致痫性局灶性脑皮质发育不良的影像特征及MR扫描方案优化

MR imaging features of epileptogenic focal cortical dysplasia and optimization of scanning protocols

目的通过致痫性局灶性脑皮质发育不良（FCD）的MR影像与病理对照研究，分析FCD的发病特点，探讨最佳MR扫描方案。方法回顾性分析经手术病理证实的36例（40个病灶）FCD患者的MR表现及扫描方法。根据Palmini病理分型，将FCD的病理结果分为FCDI型及FCDⅡ型。观察各型病灶在脑内分布及合并海马硬化的比例，采用Fisher精确概率法比较FCD各型在脑内分布的差异性。将轴面FSET2WI、SET1WI、液体衰减反转恢复（FLAIR）序列定义为常规扫描，在常规扫描的基础上增加斜冠状面FSET2WI及FLAIR定义为优化扫描。患者均进行常规扫描和优化扫描，评价FCD主要征象在各扫描方位、序列的显示情况，应用McNemar检验比较2种扫描方法对不同部位FCD及海马硬化显示的差异性。结果36例（40个病灶）FCD患者中，病变位于颞叶29个（72．5％），额叶9个（22．5％），顶叶2个（5．0％）。FCDⅠ型27例（29个病灶），颞叶病灶25个；FCDⅡ型10例（11个病灶），颞叶病灶4个，2种类型在病灶分布上差异有统计学意义（P=0．002）。合并海马硬化者共14例，其中，FCDⅠ型13例，FCDⅡ型1例。优化扫描对于颞叶FCD、海马硬化的显示优于常规扫描，颞叶FCD的阳性率从44．8％（13／29）提高到65．5％（19／29），合并海马硬化FCD的阳性率从42．9％（6／14）提高到85．7％（12／14），差异有统计学意义（颞叶FCDχ^2=4．167，P=0．031；海马硬化χ^2=4．167，P=0．031），对额叶FCD病灶的显示2种扫描方式间差异无统计学意义（χ^2=0．304，P=1．000）。结论FCD好发于颞叶，其次是额叶，FCDⅠ型好发于颞叶，且更容易合并海马硬化。垂直于海马长轴的全颞叶优化扫描是针对FCD发病特点的优化扫描序列组合，有助于病灶的显示。

Objective To analyze the MR imaging features of epileptogenic focal cortical dysplasia （FCD） and to optimize the scanning protocols by correlating MRI appearance with pathological findings. Methods MRI findings and the relative scanning protocols in 36 patients with surgically and pathologically proved 40 lesions of FCD were retrospectively analyzed. According to Palmini classification system, all 40 lesions were pathologically categorized as FCD type Ⅰ （including FCD Ⅰ A and FCD ⅠB） and FCD type Ⅱ （including FCD ⅡA and FCD ⅡB ） . The distribution of cerebral or dual lesions accompanied hippocampal sclerosis were observed. Differences of the distribution of cerebral in FCD type I and FCD type Ⅱ were compared by using Fisher exact probabilities. MR scans in all patients consisted of routine and optimized protocols. Axial FSE T2WI, axial SET1WI and axial FLAIR were recognized as routine scanning protocols,while adding oblique coronal FSE T2WI and FLAIR were recognized as optimization scanning protocols. Both routine and optimization scanning protocols were performed in all patients. The conspicuity of main findings of FCD on different imaging plane and sequences of each protocol were assessed. The detection of cerebral or dual lesion and the accompanied hippocampal sclerosis were compared between the routineprotocol and the optimized protocol by using McNemar test. Results Forty lesions were found in 36 cases with FCD, 29 had temporal lobe lesion （72. 5% ）, 9 had frontal lobe lesion（22. 5% ） and 2 had parietal lobe lesion （5.0%）. According to Palmini classification system, 29 lesions in 27 patients （72. 5% ） were FCD type I , Ⅱ lesions in 10 patients （27.5%） were FCD type Ⅱ lesions. There were 25 temporal lobe lesions in FCD type I ,while 4 in FCD type Ⅱ. There were statistically significant differences between FCD type I group and FCD type Ⅱ group in the distribution of cerebral （P = 0. 002）. Fourteen cases were found to have hippocampal sclerosis simultaneously, with 13 cases found in FCD type I patients and 1 case in typeⅡ patients. The detection rate of temporal lobe lesion was 65.5 % （19/29）and 44. 8 % （13/29） respectively on optimized protocol and routine protocol. There was statistically significant difference （χ^2 = 4. 167, P = 0. 031 ）. The detection rate of hippocampal sclerosis was 85.7% （12/14） and 42. 9% （6/14） respectively on optimized protocol and routine protocol respectively. There was statistically significant difference （χ^2 = 4. 167,P = 0. 031 ）. The detection rate of frontal lobe lesion showed no statistically significant difference between optimized protocols and routine protocols （ χ^2 = 0. 304, P = 1. 000 ） . Conclusions FCD was frequently involved the temporal lobe, followed by the frontal lobe. FCD type I lesion was frequently found in the temporal lobe, with a higher incidence of concomitant hippocampal sclerosis. The optimized whole temporal lobe scanning with imaging plane perpendicular to the hippocampus long axis was a highly desired scanning protocol specifically for FCD, which is helpful for the detection of the FCI） lesions.