| ObjectiveTo evaluate the performance of the susceptibility weighted imaging(SWI)in visualizing the normal structures of the fetal vertebrae and further diagnosing its relative anomalies,and compare with HASTE,True FISP.Materials and MethodsThe institutional review board approved this retrospective study and waived the requirement for patient informed consent.This study included 150 pregnant women(mean age,31.3years±5.2;range age,20-45 years;median age,31.0 years;mean gestational age,28.5±4.5W;range gestational age,20W-39.4W;median gestational age,28W)who were suspected fetal spinal anomalies by ultrasonography(US)screening and underwent 1.5-T fetal spine MRI.The MRI protocol was as follows:half-fourier acquisition single-shot turbo spin--echo(HASTE),true fast imaging with steady-state(TrueFISP)and SWI sequences.Two radiologists independently evaluated all images quality of the fetal vertebra in four segments of three sequences(HASTE,True FISP and SWI),using a 4-point grading scale:4 denoted an excellent depiction of anatomy and high tissue contrast,with the image thereby being optimal for diagnostic purposes;3 indicated clear visualization of the structure and reduced tissue contrast,with image-based diagnosis feasible;2 denoted suboptimal visualization of structures,with diagnosis still possible;1 indicated a poor depiction,with the image therefore not being adequate for diagnostic purposes;and 0 indicated that no structure was seen,which was assessed by kappa statistics.The image scores in HASTE,True FISP and SWI were compared by using Kruskal-Wallis test.The cervical(n=18),thoracic(n=62),lumbar(n=46)and sacral(n=24)segments,respectively of 150 cases were evaluated in HASTE,True FISP and SWI images and then assessed by ANOVA analysis.The diagnostic accuracy of the three sequences in 77 cases were compared with relative follow-up results after birth(n=59)or post-mortem imaging examination(n=18)respectively.ResultsThere was good agreement in two radiologists for the three sequences(SWI,kappa=0.887;True FISP,0.720;HASTE,0.767).The mean scores of fetal vertebra evaluated by two radiologists were significantly different between and among SWI,True FISP and HASTE(3.70 ± 0.47 Vs.2.80 ± 0.79 Vs 2.06 ± 0.88;F=216.995,P<.001),and the same was true for the cervical(1.28 ± 0.97 Vs.1.86± 0.78 Vs.3.67 ± 0.49,F=46.667,P<.001),thoracic(1.97 ± 0.79 Vs.2.72 ± 0.76 Vs.3.71 ±0.44;F= 101.885,P<.001),lumbar(2.20 ± 0.76 Vs.2.92±0.65 Vs.3.71±0.44,F=69.841,P<.001)and sacral(2.44 ± 0.70 Vs.2.96±0.46 Vs.3.69± 0.62,F=26.093,P<.001)segments.The scores of SWI was higher than that of True FISP(P=0)and HASTE(P=0)in each segment,and similar findings were noted between True FISP and HASTE(P=0)in each segment except for the cervical part.There were no statistic difference between True FISP and HASTE in the cervical part,and neither of them could meet the requirement for clinical diagnosis(score<2),while,the SWI(score>2)satisfied clinical diagnosis favorably.A total of 77 fetal vertebral anomalies were identified by follow-up after birth(n=59)or post-mortem imaging examinations(n=18)including hemivertebra(n=28),block vertebrae(n=4),butterfly vertebrae(n=6),multiple vertebral malformations(n=23),spinal dysraphism(n=12),caudal regression syndrome(n=4).The diagnostic accuracy of SWI,True FISP and HASTE was 94.81%(73/77),62.34%(48/77)and 31.17%(24/77),respectively.The diagnostic accuracy of SWI was the best among those of True-FISP and HASTE(?2=22.215,P=0;?2=64.174,P=0).ConclusionSWI shows the optimal performance in depiction of fetal vertebra and vertebral anomalies than True FISP and HASTE,especially in depiction of cervical vertebra. |
PDF Full Text Request