Chiari Malformation and Hindbrain Descent: Characterization and New Classification Based on Mechanism and Pathogenesis, and Surgical Management
Article information
Abstract
H. Chiari described 4 types of abnormal development of the posterior fossa, which were subsequently classified as Chiari malformation types I, II, III, and IV. Many issues in neurosurgery concerning classification and surgical management are without evolving concepts. This review aims to clarify the mechanisms and pathogenesis underlying hindbrain (the brain stem and cerebellum) descent, classify them accordingly, and discuss appropriate surgical management. We propose a classification of 4 independent pathogenic mechanisms: (1) constriction in the posterior cranial fossa (PCF) due to underdevelopment of the occipital bone; (2) enlargement of hindbrain; and (3) traction caused by tethering lesions. We examine the pathogenesis of hindbrain descent from embryological perspectives and neuroradiological findings, with a particular focus on lesser-known mechanisms. Additionally, another fourth mechanism is proposed: (4) instability at the craniocervical junction. We suggest a novel classification for Chiari malformation type I based on the underlying pathogenesis, guided by morphometric (occipital bone size) and volumetric (PCF volume) analyses. Furthermore, it delves deeper into their pathogenesis by drawing on insights from developmental biology, genetic studies, and experimental research. Surgical management is tailored to the underlying mechanism, and we proposed the algorithm for decision of surgical intervention. For crowding of the PCF due to underdevelopment of the occipital bone, posterior fossa decompression is the appropriate surgical intervention. For craniocervical instability, occipitocervical fixation is recommended. We also review the recent literature on surgical outcomes associated with each treatment approach. Finally, we highlight current genetic research related to the pathogenesis of hindbrain descent.
BACKGROUND AND INTRODUCTION
In articles published in 1881 and 1886, H. Chiari described 4 types of abnormal development of the posterior fossa, which have subsequently been classified as Chiari malformation types I, II, III, and IV [1,2]. Chiari malformation type II (CM-II) is characterized by a suboccipital meningoencephalocele, while type IV involves cerebellar agenesis. However, both anomalies are extremely rare and are not discussed further in this review. The designation “CM-II” should be reserved for patients with concomitant spina bifida cystica. Arnold’s name was also added to the term for this complex abnormality of the posterior fossa structures and the content of the upper cervical spinal canal, i.e., “Arnold-Chiari malformation;” however, this condition is now generally referred to simply as “Chiari malformation [3].”
Chiari’s original articles focused on hydrocephalus, and it remains unclear whether the hindbrain herniation described was the cause or result of hydrocephalus [1,2]. What has come to be known as Chiari malformation type I (CM-I) is characterized by herniation of the cerebellar tonsils through the foramen magnum. There appears to be no logical reason to associate Arnold’s name with this particular malformation. Moreover, many cases of CM-I are not only acquired but also potentially reversible; thus, the term “hindbrain descent” might better describe this condition. Nevertheless, the designation CM-I has become a well-recognized term, clearly understood in diagnostic and surgical treatment contexts [4,5].
Since the introduction of magnetic resonance imaging (MRI) as a diagnostic modality in the 1980s, it has rapidly replaced myelography and become the most widely used neuroimaging technique. Some studies have revealed that even asymptomatic individuals may present with a mild degree of tonsillar herniation (TH). As a general rule, neuroradiologists suggest a diagnosis of CM-I when the tip of one or both tonsils extends more than 3–5 mm below the basion-opisthion line (McRae line) [6,7]. In healthy adults, the cerebellar tonsils are usually located above the foramen magnum; however, textbooks continue to define TH into the spinal canal as >3–5 mm beyond McRae line [8].
According to the degree of TH, several subtypes have been introduced, including CM-0, CM-0.5, and CM-1.5. Iskandar et al. described a group of pediatric patients with syringomyelia, but without TH, who improved following posterior fossa decompression and C1 laminectomy, suggesting a Chiari-like pathophysiology [9-13]. They hypothesized that a small posterior fossa and a tight cisterna magna could explain this condition, leading to the introduction of CM-0 and CM-0.5 [9-13]. Tubbs et al. later coined the term CM-1.5 to define a phenotype intermediate between CM-I and CM-II, but not associated with neural tube defects [14]. The hallmark of CM-1.5 is the presence of TH along with caudal descent of the brainstem, with the obex located below McRae line [14]. Despite presenting with symptoms similar to those of CM-I, these variants may require different surgical strategies, making accurate classification critical [9-15].
What follows is a discussion of the pathogenesis, classification, and surgical management of CM-I. Few topics in neurosurgery remain as debated as the classification and treatment of this condition. This review focuses on: (1) clarifying the mechanisms behind hindbrain descent; (2) proposing a new classification of CM-I based on morphometric and volumetric studies; (3) suggesting surgical approaches tailored to the underlying pathogenesis; and (4) reviewing recent genetic research related to the mechanisms of hindbrain descent.
HYPOTHESES REGARDING THE CAUSES AND MECHANISMS OF CM-I INVOLVING THE HINDBRAIN DESCENT
If we consider the mechanism of hindbrain descent in a simplified manner, the following hypotheses can be proposed:
Mechanism #1: Constriction of the posterior cranial fossa (PCF). The PCF, which serves as a container for the hindbrain, is too small, causing the normally developed hindbrain to herniate into the spinal canal.
Mechanism #2: Overgrowth of the hindbrain. Although the PCF continues to grow and develop normally, the hindbrain is too large, leading to its herniation into the spinal canal.
Mechanism #3: Traction due tethering. The hindbrain is pulled downward from below (caudally), resulting their herniation into the spinal canal.
Mechanism #4: “Cranial settling” and “posterior gliding” of occipital condyle due to instability and hypermobility at the occipito-atlas-axis joints (craniocervical junction [CCJ]).
This article reviews these 4 hypotheses about mechanism of the hindbrain descent, including past reports and the author’s own findings. Furthermore, it delves deeper into their pathogenesis by drawing on insights from developmental biology, genetic studies, and experimental research. At first, we tried to clarify the mechanism of the hindbrain descent based on the volumetric and morphometric analyses of PCF and CCJ, and discussed the pathogenesis of the hindbrain descent induced from the phenotype clarification.
MATERIALS: VOLUMETRIC AND MORPHOMETRIC ANALYSES OF PCF AND CCJ
1. Normal Controls and Inclusion Criteria
This study recruited 150 normal volunteers (controls) with no neurological symptoms or abnormalities in the neural axis (aged 4–49 years; mean, 18.7 years; 64 males and 86 females). In addition, 449 patients with CM-I (cerebellar tonsil herniation ≥5 mm from the McRae line, i.e., between the basion and opisthion) presence of brainstem symptoms and/or myelopathy were included (aged 4–49 years; mean, 18.1 years; 204 males and 245 females) and 37 patients with CM-borderline (cerebellar tonsil herniation <5 mm from the McRae line, but presence of brainstem symptoms and/or myelopathy) (aged 5–45 years; mean, 18.4 years; 15 males and 22 females). The distribution and mean of age and sex were not significantly different between normal controls and CM-I group. These patients were diagnosed and treated between April 2006 and March 2020. All of normal controls and patients were selected from Asian (Asian American, Asian European, Chinese, Korean, and Japanese), considering ethnic variation. Both normal controls and patients were managed in The Chiari Institute North Shore University Hospital, New York USA, Osaka City University Hospital, Osaka City General Hospital and Moriguchi Ikuno Memorial Hospital.
2. Exclusion Criteria
The cases who had myelopathy and/or radiculopathy due to disc hernia, spondylotic change and ossification of the longitudinal ligament were excluded. The cases older than 50 years old was excluded, because they were affected the factors of ageing of the brain and degeneration of the bony structures etc. The infant and toddler younger 4 years old were excluded, because in this age the development and growth is very rapid and there is risk by radiation.
METHODS: VOLUMETRIC AND MORPHOMETRIC ANALYSES OF PCF AND CCJ
1. Volumetric and Morphometric Analyses of PCF and Hindbrain (Figs. 1 and 2)
Volumetric calculation of the posterior cranial fossa (PCF). (A) Three-dimension computed tomography (3D-CT) reconstructed image of the PCF using Osirix software. (B) Two-dimensional (2D)-CT images of the PCF. Red areas indicate the volume of PCF. (C) 2D-CT image demonstrating the area of the inferior outlet of the foramen magnum at the level of the basion and opisthion. (D) 2D-CT image demonstrating the area of the superior outlet of the foramen magnum at the level of the jugular tubercle. The PCF was defined as the almost circular space bounded by the tentorium cerebelli, occipital bone, clivus, petrous bone, and petrous ridges. The ridges of the petrous bones form the anterolateral border of the cavity, and their connection to the posterior clinoids (posterior petroclinoid ligament) forms the anterior border. The caudal end of PCF was defined as the foramen magnum, including the McRae line. McRae line was defined as the line between the basion and opisthion. The volume of brain in PCF was calculated as the neural content of the PCF, including the cerebellum, mesencephalon, pons, and medulla (blue areas in A, and red areas in B). A, anterior; P, posterior; L, left; R, right; S, superior; H.C., hypoglossal nerve canal; J.F., jugular foramen; J.T., jugular tubercle; O.C., occipital condyle.
Morphometric measurements of enchondral parts of occipital bone. (A) Two-dimension computed tomography (2D-CT) midline sagittal image demonstrating morphometric measurements of the basiocciput (clivus) (left black double arrow) and supraocciput (right black double arrow). (B) 2D-CT coronal image at the hypoglossal nerve canal (hgc) and jugular tubercle (jt). Demonstrating morphometric measurements of the exoocciput (condyle) (double lack arrows). jf, jugular foramen; O.C., articular process of occipital condyle (white dotted lines). (C) Magnetic resonance imaging midline sagittal image demonstrating morphometric measurements of the axial length of the brain stem (BSL) (large black dotted line), medullary height (MH) (black small dotted line), and the position of the 4th ventricle (4VH) (black small dotted line). IOP, the internal occipital protuberance; TDS, top of tuberculum sellae (TS); Twining line, the line between the TS and internal occipital protuberance (IOP); McRae line, the line between the basion and opisthion; TH, tonsilar herniation (black arrow). Measurements included the axial length of the clivus (basiocciput and basisphenoid) from the top of the dorsum sellae to the basion; the axial length of the supraocciput from the center of the IOP to the opisthion; and the axial length of the occipital condyle (exocciput) from the top of the jugular tubercle to the bottom of the occipital condyle and the supraocciput (distance between the opisthion and the center of the IOP): these 3 measurements were unified as occipital bone size. The axial BSL was defined between the midbrain-pons junction and medullo-cervical junction. MH was defined the vertical distance between the ponto-medullary junction and McRae line. The position of the 4VH was defined as the vertical distance between the fastigium (transverse summit of the roof of fourth ventricle) of the 4th ventricle and Twining line. TH means the length of herniation of the cerebellar tonsils.
We used 2D- or 3-dimensional (3D)-CT reconstructed images with Osirix software (free access) to calculate the volume of PCF (PCFV) (Fig. 1A and B), volume of surrounding area of foramen magnum (VAFM, i.e., total volume between the inferior and superior outlets of the foramen magnum) (Fig. 1C and D) [16-18]. PCFV was divided into the volume of the area of above Twining line and below Twining line. The volume of the brain in the posterior cranial fossa (PBFV) was calculated on MRI axial and sagittal images, excluding the 4th ventricle (4VH), herniated cerebellar tonsils, and medulla oblongata.
On 2D-CT coronal images at the occipital condyle and midline sagittal plane, the axial length of the basiocciput (clivus), exocciput (condyle), and supraocciput as enchondral parts of the occipital bone (occipital bone size) were measured (Fig. 2A and B). Using MRI midline image, the axial length of the brain stem, medullary height (MH), and the position of 4VH were measured (Fig. 2C) [6-18]. CM-I was categorized on the basis of multiple analyses of the results. The radiographic analysis was performed using Osirix software.
2. Morphometric Measurements and Evaluation of CCJ Instability (Fig. 3)
Mporphometric measurement of craniocervical junction (CCJ) in the cases with hereditary disorders of connective tissue. (A) Two-dimension computed tomography (2D-CT) midline image of CCJ demonstrating reference lines. Line A means plane of the posterior surface of anterior arch of atlas. Line B means line between the lowest point of anterior arch of atlas and the lowest point of posterior arch of atlas. Line C means superior plane of the clivus. Line D means plane of posterior surface of the dens. Line E means plane of the top of anterior arch of atlas and the lowest point of posterior arch of atlas. (B and C) 2D-CT midline image demonstrating morphometric measurements. ADI, interval between anterior arch and dens; BAI, interval between basion and line A; BDI, interval between basion and top of dens; DAI, interval between top of dens and line B; AXA, angle between axis (line D) and atlas (line E); CXA, angle between clivus (line C) and axis (line D).
CCJ instability was diagnosed using a morphometric study and craniocervical traction test, which were described by the authors and Goel et al. in past reports and a textbook [16,18-23]. For the craniocervical traction test, a tong was attached to the skull under intravenous anesthesia in the supine position and morphometric measurements were performed. Then, the patient was placed in an upright position and craniocervical instability was assessed on the basis of neurological symptoms, followed by repeat morphometric measurements. Cases with displacement of the occipito-atlanto-axial joints (>1.5 standard deviation [SD]) were defined as having instability. Craniocervical traction using 10–15 kg was applied, which resolved the neurological symptoms and led to a reduction of CCJ.
3. Statistical Analyses
The patient groups were compared with normal controls between groups divided into every 3 years old group. SPSS ver. 11.0 (SPSS Inc., USA) was used for statistical analyses. The means were compared between patients and normal controls using the Mann-Whitney U test. A p-value <0.01 was used to determine significance. For multiple analyses, heavy palindromic tests were used. Abnormalities of PCFV, VAFM, and occipital bone size were defined as values below 1.5 SD.
RESULTS: VOLUMETRIC AND MORPHOMETRIC ANALYSES OF PCF AND CCJ
The data along the development and growth in normal controls matched the previous reports [24,25]. Using multivariate regression analysis with PCFV, VAFM, and the size of the cartilaginous components of the occipital bone as explanatory variables, CM-I can be classified into 3 independent groups: CM-I types A, B, and C, and CM-borderline. The characteristics of these groups are summarized in Tables 1, 2 and Fig. 4 [16-18]. CM-I types A (160 cases), B (156 cases), and C (131 cases). CM-borderline (37 cases) indicated borderline small occipital bone size and small VAFM (-1.0 to -1.5 SD).
Subtypes of CM-I classified based on the morphometric analyses, and the cause of hindbrain descent and surgical interventions
Other mechanisms of the brainstem and cerebellum in CM-I type A, CM-borderline, and surgical intervention including secondary hindbrain descent
Illustrative cases of Chiari malformation type I (CM-I) subtypes and Chiari malformation type II (CM-II). (A) CM-I type A: normal volume of the postrior cranial fossa (PCFV), normal volume of the surrounding area of foramen magnum (VAFM), normal occipital bone size. (B) CM-I type B: normal PFCV, small VAFM and occipital bone size, small of the volume of the below Twing’s line (white doted trapezoid area) and compensatory expansion of the above Twing’s line area (black doted trapezoid area). (C) CM-I type C: small PCFV, VAFM and occipital bone size, dowmward dispalcemet of the brain stem and cerebellum (doted arrows), elongation of the brain stem (white arrow). (D) CM-II: small PCFV, VAFM, and occipital bone size, downwad displacement of the btrain stem and cerebellum (white doted arrow) and elongation of the brain stem (white arrow). (E) CM-I type A: The case associated with hereditary disorders of connective tissue (HDCT): normal volume of the PCFV, normal VAFM, normal occipital bone size. Magnetic resonance (MR) midsagittal image in supine position showing normal BDI (7.7 mm), normal BAI (3.5 mm), normal CXA (141°), large retro-odontoid pannus (asterisk), and low-lying cerebellar tonsils. (F) CM-I type A: The case associated with tethered cord syndrome (TCS): mall PCFV, VAFM and occipital bone size, downwad displacement of the btrain stem and cerebellum (white doted arrow), elongation of the brain stem (white arrow) and large foramen magnum (FM). MR midsagittal image demonstrating the character of CM- I type A with TCS. PCFV, VAFM and occipital bone size are normal size. MR image shows elongation and downward displacement of the brain stem and cerebellum and large supuracerebellar cistern (double asterisk). BDI, the distance between basion and top of dens; BAI, the distance between basion and anterior arch of atlas; CXA, the angle between clivus and axial.
1. Results of Volumetric and Morphometric Analyses of PCF and CCJ
PCFV was not significantly different between the cases and controls in all of age groups. Of the measured items, only PCFV, occipital bone size, and VAFM were significantly different between the cases and healthy controls. In CM-I type B, C and CM-borderline, PBFV/PCVF ratio in older than 4 years old age groups were signify larger than them in normal controls.
CM-I type A had normal PCFV, VAFM, and occipital bone size (Fig. 4A) in all of age groups. CM-I type B had normal PCFV, small VAFM and occipital bone size, large volume of PCFV above Twining line, and small PCFV below Twining line (Fig. 4B) in all of age groups. CM-I type C had small PCF, VAFM, and occipital bone size in all of age groups (Fig. 4C). In CM-I type A, MH was decreased and 4VH was increased compared to controls in older than 8 years old. In CM-I type C, MH was decreased and the position of 4VH was increased compared to controls in all of age groups since birth. CM-borderline had normal PCFV and occipital bone size, but borderline small VAFM after older than 8 years old (Fig. 4E).
2. Results of Morphometric Analyses of CCJ in Craniocervical Traction Test (Table 3)
Results of morphometric measurements of craniocervical junction in craniocervical traction test
The distance between anterior arch of atlas and dens (DAI) was used to evaluate the atlanto-axial horizontal dislocation. Change in ADI of >4 mm between flexion and extension is an accepted index of atlanto-axial horizontal dislocation [16,18-23]. The distance between dense and anterior arch of atlas (DAI) was used to evaluate the vertical basilar invagination. DAI>15 mm from the top of dens above McRae line is an accepted index of basilar invagination [16,18-23]. Furthermore, the morphometric analyses during craniocervical traction test revealed BDI (the distance between basion and top of dens), BAI (the distance between basion and anterior arch of atlas), and CXA (the angle between clivus and axial) as reliable and reproducible measurements in patients with instability of CCJ.
In the supine position, BDI, BAI, and CXA were normal. However, in the sitting position, the BDI decreased, BAI increased, and CXA decreased. During traction, these values returned to the normal range. In the follow-up evaluation of patients who underwent joint stabilization, normal values of BDI, BAI, and CXA were maintained. However, in patients without joint stabilization, these values were abnormal.
In 70 cases with instability, hereditary disorders of connective tissue was present in 28 cases, bony anomalies (including basilar invagination) in 43 cases and trauma in 18 cases. Out of CM-I type A, CM-I type B, CM-I type C, and CM-borderline, 58 (36.3%), 3 (1.9%), 5 (3.8%), and 4 (10.8%) had instability, respectively.
The cases with instability in CM-I type A were the most, there was significant more than other groups (p<0.01).
DISCUSSION
1. CM-I: Mechanism #1. Constriction of the PCF, CM-I Types B and C
To provide support for Mechanism #1, in which the PCF is abnormally small, resulting in the herniation of the hindbrain into the spinal cord, it is necessary to demonstrate that PCFV is reduced, while the volumes of the brainstem and cerebellum remain within normal limits. There are many reports which described that PCFV in CM-I is small [26-32].
The total volume of the posterior fossa including the area enclosed by the cerebellar tentorium did not differ significantly from that of the control group, the volume of the portion bounded by the occipital bone and located caudal to Twining line was significantly reduced (Fig. 4B and C) [16-18]. In contrast, the volumes of the brainstem and cerebellum were not significantly different compared to the normal group [16-18].
Furthermore, how the narrowing of PCF in CM-I occurs from both anatomical and embryological perspectives was considered. According to reports by the anatomist, embryologist, and teratologist O’Rahilly, the occipital bone develops from 3 cartilaginous components, i.e., the basiocciput (clivus), exocciput (condyle), and supraocciput (convexity) [33-35].
1) Neuroradiological considerations
Morphometric measurements were performed on the neuroradiological images. These 3 parts (clivus, condyle, and supraocciput) of the occipital bone were measured using x-ray and computed tomography scans from patients [16-18], and all were found to be hypoplastic.
Moreover, as the cerebellar tentorium steepened, it compensated for the narrowing of the bony elements caudal to Twining line, resulting in no significant difference in the overall volume of the PCF compared to control groups (Fig. 4B and C) [16-18]. This phenomenon was named “remodeling of the posterior fossa.” However, the ratio of the PCF volume to the volumes of the brainstem and cerebellum—as proposed in Hypothesis #1—was indicated a constricted PCF [16-18].
2) Experimental considerations
Marin-Padilla and colleagues reported that the excessive administration of retinoic acid to pregnant hamsters induced encephalocele and Chiari malformation in their fetuses [36-49]. Morphological measurements of the occipital bone, brainstem, and cerebellum were conducted in these specimens. Retinoic acid affected the expression of Hox genes, which regulate regionspecific morphological differentiation, leading to backward displacement [40-44].
This morphological transformation of the vertebrae resulted in severe malformations of the occipital bone—located at the most cranial end and at the craniovertebral junction. In contrast, excessive retinoic acid exposure led to encephalocele, herniation of the hindbrain into the spinal canal (Chiari malformation), and severe malformation of the cranial base derived from the paraxial mesoderm [40-44]. The details of this phenomenon were described, including genetic analysis.
3) Skull base abnormalities in the context of embryological mesodermal defects
Patients with CM-I of the atlas, axis, and/or subaxial cervical vertebrae include those with Klippel-Feil deformity, complete or partial atlanto-occipital assimilation, and hypoplasia, dyspla-sia, or agenesis of the atlas, among others. These findings support a mesodermal disorder and abnormal segmentation of the somite-derived cervical vertebral bodies as underlying causes of CM-I [45,46]. Such bony anomalies likely contribute to the narrowing of the PCF.
4) Genetic research: insufficiency of PCF development
Marin-Padilla and Marin-Padilla demonstrated one potential mechanism for CM-like development using a hamster model [36-39]. Additionally, CM-I has been reported in patients with mutations in retinoic acid receptor beta [47]. Retinoic acid, a metabolite of dietary vitamin A, plays multiple roles in embryonic development and cellular differentiation [48,49]. Furthermore, retinoic acid regulates fibroblast growth factor (FGF) signaling during development [50]. Similarly, gain-of-function mutations in FGF receptor 3 result in defects in posterior fossa morphology in mouse models, including alterations in the shape and size of the foramen magnum [51]. FGF receptor mutations are also linked to disorders such as craniosynostosis and achondroplasia.
Craniosynostosis, the premature closure of one or more sutures between the skull plates, is often linked with CM-I [26]. It can occur as part of syndromes that are also associated with CM-I, such as Crouzon syndrome [26]. Specifically, recurrent mutations affecting FGF signaling in craniosynostosis may also contribute to CM-I.
Beyond retinoic acid and FGF signaling, other genes involved in related developmental processes may play a role. At least 7 individual cases of mutations in the ETS2 repressor factor gene have been associated with CM-I and craniosynostosis [51]. ETS2 repressor factor may act downstream of FGF signaling during mesodermal induction [52]. A small posterior fossa may also be linked to other craniocervical bone anomalies, such as occipital dysplasia, commonly resulting in an underdeveloped occipital bone and reduced posterior fossa size [17,29].
Recently, Brockmeyer et al. [53] built on data from Abbott et al. [54] using the Utah Population Database to identify a candidate variant in homeobox C4 that may affect the development of the junction between the mesoderm and neural crest, leading to a decreased clivo-axial angle and CM-I. Matsuoka et al. [55] proposed that widening of the foramen magnum, caused by post-otic neural crest–derived mesenchymal stem cells being respecified from bone to connective tissue, may also contribute to CM-I.
2. CM-I: Mechanism #2, CM-I Type A
Embryogenic growth dysregulation is a rare disorder that results in the overgrowth of multiple somatic tissues, including the central nervous system. Macrocerebellum is characterized by an abnormally large cerebellum with preserved overall morphology [56,57]. Evaluation of cerebellar architecture in these disorders reveals that the cerebellar hemispheres are enlarged due to increased cerebellar gray matter [58]. A developmental excess of neural tissue within a normal-sized posterior fossa may cause the cerebellar tonsils to descend. In this case, PBFV is large, but occipital bone size, PCFV are normal, and so PBFV/PBFV is large. Therefore, these cases were classified into CM-I type A.
In macrocephaly-capillary malformation syndrome, approximately 50% of patients exhibit progressive tonsillar descent associated with rapid brain growth and increased crowding of the posterior fossa during infancy [58]. In a longitudinal study of 67 patients, Conway et al. [58] found that although newborns had normal cerebellar size, they developed cerebellar overgrowth and tonsillar descent during the first year of life, despite normal posterior fossa volume. Therefore, in these syndromes, cerebellar overgrowth results in a structural mismatch between the skull and neural tissue.
3. CM-I: Mechanism #3, CM-I Type A
Mechanism #3 proposes that the brainstem and cerebellum are pulled downward from below (caudally), resulting in their herniation, e.g., tethered cord syndrome and CM-II.
Milhorat, Bolognese, and Nishikawa reported Chiari malformation in patients with tethered cord syndrome [59,60]. In several cases, sectioning of the filum terminal resulted in an upward shift of the brainstem and cerebellum. They also documented cases of occult tethered cord syndrome and scoliosis associated with Chiari malformation, in which sectioning of the filum terminal led to an improvement in brainstem and cerebellar descent. Surgical outcomes were reported and published [59,60].
Although circumstantial evidence has accumulated, the traction hypothesis (hypothesis #3) remains unproven. Animal studies involving cats and dogs have been conducted, but when slight tension was applied to the caudal lumbar nerve roots, the brainstem and cerebellum did not descend, likely due to the elasticity of the spinal cord.
In contrast, in CM-II, the brainstem and cerebellum descend caudally from the fetal stage, presumably due to traction from a meningocele in combination with hydrocephalus (Fig. 4C) [36-39,61].
4. CM-I: Mechanism #4, CM-I Type A
1) Neuroradiological considerations
Milhorat, Bolognese, and Nishikawa reported patients with hereditary connective tissue disorders who experienced symptoms such as dizziness, limb numbness, swallowing difficulties, and shortness of breath when sitting or standing [62,63]. These symptoms were relieved by traction. Using open MRI, these cases were examined in the sitting position. While these phenomena did not occur in the supine position, the cerebellar tonsils were found to descend into the spinal canal when the patients were seated (Fig. 5), which was termed “cranial settling.” In these patients, the volume of the PCF and size of the cartilaginous components of the occipital bone were within normal ranges. Morphological measurements revealed “cranial settling” and a “posterior gliding” phenomenon (Fig. 5), in which the occipital condyles slid backward relative to the atlanto-occipital joint. Upon applying cranial traction, these neurological symptoms disappeared, and measurements showed. This condition of at the occipito-atlas-axis joints was observed frequently in patients with hereditary connective tissue disorders, as the pilot data were reported by Milhorat, Bolognese, and Nishikawa [62,63].
Morphometric measurement of craniocervical junction in the cases with hereditary disorders of connective tissue. (A) Photo showing the bony model of left occipitto-atlantal joint. The occipital condyle glides to posterior against atlas on sitting position. Red dotted line: the margin of atlanto-occipital joint. Blue arrow: posterior gliding and anterior flex of the occipital condyle. (B and C) The results midsagittal reconstructed 2-dimension computed tomography scanning with the patient in the supine position, showing BAI of 2.0 mm (lower left) and normal articulation of occipital condyle (OC) (lower left). (D and E) The results plain radiography showing an increase of BAI (5.0 mm) in the sitting position and a decrease of BAI (1.5 mm) during traction with a 20-lb (9.1 kg) weight. The occipito-atlantal joint (outline by dotted line) glides posteriorly in the sitting position (curved arrow) and its reduced by traction (straight arrow). Line A, plain of the posterior surface of the anterior arch of the atlas.
2) Genetic research: insufficiency of connective tissue
In addition to the skull, connective tissue plays a crucial role in supporting the cerebellum. Boyles et al. [64] performed linkage analysis on 23 families with 71 affected individuals. Of the 10,000 single nucleotide polymorphisms analyzed, they identified possible associations with chromosomal regions 15q21.1–22.3 and 9q21.33–33.1.64 The 15q21.1–22.3 region contains the fibrillin 1 gene, which is linked to Marfan syndrome and Shprintzen-Goldberg syndrome, both known to co-occur with CM-I [65]. Other connective tissue disorders, such as Ehlers-Danlos syndrome type 3 (hypermobility type), Loeys-Dietz syndrome, Stickler syndrome, and self-identified hypermobility, have also been associated with CM-I [65,66]. These associations may be related to increased tissue mobility and decreased stability, which could facilitate cerebellar herniation. Mutations in collagen type IV alpha 1 chain have likewise been associated with CM-I [67].
DECISION OF SURGICAL INTERVENTIONS: ALGORITHM FOR CHOOSING SURGICAL INTERVENTION AND TREATMENT TO ASSOCIATED SYRINGOMYELIA
1. Algorithm for Choosing Surgical Intervention
We propose the algorism for surgical; intervention, the important measurements are the occipital bone size, PCFV and VAFM (Fig. 6).
Algorithm for descicion of surgical management. Blue words: morohometric parameters for classification, Green words: the pathogenesis/mechanism of hindbrain descent, Red words: surgical intervention. SD, standard deviation; CM-I, Chiari malformation type I; PCFV, the volume of the posterior cranial fossa; VAFM, the volume of foramen magnum; CCF, cranioicervical fixation; ESCP, expansive suboccipital cranioplasty; FMD, foramen magnum decompression.
The primary goal of surgical treatment for Chiari malformations and its related disorders is decompression of the brainstem and cerebellum, as well as the resolution and normalization of cerebrospinal fluid (CSF) flow obstruction at the foramen magnum. To achieve these goals, fundamental surgical intervention must address the underlying pathogenesis and mechanisms of hindbrain descent. Therefore, we selected foramen magnum decompression (FMD) for CM-I type B and CM-borderline, because VAFM and occipital bone size were small in these pa-tients [68]. The authors suggested that FMD is adequate for CM-I type B and CM-borderline with small VAFM but normal PCFV. Conversely, in CM-I type C, it is necessary to expand the entire PCF and remodel the relationship between the brain and occipital bone. According to Sakamoto et al. [69], we selected expansive suboccipital cranioplasty (ESCP) for CM-I type C because FMD was not adequate under the setting of small VAFM, small PCFV, and small occipital bone size. In CM-I type A, other surgical methods that can treat ptosis of the brainstem and cerebellum must be selected. Craniocervical posterolateral fixation (CCF) should be selected for cases with CCJ instability causing functional cranial settling [19,62,70-72]. Initially, it is important to identify the CCJ with instability. Next, it is important to select the appropriate methods of screw insertion. The most important anchor for screw insertion is C2 and pedicle screw for C2 is the ideal.
2. Surgical Intervention to Syringomyelia Associated With CM-I
Among the disorders associated with Chiari malformation, syringomyelia is the most common and clinically significant condition. In cases of syringomyelia, obstruction of CSF flow at the foramen magnum—caused by hindbrain descent, as seen in CM-I—is considered to be the main pathophysiological mechanism [73,74]. Clinically, the main symptoms are spinal cord-related due to syringomyelia, but neurological symptoms resulting from compression of the brainstem and cerebellum, such as cranial nerve palsies and headaches due to CSF flow disturbances, are also frequently observed.
MATERIALS AND METHODS: SURGICAL INTERVENTION FOR CM-I, CM-BORDERLINE, AND INSTABILITY
1. Surgical Indications and Patient Population
Surgery was performed in cases of myelopathy, upper cervical cord symptoms, brainstem symptoms, and Japanese Orthopaedic Association (JOA) score <14 [75].
In total, 484 patients underwent surgical treatment: 302 underwent FMD, 104 underwent ESCP, and 70 underwent CCF. 10 patients underwent ventricle peritoneal shunt and 10 underwent section filum terminal and/or lysis of adhesion of the spinal cord.
Patient group for surgical interventions was same as that for the volumetric and morphometric study. The patients were treated between April 2006 and March 2020 at the institute or hospitals as described above.
2. Operative Procedures for CM-I Types B and C, and CM-Borderline
FMD consists of craniectomy of the surrounding area of the foramen magnum (2–3-cm square), decompression of the brainstem and cerebellum, and creation of major cisterns and flow out of CSF from the foramina of Magendie and Luschka [68]. ECSP involves extensive decompression to expand the surrounding area of foramen magnum and the entire PCF (along the transverse sinus and sigmoid sinus), C1 laminectomy, dural plasty, and osteoplasty to maintain the PCF expansion [69]. ESCP can achieve normalized PCFV and major cisterns, which is adequate for the decompression of the brainstem and cerebellum, and to remodel the appropriate positional relationship between the brainstem, cerebellum, and occipital bone by expanding the entire PCF, including C1.
In children younger than 12 years, the cranium and brain are under development, so the possibility of disproportion between the cranium and brain changes over time. For patients younger than 12 years, FMD was performed first.
3. Operative Procedures for CM-I Type A and Instability
CCF was performed for CCJ instability [19,70]. If the patient had instability at CCJ joints, occipitocervical posterolateral fixation (OCF) should be performed. For patients with instability at the atlanto-axial joints alone, C1–2 posterolateral fixation (C1–2 FIX) should be performed. In cases where pedicle screw could be inserted for C2, fixation of C2 with screws was adequate. In cases where screw insertion was not possible, it was necessary to fix C2 to C3 or C4.
CCF was performed in a total of 70 cases, of which C1–2 FIX was performed in 36 cases and OCF in 34 cases. For cases (7 cases) with CCJ instability associated with CM-I types B and C, and CM-borderline, staged CCF combined with ESCP or FMD was performed.
In children younger than 12 years, the relationships between the anatomical structures are under development, so there is a possibility of overestimating the instability. In children, the loss of function of occipito-atlanto-axial joints induces significant handicap. Therefore, CCF was not performed in children younger than 12 years.
4. Neurological and Neuroradiological Examinations, Follow-up, and Statistical Analyses
Postoperatively, the neurological symptoms and signs, JOA score, recovery rate of JOA score [75] (JOA score RR) ([postoperative points–preoperative points]/[full points–preoperative point] ×100%) , and neuroradiological findings (cervical spine dynamic x-ray, 2D-CT, and MRI of the cervical spine) were examined every 3 months. Persistent syringomyelia was defined by confirmation of syringomyelia larger than 2 mm in diameter. Surgical outcome was assessed using chi-square and Fisher tests. A p-value <0.01 was used to determine significance.
RESULTS: SURGICAL INTERVENTION FOR CM-I, CM-BORDERLINE, AND INSTABILITY
The follow-up duration was 24–216 months (mean, 88.5 months). The results were based on the most recent data. In total, 58 patients were lost to follow up after more than 3 years of operation, in whom, the results of the final examination were estimated.
1. Results of JOA Score and RR in FMD and ESCP (Table 4)
JOA score and its recovery rate after FMD and ESCP
In 12 and older than 12 years old adolescence and adult cases, the improvement rate of neurological symptoms and signs in FMD was 87.8%. The JOA score RR in FMD was 58.7 %, while 92.2% of cases showed improvement or stabilization of neurological symptoms. In 14 cases (7.8%), the neurological symptoms deteriorated during follow-up. In 10 cases out of 127 cases (7.9%) with syringomyelia, the symptoms persisted. The improvement rate of neurological symptoms and signs in ESCP was 88.2%. The JOA score RR in ESCP was 60.2%, and 98.0% of cases had an improvement or stabilization of neurological symptoms. In 2 cases (2.0%), the neurological symptoms deteriorated during follow-up. In 1 case out of 38 cases (2.6%) with syringomyelia, the symptoms persisted. There was no significant difference between the FMD and ESCP groups in terms of improvement or stabilization of neurological symptoms and the JOA score RR. Persistent syringomyelia and deteriorated cases of neurological symptoms and/or signs were significantly more common in patients who underwent FMD compared to ESCP (p>0.01).
In younger 12 years old child, improvement rate of neurological symptoms and/or signs were significantly lower in CM-I types A (68.9%) and C (69.0%) than them in CM-I type B (80.5%) and CM-borderline (85.7%) (p>0.01). JOA score RR was significantly lower in CM-I type C (48.6%) than them in CM-I types A (57.2%), B (58.2%), and CM-borderline (54.3%) (p> 0.01). Persistent syringomyelia and deteriorated cases of neurological symptoms and/or signs were significantly more common in patients of CM-I types A (33.3%, 31.1%) and C (50.0%, 34.5%) in patients of CM-B (3.3%, 4.2%) and CM-borderline (0%, 0%) (p>0.01).
2. Results of JOA Score and RR in CCF (Table 5)
JOA score and its recovery rate after craniocervical fixation
The JOA score RR among the 70 cases who underwent CCF was 69.7%. In total, 46 cases (67.1%) had complete bony fusion, 59 cases (84.3%) had stabilized joints, and 11 cases (15.7%) had incomplete stabilization. Syringomyelia resolved in all 21 cases.
Preoperative JOA score in C1–2 FIX was significantly lower than that in OCF. In 36 cases who underwent C1–2 FIX, the JOA score RR was 78.7%. In total, 26 cases (72.2%) had complete bony fusion, 32 cases (88.9%) had stabilized joints, and 4 cases (11.1%) had incomplete stabilization. The JOA score RR in the 34 cases who underwent OCF was 63.5%. In total, 21 cases (61.8%) had complete bony fusion, 27 cases (79.4%) had stabilized joints, and 7 cases (20.6%) had incomplete stabilization. In cases who underwent C1–2 FIX, the JOA score RR, rate of stabilization, and bony fusion of joints was higher than those in cases who underwent OCF.
3. Complications and Side Effects
In the FMD and ESCP groups, there was no mortality or permanent morbidity. Transient morbidity occurred in 10 cases (2.5%). Complications were observed in 6 cases (2.6%) with FMD and 4 cases (3.8%) with ESCP. Two cases (0.7%) with FMD had wound infection, in 2 case (0.7%), in 1 case (0.3%) with FMD had arachnoid adhesion, in 1 case (0.3%) with FMD had CSF leakage. Two cases (3.8%) with ESCP had wound infection, in 1 case (1.0%), cerebellar slugging occurred without neurological symptoms, in 1 case (1.0%) with ESCP had CSF leakage. Repetitive and persistent neck pain continued in 3 cases (2.9%) with ESCP. There were significantly more complications and side effects in ESCP than FMD.
In both C1–2 FIX and OCF, there was no mortality or permanent morbidity. Transient morbidity occurred in 2 cases (2.9%). The complications included transient swallowing disturbance in 1 case (1.4%) in OCF and injury to the vertebral artery without neurological symptoms in 1 case (1.4%) in C1–2 FIX. Repetitive and persistent neck pain continued in 3 cases (8.8%) of OCF and functional loss of visual fields occurred in 2 cases (5.9%). There were significantly more complications and side effects in OCF than C1–2 FIX.
DISCUSSION: CLINICAL IMPACT BY APPLICATION OF THE NEW CLASSIFICATION SYSTEM
Compared to previous reports involving only FMD, our results showed improved outcomes with fewer reoperations. In particular, the improvement rate for brainstem compression symptoms was markedly better, suggesting effective brainstem decompression and normalization of CSF flow dynamics. When posterior fossa narrowing is confined to the foramen magnum region, FMD alone appears sufficient for decompression and CSF flow restoration. Additionally, morbidity and complication rates were lower in our cohort [76-79]. These findings support the appropriateness of selecting the surgical procedure (ESCP or FMD) based on morphometry [80-82].
C1–2 FIX and OCF both resulted in high rates of neurological symptom improvement, joint stabilization, and/or bony fusion [80-84]. Our outcomes surpassed those of previous reports, supporting the effectiveness of CCF in managing “cranial settling” and “posterior gliding” due to CCJ instability/hypermobility. However, OCF was associated with greater functional loss and more frequent complications due to fixation of the occipitoatlanto-axial joints than C1–2 FIX. Thus, fixation involving the occipital bone requires careful consideration. The CCJ at traction test has proved useful in detecting occipito-atlanto-axial instability/hypermobility and assessing the risks of functional loss after fixation [80-82].
CONCLUSIONS
We reviewed 4 existing hypotheses regarding the causes of hindbrain descent, along with the supporting studies for each. Through morphological measurements of PCF and CCJ, we were able to infer the mechanisms and pathogenesis of hindbrain descent. This contributes to the selection of appropriate surgical treatment strategies.
Morphological differences are considered phenotypes of the underlying genes. On the basis of morphological measurements and analyses of PCF, CCJ, hindbrain, Chiari malformation can be classified according to its causes. Many diseases involving hindbrain descent, as well as malformations and instability of the CCJ, are believed to originate during fetal development or even earlier at the genetic level.
As studies continue to reveal an increasing number of potential pathogenic associations, further approaches to understand the functional consequences of gene variants will be essential. We anticipate continued advances in our understanding of the genetics and etiology of CM-I, which we hope will lead to improved treatment selection and better clinical care for patients with CM-I.
Notes
Conflict of Interest
The authors have nothing to disclose.
Funding/Support
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Acknowledgments
The authors deeply thank Professor Emeritus Thomas H. Milhorat, M.D., Feinstein Institutes for Medical Research, Northwell Health, Manhasset, New York, for leading this review and proving support.
Research Ethics
This review has approved by the Institutional Review Board of Osaka Metropolitan University (OMH-2018-4243) and Moriguchi Ikuno Memorial Hospital (MIMH-2018-1). A written informed consent should be obtained from all subjects.
Author Contribution
Conceptualization: MN, PAB; Formal analysis: HS, MH; Investigation: MN, MY, KN, NK, HI; Methodology: MN, MY, KN, NK; Project administration: MN, TG; Writing – original draft: MN, PAB; Writing – review & editing: MN, TG, MH, KO.