Use of deep learning in the MRI diagnosis of Chiari malformation type I.
Ontology highlight
ABSTRACT: Purpose
To train deep learning convolutional neural network (CNN) models for classification of clinically significant Chiari malformation type I (CM1) on MRI to assist clinicians in diagnosis and decision making.Methods
A retrospective MRI dataset of patients diagnosed with CM1 and healthy individuals with normal brain MRIs from the period January 2010 to May 2020 was used to train ResNet50 and VGG19 CNN models to automatically classify images as CM1 or normal. A total of 101 patients diagnosed with CM1 requiring surgery and 111 patients with normal brain MRIs were included (median age 30 with an interquartile range of 23-43; 81 women with CM1). Isotropic volume transformation, image cropping, skull stripping, and data augmentation were employed to optimize model accuracy. K-fold cross validation was used to calculate sensitivity, specificity, and the area under receiver operating characteristic curve (AUC) for model evaluation.Results
The VGG19 model with data augmentation achieved a sensitivity of 97.1% and a specificity of 97.4% with an AUC of 0.99. The ResNet50 model achieved a sensitivity of 94.0% and a specificity of 94.4% with an AUC of 0.98.Conclusions
VGG19 and ResNet50 CNN models can be trained to automatically detect clinically significant CM1 on MRI with a high sensitivity and specificity. These models have the potential to be developed into clinical support tools in diagnosing CM1.
SUBMITTER: Tanaka KW
PROVIDER: S-EPMC9271110 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Chiari malformation is a congenital deformity leading to herniation of cerebellar tonsils. Headache is a typical symptom of this condition, but patients with Chiari malformation often present with double vision and vertigo. Examination of eye movements in such patients often reveals nystagmus and strabismus. Eye movement deficits in the context of typical symptomatic presentation are critical clinical markers for the diagnosis of Chiari malformation. We will review eye movement deficits that seen in patients with type 1 Chiari malformation. We will then discuss the underlying pathophysiology and therapeutic options for such deficits.
| S-EPMC4733655 | biostudies-literature
Project description:BackgroundChiari malformation type 1 affects approximately one in 1,000 people symptomatically, although one in 100 meet radiological criteria, making it a common neurological disorder. The diagnosis of underlying conditions has become more sophisticated, and new radiological markers of disease have been described. We sought to determine the prevalence and impact of additional comorbidities and underlying diagnoses in patients with Chiari malformation type 1 on symptomatology and surgical treatment.MethodsA retrospective review of 612 pediatric patients with a Chiari malformation type 1 diagnosis and imaging data evaluated between 2008 and 2018 was performed. Because of extensive clinical heterogeneity, patients were separated into four categories based on their primary comorbidities (nonsyndromic, central nervous system, skeletal, and multiple congenital anomalies) to identify associations with age of onset, radiographic measurements, syringomyelia, and surgical treatment.ResultsThe largest group had nonsyndromic Chiari malformation type 1 (70%) and the latest age at diagnosis of any group. In the syndromic group, 6% were diagnosed with a known genetic abnormality, with overgrowth syndromes being the most common. Patients with multiple congenital anomalies had the earliest Chiari malformation type 1 onset, the most severe tonsillar ectopia and obex position, and were overrepresented among surgical cases. Although there were no statistically significant differences between groups and rates of syrinx, we observed differences in individual diagnoses.ConclusionThe underlying diagnoses and presence of comorbidities in patients with Chiari malformation type 1 impacts rates of syringomyelia and surgery. Although most Chiari malformation type 1 cases are nonsyndromic, clinical evaluation of growth parameters, scoliosis, and joint hypermobility should be routine for all patients as they are useful to determine syringomyelia risk and may impact treatment.
| S-EPMC7156318 | biostudies-literature
Project description:Outcomes research on Chiari malformation type 1 (CM-1) is impeded by a reliance on small, single-center cohorts.To study the complications and resource use associated with adult CM-1 surgery using administrative data.We used a recently validated International Classification of Diseases, Ninth Revision, Clinical Modification code algorithm to retrospectively study adult CM-1 surgeries from 2004 to 2010 in California, Florida, and New York using State Inpatient Databases. Outcomes included complications and resource use within 30 and 90 days of treatment. We used multivariable logistic regression to identify risk factors for morbidity and negative binomial models to determine risk-adjusted costs.We identified 1947 CM-1 operations. Surgical complications were more common than medical complications at both 30 days (14.3% vs 4.4%) and 90 days (18.7% vs 5.0%) postoperatively. Certain comorbidities were associated with increased morbidity; for example, hydrocephalus increased the risk for surgical (odds ratio [OR] = 4.51) and medical (OR = 3.98) complications. Medical but not surgical complications were also more common in older patients (OR = 5.57 for oldest vs youngest age category) and male patients (OR = 3.19). Risk-adjusted hospital costs were $22530 at 30 days and $24852 at 90 days postoperatively. Risk-adjusted 90-day costs were more than twice as high for patients experiencing surgical ($46264) or medical ($65679) complications than for patients without complications ($18880).Complications after CM-1 surgery are common, and surgical complications are more frequent than medical complications. Certain comorbidities and demographic characteristics are associated with increased risk for complications. Beyond harming patients, complications are also associated with substantially higher hospital costs. These results may help guide patient management and inform decision making for patients considering surgery.
| S-EPMC4506249 | biostudies-literature
Project description:OBJECT:Chiari malformation Type I (CM-I) is characterized by hindbrain deformity. We investigated the effects of craniocervical decompression surgery on the anatomical features of hindbrain deformity with a prospective MRI study of patients with CM-I. METHODS:A prospective longitudinal study was conducted in 48 patients with CM-I (39 with syringomyelia) treated with craniocervical decompression. Clinical examinations and cervical MRI were performed before surgery and 1 week, 3-6 months, and annually after surgery. Hindbrain deformity was defined by tonsillar ectopia, pointed cerebellar tonsils, and/or cervicomedullary protuberance. The length of the clivus, basiocciput (sphenooccipital synchondrosis to basion), supraocciput (internal occipital protuberance to opisthion), and anteroposterior (AP) width of CSF pathways at the foramen magnum were measured and compared with those from 18 healthy volunteers (control group). RESULTS:Before surgery, the patients' posterior fossa bones were short and their CSF pathways were narrow. All patients had tonsillar ectopia (mean [± SD] 12.3 ± 5.1 mm; normal 0.3 ± 1.0). The majority of patients had pointed tonsils and more than two-thirds exhibited a cervicomedullary protuberance. Clivus and basiocciput lengths were significantly shorter than the values obtained in the control group. However, the supraocciput length did not differ significantly from control measurements. The mean bulbopontine sulcus distance superior to the basion was 9.5 ± 2.6 mm (vs 13.6 ± 2.8 mm in controls; p < 0.0001). The AP widths of the CSF pathways at the level of the foramen magnum were significantly narrowed. After surgery, CSF pathways significantly expanded both ventrally and dorsally. By 3-6 months after surgery, pointed tonsils became round, cervicomedullary protuberance disappeared, and tonsillar ectopia diminished by 51% (to 6.0 ± 3.3 mm; p < 0.0001). CONCLUSIONS:The cerebellar tonsils and brainstem assumed a normal appearance within 6 months after craniocervical decompression. These findings support the concept that the CM-I is not a congenital malformation of the neural elements but rather an acquired malformation that arises from pulsatile impaction of the cerebellar tonsils into the foramen magnum. Clinical trial registration no.: NCT00001327.
| S-EPMC3786329 | biostudies-literature
Project description:This study aimed to summarize the clinical features, diagnosis, and treatment of Chiari malformation type I- (CM-1-) associated syringobulbia. We performed a literature review of CM-1-associated syringobulbia in PubMed, Ovid MEDLINE, and Web of Science databases. Our concerns were the clinical features, radiologic presentations, treatment therapies, and prognoses of CM-1-associated syringobulbia. This review identified 23 articles with 53 cases. Symptoms included headache, neck pain, cranial nerve palsy, limb weakness/dysesthesia, Horner syndrome, ataxia, and respiratory disorders. The most frequently involved area was the medulla. Most of the patients also had syringomyelia. Surgical procedures performed included posterior fossa decompression, foramen magnum decompression, cervical laminectomy, duraplasty, and syringobulbic cavity shunt. Most patients experienced symptom alleviation or resolution postoperatively. A syringobulbic cavity shunt provided good results in refractory cases. Physicians should be aware of the possibility of syringobulbia in CM-1 patients, especially those with symptoms of sudden-onset brain-stem involvement. The diagnosis relies on the disorder's specific symptomatology and magnetic resonance imaging. Our review suggests that the initial therapy should be posterior fossa decomposition with or without duraplasty. In refractory cases, additional syringobulbic cavity shunt is the preferred option.
| S-EPMC6444244 | biostudies-literature
Project description:The management of Chiari I malformation (CMI) is controversial because treatment methods vary and treatment decisions rest on incomplete understanding of its complex symptom patterns, etiologies, and natural history. Validity of studies that attempt to compare treatment of CMI has been limited because of variable terminology and methods used to describe study subjects. The goal of this project was to standardize terminology and methods by developing a comprehensive set of Common Data Elements (CDEs), data definitions, case report forms (CRFs), and outcome measure recommendations for use in CMI clinical research, as part of the CDE project at the National Institute of Neurological Disorders and Stroke (NINDS) of the US National Institutes of Health. A working group, comprising over 30 experts, developed and identified CDEs, template CRFs, data dictionaries, and guidelines to aid investigators starting and conducting CMI clinical research studies. The recommendations were compiled, internally reviewed, and posted online for external public comment. In October 2016, version 1.0 of the CMI CDE recommendations became available on the NINDS CDE website. The recommendations span these domains: Core Demographics/Epidemiology; Presentation/Symptoms; Co-Morbidities/Genetics; Imaging; Treatment; and Outcome. Widespread use of CDEs could facilitate CMI clinical research trial design, data sharing, retrospective analyses, and consistent data sharing between CMI investigators around the world. Updating of CDEs will be necessary to keep them relevant and applicable to evolving research goals for understanding CMI and its treatment.
| S-EPMC7054710 | biostudies-literature
Project description:Background and purposeInvasive pressure studies have suggested that CSF flow across the foramen magnum may transiently decrease after coughing in patients with symptomatic Chiari I malformation. The purpose of this exploratory study was to demonstrate this phenomenon noninvasively by assessing CSF flow response to coughing in symptomatic patients with Chiari I malformation by using MR pencil beam imaging and to compare the response with that in healthy participants.Materials and methodsEight symptomatic patients with Chiari I malformation and 6 healthy participants were studied by using MR pencil beam imaging with a temporal resolution of ∼50 ms. Patients and healthy participants were scanned for 90 seconds (without cardiac gating) to continuously record cardiac cycle-related CSF flow waveforms in real-time during resting, coughing, and postcoughing periods. CSF flow waveform amplitude, CSF stroke volume, and CSF flow rate (CSF Flow Rate = CSF Stroke Volume × Heart Rate) in the resting and immediate postcoughing periods were determined and compared between patients and healthy participants.ResultsThere was no significant difference in CSF flow waveform amplitude, CSF stroke volume, and the CSF flow rate between patients with Chiari I malformation and healthy participants during rest. However, immediately after coughing, a significant decrease in CSF flow waveform amplitude (P < .001), CSF stroke volume (P = .001), and CSF flow rate (P = .001) was observed in patients with Chiari I malformation but not in the healthy participants.ConclusionsReal-time MR imaging noninvasively showed a transient decrease in CSF flow across the foramen magnum after coughing in symptomatic patients with Chiari I malformation, a phenomenon not seen in healthy participants. Our results provide preliminary evidence that the physiology-based imaging method used here has the potential to be an objective clinical test to differentiate symptomatic from asymptomatic patients with Chiari I malformation.
| S-EPMC7960326 | biostudies-literature
Project description:Sleep apnea represents a relative indication for posterior fossa decompression in pediatric patients with Chiari malformation type 1. Duraplasty was associated with improvement of sleep apnea in 100% of patients and dural splitting with improvement in 50% of patients. Duraplasty and dural splitting were associated with a similar reduction in tonsillar herniation on radiographic imaging of 58% (37% excluding tonsillectomy) and 35%, respectively. Longitudinal follow-up studies of patients with either neurologic deficits or severe symptoms will further elucidate the natural history of Chiari malformation type 1 and more appropriately gauge the risk-benefit tradeoff of surgical intervention.
| S-EPMC4584091 | biostudies-literature
Project description:Chiari malformation type I (CMI) is a disorder characterized by hindbrain overcrowding into an underdeveloped posterior cranial fossa (PCF), often causing progressive neurological symptoms. The etiology of CMI remains unclear and is most likely multifactorial. A putative genetic contribution to CMI is suggested by familial aggregation and twin studies. Experimental models and human morphometric studies have suggested an underlying paraxial mesoderm insufficiency. We performed a case-control association study of 303 tag single nucleotide polymorphisms (SNP) across 58 candidate genes involved in early paraxial mesoderm development in a sample of 415 CMI patients and 524 sex-matched controls. A subgroup of patients diagnosed with classical, small-PCF CMI by means of MRI-based PCF morphometry (n?=?186), underwent additional analysis. The genes selected are involved in signalling gradients occurring during segmental patterning of the occipital somites (FGF8, Wnt, and retinoic acid pathways and from bone morphogenetic proteins or BMP, Notch, Cdx and Hox pathways) or in placental angiogenesis, sclerotome development or CMI-associated syndromes. Single-marker analysis identified nominal associations with 18 SNPs in 14 genes (CDX1, FLT1, RARG, NKD2, MSGN1, RBPJ1, FGFR1, RDH10, NOG, RARA, LFNG, KDR, ALDH1A2, BMPR1A) considering the whole CMI sample. None of these overcame corrections for multiple comparisons, in contrast with four SNPs in CDX1, FLT1 and ALDH1A2 in the classical CMI group. Multiple marker analysis identified a risk haplotype for classical CMI in ALDH1A2 and CDX1. Furthermore, we analyzed the possible contributions of the most significantly associated SNPs to different PCF morphometric traits. These findings suggest that common variants in genes involved in somitogenesis and fetal vascular development may confer susceptibility to CMI.
| S-EPMC3578784 | biostudies-literature
Project description:Arnold Chiari malformation is one of the commonest cause of congenital hydrocephalus. Cause of fetal development of cerebellar tonsils remains unknown and may be diagnosed at later in life. The association of Arnold Chiari malformation with acromesomelic dwarfism is not known. We report male infant diagnosed with acromesomelic dwarfism at end of gestation period on basis of antenatal ultrasonography findings. An ultrasound scan of infant head at fifth month of birth was performed in view of increasing head circumference that revealed aqueductal stenosis with dilated posterior horn of lateral ventricles in brain.
| S-EPMC8017357 | biostudies-literature