Application Of Magnetic Resonance Imaging In Prenatal Diagnosis Of Fetal Central Nervous System

Objective: To evaluate the application of magnetic resonance imaging in prenatal diagnosis of fetal central nervous system.Materials and Methods: This study was approved by the hospital ethics committee and informed consents of patients were obtained. According to the inclusion and exclusion criteria, pregnant women who applied to our department for fetal brain MRI examination were scanned on a 1.5T superconducting MR scanner (Achieva Nova Dual, Philips, Netherlands) with two fast scan sequences (Balance FFE and Single Shot TSE). Image quality were assessed and compared by three radiologists between the two sequences. Comparison was also performed between our images and recently published images in literatures. Cases which met the standards of radiology (normal fetal brain anatomy and good image quality) and clinic (follow-up proven normal fetal brain) were selected. Major lines of fetal brain were measured in coronal and sagittal images, including cerebral frontal-occipital diameter (cFOD), cerebral biparietal diameter (CBD) and length of the corpus callosum (LCC). Relationships between the median values of those indexes and gestational age (GA) were analyzed by using SPSS 13.0 statistical software with linear regression and quadratic polynomial function fitting. The function with a higher R2 value was adopted as an optimal regression model. And a series of regression equations between the major lines of the fetal brain and GA were established. Furthermore, a number of fetal brain malformations were analyzed by using these equations. Theoretical values of cFOD, CBD and LCC were calculated according to GA and then compared with actually measured values with paired t test. As to some types of fetal brain abnormalities with only several cases, actually measured values were assessed according to the normal range of the index.Results: From January 2009 to January 2011, a total of 450 pregnant women applied to our department for fetal brain MR examination. 445 women with 450 fetuses (5 twin pregnancy) were included. Among them 30 cases underwent repeated MR examination due to significant fetal movement artifacts, resulting of 10 successes. A total of 425 cases were successfully performed. Comparison of image quality between two sequences with chi-square test showed that image quality of Balance FFE sequence was significantly better than that of Single Shot TSE (p <0.05). Also image quality of our study was better than those images published in recent literatures. A total of 156 cases of normal fetal brain were included, the median, maximum and minimum values of cFOD, CBD and LCC all increased with the progress of GA. Correlation and regression analysis showed that quadratic polynomial function fitting was better than linear regression. The regression equations of median cFOD, CBD, LCC and GA were as follows: cFOD = -0.168 GA 2 + 13.176 GA - 152.722 (r 2 = 0.813, p <0.01); CBD = -0.172 GA 2 + 13.363 GA - 176.276 (r 2 = 0.834, p <0.01); LCC = -0.036 GA 2 + 3.172 GA - 29.321 (r 2 = 0.640, p <0.01). In hydrocephalus (7 cases) and posterior fossa arachnoid cyst (6 cases), there were no significant differences between the measured and theoretical values of cFOD, CBD and LCC with paired t test (p > 0.05). In corpus callosum hypoplasia (5 cases) measured values of LCC and cFOD was significantly less than theoretical values (p <0.05). In Dandy-Walker malformation (6 cases) measured values of LCC were significantly smaller than theoretical values (p <0.05). Additionally, in brain parenchymal arachnoid cyst (1 case), septum pellucidum cyst (1 case), septum pellucidum cyst associated with ventricular cyst (1 case), periventricular haemorrhage (1 case), cerebellar hemisphere hypoplasia (1 case), measured values of cFOD, CBD and LCC were within the normal range, while in four-permian pool cyst (1 case) measured values of cFOD, CBD, and LCC were smaller than theoretical values. In cranial malacia (n = 1) measured LCC appeared smaller while measured CBD and cFOD seemed larger than the theoretical values. In cerebral malacia combined with ventricular dilatation (1 case), measured CBD was larger and LCC was smaller than theoretical values. In intraventricular hemorrhage (n = 1), measured cFOD and CBD values were smaller than theoretical ones. In sacrococcygeal cystic occupation (1 case), measured cFOD was smaller than theoretical value. In lissencephaly (1 case) and porencephalia (n = 1) measured values of three major lines were all smaller than theoretical values.Conclusion: MR scan sequences for fetal brain were successfully established in this study, providing a high success rate of examination and well image quality. A second-degree polynomial functional rather than linear relationship between the median values of cFOD (mm), CBD (mm) and LCC (mm) and GA (weeks) was revealed. Abnormal changes of major lines in a number of fetal central nervous system malformations were initially evaluated with those equations. Major morphologic indexes can provide valuable information for clinical evaluation of development status of the fetal brain.