Paul Strickland Scanner Centre, Mount Vernon Cancer Centre, Rickmansworth Rd, Northwood, Middlesex, HA6 2RN, UK.
AJR Am J Roentgenol. 2013 Jan;200(1):163-70. doi: 10.2214/AJR.11.8185.
The purposes of this study were to observe the relation between signal intensity (SI) on MR images with a high b value and the apparent diffusion coefficient (ADC) of bone marrow on body diffusion-weighted MR images, to determine cutoff values that enable separation of malignant and normal bone marrow, and to identify the upper ADC values of untreated multiple myeloma lesions and bone metastatic lesions of breast cancer.
Retrospective evaluations of 16 patients without bone disease, 21 patients with untreated metastases of breast cancer, and 12 patients with myeloma undergoing body diffusion-weighted MRI were performed (b values, 50 s/mm(2) and 800 or 900 s/mm(2)). Normal yellow and red bone marrow regions were compared with metastatic breast and myeloma bone marrow lesions (one to five regions of interest per patient). SI values were normalized to kidney, muscle, and spinal cord SI. Signal-to-noise ratio and ADC for each lesion were recorded. Nonparametric, receiver operating characteristic, and nonlinear regression analyses were performed.
Yellow bone marrow and red bone marrow ADC values were lower than the tumor values (p < 0.001; area under the curve, 0.94; cutoff, 774 μm(2)/s). Tissue-normalized SI and the signal-to-noise ratio of normal bone marrow were also lower than those in tumor regions (p < 0.001; area under the curve, 0.86-0.88). Second-order polynomial curve fitting between SI and ADC was observed (muscle normalized SI, R(2) = 0.4). The 95th percentile and maximum values for mean tumor ADC distribution were 1209 μm(2)/s and 1433 μm(2)/s.
Both tissue-normalized SI and ADC measurements allow differentiation between normal bone marrow and tumors of myeloma and breast cancer. The presence of a nonlinear relation between bone marrow SI and ADC values enables definition of an upper limit of ADC value for untreated myeloma lesions and metastatic lesions of breast cancer.
本研究旨在观察高 b 值下磁共振图像信号强度（SI）与体部扩散加权磁共振图像骨髓表观扩散系数（ADC）之间的关系，确定能够区分恶性和正常骨髓的截断值，并确定未经治疗的多发性骨髓瘤病变和乳腺癌骨转移病变的 ADC 值上限。
对 16 例无骨病患者、21 例未经治疗的乳腺癌转移患者和 12 例多发性骨髓瘤患者进行了回顾性评估（b 值为 50 s/mm²和 800 或 900 s/mm²）。将正常的黄骨髓和红骨髓区域与乳腺癌转移和多发性骨髓瘤骨病变（每位患者 1 至 5 个感兴趣区）进行比较。SI 值与肾脏、肌肉和脊髓 SI 进行归一化。记录每个病变的信噪比和 ADC。进行了非参数、接收者操作特征和非线性回归分析。
黄骨髓和红骨髓的 ADC 值低于肿瘤值（p<0.001；曲线下面积，0.94；截断值，774 μm²/s）。正常骨髓的组织归一化 SI 和信噪比也低于肿瘤区域（p<0.001；曲线下面积，0.86-0.88）。观察到 SI 与 ADC 之间的二次多项式曲线拟合（肌肉归一化 SI，R²=0.4）。平均肿瘤 ADC 分布的第 95 百分位数和最大值分别为 1209 μm²/s 和 1433 μm²/s。
组织归一化 SI 和 ADC 测量均可区分多发性骨髓瘤和乳腺癌的肿瘤与正常骨髓。骨髓 SI 和 ADC 值之间存在非线性关系，这使得能够为未经治疗的多发性骨髓瘤病变和乳腺癌骨转移病变定义 ADC 值上限。
