Objective
The objective of the study was to evaluate cortical development parameters by magnetic resonance imaging (MRI) in late-onset intrauterine growth–restricted (IUGR) fetuses and normally grown fetuses.
Study Design
A total of 52 IUGR and 50 control fetuses were imaged using a 3T MRI scanner at 37 weeks of gestational age. T2 half-Fourier acquisition single-shot turbo spin-echo anatomical acquisitions were obtained in 3 planes. Cortical sulcation (fissures depth corrected by biparietal diameter), brain volumetry, and asymmetry indices were assessed by means of manual delineation and compared between cases and controls.
Results
Late-onset IUGR fetuses had significantly deeper measurements in the left insula (late-onset IUGR: 0.293 vs control: 0.267; P = .02) and right insula (0.379 vs 0.318; P < .01) and the left cingulate fissure (0.096 vs 0.087; P = .03) and significantly lower intracranial (441.25 cm 3 vs 515.82 cm 3 ; P < .01), brain (276.47 cm 3 vs 312.07 cm 3 ; P < .01), and left opercular volumes (2.52 cm 3 vs 3.02 cm 3 ; P < .01). IUGR fetuses showed significantly higher right insular asymmetry indices.
Conclusion
Late-onset IUGR fetuses had a different pattern of cortical development assessed by MRI, supporting the existence of in utero brain reorganization. Cortical development could be useful to define fetal brain imaging-phenotypes characteristic of IUGR.
Late-onset intrauterine growth restriction (IUGR) accounts for the majority of clinical forms of growth restriction and affects approximately 10% of the general population. For a long time considered as a relatively benign condition with a majority of constitutionally small fetuses, recent evidence has demonstrated that late-onset growth restriction is strongly associated with adverse pregnancy and long-term outcomes. Abnormal neurodevelopment in newborns and children is among the most relevant consequences of late-onset IUGR.
Neurodevelopmental deficits associated with this condition are thought to be the result of brain reorganizational changes, as suggested by studies demonstrating differences in brain metabolism and microstructure, morphology, and connectivity. However, the impact of late-onset IUGR on brain development is still poorly characterized. There is a need to develop imaging biomarkers that help identifying the patterns of neurodevelopment associated with sustained undernutrition in utero.
Evaluation of cortical development (CD) may provide valuable information about the influence of prenatal conditions on brain maturation. Cortical sulcation is a continuous process that occurs mainly during fetal life and results in the intricate array of brain fissures and sulci as present in term fetuses. The progress of sulcation can be used as a reliable estimate of gestational age and a good marker of fetal cortical maturation.
An altered convolution pattern has been demonstrated in preterm newborns diagnosed with severe early-onset IUGR, but the existence of differences in late-onset IUGR fetuses has not been evaluated. Aside from sulcation, normal CD during intrauterine life results in a physiological brain asymmetry. Abnormal brain asymmetry has been described in childhood conditions characterized by altered neurodevelopment, such as autism spectrum disorders, attention deficit-hyperactivity disorder (ADHD), dyslexia, and schizophrenia. It seems plausible that term late-onset IUGR fetuses could show differences in their CD pattern because the most relevant changes in brain sulcation and expression of asymmetry occur during late third trimester.
The aim of this study was to evaluate the existence of individual or combined CD differences in fetuses with late-onset IUGR compared with controls. We evaluated the depth of brain fissures, brain volumetries, and the asymmetry indices by fetal brain magnetic resonance imaging (MRI) in late-onset IUGR and adequate-for-gestational-age (AGA) fetuses at term.
Materials and Methods
Subjects
This study is part of a larger prospective research program on IUGR involving fetal, neonatal, and long-term postnatal follow-up. The specific protocol of this study was approved by the institutional ethics committee (institutional review board no. 2008/4422), and all participants gave written informed consent.
A total of 102 singleton fetuses were included in our cohort, classified in 52 late-onset IUGR and 50 AGA fetuses. Pregnancies were dated according to the first-trimester crown-rump length measurement. IUGR was defined by an estimated and postnatally confirmed fetal weight less than the 10th centile, according to local standards with normal umbilical Doppler (umbilical artery pulsatility index less than the 95th centile). AGA subjects were defined as normal term fetuses with an estimated and postnatally confirmed fetal weight of the 10th centile or greater, according to local standards. Exclusion criteria included congenital malformations, chromosomal abnormalities, perinatal infections, chronic maternal pathology, contraindications for MRI, and noncephalic presentations.
Clinical and ultrasound data
IUGR patients were followed up in our fetal growth restriction unit from diagnosis until delivery. The entire sample underwent serial Doppler ultrasound scans using a General Electric Voluson E8, 6-2 MHz curved-array transducer (GE Medical Systems, Zipf, Austria). Fetal umbilical artery and middle cerebral artery pulsatility index (PI) were measured in 3 or more consecutive waveforms, with the angle of insonation as close as possible to 0° and in the absence of fetal or maternal movements and used for the calculation of the cerebroplacental ratio (CPR). Abnormal CPR was defined as a value below the fifth centile according to previously published reference values. Uterine artery (UtA) PI was measured transabdominally as previously described, and values were considered abnormal when greater than the 95th centile. All fetuses in this study had at least 1 scan within 1 week of delivery. Maternal and perinatal data were prospectively recorded in all study patients.
Fetal MRI imaging acquisition
MRI was performed at 37 weeks of gestational age on a clinical magnetic resonance system using the IDIBAPS image platform, operating at 3.0 Tesla (Siemens Magnetom Trio Tim syngo MR B15; Siemens, Munich, Germany) without fetal sedation and following the American College of Radiology guidelines for the use of medical imaging during pregnancy and lactation. A body coil with 8 elements was wrapped around the mother’s abdomen. Routine fetal imaging took from 15 to 20 minutes and consisted of single-shot, fast spin echo T2-weighted sequences (repetition time 990 milliseconds, echo time 137 milliseconds, slice thickness 3.5 mm, no gap, field of view 260 mm, voxel size 1.4 × 1.4 × 3.5mm, matrix 192 × 192, flip angle 180°, and acquisition time 24 seconds) acquired in the 3 orthogonal planes. If the quality of the images was suboptimal, sequences were repeated.
Structural MRI images were reviewed for the presence of anatomical abnormalities by an experienced specialist in neuroradiology blinded to group membership.
Fetal imaging after processing and delineation
Offline analyses of brain biometric and volumetric measurements were performed using the semiautomatic Analyze 9.0 software (Biomedical Imaging Resource; Mayo Clinic, Kansas City, KS) by 2 experienced examiners blinded to group membership. Cortical fissure delineations in the fetal MRI were performed adapting previously described methodology to assess CD on prenatal ultrasound. All volumetric estimations were obtained using Cavalieri’s principle by a multiplanar analysis considering a slice thickness of 3.5 mm with no gap interval between them.
Both biometric and volumetric measurements showed optimal quality to perform CD analysis in 97% of the cases and in 98% of controls.
Brain biometric and sulcation analysis
Cortical fissure depths were measured bilaterally as shown in Figure 1 and corrected by biparietal diameter (BPD), obtaining a ratio (fissure/BPD) for each fissure measurement to perform the statistical analysis as follows.
- •
BPD was measured in the transthalamic axial plane as described for ultrasound by the International Society of Ultrasound in Obstetrics and Gynecology.
- •
Parietoccipital fissures were measured in an axial slice above the transthalamic plane used for the BPD assessment, tracing a perpendicular line from the longitudinal fissure to the apex of the parietoccipital fissures.
- •
Insular depths were measured in the axial slice located immediately below the anterior commissure and the cavum septum pellucidum, tracing a perpendicular line from the median longitudinal fissure to the most external border of the insular cortex.
- •
Lateral fissure depths were measured in the same plane described above, with a continuing line starting from the most external border of the insular cortex to the interface conformed between the subarachnoid space and the skull.
- •
Cingulate fissures were measured in the midcoronal plane, tracing a perpendicular line from the median longitudinal fissure to the apex of the cingulate fissures.
- •
Calcarine fissures were measured in the coronal transcerebellar plane as described in ultrasound by the International Society of Ultrasound in Obstetrics and Gynecology, tracing a perpendicular line from the median longitudinal fissure to the apex of the calcarine fissures.
Brain volumetric analysis
For the statistical analysis, total intracranial volume and total brain volume were adjusted by birthweight centile, and the opercular volumes were adjusted by total brain volume (TBV) ( Figure 2 ).
- •
Total intracranial volume (TIC) was successively delineated including the extra- and intraventricular cerebrospinal fluid (CSF), cerebral, cerebellar, and brain stem.
- •
TBV was successively delineated including intraventricular CSF, cerebral, cerebellar, and brain stem parenchymal volumes but excluding the extraventricular CSF volume.
- •
Opercular volumes were delineated bilaterally in all the slices in which the operculums were identified, following the opercular cortex until the external borders of the lateral fissures, thus closing the volume by a straight line joining the anterior and posterior edges of the fissure. A methodology reported by Nakamura et al was adapted and followed.
Brain asymmetry
To assess the degree of brain asymmetry, we applied a previously reported asymmetry index expressed by the following: asymmetry index = (R–L)/(R+L) to compare right (R) and left (L) fissures and opercular measurements.
Interobserver variability
To establish both biometric and volumetric measurement reproducibility, a total of 15 cases were completely assessed independently by 2 experienced examiners blinded to group membership. Bland-Altman plots were used to assess interobserver variability with a significance level set at 5% ( P < .05) for all measurements.
Statistical analysis of clinical and CD data
Student t test for independent samples and Pearson’s χ 2 tests were used to compare quantitative and qualitative data, respectively, between late-onset IUGR and controls. Multivariate analyses of covariance were conducted for biometric and volumetric measurements, adjusting by sex, gestational age at MRI, and maternal body mass index as covariants. Furthermore, a classified model was constructed by means of a logistic regression including all cortical parameters to explore the existence of combined patterns that classified the study groups better than the top individual parameter.
Finally, a subanalysis was conducted to explore the existence of differences in relation with previously described predictors of poorer perinatal outcome, namely abnormal CPR, UtA Doppler before birth, or a birthweight less than the third centile in the last scan before delivery.
The software package SPSS 17.0 (SPSS, Chicago, IL) and the MedCalc 8.0 (Broekstraat, Belgium) were used for the statistical analyses.
Results
Clinical characteristics in the study population
All 102 fetuses from our sample (52 late-onset IUGR and 50 AGA) underwent fetal MRI at 37 weeks of gestational age. Maternal characteristics and time of MRI scans did not differ between cases and controls ( Table 1 ). As expected, IUGR fetuses were delivered earlier with higher rates of labor induction, emergency cesarean section, neonatal acidosis, and lower Apgar scores at 5 minutes of life ( Table 2 ).
| Charactertistic | IUGR (n = 52) | AGA (n = 50) | P value a |
|---|---|---|---|
| Maternal age, y | 31.8 ± 6.2 | 32.2 ± 4.3 | .64 |
| Maternal smoking | 15.4% | 6% | .13 |
| Maternal height, m | 1.58 ± 0.05 | 1.65 ± 0.06 | < .01 |
| Maternal weight, kg | 56.9 ± 10.1 | 63.2 ± 11.3 | .01 |
| Maternal BMI, kg/m 2 | 22.7 ± 3.7 | 23.2 ± 3.9 | .49 |
| Primiparity | 65% | 64% | .88 |
| Caucasian ethnicity | 79% | 68% | .18 |
| GA at MRI, wks | 37.5 ± 0.8 | 37.7 ± 0.8 | .30 |
a Student t test for independent samples or Pearson’s χ 2 test.
| Variable | IUGR (n = 52) | AGA (n = 50) | P value a |
|---|---|---|---|
| GA at birth, wks | 38.8 ± 1.0 | 40.0 ± 1.0 | < .01 |
| Birthweight, g | 2488 ± 247 | 3452 ± 311 | < .01 |
| Birthweight percentile | 3.9 ± 6.6 | 55.2 ± 24.2 | < .01 |
| Male sex | 62% | 52% | .34 |
| Labor induction | 77.1% | 14.3% | < .01 |
| Emergency cesarean section | 31% | 6% | < .01 |
| Neonatal acidosis b | 19.6% | 5% | .04 |
| Apgar score less than 7 at 5 minutes | 5.8% | 0% | .04 |
| NICU stay length, d | 0.3 | 0 | .09 |
a Student t test for independent samples or Pearson’s χ 2 test
b Neonatal acidosis: umbilical artery pH less than 7.15 and base excess greater than 12 mEq/L.
Interobserver agreement
Overall CD measurements showed a good interobserver reproducibility. Regarding the sulcation analysis, the coefficients of variation were 4.4% for the left insular depth, 13.5% for the right insular depth, 19.7% for the left lateral fissure depth, 18.7% for the right lateral fissure depth, 21.5% for the left parietoccipital fissure depth, 22.6% for the right parietoccipital fissure depth, 25.2% for the left cingulate fissure depth, 30.7% for the right cingulate fissure depth, 32.6% for the left calcarine fissure depth, and 21.3% for the right calcarine fissure depth . The coefficient of variation for the volumetric analysis was 3.4%.
Brain biometric and sulcation analysis
IUGR fetuses showed smaller BPD measurements (IUGR: 94.33 mm ± 2.76 vs AGA: 100.63 mm ± 3.1; P < .01). IUGR fetuses showed significantly deeper fissure measurements in the right and left insula and the left cingulate fissure ( Table 3 ).
| Variable | IUGR (n = 52) | AGA (n = 50) | P value a |
|---|---|---|---|
| LID/BPD | 0.293 ± 0.052 | 0.267 ± 0.029 | .02 |
| RID/BPD | 0.379 ± 0.066 | 0.318 ± 0.075 | < .01 |
| Left lateral fissure/BPD | 0.154 ± 0.042 | 0.151 ± 0.024 | .81 |
| Right lateral fissure/BPD | 0.151 ± 0.037 | 0.154 ± 0.042 | .82 |
| Left parietooccipital fissure/BPD | 0.139 ± 0.029 | 0.132 ± 0.036 | .18 |
| Right parietooccipital fissure/BPD | 0.138 ± 0.029 | 0.132 ± 0.036 | .22 |
| Left cingulate fissure/BPD | 0.096 ± 0.016 | 0.087 ± 0.019 | .03 |
| Right cingulate fissure/BPD | 0.093 ± 0.019 | 0.088 ± 0.018 | .19 |
| Left calcarine fissure/BPD | 0.180 ± 0.039 | 0.182 ± 0.040 | .56 |
| Right calcarine fissure/BPD | 0.179 ± 0.038 | 0.187 ± 0.041 | .14 |
a GLM adjusted for sex, gestational age at MRI, and maternal BMI.
Brain volumetric analysis
The IUGR group had smaller TIC and TBV (441.25 cm 3 ± 56.01 vs 515.82 cm 3 ± 36.63; P < .01; and 276.47 cm 3 ± 52.61 vs 312.07 cm 3 ± 40.85; P < .01, respectively) and smaller left opercular volumes (2.52 cm 3 ± 0.69 vs 3.02 cm 3 ± 0.71; P < .01) ( Table 4 and Figure 3 ).
