Pontocerebellar hypoplasia: a review from 1912 to 2022

Abstract

Pontocerebellar hypoplasia is a rare neurodevelopmental disorder that results from differences in formation and function of the pons, cerebellum and cerebrum. It can be diagnosed prenatally or postnatally with a combination of clinical, neuroimaging and genetic data obtained over time. The diagnosis of pontocerebellar hypoplasia usually portends severe developmental delay, epilepsy and/or neurodegeneration in childhood. Here we perform a comprehensive review with the primary goal of evaluating published evidence addressing the clinical and genetic features of pontocerebellar hypoplasia by type and subtype. Secondly, we summarize neurodiagnostic patterns of pontocerebellar hypoplasia and demonstrate its spectrum. Finally, we provide recommendations in diagnosis, prognosis and management for the neurologist. To address these goals, we performed an extensive review of published literature from 1912 to 2022. We identified 191 publications by combining search results from PubMed, OMIM and cross-referenced bibliographies. Publications on developmental neuroanatomy, not pertaining to pontocerebellar hypoplasia or published in a foreign language were excluded. We performed both qualitative (1912–1993) and quantitative (1993–2022) analyses to understand the current classification of this disease as it pertains to genetic and neurodiagnostic features of pontocerebellar hypoplasia by type and subtype. Our review shows that the most reported types of pontocerebellar hypoplasia are 1, 2 and 6; less frequently described are 3, 4 and 9. Very few cases are described for all other subsequent pontocerebellar hypoplasia types. Mutations in TSEN54, RARS2, EXOSC3 and AMPD2 (genes that regulate RNA processing and basic cellular metabolism) are the most frequently reported pathological mutations in pontocerebellar hypoplasia. The neuroradiographic features of pontocerebellar hypoplasia are complex and evolve over time, affecting the pons, cerebellum, vermis, cortex and cerebral white matter. In conclusion, pontocerebellar hypoplasia is a rare neurodevelopmental disorder, often the result of genetic dysfunction in basic neural metabolism. The diagnosis conveys significant implications for the affected individual and their families and requires a combination of clinical, neuroradiographic, and genetic testing to best inform type/subtype categorization of pontocerebellar hypoplasia.

Kukulka et al. report that pontocerebellar hypoplasia is most often linked to mutations in TSEN54, RARS2, EXOSC3 and AMPD2, with Types 1, 2, and 6 being most common. They propose a dyadic gene-based classification to improve diagnostic clarity, guide clinical management and highlight research opportunities.

Graphical Abstract

Graphical Abstract.

Introduction

Pontocerebellar hypoplasia (PCH) is a clinical disorder that results from abnormal growth or maturation of the pons and cerebellum due to differences in neurodevelopmental programming. Concurrently, it can also indicate a relative lack of age-appropriate volumetric growth of the pons and cerebellum without known clinical implications. Neuroanatomical disorders of the pons and cerebellum likely result from chromosomal/genetic abnormalities affecting neural migration or metabolic derangements in the metencephalon at or after the sixth week of gestation. In the interest of lexicon clarity and consistency, there is a movement in the field to reserve the term ‘PCH’ for disorders with suspected or confirmed specific genetic mutations carrying a clinically significant phenotype. While the aetiology of PCH is rooted in numerous genetic mutations, the contribution of exogenous, endogenous, and behavioural factors constituting the neural exposome remains poorly understood, necessitating further investigation.¹

Recent advances in genetic sequencing have facilitated growing sub-categorization in diagnosing PCH as a disorder with sometimes variable clinical significance. Current literature has expanded to 16 PCH types, with additional A–F subtypes for PCH 1 and 2, yielding 26 different disease categories. With few exceptions, we now recognize that sequencing data are not always clinically predictive because mutations in the same gene can lead to different PCH subtypes (allelic heterogeneity) and mutations in different genes can lead to the same PCH subtype (locus heterogeneity).^2-6 In parallel, advances in neuroimaging and quantitative volumetric analyses have provided greater resolution of posterior fossa structures, further increasing diagnostic suspicion for more PCH subtypes. The heterogeneity and rarity of PCH have limited our ability to make substantive progress in the field towards disease-alleviating therapies. As a result, there is increasing difficulty in identifying, diagnosing and managing patients and counselling their families.

To bridge the gap and advance the field further, we performed the most comprehensive qualitative and quantitative literature review of PCH to date. We defined PCH as a constellation of specific radiographic, clinical and genetic features, currently recognized by Online Mendelian Inheritance in Man® (OMIM) and not explained by another disease aetiology. We did not include otherwise known genetic syndromes with an associated radiographic feature of hypoplastic pons and cerebellum. This review aims to summarize over 100 years of PCH literature (from 1912 to 2022), offer a comprehensive overview of the historical and current categorization of PCH as a disease, and provide the most up-to-date phenotypic, genetic and neuroimaging characteristics of the disease by type and subtype. In addition, this work highlights research opportunities and serves as an educational resource not only neonatal neurologists but for the broader neurology community—especially in light of the recent United Council for Neurologic Subspecialties approval of a neonatal neurocritical care neurology fellowship.

Methods

To conduct this narrative review, we began by searching PubMed using the term ‘pontocerebellar hypoplasia’ and did not exclude any search terms (Fig. 1), which yielded 189 relevant publications. Reviewing each of these, we excluded publications addressing only developmental neuroanatomy (n = 13), text not written in English (n = 6) or when the diagnosis of PCH was only discussed as a differential diagnosis (n = 6). Each manuscript’s bibliography was then reviewed and cross-referenced with OMIM, yielding 19 additional publications. We subsequently expanded the timeline to include manuscripts until 2022, thus adding eight more relevant publications. In total, we identified and comprehensively reviewed 191 publications between 1912 and 2022 for this review. To avoid misclassification bias, review and classification of the literature was performed by one individual, and then critically reviewed and evaluated by another individual. We extracted information on reported cases, genetic aetiologies and neuroradiographic features by PCH type and subtype.

Figure 1.

Literature selection. A PubMed search for the term ‘pontocerebellar hypoplasia’ identified 189 publications. Excluded manuscripts included those addressing (i) developmental neuroanatomy of the pons and cerebellum, (ii) clinical cases identifying PCH only as differential diagnosis and (iii) publications not written in English. Using these criteria, we identified 164 publications between 1977 and 2020. Cross-referencing OMIM, available bibliographies and extending the publications to 2022 yielded a total of 191 manuscripts (1912–2022). The term PCH was coined in 1993. As a result, we divided the literature into a set of publications for qualitative (1912–1993) and quantitative analyses (1993–2022). OMIM, Online Mendelian Inheritance in Man®; PCH, pontocerebellar hypoplasia.

The term PCH was coined by Barth et al. in 1993, which corresponded with the beginning of notable advances in neuroimaging and genetic sequencing.^7,8 We qualitatively analysed 19 manuscripts prior to 1993 and proposed a current PCH categorization (Table 1). Manuscripts, including and after 1993, were incorporated in the quantitative analyses. When available, information on the observed inheritance pattern and availability of translational research systems was obtained. Publications addressing more than one form of PCH were included with each type addressed.

Table 1.

Historical review of the PCH literature

Authors	Year	Significant events	Initial classification	Proposed classification
Vogt and Astwazaturow9	1912	1st neuropathological account of PCH	Combination of neocerebellar and pontine hypoplasia	Unavailable
Brun10	1917	Combination of hypoplasia + neuronal atrophy + choreiform movements	Neocerebellar hypoplasia	PCH 2
Brouwer11	1924	1st neuropathological report in a patient Attempt to name the disorder	Hypoplasia ponto-neocerebellaris	PCH 2
Koster12	1926	Further neuropathological evaluation	Hypoplasia ponto-neocerebellaris	PCH 2
Biemond13	1955	Dentate malformation must be considered the core and cause of the ponto-neocerebellar hypoplasia	Ponto-neocerebellar hypoplasia	PCH 1
Norman and Urich14	1958	Validation of Brouwer’s principle of systemic degeneration; new link between malformation and abiotrophy	Cerebellar hypoplasia associated with systemic degeneration in early life	PCH 4
Norman15	1961	1st description of PCH + anterior horn disease	Cerebellar hypoplasia in Werdnig–Hoffmann disease	PCH 1
Weinberg and Kirkpatrick16	1975	Supportive case for Norman and Urich	Cerebellar hypoplasia in Werdnig–Hoffmann Disease	PCH 1
Goutières, Aicardi, and Farkas17	1977	Authors proposed the disease is distinct from Werdnig–Hoffmann disease	Anterior horn cell disease associated with PCH in infants	PCH 1
Peiffer and Peiffer18	1977	1st description of PCH + microcephaly + chorea	Dysplastic neocortical architecture	PCH 2
Steiman et al.19	1980	Mild PCH + anterior horn disease Attempt to name the disorder	Infantile neuronal degeneration	PCH 1
Herrick et al.20	1983	Support for progressive multisystem atrophy; thought to represent autosomal recessive inheritance	Could represent ponto-neocerebellar hypoplasia or olivo-ponto-cerebellar atrophy	PCH 4
de León et al.21	1984	Attempt to name the disorder	Amyotrophic cerebellar hypoplasia	PCH 1
Kawagoe and Jacob22	1986	Disease as a result of inactivity-abiotrophia caused by the hypoplastic and insufficiently functional neocerebellar cortex	Neocerebellar hypoplasia with systemic combined olivo-ponto dentatal degeneration	PCH 4
De Caro et al.23	1987	Histological determination of the pathology and chronology of lesions	Bulbopontocerebellar hypoplasia with aplasia of the inferior olivary nucleus	PCH 4
Barth et al.2	1990	Study confirmed the disease as an inherited neurodegenerative disease	Systemic atrophy with early onset	PCH 2
Kamoshita et al.24	1990	Attempt to name the disorder	Norman syndrome	PCH 1
Barth25	1992	Review of previously published literature plus case report of siblings with PCH	Pontocerebellar hypoplasias	PCH 1
Albrecht et al.3	1993	Report of previously not known familial infantile encephalopathy with olivopontocerebellar hypoplasia	Fatal infantile encephalopathy with olivopontocerebellar hypoplasia and microencephaly	PCH 4
Barth7	1993	Proposed delineation among PCHs Attempt to name the disorder	PCH Type-1 and PCH Type-2	PCH 1 and PCH 2

The first manuscript describing PCH was published in 1912, and since then several publications have described the clinical entity using different nomenclature. Using available information, we attempt to classify each of these based on current disease classification. Most cases were PCH Types 1, 2 and 4.

To craft a comprehensive review, we meticulously categorized publications by the significance of their contribution to the field. Categories included standalone case report/series and those with histopathological, biochemical, prenatal, genetic or radiographic data. This was intended to address the multidisciplinary approach to the disease that included contributions from laboratory researchers, pathologists, neurologists, geneticists and radiologists. Reviews, commentaries, basic research, morphometric and natural history studies were also evaluated. The literature was then chronologically evaluated and sorted by PCH type, subtype and genetic aetiology. Focusing on neuroimaging data, we analysed the range and commonality of radiographic characteristics by disease type and subtype and created a comprehensive table summarizing such findings. To demonstrate the intricacy of PCH diagnosis as a disorder, we provide a radiographic example of cerebellar hypoplasia as an isolated imaging finding and as PCH 6 over time. This study was completed from 2019 to 2023 under IRB approval that qualified with research exemption (protocol number 12020) at Children’s National Hospital.

Results

Literature review

The first description of a hypoplastic pons and cerebellum was in 1912 by Vogt and Astwazaturow (historical perspective since 1912 reviewed in Table 1 and Supplementary Fig. 1).⁹ Applying current classification, cases published prior to 1993 could be consistent with the diagnosis of PCH Types 1, 2 and 4, yet classification was challenging due to changes in nomenclature (Table 1).^2,3,26-32 Since 1993, research on PCH further evolved with a notable increase in publications including genetic sequencing starting in 2011, which continued through 2022 (Supplementary Fig. 2). As demonstrated by Supplementary Fig. 2, organized in 5-year brackets, we see a trend for more research publications with novel model systems (3 publications in 2008–2012 increased to 12 in 2018–2022). It is important to highlight that there has only been one natural history study on PCH 2A, published in 2014,³³ and only two studies with sole focus on prenatal investigation of PCH 2, published in 2010.^34,35 Over the last century, there are only two studies with extensive biochemical analyses of PCH: one investigated the enzyme activities of the respiratory chain in PCH6 and OXPHOS complex,³⁶ and the other evaluated mitochondrial dysfunction with recessive mutations in EXOSC3 with PCH. Publications discussing radiographic patterns have been rare with a total of six case reports focusing on MRI data and three morphometric studies (1992–2022). While the number of PCH publications grew in 2008, so did the classification of PCH. This growth required a more detailed examination of individual types, subtypes and genetic causes.

Inconsistencies in PCH categorization

We found discrepancies in categorizing disorders with the same genetic aetiology. Often, different names were used. For example, progressive cerebello-cerebral atrophy (PCCA) Type 1 and PCH 2D are both due to SEPSECS mutations, and PCCA Type 2 and PCH 2E are due to VPS53 mutations.³⁷ Some have reported new mutations under PCH subtypes that are still not widely recognized or included in OMIM, such as PLA2G6 as a PCH 1 subtype³⁸ or CASK mutations as part of PCH 3.³⁹ Other notable discrepancies are between the literature and OMIM. Although the TSEN c.919G>T mutation is listed for PCH Types 1, 2, 4 and 5, the connection between TSEN mutation and PCH 1 is not published in OMIM.⁴⁰ In addition, in 2017, van Dijk et al. classified the SLC24A46 mutation as PCH 1D;⁴¹ however, subsequent literature and OMIM refer to it as PCH 1E.

Clinical features and associations with PCH

As reported, PCH is clinically heterogeneous. In fact, there is no single clinical feature shared in the diagnosis of PCH, neither among the 26 subtypes nor within the 16-type classification. However, a few unifying features are identified. First, many PCH types have an onset early in life, from prenatal to infancy. Second, many have developmental delays and/or abnormal muscle tone (hypertonia or hypotonia). Third, many affected patients develop epilepsy, which might be refractory to treatment. To provide a comprehensive review of current categorization of PCH, we incorporated previously recognized clinical features of each disorder (Table 2; see Supplementary Table 1 for the complete table). As displayed in the table, the diagnosis of PCH by subtype is made by clinically and/or radiologically distinguishing features. For instance, PCH 1 has been associated with polyneuropathy and contractures, while PCH 2 is recognized by extrapyramidal dyskinesias and abnormal movements (of eyes or limbs). Similarly, PCH 4 is distinguished by fatal/neonatal apnoeas, whereas there is a serum/CSF lactic acidosis in PCH 6. Due to the XY reversal in PCH Type 7, affected infants can have ambiguous genitalia.⁵⁰ Furthermore, not all types have been known to be neurodegenerative in nature, including PCH 8 and PCH 11.

Table 2.

Summary of PCH Categorization

PCH phenotype	Locus	IHT	Gene	Gene function	First report in PCH	Gene identification in PCH	Clinical features	PCH animal models
PCH Type 1
Type 1A	14q32.2	AR	VRK1	Neuronal migration	Norman (1961)42	Renbaum et al. (2009)15	Onset prenatal to late infancy, SMA phenotype, polyneuropathy	Mouse: Vinograd-Byk et al. (2015)43
Type 1B	9p13.2	AR	EXOSC3	mRNA degradation	Ryan et al. (2000)26	Wan et al. (2012)44	Neonatal onset, early neonatal apnoea, oculomotor apraxia (+/−), optic atrophy (+/−), tongue fasciculations, contractures, axonal motor neuropathy	None
Type 1C	13q13.3	AR	EXOSC8	mRNA degradation	Boczonadi et al. (2014)28	Boczonadi et al. (2014)28	Onset in first months of life, respiratory failure, contractures, SMA phenotype, atrophy, hearing impairment	Zebrafish: Boczonadi et al. (2014)28
Type 1D	4q27	AR	EXOSC9	mRNA degradation	Burns et al. (2018)31	Burns et al. (2018)31	Onset at birth to early infancy, reduced foetal movements, early neonatal apnoea, contractures, fasciculations, impaired pursuit, axonal motor neuronopathy	Zebrafish: Burns et al. (2018)31
Type 1E	5q22.1	AR	SLC25A46	Mitochondrial fission/fusion	Wan et al. (2016)30	Wan et al. (2016)30	Prenatal onset, neonatal lethal, polyhydramnios, early neonatal apnoea, contractures, optic atrophy (+/−), polyneuropathy	Zebrafish: Wan et al. (2016)30
Type 1F #	10q24.1	AR	EXOSC1	mRNA degradation	Somashekar et al. (2021)32	Somashekar et al. (2021)32	Onset in infancy, developmental delays, blue sclera, microcephaly, dysmorphic facies, hypotonia, diminished reflexes	None
PCH Type 2
Type 2A	17q25.1	AR	TSEN54	tRNA splicing	Barth et al. (1990)2	Budde et al. (2008)4	Onset at birth, impaired swallowing, central visual impairment, hypertonia at birth, extrapyramidal dyskinesia	None
Type 2B	3p25.2	AR	TSEN2	tRNA splicing	Barth et al. (1990)2	Namavar et al. (2011)6	Onset at birth, central visual impairment, hyperkinetic involuntary movements	None
Type 2C #	19q13.42	AR	TSEN34	tRNA splicing	Barth et al. (1990)2	Budde et al. (2008)4	Onset at birth, extrapyramidal dyskinesia	None
Type 2D	4p15.2	AR	SEPSECS	Selenocysteine synthesis catalyst	Ben-Zeev et al. (2003)27	Agamy et al. (2010)45	Onset in infancy, contractures, sleep disturbances, irritability, oedema of face and limbs, polyneuropathy, optic atrophy (+/−)	None
Type 2E	17p13.3	AR	VPS53	Retrograde transport of endosomes to Golgi	Feinstein et al. (2014)29	Feinstein et al. (2014)29	Onset in infancy, progressive, failure to thrive, gaze-evoked nystagmus (+/−), optic atrophy (+/−), distal limb oedema (+/−)	None
Type 2F	1q25.3	AR	TSEN15	tRNA splicing	Barth et al. (1990)2	Breuss et al. (2016)46	Onset at birth, poor or absent fixation	None
PCH Type 3
Type 3 #	7q21.11	AR	PCLO	Regulation of synaptic protein/vesicle formation	Rajab et al. (2003)47	Ahmed et al. (2015)48	Onset at birth, neonatal hypotonia, optic atrophy	Rat: Falck et al. (2020)49
PCH Type 4
Type 4	17q25.1	AR	TSEN54	tRNA splicing	Albrecht et al. (1993)3	Budde et al. (2008)4	Prenatal onset, death in infancy, polyhydramnios, early neonatal apnoea, contractures	None
PCH Type 5
Type 5 #	17q25.1	AR	TSEN54	tRNA splicing	Patel et al. (2006)5	Namavar et al. (2011)6	Prenatal onset, death in neonatal period, seizure including seizure-like activity in utero starting around 18 weeks gestation, polyhydramnios, early neonatal apnoea	None

IHT, inheritance; AR, autosomal recessive; SMA, spinal muscular atrophy .

A subset of the PCH type/subtypes (1A–16) are listed with its locus, associated genetic aetiology, protein function, inheritance pattern, OMIM IDs, publications first reporting the disease and associated genetic aetiology, and availability of translational/model research systems. A hashtag (#) indicates a provisional relationship between the phenotype and gene per OMIM. A complete version of this table, including the full list of PCH type/subtypes is provided in Supplementary Table 1.

Of importance to management are also reports of unique clinical associations seen in patients with a PCH diagnosis. For instance, PCH 2 was reported in a case of recurrent rhabdomyolysis.⁵¹ PCH 3 was reported in conjunction with severe Vitamin A deficiency⁵² and separately in a case of tetralogy of Fallot.⁵³ Infantile spasms or early myoclonic encephalopathy were identified in a patient with PCH 6,⁵⁴ and Stickler Syndrome Type 2 was identified in patient with PCH 9.⁵⁵ Radiographically, PCH 1B was reported with rhombencephalosynapsis and microlissencephaly.⁵⁶ In an effort to avoid anchoring bias, recognizing these features and associations can be useful in diagnosing, managing and treating affected patients.

Most frequently reported PCH type/subtype and associated genetic aetiologies

PCH 1 (n = 47), 2 (n = 45) and 6 (n = 25) and Subtypes 1-B (n = 14), 2-A (n = 9) and 1-E (n = 5) are among the most frequently reported (Fig. 2A and B). Types 12–16 are described in less than a total of five publications over the past 29 years, starting with manuscripts after 2018, with emphasis on basic research or case report/series with contributory novel sequencing. As with other diseases, mutations in the same gene can lead to different types of PCH (allelic heterogeneity) and mutations in different genes can lead to the same PCH type (locus heterogeneity).^2-6 The most frequently reported genetic aetiology among 24 candidate genes reported include TSEN54 (n = 20), RARS2 (n = 19), EXOSC3 (n = 15) and AMPD2 (n = 9) (Fig. 2C). The 20 remaining genes are each reported in five citations or less (Fig. 2C). Reported gene function often involves RNA synthesis or processing (see Gene function column, Table 2). Chromosome 17 includes most PCH loci (2A, 2E, 4, 5 and 12, which result from mutations in TSEN54, COASY and VPS53) (Table 3). To complement genetic and clinical findings, preclinical models, including yeast, zebrafish, mice and rats, have been developed for several PCH types and subtypes, providing early insights into disease mechanisms (see PCH animal models column, Table 2).

Figure 2.

Studies performed on PCH by type/subtype and reported genetic aetiologies. The figure represents summed values of publications on PCH (y-axis) sorted by (A) PCH type (x-axis), (B) PCH subtype (x-axis) and (C) genetic aetiology. Staggered columns are colour-coded by the type of study published. The most frequently reported PCH types are 1, 2 and 6. Less frequently reported are 3, 4 and 9. The most frequently reported subtypes are 1-B, 2-A and 1-E. There are 24 associated genetic aetiologies, and the most reported mutations include the TSEN54, RARS2, EXOSC3 and AMPD2 genes. To support broader accessibility, we have incorporated shape-based annotations: ■ (square for red), ● (circle for light green), and ▴ (triangle for dark green).

Table 3.

Breakdown of PCH subtypes by locus and gene

Locus	PCH Subtype	Gene
Chromosome 1
1p13.3	PCH Type 9	AMPD2
1q25.3	PCH Type 2F	TSEN15
1p34.1	PCH Type 7	TOE1
Chromosome 3
3q12.1-q12.2	PCH Type 11	TBC1D23
3p25.2	PCH Type 2B	TSEN2
Chromosome 4
4p15.2	PCH Type 2D	SEPSECS
4q27	PCH Type 1D	EXOSC9
Chromosome 5
5q22.1	PCH Type 1E	SLC25A46
Chromosome 6
6q15	PCH Type 6	RARS2
6p21.2	PCH Type 14	PPIL1
6q21	PCH Type 15 #	CDC40
Chromosome 7
7q21.11	PCH Type 3 #	PCLO
Chromosome 9
9p13.2	PCH Type 1B	EXOSC3
Chromosome 10
10q23.2	PCH Type 16	MINPP1
10q24.1	PCH Type 1F #	EXOSC1
Chromosome 11
11q12.1	PCH Type 10	CLP1
11q13.1	PCH Type 13	VPS51
Chromosome 13
13q13.3	PCH Type 1C	EXOSC8
Chromosome 14
14q32.2	PCH Type 1A	VRK1
Chromosome 16
16q24.3	PCH Type 8	CHMP1A
Chromosome 17
17p13.3	PCH Type 2E	VPS53
17q21.2	PCH Type 12	COASY
17q25.1	PCH Type 2A	TSEN54
17q25.1	PCH Type 4	TSEN54
17q25.1	PCH Type 5 #	TSEN54
Chromosome 19
19q13.42	PCH Type 2C #	TSEN34

Ordered by chromosome number and locus, this table highlights that many PCH types/subtypes are linked to single loci. A hashtag (#) indicates a provisional relationship between the phenotype and gene per OMIM.

Analysis of the neuroimaging literature

We found that underdevelopment of the pons and cerebellum is usually, but ‘not invariably’, present to some degree in newborns. Reductions in volume may not be present, qualitatively or quantitatively, in the early and mid-prenatal period.^6,57-59 As a result, it is difficult to exclude the diagnosis of PCH in the prenatal and neonatal stages, and follow-up imaging is often needed, especially when the diagnosis is postulated with clinical course or sequencing data. Summarized below are key imaging considerations.

Utility of prenatal ultrasound in diagnosing PCH

Prenatal ultrasound is widely accessible, costs less than MRI and is routinely used in clinical practice, often providing first views of the developing brain. However, ultrasound may be unreliable and insufficiently sensitive for the diagnosis, especially at earlier gestational ages. In most cases, prenatal ultrasounds were either normal or revealed an enlarged cisterna magna in isolation or hypoplastic posterior fossa structures with or without cerebral abnormalities.^6,60 There were exceptions in the literature (e.g. foetal ultrasound performed at 20 and 28 weeks’ gestation revealed PCH and microcephaly in foetuses with COASY gene defects corresponding to PCH 12).^5,61,62 However, in a series of five cases imaged between 22 and 25 weeks and confirmed postnatally, ultrasound did not show diagnostic features of PCH.⁵⁸ Among these cases, cerebellar hypoplasia was first identified only at 30 weeks in a single patient. Similarly, in a twin gestation with documented TSEN54 mutations at 20 weeks, cerebellar hypoplasia was first detected at 31 weeks.³⁴ As a result, diagnosing PCH sonographically may not be feasible until later in the third trimester. Thus, follow-up evaluation should be considered, especially when there is an accompanying suspicion for a neurogenetic disorder.

Prenatal MRI

Compared with ultrasound, prenatal MRI is most useful in evaluating structural details of the posterior fossa parenchyma and its coverings.⁶³ It complements and often enhances sonographic assessments with its superior signal-to-noise ratio. Furthermore, prenatal MRI allows for a more complete and uniform view of the entire brain. While volumetric measurements (3D) may be superior to 2D measurements, they require thinner section acquisition (generally <2 × 2 × 2 mm voxel size) that extends the duration of the scan and, consequently, increases the risk of foetal motion artefact.^64-66 Compared to conventional 1.5T MRI, the 3T prenatal MRI usually offers better anatomic detail of the cerebellum and brainstem, but at the expense of prolonging scan time to offset the increased specific absorption rate.^67,68

There is scarce literature regarding diagnosis and assessment of PCH using prenatal brain MRI. However, in some instances, prenatal MRI can be diagnostic. For example, PCH, corpus callosum hypoplasia and cerebral white matter hypoplasia associated with AMPD2-related PCH 9 was diagnosed at 24 and 30 weeks’ gestation.⁶⁹ The cardinal feature of PCH 9, a midbrain with ‘figure of 8’ appearance, was identified at 30 weeks.⁶⁹

Post-natal MRI

Post-natal non-sedate brain MRI is useful in identifying, following-up and confirming a suspected diagnosis of PCH, even when the genetic aetiology is uncertain. In the neonatal period, this can be done ‘feed-and-bundle’ style without anaesthesia. A standard brain MR protocol at Children’s National Hospital includes volumetric T1-weighted imaging (T1WI), axial T2, axial fluid-attenuated inversion recovery (FLAIR) (>1 year) or proton density (<1 year), axial diffusion tensor imaging, axial 3D gradient echo and coronal fat sat T2-weighted imaging (T2WI). A diffusion-weighted sequence is useful in excluding ischaemic or infectious PCH mimics (Fig. 3) and in considering RARS2-related aetiologies (Fig. 3). In such cases, loss of RARS2 leads to insufficient mitochondrial tRNS-arg charging and a phenotype similar to patients who have a genetic difference affecting the mitochondrial respiratory chain. Thus, diffusion-weighted imaging (DWI) is useful in detecting neurometabolic decompensation for at least some PCH cases.⁷⁰ Diffusion tensor imaging is an advanced MR technique that can quantify the degree and extent of white matter fibre deficiency and is potentially useful in distinguishing PCH from non-progressive hypoplasia and progressive atrophy.⁷¹

Figure 3.

Neuroimaging findings of PCH. (A) Sagittal T1WI (TR/TE ms, 12/5) and (B) coronal T2WI (TR/TE ms, 2502/44) from a 3-month-old, former 27-week premature infant, with resulting cerebellar injury, show the ‘dragonfly’ sign of cerebellar hypoplasia. Note visible volume loss, more pronounced in the cerebellar hemispheres than the vermis. Post-haemorrhagic hydrocephalus is also present along with hemosiderin staining from prior haemorrhage seen along the under-surface of the right cerebellar hemisphere (arrow, B). Images (C–E) demonstrate radiographic findings in a patient with a RARS2 mutation (PCH6). Sagittal T2WI (TR/TE ms, 5000/161) (C) from a foetal MR at 22 weeks gestation shows the cerebrum, cerebellum and brainstem to be of age-appropriate size and morphology. Post-natal sagittal T1WI MR exams (TR/TE ms, 8–11/3–5) at 7 months (D) and 10 years (E) reveal development of PCH with progressive cerebellar atrophy and lack of appropriate pontine growth. Images were obtained from patients seen at CNH under an IRB-approved protocol.

Pathognomonic imaging features of PCH

There can be distinguishing features among PCH types and these warrant specific mention when the diagnosis is suspected (Table 4). For example, cerebellar hypoplasia without pontine hypoplasia is often found in PCH 1.^59,72,73 Anterior horn involvement in PCH 1 can theoretically lead to reduced spinal cord volume in mimicking spinocerebellar ataxia; however, we did not find such examples by MRI. TSEN54 mutations are strongly associated with the appearance of a ‘dragonfly’ cerebellum, showing disproportionate hypoplasia of the hemispheres (dragonfly wings) compared to the vermis (dragonfly body) in the coronal plane.^6,59,74,75 However, the ‘dragonfly’ appearance is neither completely sensitive nor specific, as it may be absent in TSEN54 mutations, present in other PCH types (1, 9, 11 and 16) or present in acquired forms of posterior fossa hypoplasia, especially in prematurity-related injury (Fig. 3).^76-79

Table 4.

Range and commonality of neuroimaging attributes by PCH type

Type	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
Age	P-I	P-I	N	P	P	N-I	N	I-C	P-C	I-C	I-C	P	I-C	P	I	P-I
PH	+/−#	+/−	+*	+/−*	+/−*	+/−	+*	+*	+*	+/−*	+*	+*	+/−*	+	+*	+
CH	+/−*	+/−	+*	+*	+/−*	+/−*	+*	+*	+*	−#	+*	+*	+/−*	+	+*	+/−*
VH	+/−	+/−	+/−	+/−	+/−*	+/−*	+*	+*	+/−	−#	+/−	+*	+	+	+*	+/−*
BA	+/−	+/−	−	+*	+/−	+/−	+	-	+/−	-	-	-	+/−*	+/−	+/−	+/−
CA	+/−	+/−	+	+/−	+/−	+/−	+*	+/−	+/−	+/−	+*	-	+	+/−	+/−	+/−
OA	+/−	+/−	-	+/−*	+/−*	+/−	+*	-	+/−	-	-	-	+/−	+/−	-	+/−
WMv	+/−*	+/−	+/−	+/−	-	+/−	+*	+*	+*	+*	+/−	+	+	+	+/−	+/−*
WMm	+/−*	+/−	−#	+/−	+/−	+/−	-	-	+/−*	+/−*	−#	-	+/−	+/−#	+/−	+/−*
CCH	+/−*	+/−	+/−	-	-	+/-	+*	+*	+*	+*	+/−*	+	+	+	+*	+/−*
CeCo	+/−*	+/−	+/−	+/−	+/−	+/−	+/−	-	+/−	+/−	+/−	-	+/−*	+/−	+/−	+/−
ONA	+/−	−#	+/−*	-	-	+/−	+/−	-	-	-	+/−#	-	-	-	-	-
BG	+/−	+/−	-	+/−	+/−	+/−#	-	-	+	-	-	-	+/−	+/−	-	+
Ccyst	+/−	+/−	-	-	-	+/−	-	-	-	-	-	-	-	-	-	-

N, neonate; I, infant; P, prenatal; C, child; PH, pontine hypoplasia; CH, cerebellar hemisphere hypoplasia; VH, vermian hypoplasia; BA, brainstem atrophy; CA, cerebellar atrophy; OA, optic atrophy; WMv, cerebral white matter volume loss or hypoplasia; WMm, white matter myelination deficiency; CCH, corpus callosum hypoplasia or hypogenesis; CeCo, Cerebral cortex volume loss; ONA, olivary nuclear atrophy; BG, basal ganglia volume loss; Ccyst, cerebellar cysts. + = present; - = absent or not reported; +/−, present or absent.

The symbol # denotes that key finding is negative, and * indicates key finding is positive; no special symbol indicates reporting variability and thus not key distinguishing attribute.

Alternative patterns of cerebellar involvement range from diffuse to focal, affecting the vermis and leading to a ‘butterfly’ appearance of the cerebellum in the coronal plane as seen in PCH 3 and PCH 6.^48,80  RARS2-related PCH 6 generally manifests with progressive neurodegeneration, starting with normal volumes in the neonatal period and progressing to moderate or marked diffuse reductions in volume postnatally.^54,80,81 Reduced white matter diffusion is sometimes seen as well.⁸¹ As noted, AMPD2-related PCH 9 almost always manifests a ‘figure of 8’ appearance of the midbrain in the axial plane due to hypoplastic shortening of the cerebral peduncles.^{59,69,79,82,83} In PCH 10, due to mutations in CLP1, there is often only mild pontine hypoplasia with relative or complete sparing of the cerebellum in concert with reduced cerebral white matter volume.^59,84-86  TBC1D23-related PCH 11 is typified by non-progressive but marked pontine hypoplasia, moderate cerebellar hemisphere hypoplasia with associated lateral folial predominant volume loss, mild vermis hypoplasia and marked corpus callosum deficiency ranging from severe hypoplasia to agenesis.^77,78 Congenital microcephaly with progressive atrophy and simplified gyral pattern typifies PCH 13 due to mutations in PS51.^87,88 Selective basal ganglia hypoplasia and/or atrophy is common in PCH 16.^89,90

Emerging PCH-associated genes

Since the end of our review period, at least seven publications have described additional genes implicated in PCH or PCH-like phenotypes. Newly associated genes include ATAD3A, INPP4A (in two separate reports), WDR11—a component of the FAM91A1 complex, which directly interacts with TBC1D23, a previously established PCH gene—as well as FTH1, MPDU1 and RAF1.^91–95 Together, these findings underscore the growing genetic heterogeneity of PCH.^89,90

Discussion

A historically informed comprehensive review

Since 1993, ‘PCH’ has been used to describe a neurodevelopmental disorder affecting the size of the pons and cerebellum with a clinical course that includes severe neurodevelopmental delays or neurodegeneration. Greater resolution of posterior fossa structures has heightened interest in PCH, which has serious implications. Significant clinical, neuroradiographic and genetic overlap among the 26 categories of PCH has likely impeded a neurologist’s ability to have more clarity in prognosis while counselling families. These clinical needs might impact patient care. Therefore, there is a significant need for a clarified PCH definition to guide diagnosis and management—from prenatal screening to post-natal care.

Considerations in developing a differential

Several features of PCH are also consistent with other prenatal and acquired disorders that lead to substandard growth of the cerebellum and/or pons. In fact, disruptive events of sufficient severity during the course of cerebellar development, including infections, toxins, hypoxic-ischaemic injury, haemorrhage and metabolic disturbances can lead to cerebellar or PCH.⁹⁶ However, PCH is usually not found in isolation. Chromosomal trisomies, such as Down syndrome, often result in underdevelopment of the pons and cerebellum.⁹⁷ Pontocerebellar disruption due to prematurity can lead to a hypoplastic pons and cerebellum, which might not progress to PCH as a disease, but their imaging may be viewed as ‘lesional mimicry.’⁹⁸ Rhombencephalosynapsis, pontine tegmental cap dysplasia, ciliopathies, tubulinopathies and alpha dystroglycanopathies all commonly have some degree of cerebellar and brainstem hypoplasia, though additional structural abnormalities specific to these disorders usually aid in the distinction.^59,99 Congenital disorders of glycosylation may also present with cerebellar and pontine hypoplasia, often with associated cerebellar volume loss and cortical signal abnormality.^59,99,100

PCH mimics

Radiographic and histological features of PCH can lead to diagnostic challenges. For example, a loss of AGTPBP1 mimics PCH Type 1,¹⁰¹ yet this gene is not currently recognized as an aetiology. Similarly, a homozygous deletion in WDR81 gene¹⁰² or loss of HEATR5B protein¹⁰³ have been associated with PCH, but not included as a cause. Further, a foetal case of Bainbridge–Ropers Syndrome was later identified as PCH Type 1.¹⁰⁴ Radiographic features of PCH have also been identified due to mutations in the very low-density lipoprotein receptor [cerebellar ataxia, mental retardation and disequilibrium syndrome 1 (CAMRQ1)]^105,106 and in DKC1 (Hoyeraal–Hreidarsson Syndrome).¹⁰⁷ More genes are known to cause similar radiographic features but have not been classified as a PCH disorder, including PPP2R1A gene,^108,109  PLA2G6³⁸ and CASK gene.¹¹⁰ An overlap in clinical presentation was also reported in PCH and progressive encephalopathy with oedema, hypsarrhythmia and optic atrophy;¹¹¹ Emanuel syndrome and auditory neuropathy spectrum disorder.¹¹²

Integrating neuroimaging and genetic testing towards diagnosing PCH

The diagnosis of PCH is associated with clinical, neuroimaging and genetic characteristics. Among these, neuroimaging is most reliable, especially when followed serially. The clinical course can be variable, and genetic sequencing continues to show an expanding array of aetiologies and associated molecular pathways. Thus, it might be difficult to exclude the diagnosis at time of presentation and with genetic testing alone. Among such variability, imaging characteristics define PCH. Reduced volume of the brainstem and/or cerebellum corresponding to parenchymal hypoplasia with or without associated volume loss is mandatory for the diagnosis. Importantly, serial imaging enhances diagnostic credibility by demonstrating reductions in age-appropriate interval growth or volume loss over time. Greater specificity for the diagnosis is possible when other features are present or develop, including (i) reduced cerebral volume and hypoplastic corpus callosum; (ii) callosal agenesis or dysgenesis; (iii) simplified gyral pattern; (iv) selective volume loss or atrophy of the basal ganglia; (v) deficient myelination; and (vi) olivary hypoplasia, cerebellar cysts and optic hypoplasia and/or volume loss.^35,76

Disambiguating PCH towards diagnostic and prognostic clarity

Various terms are used to refer to the diagnosis of PCH, leading to uncertainty and ambiguity within the field. Some classifications with highly overlapping clinical, imaging and genetic features include ‘olivopontocerebellar hypoplasia’, ‘congenital olivopontocerebellar atrophy,’ ‘cerebellar hypoplasia’, ‘pontocerebellar atrophy’, ‘pontine and cerebellar hypoplasia’ and ‘PCCA’, among others. In referring to the disease spectrum as a whole, we propose the term pontocerebellar hypoplasia spectrum disorder (PHSD). We suggest adding the word ‘spectrum’ to represent the range of overlapping features, which are not always pathognomonic for a disease type/subtype.

We agree with efforts to associate types of PCH by genetic aetiology, as in ‘TSEN mutation spectrum disorders’ (to include PCH Types 2, 4 and 5),⁷⁴ or to classify PCH Type 6 as a RARS2-associated phenotype with early onset mitochondrial encephalopathy.⁷⁰ As such, we propose the introduction of a dyadic naming strategy (e.g. RARS2-associated PHSD) to replace the naming conventions that rely on numerical designations (e.g. PCH 1, PCH 2). Traditional naming strategies fail to capture the genetic heterogeneity of these conditions, whereas dyadic naming links the disorder directly to the specific gene involved in its pathogenesis. This enhances clarity in diagnosis and guides targeted research while also resolving inconsistencies in current classification systems by providing a stable framework. When the aetiology is not genetic, abnormal growth of the pons or cerebellum can be defined by the primary aetiology and termed ‘lesional mimicry’.⁹⁸

Practical recommendations in pre-, peri- and post-natal diagnosis and management

Based on our review, we propose the following considerations in diagnosis and management (Fig. 4). Prenatally, concern for PCH is often a by-product of antenatal anatomy ultrasounds during the second trimester. In such cases, initial MR imaging between 20and 25 weeks’ gestation is useful to confirm the findings. Typical MRI evaluation by a paediatric neuroradiologist is needed, and 3D measurements can provide additional supportive information. If the findings are confirmed, repeat imaging between 30 and 34 weeks provides the neurologist and expectant families an opportunity to further consider post-natal management. Follow-up evaluation and imaging, when feasible, is recommended because (i) hypoplasia can become apparent over time due to differences in typical regional growth compared with other brain structures, (ii) later-onset molecular dysfunction can lead to neurodegeneration and regression of volume or (iii) diminished volume can be undetectable at younger gestational ages due to technical limitations in the resolution of imaging. If imaging shows arrested growth or these progressive changes, a diagnosis of pre/post-natal PHSD should be considered. However, in some genetic subtypes, minimal interval change may limit the utility of repeat imaging in the foetal period, which may also be restricted by insurance coverage or access.

Figure 4.

Proposed diagnostic algorithm for PHSD. Recommended diagnostic approach for PHSD according to the proposed naming strategy. MRI, magnetic resonance imaging; PCH, pontocerebellar hypoplasia.

Following imaging findings suggestive of PCH, genetic testing should be considered in the prenatal period or postnatally based on family preferences and joint discussions that includes the obstetricians or maternal-foetal specialists and the neurologist. Whole-exome sequencing is recommended because of expected difficulty in distinguishing among PCH types and known genetic heterogeneity; genetic sequencing of individual candidate genes might have lower yield, thus prolonging the diagnostic odyssey. If a genetic cause if identified, the condition should be named as (gene)-associated PHSD.

In the peri- and neonatal period, management can include intensive care to support ventilation and nutrition, additional imaging and referrals to subspecialists and therapists. Infants with PCH can have persistent needs for life-sustaining care and require evaluations for encephalopathy and/or hypotonia. Ultrasound should be obtained initially to evaluate midline, supratentorial structures; greater resolution of the posterior fossa is usually obtained with an MR imaging (preferably non-sedate). As with prenatal management, if clinical diagnosis remains unknown, we recommend follow-up interval imaging, usually 3–6 months after the initial study. When there is a high index of suspicion for PCH based on initial or follow-up imaging, genetic testing is recommended.

Currently, there are no treatments for PCH and affected infants and children are supported symptomatically. Affected individuals may have epilepsy, and their seizures can be refractory to treatment. Greater benefit for affected families might ensue from combination of these supportive measures and prognostic counselling towards anticipating needs and planning care. Clinical consultations often include referral for physical medicine and rehabilitation, gastroenterology, ophthalmology, otolaryngology and palliative care. Frequently, there is a need for longer-term support by physical, occupational and speech therapists, nutritionists and audiologists.

A much-needed path forward

The majority of PCH studies were retrospective case reports/series, with only a few prospective natural history studies. This persistent gap in the field might bias our prognostic information towards more severely affected children. Thus, there is a pressing need to prospectively consider the clinical, neuroimaging or genetic correlates of long-term outcomes. As a rare, autosomal recessive, neurogenetic disorder, such studies are crucial to systematically characterize PCH by type and correlate types with neurodevelopmental outcomes. Ideally, this would be a multi-centre study, including participants nationally and internationally. In parallel, more basic and translational research is needed to understand the molecular pathophysiology of PCH, as we build a foundation towards developing disease-alleviating (versus. prolonging) therapies.

Overall, we found fewer publications addressing recently described PCH types, especially 10–16. This skew might result from several factors, including (i) true differences in disease frequency; (ii) challenges in diagnosing certain types of PCH due to overlap in clinical, neuroimaging and genetic features; (iii) complexities in differentiating the diseases as they evolve with time and/or (iv) need to accumulate more descriptions with time. Prospective studies could also resolve some of these questions.

Conclusion

Accurate diagnosis of PCH is usually made with a combination of clinical, neuroradiographic and genetic sequencing data obtained pre- or postnatally through follow-up. To address the growing complexity of PCH classification, we propose adopting the term ‘PHSD’ and implementing a dyadic naming system that links disorders directly to their genetic aetiology. This approach replaces the traditional numerical classification system, which has become increasingly complex with 16 types (categorized as 1–16), 8 subtypes (Types 1 and 2 each have sub-categorization spanning A–F) and over 24 genetic aetiologies. The new system provides greater diagnostic clarity and better reflects our current understanding of the genetic heterogeneity underlying these disorders. There are no treatments for PCH, and neurological care is supportive, personalized and integrated with other medical/therapeutic teams. Moving forward, there is a pressing need for prospective natural history studies to systematically characterize outcomes across the spectrum of PHSD. Future work should also empirically address whether the current categorization system can be reorganized based on a combination of clinical, genetic and neurodiagnostic features.

Supplementary material

Supplementary material is available at Brain Communications online.

Funding

This work was supported by extramural funding from the National Institute of Health to YAK (K08NS119882).

Competing interests

The authors report no conflicts of interest.

Data availability

Data sharing is not applicable to this article as no new data were created or analysed in this study.