 加权 (T2w) MRI 腹部/骨盆的额外样本，使用 MRI 评估观察者内 SMA 的变异性以及 MRI 和 CT 之间 SMA 的一致性。使用 Slice-O-Matic 对 SMA 进行分割。使用组内相关系数 (ICC) 评估 SMA 再现性。使用 Bland-Altman 图和 Pearson 相关系数分析 SMA 一致性。 L3 中期 CT 和 MRI SMA 的观察者内部变异性较低，ICC 分别为 0.998 和 0.985。 Bland-Altman 分析显示，T2w MRI 相对于 CT 的偏差为 0.74%。 Pearson 相关系数为 0.997 (p < 0.0001)，表明 CT 和 T2w MRI 之间具有很强的相关性。 CT 和 T2w MRI 均可对 L3 中部的腹部 SMA 进行可重复的分割，两种模式之间具有很强的相关性。 T2w MRI 可与 CT 互换使用来评估 SMA 和肌肉减少症。这一发现具有重要的临床意义。
Sarcopenia is associated with poor outcomes in a variety of conditions, including malignancy. Abdominal skeletal muscle area (SMA) segmentation using computed tomography (CT) has been shown to be an accurate surrogate for identifying sarcopenia. While magnetic resonance imaging (MRI) segmentation of SMA has been validated in cadaver limbs, few studies have validated abdominal SMA segmentation using MRI at lumbar level mid-L3. Our objective was to assess the reproducibility and concordance of CT and MRI segmentation analyses of SMA at mid-L3. This retrospective analysis included a random sample of 10 patients with renal cell carcinoma (RCC) and CT abdomen/pelvis, used to assess intra-observer variability of SMA measurements using CT. An additional sample of 9 patients with RCC and both CT and T2-weighted (T2w) MRI abdomen/pelvis was used to assess intra-observer variability of SMA using MRI and concordance of SMA between MRI and CT. SMA was segmented using Slice-O-Matic. SMA reproducibility was assessed using intraclass correlation coefficient (ICC). SMA concordance was analyzed using Bland-Altman plot and Pearson correlation coefficient. The intra-observer variability of CT and MRI SMA at mid-L3 was low, with ICC of 0.998 and 0.985, respectively. Bland-Altman analysis revealed bias of 0.74% for T2w MRI over CT. The Pearson correlation coefficient was 0.997 (p < 0.0001), demonstrating strong correlation between CT and T2w MRI. Abdominal SMA at mid-L3 is reproducibly segmented for both CT and T2w MRI, with strong correlation between the 2 modalities. T2w MRI can be used interchangeably with CT for assessment of SMA and sarcopenia. This finding has important clinical implications.