MRI超短回波时间成像序列结合T2*映射技术用于在体耳软骨无创检测

提出使用磁共振图像(MRI)超短回波时间(UTE)成像序列结合T2*映射技术, 对耳软骨T2*值进行在体无创检测, 探究既能对耳软骨进行非侵入性成像又能定量评估在体耳软骨生物成分的方法, 为耳软骨再造和修复整形手术评价的标准化提供新思路。首先, 使用1个UTE和5个短回波时间(TE)的成像组合序列采集30名志愿者右侧外耳的MRI图像;然后, 利用采集到的每名志愿者的图像进行组内刚性配准及手动分割耳软骨与外耳轮廓(包含耳软骨及周围组织, 如皮肤、脂肪和其他软组织)的预处理;接下来, 分别运用单指数和双指数衰减模型在分割出来的耳软骨和外耳区域进行T2*值测量;最后, 分别使用这两种模型拟合耳软骨信号强度随回波时间变化的衰减曲线, 并比较拟合模型的准确性。结果显示, 在30例右耳的单成分分析(单指数模型)实验中, 外耳的T2*_m平均值为(49.269±16.979)ms, 耳软骨的T2*_m平均值为(23.799±9.629)ms。在双成分分析(双指数模型)中, 外耳的短成分T2*_s平均值为(11.713±3.111)ms, 长成分T2*_l平均值为(65.128±13.132)ms, 耳软骨的短成分T2*_s平均值为(5.577±1.830)ms, 长成分T2*_l平均值为(30.628±8.413)ms。统计分析显示, 单成分分析计算得到的T2*_m, 和双成分分析计算得到的T2*_s、T2*_l, 在外耳和耳软骨区域均存在显著差异(P <0.05)。在曲线拟合中, 双指数模型优于单指数模型(R²bi]=0.999±0.001 vs R²mono]=0.905±0.014, P<0.05)。实验结果表明, 超短回波时间成像序列结合T2*映射技术对在体耳软骨进行T2*值无创检测具备可行性, 有望为软骨组织工程和3D生物打印技术制作的耳软骨-支架复合物应用在耳廓修复和再造中提供医学影像学的支撑, 也为小耳畸形外科整形手术术后定量评估耳软骨提供一种可行性方案。

Non-Invasive Measurement of Auricular Cartilage in vivo by UTE T2* Mapping

This paper proposed to utilize MRI ultra-short echo time (UTE) imaging in conjunction with T2* mapping for image-based analysis ofin vivo auricular cartilage, exploring the technical feasibility of not only non-invasive imaging but also quantitative measurement of the biological components of auricular cartilage and providing innovative ideas for standardization and evaluation of advanced biotechnology-based microtia reconstruction and surgery. Firstly, in the presented method, 30 volunteers were imaged for the right-sided ear using the sequence of multiple echo times (TE) containing one UTE and five short TEs. Secondly, these multiple images for each volunteer were preprocessed with intra-rigid registration and then manual segmentation of auricular cartilage and the external ear (containing not only auricular cartilage but also surrounding tissues, e.g. skin, fat, and other soft tissues). Next, for each volunteer, the component analysis was carried out to calculate the T2* values in ROIs of auricular cartilage and external ear respectively, using both the mono- and bi-exponential models (short component T2* and long component T2*). Finally, both exponential models were used to fit the curves of the auricular cartilage intensityversus echo time. In the experiment of mono-component analysis of the 30 right-sided ears, the mean T2* value of external ear was(49.269±16.979) ms and the mean T2* value of segmented auricular cartilage was (23.799±9.629) ms. In the bi-component analysis, the mean short component T2* was (11.713±3.111) ms and the mean long component T2* was (65.128±13.132) ms for the external ear, while for the segmented auricular cartilage, the mean short component T2* was(5.577±1.830) ms and the mean long component T2* was(30.628±8.413) ms. There was a significant difference of T2*values calculated by both component analyses between the external ear and the auricular cartilage (P<0.05). The model with bi-exponential fitting outperformed the one with mono-exponential fitting with a better fitted curve and a calculated value of R² bi]=0.999 ± 0.001 vs R² mono]=0.905±0.014 (P <0.05). Our preliminary results demonstrated that the proposed UTE T2* mapping has shown to be a feasible non-invasive means for quantifying the auricular cartilage in vivo and a potential tool used to image and evaluate the complex of auricular cartilage with biomaterials in reconstructive surgery for microtia using tissue engineering or 3D bioprinting technique in the future.