Classic “PCH” genes are a rare cause of radiologic pontocerebellar hypoplasia

Background: The term Pontocerebellar Hypoplasia (PCH) was initially used to designate a heterogeneous group of fetal-onset genetic neurodegenerative disorders. As a descriptive term, PCH refers to pons and cerebellum of reduced volume. In addition to the classic PCH types, many other disorders can result in a similar imaging appearance. Objective: To review imaging, clinical and genetic features and underlying etiologies of a cohort of children with PCH on imaging. Methods: We systematically reviewed brain images and clinical charts of 38 patients with radiologic evidence of PCH. Results: Our cohort included 21 males and 17 females, with ages ranging between 8 days to 15 years. All individuals had pons and cerebellar vermis hypoplasia, and 63% had cerebellar hemisphere hypoplasia. Supratentorial anomalies were found in 71%. An underlying etiology was identi�ed in 65% and included chromosomal (21%), monogenic (34%) and acquired (10%) causes. Only one patient had pathogenic variants in a “classic” PCH gene. Outcomes were poor regardless of etiology, though no one had regression. Approximately one third of patients deceased at a median age of 8 months. All individuals had global developmental delay, 50% were non-verbal, 64% were non-ambulatory and 45% required gastrostomy feeding. Conclusion: Radiologic PCH has heterogenous etiologies and the “classic” PCH genes underlie only a minority of cases. Broad genetic testing, including chromosomal microarray and exome or multigene panels, is recommended in individuals with PCH-like imaging appearance. Our results strongly suggest that the term PCH should be used to designate radiologic �ndings, and not to imply neurogenerative disorders.


Introduction
The term Pontocerebellar hypoplasia (PCH) was initially used by Peter Barth in 1990 to designate a heterogeneous group of prenatal-onset genetic neurodegenerative disorders with autosomal recessive inheritance [1].The rst attempt to classify these disorders was made in 1993 and included only two subtypes [2].Since then, signi cant progress in the domains of neuroimaging and neurogenetics has resulted in an increasing number of new subtypes added to the original classi cation.Currently, there are 17 types of PCH described associated with over 20 different genes, many of which have important functions in RNA processing and protein translation, and mitochondrial regulatory chain regulation [3].Patients on the severe end of the clinical spectrum present with severe hyperre exia, contractures, a progressive course and early death due to central respiratory failure [4].Patients on the less severe end of the spectrum present with early hyperre exia, developmental delay, feeding problems, and eventually develop spasticity and involuntary movements.The characteristic MRI ndings are cerebellar hypoplasia with superimposed atrophy, signi cant reduction of pontine prominence, normal or slightly reduced posterior fossa size and variable cerebral involvement including atrophy and delayed myelination [5].
As a descriptive term, PCH refers to pons and cerebellum of reduced volume on brain imaging, not necessarily implicating that these structures are affected equally.In fact, in addition to the classic PCH types, many other disorders can result in a similar radiologic appearance, including acquired etiologies as described in cerebellar disruptions associated with extreme prematurity.Other genetic etiologies of PCH include neurometabolic disorders (such as congenital disorder of glycosylation type 1A), chromosomal abnormalities, and monogenic conditions such as tubulinopathies, CASK, RELN, and VLDRL -related disorders [5,6].Clinically, the presentation of these other "non-classic" PCH is also variable, with most showing a static course over time [4].However, it is yet not clear what proportion of radiologic PCH is caused by classic PCH genes in comparison with the other conditions that may result in similar imaging ndings.
Our study aims to review imaging, clinical and genetic features as well as the underlying etiologies of a cohort of 38 children in whom brain imaging revealed radiologic evidence of PCH.

Study design
We performed a retrospective chart review of patients with radiologic evidence of pontocerebellar hypoplasia from two pediatric tertiary care centers (Montreal Children's Hospital and Sainte Justine's Hospital, Quebec).This study was approved by our Institutional Research Ethics Board.(MP-37-2021-6580).

Patients
PCH cases diagnosed between 1995 and 2021 at our two institutions were identi ed by searching radiology reports for description of hypoplasia of the pons and cerebellum (Montreal Children's hospital) or discharge summaries for a diagnosis of q brain malformation (Sainte Justine Hospital).Individuals who did not ful ll the radiologic criteria for PCH as outlined below, were excluded.

Clinical data
Demographic data including gender, age and clinical data such as perinatal history, neurodevelopmental outcomes, physical examinations, and genetic results were collected retrospectively from the medical records.

Neuroimaging review
Magnetic resonance imaging (MRI) is the modality of choice for the detailed evaluation of the posterior fossa of the brain.All children included in this study had either a 1.5T or 3T MRI imaging.Most patients (34/38) had complete MRI studies (coronal, sagittal, and axial sections) which were re-evaluated by an experienced pediatric neuroradiologist (CSM).For the remaining 4 patients, neuroimages were unavailable and the full radiologic report was used to assess for the inclusion criteria.
The mid-sagittal view was used to assess the pons, midbrain and cerebellar vermis.The normal craniocaudal (CC) length of the pons is approximately 1.5 to 2 times the CC length of the midbrain.The pontine protuberance is considered normal when its anterior-posterior (AP) diameter is bigger than the AP diameter of the midbrain [4][5][6].In our study, the pons was considered hypoplastic if the ratio of the length of CC pons:CC midbrain was less than 1.5 and/or the AP dimension length of the pons was smaller than the AP diameter of the midbrain.(Figure 1a).
Although there are no standard values de ning vermis hypoplasia, a recent paper provided reference biometric data using simple and reproducible measurements [7].We plotted vermis height (VH) and AP diameter (AP-V) values against the reference biometric data for patients aged 3 months or more (available online http://ci2c.fr/vermis.php); the vermis was considered hypoplastic if the VH or AP-V were less than the 3rd percentile based on age and gender.For patients younger than 3 months, the vermis assessment was based on the experience of the neuroradiologist.(Figure 1b) The morphology and size of the cerebellar hemispheres were assessed using the axial and coronal images.In axial view, the normal cerebral folia run parallel towards the calvarium in an "onion-like" con guration [4].The cerebellar malformation can be classi ed as hypoplastic (i.e., reduced cerebellar volume), dysplastic (i.e. with abnormal cerebellar foliation, ssuration, and architecture) or hypodysplastic (i.e.combined features of hypoplasia and dysplasia) [5].Two patterns have been used to describe the appearance of the hypoplastic cerebellum, appreciated on the coronal view: the dragon y pattern, with more severe involvement of the cerebellar hemispheres relative to the vermis [5]; and the butter y pattern, with proportionate involvement of both hemispheres and vermis [8].
We also systematically noted presence of cerebellar atrophy, supratentorial abnormalities, white matter, cortex, corpus callosum, basal ganglia and ventricular anomalies.

Demographic data
We identi ed 49 cases (including 8 cases with prenatal MRI) with a description of PCH in their radiology reports.Of these, 11 were excluded (7 did not ful l our radiologic criteria for PCH, 3 had no clinical data, 1 underwent termination of pregnancy).Of the total of 38 cases included.There were 21 males and 17 females, with ages ranging between 8 days to 15 years.The average age at the last assessment was 4 years (median of 3 years, ranging from 8 days-15 years).The median follow-up period was 2 years and 10 months (ranging from 13 days to 15 years).Fourteen individuals (36%) were deceased, with an average age of death of 17 months (median of 8 months, ranging from 8 days -5 years).The main causes of death were related to complications expected within these patients' clinical presentations: uncontrolled seizures, recurrent episodes of aspiration pneumonias and cardiopulmonary deterioration.

Imaging Findings
A summary of the imaging ndings is presented in Table 1.All 38 individuals in our cohort had hypoplasia of the pons and the cerebellar vermis (38/38, 100%).The VH and AP-V values were plotted against the reference biometric data [7] in the 14 individuals with brain imaging performed after age 3 months.All (14/14, 100%) had VH less than 3 rd percentile, while 11 individuals (11/14, 78,5%) had AP diameter less than the 3 rd percentile.

Identi ed etiologies
An underlying etiology was identi ed in 65% (25/38) of cases (see Table 2 and supplementary table 1): 21% (8/38) had chromosomal abnormalities, 34% (13/38) had monogenic disorders and 10% (4/38) had acquired etiologies (prematurity n=3 and hypoxic-ischemic encephalopathy n=1).Interestingly, only one of our patients had pathogenic variants in classic PCH genes.There were 3 individuals with causal POMGnT1 pathogenic variants, and two with CASK gene pathogenic variants.In addition, causal pathogenic variants were identi ed in a single patient for the each of following genes: AIMP1, ASPM, CHD7, DHCR7, NFIX, OFD1 and VLDLR.The individual with causal AIMP1 variants was previously published [9].There were 13 individuals with no identi able etiology (34%, 13/38).All these individuals underwent genetic testing, though there was wide variability in which tests were performed.Karyotype or chromosomal microarray (CMA) was performed in 11, PCH gene panel in 3, and other genetic test (e.g.speci c gene, other gene panel or FISH studies) in 7.Among these patients, there was one individual with cerebellar disconnection syndrome who had multiple congenital anomalies (urinary bladder diverticulum, pectus carinatum, fused ribs, and several hemivertebra) and negative research whole exome sequencing and CMA.Furthermore, one individual was diagnosed with Chiari-II malformation and another one with Pierre-Marie-Robin syndrome.

Pregnancy and perinatal data
At the time of delivery, the average maternal age was 30 years and 7 months (median 31, ranging from 18 -43 years).Six of the mothers were diagnosed with gestational diabetes mellitus, 2 of whom required insulin treatment.Another mother was previously diagnosed with diabetes mellitus type II and also required insulin.Other maternal co-morbidities included antiphospholipid syndrome (n=2), and Oral-Facial-Digital syndrome (n=1).
Prenatal MRIs were performed for seven foetuses due to abnormal ultrasound ndings, and PCH was reported in only four.One foetus had a prenatal diagnosis of trisomy 21 based on amniocentesis karyotype.

Developmental and Clinical Outcomes
Developmental and clinical outcomes are also summarized in Table 3. Outcomes were very poor, as all of the individuals in our cohort had a global developmental delay.None of them presented developmental regression, 50% were non-verbal (13/26) and 64% were non-ambulatory (18/28).In addition, 45% of individuals (14/31) required feeding through a gastrointestinal tube.
Epilepsy was diagnosed in 36% (13/36) of individuals.Of these, 6 had drug-resistant seizures and were treated with two or more antiseizure medications.
Musculoskeletal anomalies were noted in 18% (7/38) of individuals.The most common ones were arthrogryposis (n=2) and spinal stenosis (n=2).Clinodactyly and bilateral metatarsus adductus were present in one individual each.
Renal anomalies were present in 13% (5/38).Horseshoe kidney, karyomegalic interstitial nephritis, bilateral congenital hydronephrosis, pyelectasis, and urinary bladder diverticulum, were each noted in one individual.Genital abnormalities were the less common, found in only 8% (3/38) of our cohort, and included ambiguous genitalia, cryptorchidy, and hypoplastic labia minora noted in one patient each.

Correlation between etiology, radiologic ndings and clinical outcomes
In general, there was no correlation between etiology, radiologic ndings and outcomes (note that sample size precluded reliable statistical testing).
The underlying etiology could at times be suspected on brain imaging based on the pattern recognition of constellations of ndings.For example, MEB disease were often suspected based on the appearance of the brainstem (e.g.midline pontine cleft) and its associated cortical anomaly; acquired causes were suspected based on the presence of blood, or the pattern of injury of the cerebellum.However, for the great majority of cases, the underlying etiology was not readily apparent on brain imaging.Note that one of the only two individuals with a dragon y appearance on the coronal section (Figure 2 b,n) had TSEN54 pathogenic variants, suggesting that the dragon y appearance of the cerebellum may be more suggestive of a "classic PCH" disorder.
Post-natal admission and neonatal issues were frequent and common to all individuals regardless of etiology subgroup.Individuals with an acquired PCH were more likely to be born prematurely compared to the remained of the cohort (75% vs 26%), and none of them had deceased by the time of the last assessment, in comparison with an overall 36% death rate for the remaining groups.
Severities of developmental impairments appear relatively similar across etiology groups, with perhaps less severe impairment in the acquired subgroups, as these patients were less likely to be non-verbal, nonambulatory and G-tube fed compared to the remainder of the cohort, though there were small patient numbers to be con dent of this observation.In addition, similar proportion of individuals deceased in each etiologic category.

Discussion
In this study, we reviewed the clinical, radiologic and genetic ndings of a cohort of 38 individuals with radiologic evidence of PCH.We were able to identify the underlying etiology in 65%: 55% were genetic (chromosomal abnormality or single monogenic disorder) and 10% were acquired.Only one individual from our entire cohort had pathogenic variants in a "classic" PCH gene.
Our study highlights that the term pontocerebellar hypoplasia (PCH) can be misleading.Though "PCH" is traditionally used to refer to prenatal-onset genetic neurodegenerative disorders, our results show that PCH on imaging is associated with a large group of heterogeneous and static disorders, including monogenic and acquired etiologies.Furthermore, it is now well-established that some "classic PCH" subtypes, such as PCH8 and PCH11, have a nonprogressive course [8, 10,11].Therefore, the PCH designation does not imply a genetic disorder with a progressive neurodegenerative course as it was initially described.
The single patient from our cohort with a classic genetic neurodegenerative PCH subtype carried compound heterozygous pathogenic variants in TSEN54, associated with PCH types 2 and 4 [12,13]: the common missense variant c.919G > T (p.A307S) [12], as well as a frameshift variant c.1032dupC (p.Lys345GlnfsX90).This individual had a severe phenotype, including hypertonia, congenital contractures, dependence on mechanical ventilation, and early death, which was compatible with PCH4.Previous reports have found that patients who were homozygous for the common p.A307S variant had relatively milder phenotypes and a clinical diagnosis of PCH2, whereas those who were compound heterozygous showed more severe phenotypes in keeping with PCH4.[13,14] This patient was also one of the only two individuals from our cohort who had a dragon y appearance of the cerebellum on imaging.(Fig. 2a-c).In classic genetic neurodegenerative PCH, cerebellar hemispheres are usually compromised, and the vermis is relatively spared, giving the appearance of the dragon y pattern [8].This radiologic nding is more frequently seen in association with the p.A307S variant, found in our patient [14].This suggests that the dragon y pattern, although not speci c, might help to narrow down the differential diagnosis towards the classical neurodegenerative PCH subtypes, especially if there is no history of prematurity.
However, we also identi ed pathogenic variants in two genes, NFIX and DHCR7, that have not been previously associated with PCH, though were related to different cerebellar malformations.
NFIX encodes a CCAAT-binding transcription factor and pathogenic variants in this gene are associated with Marshall-Smith Syndrome, which is characterized by accelerated skeletal maturation, respiratory di culties, facial dysmorphic features, and variable intellectual disability.[18,19] Although it is not common, hypoplastic cerebellum and enlarged cisterna magna were already reported in association with this syndrome.[20] Our patient had radiologic PCH with delayed myelination.(Fig. 2d-f) Biallelic pathogenic variants in DHCR7 are associated with Smith-Lemli-Opitz syndrome (SLOS), a neurodevelopmental disorder caused by de ciency of 7-dehydrocholesterol reductase [21].It is inherited in an autosomal recessive manner and characterized by growth retardation, microcephaly, intellectual disability, distinctive facial features, postaxial polydactyly, and syndactyly of the 2nd and 3rd toes.The most common MRI brain ndings in patients with SLOS are abnormalities of the septum pellucidum and corpus callosum, but cerebellar abnormalities, such as hypoplasia of the cerebellar tonsils and/or inferior vermis have been reported [22].Our patient had hypoplastic pons, cerebellum hemispheres, and vermis with a resembling butter y appearance on the coronal view, as well as mild ventriculomegaly.(Fig. 2g-i) In our study, 21% of individuals (8/38) presented chromosomal abnormalities.Three patients had trisomy 18, one had trisomy 21, one individual had tetrasomy 9q34.11,two had Cri Du Chat Syndrome and one had trisomy 9 mosaicism.Small pons and cerebellar hypoplasia have already been recognized in individuals with trisomy 18 [23].Neuroimaging studies in individuals with Trisomy 21 show smaller overall brain volume, with disproportionately smaller cerebellar volumes and relatively larger subcortical grey matter volumes [24].In addition, hypoplastic pons and cerebellar vermis have been observed prenatally in fetuses with Trisomy 21 [25].Cri Du Chat Syndrome is also known to be associated with pontocerebellar hypoplasia [26].Trisomy 9 mosaicism, on the other hand, was previously associated with different brain abnormalities such as hydrocephalus and Dandy-Walker syndrome but, an association with PCH imaging appearance, to our knowledge, has never been reported.[27] (Fig. 2j-l) Acquired, non-genetic etiologies were also identi ed in our cohort.Prematurity is a known cause of disruptive PCH.This association is not surprising since, from 28 gestational weeks gestation to term, there is a more than a 30-fold increase in the surface of the cerebellar cortex.This rapid proliferation is highly energy-demanding and places the developing cerebellum at high risk for injury [28][29][30].Amongst individuals with acquired causes of PCH in our cohort, 3/4 were extreme or very premature babies.One patient had perinatal asphyxia with hypoxic-ischaemic encephalopathy and the MRI at day 14 of life showed hypoplastic pons and cerebellum (vermis and hemisphere) among the expected hypoxicischaemic changes.
The developmental outcome in our cohort was invariably poor as all individuals showed global developmental delay, and many individuals were non-verbal, non-ambulatory, and required feeding through G-tube.One third of individuals of our cohort deceased with a median age of death of 8 months.Poor neurodevelopmental outcome is consistent with other PCH clinical pro les, in which severe microcephaly, seizures, pyramidal/extrapyramidal involvement, and poor psychomotor development are common features [14,31].However, in contrast to many of the "classic PCH" phenotypes, none of our patients exhibited any clinical regression.Our cohort showed a high prevalence of associated cerebral malformations and extra-neural anomalies.
Our study has a few limitations.The images were unable for review for 4 individuals, in which cases the data was extracted from radiology reports.Given its retrospective nature, some clinical details were missing, especially regarding follow-up examination ndings.Finally, there was inconsistent genetic testing performed among the individuals, which may underestimate the identi cation of the underlying etiology.

Conclusion
In summary, our study reveals that radiologic PCH is associated with a broad range of etiologies.These include heterogenous monogenetic causes and chromosomal anomalies.The "classic" PCH genes underlie only a minority of cases.Aside from disorders frequently associated with PCH such as Muscleeye brain disease and CASK-related disorder, PCH can be an infrequent feature of many neurodevelopmental disorders.Broad genetic testing, including chromosomal microarray and exome sequencing or multigene next generation panels, is highly recommended in individuals with PCH-like imaging appearance.Outcomes in individuals with radiologic PCH was invariably poor, although none of the patients in this cohort presented developmental regression.Finally, the result of our study strongly indicates that the term "PCH" should be used to designate the radiologic ndings, and not to imply a neurogenerative disorder.

Declarations Ethical Approval
This study was approved by the McGill University Research Ethics Board.(MP-37-2021-6580) and the CHU Sainte Justine research Ethics Board (MEO-37-2021-2915).As per our research ethics board protocol, signed consents for participation and publication by the patients and families was not required as this retrospective study does not include information or images that may identify the subjects.

Competing interests
Authors declare no competing interests.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download. Supplementaltable1.docx

Figures Figure 1
Figures

Table 1 :
Summary of brain imaging ndings in individuals with pontocerebellar hypoplasia on imaging

Table 2 :
Etiologies identi ed in our PCH cohort Table of clinical features of our PCH cohort