多模态磁共振成像对新生儿局灶性脑白质损伤信号分型的研究
Signal classification by multimodal magnetic resonance imaging on neonatal punctate white matter lesions

摘要：目的 探讨多模态磁共振成像(MRI)(常规MRI、磁共振扩散成像、磁敏感加权成像)对局灶性白质损伤(PWML)信号分型的意义。方法?? 研究对象为2011年3月至2013年2月在西安交通大学第一附属医院GE 3.0T磁共振设备行头颅MRI检查诊断为PWML的新生儿。扫描序列包括T1加权成像(T1WI)、T2加权成像(T2WI)、扩散张量成像(DTI)及磁敏感加权成像(ESWAN)。DTI及ESWAN经过图像后处理，分别得到相应表观扩散系数(ADC)图、幅度图及相位图。观察PWML病灶在T1WI、T2WI、ADC图及幅度图的信号表现，进行信号组合分型，统计每一类型病例数。结果 ?共纳入PWML新生儿84例，早产儿47例，足月儿37例。以T1WI所示病灶为基础，进行各序列图像病灶信号一一对应观察，共发现7种信号类型，归为3组：①Ⅰ型最多，59例，占70.24%，表现为T1WI高信号、T2WI低信号、ADC图低信号、幅度图高信号；②Ⅱ型7例，占8.33%，表现为T1WI高信号、T2WI低信号、ADC图低信号、幅度图低信号；③Ⅲ～Ⅶ型即含等信号组26例，占30.95%，表现为T1WI高信号、T2WI低/等信号、ADC图低/等信号、幅度图高/等信号。比较3组不同信号类型在早产儿、足月儿的发生率，差异无统计学意义(P>0.05)。结论 MRI信号可在一定程度上反映PWML病灶的特点。多模态MRI信号分型有助于判断PWML病灶的损伤类型，对临床治疗有一定的指导意义。
ABSTRACT: Objective To explore the significance of signal classification by multimodal magnetic resonance imaging (MRI) (including conventional MRI, diffusion imaging, and susceptibility-weighted imaging) on neonatal punctate white matter lesions (PWML). Methods? ?Between March 2011 and February 2013, neonates with PWML who underwent 3.0T MRI examination were recruited. The routine protocols included T1 weighted imaging (T1WI), T2 weighted imaging (T2WI), diffusion tensor imaging (DTI), and 3D-enhanced susceptibility-weighted angiography (ESWAN). Apparent diffusion coefficient (ADC) map, magnitude image, and phase image were obtained from DTI and ESWAN, respectively. The signal of PWML was one-to-one observed on T1WI, T2WI, ADC map and magnitude image. The combined signals were classified and the number of cases of each type was counted. Results?? Totally 84 neonates with PWML were included (preterm/term, 47/37). Seven signal types were found and classified into three groups: ① Type Ⅰ was the most, 59 cases, accounting for 70.24%, with the performance of high-signal intensity on T1WI, low-signal intensity on T2WI, low-signal intensity on ADC map, and high-signal intensity on magnitude image. ②Type Ⅱ, 7 cases, accounting for 8.33%, high-signal intensity on T1WI, low-signal intensity on T2WI, low-signal intensity on ADC map, and low-signal intensity on magnitude image. ③ Type Ⅲ-Ⅶ, 26 cases, accounting for 30.95%, high-signal intensity on T1WI, low/iso-signal intensity on T2WI, low/iso-signal intensity on ADC map, high/iso-signal intensity on magnitude image. There was no significant difference in the incidence of different signal types between preterm and full-term PWML neonates. Conclusion Magnetic resonance signal may reflect the characteristics of PWML lesions. Signal classification of PWML by multimodal magnetic resonance imaging is helpful in determining injury type and has certain significance in guiding