Dr C Garel, Service d’imagerie pédiatrique, Hôpital Robert Debré, 48, boulevard Sérurier 75019 Paris, France. Email email@example.com
Objective The justification for magnetic resonance imaging (MRI) in isolated mild ventriculomegaly remains controversial. This study was undertaken to evaluate the contribution of third-trimester MRI in isolated 10- to 12-mm fetal ventriculomegaly.
Design Observational prospective cohort study.
Setting Universitary prenatal reference centre.
Population From February 2000 to May 2005, we prospectively collected data concerning fetuses referred to us for cerebral MRI following detection of ventriculomegaly by ultrasound scan (n= 310).
Methods Among these, we identified and analysed those cases in which ventriculomegaly was isolated and did not exceed 12 mm in ultrasound examinations prior to MRI scan (n= 185).
Main outcome measure Cases in which MRI provided additional information that was likely to have an impact on prenatal management were detailed.
Results During the study period, 310 MRI were performed because of fetal ventriculomegaly. Hundred and eighty-five were suspected to be isolated 10- to 12-mm ventriculomegalies in ultrasound scan and formed our database. MRI confirmed the 10- to 12-mm isolated fetal ventriculomegaly in 106 cases (57.3%) and found other abnormalities in 5 (4.7%) of these 106 cases. MRI found ventricular measurement to be less than 10 mm in 43 cases (23.3%) and more than 12 mm in 36 cases (19.4%). Among these 36 fetuses with ventricle size more than 12 mm, 6 (16.7%) had other abnormalities, whereas MRI did not find other abnormalities in the 43 cases with ventricle size below 10 mm.
Conclusion Before advantages of MRI to ultrasound examination can be demonstrated, it seems reasonable that MRI should remain an investigational tool, restricted to selected clinical situations in which the results are expected to modify case management. Where ultrasound scan suspects isolated ventriculomegaly of 10 to 12 mm, our data suggest that when the finding is confirmed with MRI this could be expected in around 5% of cases. Therefore, the policy of routine MRI in such cases should depend on prenatal centres’ priorities.
Ventriculomegaly is an excess of fluid in the lateral ventricles of the developing brain.1 It is one of the most common sonographically diagnosed fetal cerebral abnormalities and may lead to the detection of many other pathologies. When the axial diameter, measured across the atrium of the ventricle at any gestational age (GA), exceeds 15 mm, ventriculomegaly is said to be severe.1–5 It is generally agreed that 10 mm is the upper limit of the normal range for the lateral ventricles in an axial plane.1–5 Therefore, sonographically diagnosed isolated mild ventriculomegaly (IMV) of the fetus usually refers to cases in which the ventricle sizes are between 10 and 15 mm.6,7 It is the most common brain abnormality found on prenatal ultrasound scan.8
These mild cases are the most challenging ones as optimal management and counselling have not been well defined. Management of such cases usually includes targeted ultrasound screening for associated fetal cerebral and extracerebral abnormalities, amniocentesis searching for abnormal fetal karyotype and infectious diseases.6,7 MRI where available is also frequently offered at third trimester in such cases.9 However, the justification for the use of MRI is debated because of high cost, low availability and unproved greater diagnostic accuracy compared with sonography.10,11
In addition to these issues, fetuses presenting with a 10- to 12-mm ventriculomegaly are considered in a group that many feel have little morbidity. The prognosis and association with chromosomal abnormality might be different for a ventriculomegaly of 10–12 mm compared with 12–15 mm,12,13 and a large prospective study suggested that the 10-mm cutoff level was too low and 12 mm was suggested as the upper limit of the normal range.4 This study was therefore undertaken to evaluate the contribution of MRI in isolated third-trimester ventriculomegaly of 10–12 mm.
Materials and methods
From February 2000 to May 2005, we prospectively collected data in fetuses referred to us for cerebral MRI following detection of ventriculomegaly by ultrasound scan at a GA ranging from 22 to 35 weeks. GA was determined from the earliest available ultrasound scan. All subjects underwent amniocentesis for determination of the karyotype and serologic cytomegalovirus (CMV) studies. Prior to MRI, in each case, ultrasound scan was performed by a trained sonographer using 5-MHz curvilinear probe (5-MHz curvilinear abdominal probe, ATL HDI 5000, Bothell, WA, USA). Fetal brain was examined in the axial, coronal and sagittal planes to detect abnormalities of the midline, cerebral parenchyma and gyration. If the presence of corpus callosum was doubtful, colour flow imaging was used to view the pericallosal artery. Mothers’ informed consent was obtained in all cases after the nature of the procedures had been fully explained.
MRI was performed on a 1.5-T unit (Integra, Philips, Best, The Netherlands), 30–40 minutes following a fetal sedation achieved by maternal oral administration of flunitrazepam. Sedation was used in order to reduce fetal motion, which is increased during T1 sequences, due to a long acquisition time. Brain examination was performed using T1-weighted spin echo (SE), spectral presaturation inversion recovery, fat-saturated sequences (697/14/2; α= 90°; 256 × 256; field of view (FOV), 32 cm; section thickness, 4 mm; acquisition time, 2 minutes 56 seconds; 15 sections) and T2-weighted single-shot turbo SE imaging (24 617/100/1 α= 90°; 256 × 256; FOV, 280 cm; section thickness, 3 mm; acquisition time, 24 seconds). Measurement of the cerebral biometric parameters as well as ventricular diameter was obtained on these T2-weighed images. Additional diffusion tensor imaging was acquired as part of routine examination on transverse slices with a multishot echoplanar imaging sequence with diffusion gradient (amplitude: 30 mT/m, b = 700 seconds/mm2) applied in six noncolinear axes (acquisition time, 54 seconds).
Measurement of cerebral ventricle was performed by ultrasound scan and MRI as previously described:14 measurement of both atria of the cerebral ventricles was performed in the coronal plane (with good visibility of the choroid plexuses), on an axis perpendicular to that of the ventricle, at midheight of the ventricles. In ultrasound scan, calipers were positioned inside the echoes generated by the ventricular walls and in MRI, inside the hypointensity of the ventricular walls.
For the purpose of this study, only fetuses presenting with isolated 10- to 12-mm ventriculomegaly (unilateral or bilateral) in ultrasound scan using our coronal measurements on one or more scans, normal karyotype and negative screening for CMV were included and formed our database. Monthly follow-up ultrasound examinations were performed in all cases, and MRI was planned at the optimal GA of 30–32 weeks depending on MRI scan availability or as soon as possible when ventriculomegaly was diagnosed later in gestation. Magnetic resonance images were prospectively reviewed by a radiologist with experience in fetal neuroradiology and who was aware of the sonographic findings. In case of termination of pregnancy (TOP), pathological examination of the fetal brain was performed after parent’s consent. In all other cases, monthly follow-up ultrasound examinations were performed until birth, and postnatal MRI examination was offered in all cases and strongly recommended whenever prenatal MRI demonstrated abnormalities other than isolated 10- to 12-mm ventriculomegaly.
In all cases, we examined retrospectively i) if both atrial diameters in MRI were measured between 10 and 12 mm, above 12 mm or below 10 mm and ii) if MRI provided additional information that was likely to change prenatal management and its correlation with postnatal or pathological findings.
All cases, in which IMV was confirmed by MRI, underwent postnatal imaging (MRI examination at 2 months and 2 years) and clinical follow up (clinical examination by a paediatric neurologist at 2, 6, 9, 12, 18 and 24 months).
During the study period, 310 MRIs were performed because of fetal ventriculomegaly. Among these, 185 were isolated and the sizes ranged between 10 and 12 mm; they formed our database.
In this study, we focused only on these 185 borderline ventriculomegalies diagnosed by ultrasound scan; in 114 cases (61.6%; 95% CI 54.2–68.7), size of only one ventricle was measured between 10 and 12 mm (left one in 65 cases [35.1%] and right one in 49 cases [26.5%]) whereas the size of other one was less than 10 mm. Therefore, there was no significant statistical difference between left and right ventriculomegaly (P= 0.07). However, in these 114 cases with unilateral ventriculomegaly, the difference between ventricular size was less than 2 mm in 43 cases (38%). In 71 cases (38.4%; 95% CI 31.3–45.8), ventriculomegaly was bilateral, size of both ventricles being measured between 10 and 12 mm (Figure 1).
MRI was performed at a mean (±SD) GA of 33.2 weeks (±2.03, range 30–38 weeks). In all cases, the time interval between referring ultrasound scan and MRI did not exceed 2 weeks. In 43 cases (23.2%; 95% CI 17.4–30), the ventricles were measured to be less than the 10-mm threshold in MRI and no abnormality was found. MRI confirmed the presence of 10- to 12-mm isolated fetal ventriculomegaly in 106 cases (57.3%; 95% CI 49.8–64.5), 74 (70%) of which being unilateral ventriculomegaly and did not provide additional information in 101 of these 106 cases. Conversely there were associated abnormalities in five cases (4.7%). Finally, in 36 cases (19.5%; 95% CI 14–25.9), the size of the ventricles were measured to be more than 12 mm in MRI, 23 (64%) of which being unilateral ventriculomegaly. In 30 of these 36 cases, ventriculomegaly remained isolated whereas it was associated with other abnormalities in six cases (16.6%). These data are summarised in the flow chart (Figure 1).
In all the cases with no other abnormalities at MRI, follow up during the end of pregnancy as well as neonatal clinical examination were uneventful except for one case of unexplained in utero fetal death. The 11 cases where MRI did find additional abnormalities as compared with ultrasound scan are detailed in Table 1, including follow up and outcome. There were abnormalities of the midline (Figure 2), the white matter (diffuse or focal) or migration (Figure 3). TOP was performed in 6/11 cases and MRI abnormalities were confirmed by the pathological examination. Follow up is available for the five other pregnancies. Postnatal imaging confirmed antenatal MRI findings in all cases. One child (case 2) presents cognitive disorders. In case 4, subependymal heterotopias are responsible for seizures. In cases 5, 6 and 8, postnatal follow up is uneventful for these children who are still very young, under 1 year of age.
Table 1. Cases in which MRI provided additional information
GA at MRI (weeks)
Ultrasound scan findings
Ventricles at MRI
Additional findings at MRI
ADC, apparent diffusion coefficient; LA, left atrium; RA, right atrium.
Left ventriculomegaly (RA = 7 mm, LA = 12 mm)
Left ventriculomegaly (RA = 7 mm, LA = 12.5 mm)
T2 hypersignal of the frontal white matter with increased ADC and left germinolysis
TOP: left germinolysis with diffuse hypoxic–ischaemic lesions of the white matter.
Bilateral ventriculomegaly (RA = 11 mm, LA = 10.7 mm)
Left ventriculomegaly (RA = 6.5 mm, LA = 12.5 mm)
Irregularity of the right ventricle wall
Born: 3 years old. MRI at 2 years: RA = 18 mm, LA = 23 mm. Abnormal white matter and cognitive disorders.
Bilateral ventriculomegaly (RA = 11 mm, LA = 10 mm)
Right ventriculomegaly (RA = 13.7 mm, LA = 9.6 mm)
Thin and short corpus callosum
TOP: Corpus callosum abnormalities confirmed.
31 + 4
Right ventriculomegaly (RA = 10.8 mm, LA = 9.8 mm)
Right ventriculomegaly (RA = 10 mm, LA = 7.5 mm)
Multiple subependymal heterotopias
TOP refused: subependymal heterotopias confirmed by postnatal MRI. Seizures.
37 + 2
Left ventriculomegaly (RA = 9 mm, LA = 11 mm)
Left ventriculomegaly (RA = 7 mm, LA = 10 mm)
Cavitations of the white matter with increased ADC
TOP refused: cavitations confirmed by postnatal MRI. Normal follow up at 9 months.
34 + 2
Left ventriculomegaly (RA = 9 mm, LA = 10 mm)
Left ventriculomegaly (RA = 9 mm, LA = 10.6 mm)
Born: normal follow up at 1 year.
Left ventriculomegaly (RA = 5 mm, LA = 10 mm)
Left ventriculomegaly (RA = 7.7 mm, LA = 13 mm)
Multiple subependymal heterotopias
TOP: multiple subependymal heterotopias.
Bilateral ventriculomegaly (RA = 11 mm, LA = 10 mm)
Bilateral ventriculomegaly (RA = 12.6 mm, LA = 10.4 mm)
White matter T2 hypersignal with increased ADC
Born: normal follow up at 5 months.
Right ventriculomegaly (RA = 11.5 mm, LA = 5 mm)
Right ventriculomegaly (RA = 11.6 mm, LA = 5.3 mm)
Large cavitation beside the right temporal horn
TOP: abnormalities confirmed by pathologic examination.
Bilateral ventriculomegaly (RA = 10 mm, LA = 10.5 mm)
Bilateral ventriculomegaly (RA = 14.5 mm, LA = 10.5 mm)
Frontal sulcation abnormality
TOP: abnormalities confirmed by pathologic examination.
Right ventriculomegaly (RA = 11 mm, LA = 8.6 mm)
Right ventriculomegaly (RA = 11.8 mm, LA = 8.7 mm)
TOP: abnormalities confirmed by pathologic examination.
The follow up of the children for whom antenatal MRI confirmed the diagnosis of IMV is not the purpose of this study and will be detailed in another article.
Mild cerebral ventriculomegaly affects 0.15–0.7% of pregnancies.15,16 Diagnosis of IMV is a dilemma in prenatal counselling: indeed, despite normal outcome in most cases of IMV, it may also be a marker for various pathological processes that have to be looked for. Several studies have emphasised the potential role of MRI in fetal brain analysis and the information that MRI can provide as an adjunctive tool to sonography.9,17–20 Although some studies have evaluated the potential additional information that may be provided by MRI in the evaluation of ventriculomegaly,21 the contribution of MRI in the evaluation of borderline 10- to 12-mm IMV has not been specifically studied and remains controversial.
In our cohort, MRI confirmed the mild (10–12 mm) fetal ventriculomegaly in most cases (106/185). This is consistent with previous report that demonstrated the close agreement between ultrasound scan and MRI data.14 Transient variation in ventricular size throughout gestation22 as well as minor discrepancies of no more than 2 mm can easily explain those cases in which MRI found ventricles below the 10-mm threshold (43/185) or those cases in which MRI eventually found ventricles larger than 12 mm (36/185). In the former cases, however, the role of MRI was to reassure the parents regarding to the absence of brain anomalies. In the latter cases, MRI found associated abnormalities in six cases (16.6%). This emphasises that the risk of abnormality is higher in that group of fetuses with ventricle larger than 12 mm and that MRI should more readily be considered in that group.12,13
More importantly, MRI diagnosed associated abnormalities in 5 of the 106 cases where the 10- to 12-mm ventriculomegaly was confirmed. It demonstrates that, after targeted fetal ultrasound screening, one could expect that MRI will provide information in 4.7% of cases of 10- to 12-mm IMV confirmed by MRI. This is an important information because MRI is a high-cost and time-consuming examination that is not available in all prenatal centres. A study evaluating the contribution of MRI21 in 36 cases of IMV between 10 and 15 mm found MRI useful in three cases (8.3%). Other studies using ultrasound scan alone have found less association with other abnormalities when IMV was between 10 to 12 mm only12 and have found lower rates of abnormal development (3.8%) with IMV less than 12 mm than with IMV between 12 and 15 mm (14%).13 These results are consistent with ours, obtained from a much larger cohort and in which MRI was useful in 5/106 (4.7%) of the confirmed 10- to 12-mm ventriculomegalies and 6/36 cases (16.6%) in which ventricles were eventually measured between 12 and 15 mm on MRI. In our series, there was evidence of more frequent abnormalities in ventriculomegaly larger than 12 mm as compared with the 10- to 12-mm group (6/36 versus 5/106; P= 0.03 Fisher's exact test).
Unilateral mild ventriculomegaly is usually considered as a separate entity from bilateral ventriculomegaly, with a lower risk of perinatal mortality and morbidity.23–26 Our study does not confirm such findings. Overall, in the 11 cases in which MRI found associated abnormalities, there were 9/11 unilateral ventriculomegalies. In the 106 cases where 10- to 12-mm ventriculomegaly was confirmed by MRI, 5 of the 74 unilateral cases had associated abnormalities (6.7%) whereas none of the 32 bilateral cases was associated with other abnormalities (Fisher’s exact test P= 0.32). In the 36 cases with size of ventricles more than 12 mm, 4 of the 23 unilateral cases had associated abnormalities (17.4%) whereas other abnormalities were found in two of the bilateral cases (15.3%) (Fisher’s exact test P= 1). This shows that differentiation between unilateral and bilateral ventriculomegaly may be less important for outcome, especially in 10- to 12-mm ventriculomegalies in which we found a, nonsignificant, higher proportion of associated abnormalities in the unilateral cases.
In the cases with other abnormalities, it must be taken into account that these cerebral abnormalities were detected with MRI very late in pregnancy. TOP is not allowed so late in many countries and such late TOP may have poor psychological impact on parents. As some abnormalities could have been detected earlier by MRI (subependymal heterotopias, thin and short corpus callosum), we suggest that MRI might be performed earlier in some cases according to the suspected abnormalities and the country legislation. Studies need to be undertaken to see whether earlier MRI would obtain results comparable with those reported in our cohort.
Using our policy of routine MRI in sonographically diagnosed IMV of 10–12 mm, MRI made it possible to reassure a large proportion of parents, and we found associated abnormalities in 5 of the 106 cases of 10- to 12-mm ventriculomegalies confirmed by MRI. Moreover, our study shows that without offering MRI, counselling would have been done in 185 sonographically diagnosed IMV without taking into account the associated abnormalities overlooked in 11/185 (6%) cases. Therefore, all these parents should be informed that there is a 6% risk that the fetus might have an invisible (with ultrasound scan) serious problem. In some situations, such prenatal management could lead to request for TOP where the infant might well be healthy. MRI is expensive and not readily available in most centres worldwide. Large prospective studies remain mandatory to prove additional advantages of MRI in a statistically significant number of patients. Before that can be demonstrated, it seems reasonable that MRI should remain an investigational tool, restricted to selected clinical situations in which the results are expected to modify case management. When isolated ventriculomegaly of 10 to 12 mm is suspected with ultrasound scan, our data suggest that when the finding is confirmed with MRI this could be expected in around 5% of cases. Therefore, the policy of routine MRI in such cases should depend on prenatal centres’ priorities.