Prenatal diagnosis of Aicardi syndrome based on a suggestive imaging pattern: A multicenter case‐series

Abstract Objectives To characterize a suggestive prenatal imaging pattern of Aicardi syndrome using ultrasound and MR imaging. Methods Based on a retrospective international series of Aicardi syndrome cases from tertiary centers encountered over a 20‐year period (2000–2020), we investigated the frequencies of the imaging features in order to characterize an imaging pattern highly suggestive of the diagnosis. Results Among 20 cases included, arachnoid cysts associated with a distortion of the interhemispheric fissure were constantly encountered associated with complete or partial agenesis of the corpus callosum (19/20, 95%). This triad in the presence of other CNS disorganization, such as polymicrogyria (16/17, 94%), heterotopias (15/17, 88%), ventriculomegaly (14/20, 70%), cerebral asymmetry [14/20, 70%]) and less frequently extra‐CNS anomaly (ocular anomalies [7/11, 64%], costal/vertebral segmentation defect [4/20, 20%]) represent a highly suggestive pattern of Aicardi syndrome in a female patient. Conclusion Despite absence of genetic test to confirm prenatal diagnosis of AS, this combination of CNS and extra‐CNS fetal findings allows delineation of a characteristic imaging pattern of AS, especially when facing dysgenesis of the corpus callosum.


| INTRODUCTION
Aicardi syndrome (AS) is a rare developmental encephalopathy, characterized by the classic triad of infantile spasms, agenesis of the corpus callosum (ACC), and chorioretinal lacunae, as initially described by Jean Aicardi in 1965. 1 Its prevalence is estimated between 1/93,000 and 1/167,000 living births, 2 mainly affecting females, with few cases described in Klinefelter patients (47,XXY). 3,4 An X-linked dominant transmission has been postulated, but the exact genetic etiology remains unknown. 5,6 Currently, diagnosis of AS is based on a combination of clinical and imaging features. In addition to the classic triad, major features have been described, mainly using cerebral MR imaging (MRI): cortical malformations (polymicrogyria), periventricular and subcortical heterotopias, arachnoid cysts around the third ventricle, choroid plexus cysts, and optic nerve/disc coloboma or hypoplasia. Additional features can also be suggestive of the diagnosis especially when the classic triad is incomplete: cerebral asymmetry, vertebral and rib malformations (hemivertebrae, bifid ribs), microphthalmia, vascular malformations, split-brain EEG (asynchronous multifocal epileptiform discharges with interhemispheric dissociation), and facial dysmorphism including prominent premaxilla with upturned nasal tip and sparse lateral eyebrows. 6 Diagnosis of AS based on the classic triad cannot be extrapolated to prenatal diagnosis as seizures and chorioretinal lacunae cannot reliably be diagnosed before birth. Indeed, AS can only be suspected if major or/and additional features are seen in a female fetus who has normal genetic testing results. Therefore, AS is rarely diagnosed prenatally and only a small number of cases have been reported. 7 Our goal was to confirm and complete a prenatal imaging pattern suggestive of the diagnosis, based on the most frequent prenatal imaging features encountered in a large series of AS cases suspected prenatally with postnatal confirmation.

| Study settings and participants
This study is a retrospective analysis of cases of AS prenatally (Houston, USA), Massachusetts General Hospital (Boston, USA). The study was approved by the local Institutional Review Boards according to each country's requirements.

| Inclusion and exclusion criteria
Criteria for inclusion in our study were the following: (1) Suspicion of AS after a prenatal evaluation in a tertiary center; (2) Prenatal and pathological/postnatal data and imaging available for review;

| Data collection
Anonymous clinical data and prenatal/postnatal imaging were collected by the principal investigators (Léo Pomar and Yvan Vial), through a datasheet via the secure messaging system of Lausanne University Hospital. Each referent kept the source file allowing to link the identity of the patients of their center with the anonymized number used for the study.

| Analysis of imaging data
Prenatal sonographic and MR iconography was reviewed (Léo Pomar, Yvan Vial, and Laurent Guibaud) to identify features associated with AS. The quality of the examinations was evaluated based on ISUOG guidelines. [10][11][12] Ultrasound examinations were considered to be of "good" quality when the trans-vaginal or transabdominal high-frequency approach was performed using axial, sagittal and coronal sections of the fetal brain, sections of the eyeballs, optic nerve and retina, with good illustration of the anomalies. Quality was considered "sufficient" if the iconography included an examination of the brain in three orthogonal planes, as well as an examination of the eyeballs with good illustration of the anomalies found, but without detailed description of the eyes (optic disc, retina, optic nerve). The quality was considered "insufficient" when appropriate examination of the brain or of the eyes was not available.
MRIs were considered of "good" quality if a three-plane examination of the fetal brain, eyeballs, optic nerves and retina was available in at least one T2 weighted-contrast sequence (other sequences examined: T1 weighted-contrast, single-shot high resolution, diffusion weighted or tensor imaging). MRIs were considered of "sufficient" quality if a three-plane examination of the fetal brain and eyeballs was available. MRI scans were considered to be of "insufficient" quality if three-plane fetal brain and/or eyeball scans were not available.

| Prenatal imaging
All included patients were referred for suspected corpus callosum anomalies and/or cerebral cysts and/or fetal ventriculomegaly in the second or third trimester. The median gestational age at the time of the referral ultrasound was 23w (18 to 37w). Fifteen patients underwent fetal brain MRI, performed at a median gestational age of 31w (22 to 36w). Description of MR protocols, as well as data on prenatal sonographic and MR examination quality, are presented in Table S2.
Choroid plexus cysts were described in three cases (15%; Cases

| Confirmation of AS diagnosis on pathological/ post-natal data
Pathological and postnatal features are summarized in Table 1.

| Electroencephalograms (EEGs)
In the seven cases in whom data were available, EEG was abnormal.

| Genetics data
Karyotyping and CGH array, performed in 19 and five cases respectively, identified no chromosome number abnormalities nor abnormal copy number variations. Exome sequencing, performed in the five most recent cases, found no known pathogenic variants.

| DISCUSSION
The imaging findings reported in our series are in concordance with and heterotopias (periventricular, in the corona radiate and cerebellar). The majority of patients (95%) had also intra-cranial cysts, mainly interhemispheric, and fossa posterior anomalies. 13 The frequency of cerebellar anomalies was higher in their series, two third and half of cases presenting dysplastic/hypoplastic cerebellar hemispheres and enlarged cisterna magna, respectively. In the other series, the frequency of posterior fossa anomalies was close to our personal data (6%-63%). [14][15][16][17][18] The localization and description of intra-cranial cysts in our series was consistent with previous literature data, with cysts located predominantly in the interhemispheric fissure in the frontal area, or between occipital lobes or, to a lesser extent, in the posterior fossa. 13,18,19 These cysts correspond to type 2b (glioependymal, associated with heterotopia, and polymicrogyria) and 2d (arachnoid cysts with CSF-like signal), according to the classification of Barkovich et al. 20 Distortion of the interhemispheric fissure was present in most of the previously published cases, but this feature was rarely emphasized by the authors, preferring to highlight an asymmetry of the cerebral hesmispheres.
Although this feature is not specific to AS, 21 this distortion is important to underline, as an easily accessible feature since interhemispheric fissure is included in the analysis of the anterior complex when performing routine sonographic exam. 22 Frequency of cortical/ migration anomalies is close to 100% in all series, including mainly nodular and diffuse heterotopia and polymicrogyria. 13,18 In cases 3, 11, and 20, prenatal imaging identified dysgyria or delayed sulcation, which was related to polymicrogyria on postnatal imaging or autopsy.
Since postnatal imaging offers more details on cortical anomalies than prenatal imaging, one should note that polymicrogyria may vary from focal to diffuse forms but was more frequently focal in our study. Therefore, when AS is prenatally suspected, one should examine closely to the cortical ribbon to look for any focal cortical dysplasia, which can be easily overlooked. Use of DTI in Case 8 revealed disruption of cortical white matter organization close to polymicrogyria, which has been reported as a specific feature of AS by Wahl et al. in two patients. 23 In some patients, we also reported nodular heterotopia of the basal ganglia area, found in up to 76% of cases in postnatal series. 24 In our series, distinctive chorioretinal lacunae were confirmed in all cases but we may have overestimated the frequency of this anomaly, as it was a nonexclusion criterion. These lacunae were present in 88%-100% of cases in other series. 6 31 However, chorioretinal lacunae and colobomas are also described in oro-faciodigital 9 syndrome, but interhemispheric cysts are not described in this syndrome and we did not observe digital anomalies in our cases. 32 We should also underline that the small proportion of patients who underwent WES in our study did not provide any new highlights on the genetics of AS. 5,[33][34][35] The other bias related to the long study period is the heterogeneous quality of prenatal imaging, overlooking subtle prenatal features in older cases, especially regarding cortical and ocular anomalies.

| CONCLUSION
Finally, despite the lack of a genetic test to enable a prenatal diagnosis of AS, prenatal imaging can identify a combination of CNS and extra-CNS features which form a characteristic imaging pattern of this syndrome, especially when facing intra-cranial cyst associated with distortion of the inter-hemispheric fissure and callosal dysgenesis.

ACKNOWLEDGMENTS
This study and the research behind it would not have been possible without the care and imaging provided by the teams at the centres involved, and we would like to thank every health professional involved in the care of these foetuses and children.
Open access funding provided by Universite de Lausanne.