Maternal COVID-19 Infection Associated with Fetal Systemic Inflammatory Complications in COVID-19-Negative Neonates: A Case-Series.

OBJECTIVE: This study aimed to examine fetal and neonatal inflammatory and neurologic complications associated with maternal coronavirus disease 2019 (COVID-19) infection. STUDY DESIGN: Case-series using a convenience sample of neonates cared for in a large referral-based children's hospital neonatal intensive care unit between September 2021 and May 2022. RESULTS: We identified seven neonates with exposure to maternal severe acute respiratory syndrome related coronavirus 2 (SARS-CoV-2) and a presentation consistent with inflammatory complications. All had some degree of neurologic injury with neuroimaging findings including restricted diffusion indicating injury in the white matter, cortex, deep gray structures, and splenium of the corpus callosum as well as intracranial hemorrhage. In addition, many infants had cytopenia and abnormal coagulation studies. Placental pathology, when available, revealed inflammation, clot with calcifications, and hematomas with associated infarcts. CONCLUSION: Neonates born to mothers with SARS-CoV-2, even when negative for the virus themselves, may have complications consistent with a systemic inflammatory syndrome. Placental pathology as well as neurologic imaging in infants with neurologic findings may help to support this diagnosis. KEY POINTS: · A systemic inflammatory response may cause illness in babies born to mothers with a history of COVID-19.. · Inflammatory markers and placental pathology are helpful in supporting this diagnosis.. · Consider neuroimaging in infants of mothers with a history of COVID-19 with neurologic findings..

Changes to pediatric brain tumors in 2021 World Health Organization classification of tumors of the central nervous system.

New tumor types are continuously being described with advances in molecular testing and genomic analysis resulting in better prognostics, new targeted therapy options and improved patient outcomes. As a result of these advances, pathological classification of tumors is periodically updated with new editions of the World Health Organization (WHO) Classification of Tumors books. In 2021, WHO Classification of Tumors of the Central Nervous System, 5th edition (CNS5), was published with major changes in pediatric brain tumors officially recognized including pediatric gliomas being separated from adult gliomas, ependymomas being categorized based on anatomical compartment and many new tumor types, most of them seen in children. Additional general changes, such as tumor grading now being done within tumor types rather than across entities and changes in definition of glioblastoma, are also relevant to pediatric neuro-oncology practice. The purpose of this manuscript is to highlight the major changes in pediatric brain tumors in CNS5 most relevant to radiologists. Additionally, brief descriptions of newly recognized entities will be presented with a focus on imaging findings.

The current and future roles of artificial intelligence in pediatric radiology.

Artificial intelligence (AI) is a broad and complicated concept that has begun to affect many areas of medicine, perhaps none so much as radiology. While pediatric radiology has been less affected than other radiology subspecialties, there are some well-developed and some nascent applications within the field. This review focuses on the use of AI within pediatric radiology for image interpretation, with descriptive summaries of the literature to date. We highlight common features that enable successful application of the technology, along with some of the limitations that can inhibit the development of this field. We present some ideas for further research in this area and challenges that must be overcome, with an understanding that technology often advances in unpredictable ways.

Improving the detection of ventricular shunt disruption using volume-rendered three-dimensional head computed tomography.

Hydrocephalus is the most common neurosurgical disorder in children, and cerebrospinal fluid (CSF) diversion with shunt placement is the most commonly performed pediatric neurosurgical procedure. CT is frequently used to evaluate children with suspected CSF shunt malfunction to assess change in ventricular size. Moreover, careful review of the CT images is important to confirm the integrity of the imaged portions of the shunt system. Subtle shunt disruptions can be missed on multiplanar two-dimensional (2-D) CT images, especially when the disruption lies in the plane of imaging. The use of volume-rendered CT images enables radiologists to view the extracranial shunt tubing within the field of view as a three-dimensional (3-D) object. This allows for a rapid and intuitive method of assessing the integrity of the extracranial shunt tubing. The purpose of this pictorial essay is to discuss how volume-rendered CT images can be generated to evaluate CSF shunts in the pediatric population and to provide several examples of their utility in diagnosing shunt disruption. We also address the potential pitfalls of this technique and ways to avoid them.

Non-inferiority of a non-gadolinium-enhanced magnetic resonance imaging follow-up protocol for isolated optic pathway gliomas.

BACKGROUND: Pediatric patients with optic pathway gliomas (OPGs) typically undergo a large number of follow-up MRI brain exams with gadolinium-based contrast media (GBCM), which have been associated with gadolinium tissue retention. Therefore, careful consideration of GBCM use in these children is warranted. OBJECTIVE: To investigate whether GBCM is necessary for OPG MR imaging response assessment using a blinded, non-inferiority, multi-reader study. MATERIALS AND METHODS: We identified children with OPG and either stable disease or change in tumor size on MRI using a regional cancer registry serving the U.S. Pacific Northwest. For each child, the two relevant, consecutive MRI studies were anonymized and standardized into two imaging sets excluding or including GBCM-enhanced images. Exam pairs were compiled from 42 children with isolated OPG (19 with neurofibromatosis type 1), from a population of 106 children with OPG. We included 28 exam pairs in which there was a change in size between exams. Seven pediatric radiologists measured tumor sizes during three blinded sessions, spaced by at least 1 week. The first measuring session excluded GBCM-enhanced sequences; the others did not. The primary endpoint was intra-reader agreement for ≥ 25% change in axial cross-product measurement, using a 12% non-inferiority threshold. RESULTS: Analysis demonstrated an overall 1.2% difference (95% confidence interval, -3.2% to 5.5%) for intra-reader agreement using a non-GBCM-enhanced protocol and background variability. CONCLUSION: A non-GBCM-enhanced protocol was non-inferior to a GBCM-enhanced protocol for assessing change in size of isolated OPGs on follow-up MRI exams.

Are Complex Skull Fractures Indicative of Either Child Abuse or Major Trauma in the Era of 3-Dimensional Computed Tomography Imaging?

OBJECTIVE: The aim of the study was to determine whether complex skull fractures are more indicative of child abuse or major trauma than simple skull fractures. DESIGN: This is a retrospective chart and imaging review of children diagnosed with a skull fracture. Subjects were from 2 pediatric tertiary care centers. Children younger than 4 years who underwent a head computed tomography with 3-dimensional rendering were included. We reviewed the medical records and imaging for type of skull fracture, abuse findings, and reported mechanism of injury. A complex skull fracture was defined as multiple fractures of a single skull bone, fractures of more than 1 skull bone, a nonlinear fracture, or diastasis of greater than 3 mm. Abuse versus accident was determined at the time of the initial evaluation with child abuse physician team confirmation. RESULTS: From 2011 to 2012, 287 subjects were identified by International Classification of Diseases, Ninth Revision, code. The 147 subjects with a cranial vault fracture and available 3-dimensional computed tomography composed this study's subjects. The average age was 12.3 months. Seventy four (50.3%) had complex and 73 (49.7%) had simple fractures. Abuse was determined in 6 subjects (4.1%), and a determination could not be made for 5 subjects. Adding abused children from 2013 to 2014 yielded 15 abused subjects. Twelve of the abused children (80%) had complex fractures; more than the 66 (48.5%) of 136 accidentally injured children (P = 0.001; relative risk = 1.65 [1.21-2.24]). However, among children with a complex fracture, the positive predictive value for abuse was only 7%. CONCLUSIONS: Complex skull fractures frequently occur from accidental injuries. This study suggests that the presence of complex skull fractures should not be used alone when making a determination of abuse.

Focal Cerebral Arteriopathy of Childhood: Clinical and Imaging Correlates.

Background and Purpose: Focal cerebral arteriopathy (FCA) of childhood with unilateral stenosis of the anterior circulation is reported to account for up to one-quarter of childhood arterial ischemic stroke, with stroke recurrence in 25% of cases. Limited knowledge regarding pathophysiology and outcome results in inconsistent treatment of FCA. Methods: Children with arterial ischemic stroke due to FCA between January 1, 2009, and January 1, 2019, were retrospectively identified at our institution which serves the US Pacific Northwest region. Electronic health record data, including neuroimaging studies, were reviewed, and the Pediatric Stroke Outcome Measure at 1 year was determined as the primary clinical end point. Results: Fifteen children were diagnosed with FCA, accounting for 19% of children with cerebral arteriopathies (n=77). Among children with FCA, the median age at the time of stroke was 6.8 years (Q1­Q3, 1.9­14.0 years). Four (20%) patients had worsening stroke, 3 of whom had concurrent infection. Three (20%) FCA cases were treated with steroids, one of whom had worsening stroke. Median Pediatric Stroke Outcome Measure at 1 year was 1.0 (Q1­Q3, 0.6­2.0). Variability in arteriopathy severity was observed within many patients. Patients with more severe arteriopathy using the Focal Cerebral Arteriopathy Severity Score had larger strokes and were more likely to have worsening stroke. The most common long-term neurological deficit was hemiparesis, which was present in 11 (73%) patients and associated with middle cerebral artery arteriopathy and infarcts. Conclusions: FCA may be less common than previously reported. Neuroimaging in FCA can help identify patients at greater risk for worsening stroke.

MRI measures of hypothalamic injury are associated with glucagon-like peptide-1 receptor agonist treatment response in people with hypothalamic obesity.

AIM: To evaluate whether neuroimaging-delineated regions of hypothalamic injury are associated with a differential treatment response to a glucagon-like peptide-1 receptor agonist (GLP-1RA) in patients with hypothalamic obesity (HO). MATERIALS AND METHODS: We performed a prespecified secondary analysis of a randomized, multicentre, double-blind, placebo-controlled trial of people aged 10-25 years with hypothalamic injury and HO randomized to the GLP-1RA exenatide once-weekly (ExQW) or placebo for 36 weeks. Subjects underwent MRI prior to enrolment and the degree of hypothalamic damage was assessed using an integrative hypothalamic lesion score (HLS). Mammillary body (MB) damage was specifically determined. The main clinical endpoints were % change in body mass index (BMI) and change in % body fat. Nested ANCOVA models including a treatment × imaging measure interaction were compared using partial F-tests to assess whether the effect of ExQW treatment differed by severity of hypothalamic damage. RESULTS: Complete data were available in 35/42 randomized participants (placebo, n = 15; ExQW, n = 20). ExQW-treated patients with worse HLS or bilateral MB damage had greater reductions in % body fat at 36 weeks (interaction coefficient estimates for HLS: -0.9%, 95% CI -1.6% to -0.2%, p = .02; for MB damage: -7.4%, 95% CI -10.1% to -4.7%, p < .001, respectively) but not for BMI % change. Similarly, patients with more damaged and smaller MB cross-sectional areas had greater reductions in % body fat following ExQW (interaction coefficient estimate 0.3%, 95% CI 0.2%-0.4%, p < .001). CONCLUSIONS: In people with HO, greater hypothalamic damage as determined by MRI, in particular MB injury, is associated with greater reductions in adiposity following GLP-1RA treatment.

A phase 3 randomized clinical trial using a once-weekly glucagon-like peptide-1 receptor agonist in adolescents and young adults with hypothalamic obesity.

AIM: To evaluate the efficacy, safety and tolerability of a glucagon-like peptide-1 receptor agonist (GLP-1 RA) in patients with hypothalamic obesity (HO). MATERIALS AND METHODS: A two-arm, randomized, multicentre, double-blind, placebo-controlled trial was conducted in 10- to 25-year-olds with hypothalamic injury following intracranial tumour and HO. Participants were randomized to once-weekly subcutaneous injections of a GLP-1 RA exenatide 2 mg (ExQW) or placebo for 36 weeks. The primary efficacy endpoint was 36-week % change in body mass index (BMI). Secondary outcomes included change in body composition (by dual energy x-ray absorptiometry). RESULTS: Forty-two participants were randomized to ExQW (n = 23) or placebo (n = 19). Participants were 5 ± 2 years (mean ± SD) postdiagnosis and development of HO (BMI 37.3 ± 7.1 kg/m2 ). In intention-to-treat analysis, the effect of 36-week ExQW vs. placebo on % Δ BMI was not significant (estimated treatment difference -1.7 ± 1.8%, 95% CI -4.1 to 0.6%, P = .40); however, total body fat mass was reduced (estimated treatment difference -3.1 ± 1.4 kg, 95% CI -5.7 to -0.4 kg, P = .02). There was a significant reduction in waist circumference (estimated effect of treatment -3.5 [95% CI -5.5 to -1.6] cm, P = .004). All patients treated with placebo increased % of adipose tissue, while 50% treated with ExQW had reductions (P < .001). Mean HbA1c, glucose tolerance and serum lipids did not change significantly with therapy. ExQW was well tolerated. The most frequent adverse events were transient gastrointestinal disturbances (ExQW vs. placebo: nausea 6/23 vs. 3/18, vomiting 4/23 vs. 4/18 and diarrhoea 7/23 vs. 3/18). CONCLUSIONS: GLP-1 RAs are a promising and safe treatment to improve or stabilize HO in children and young adults.

Delays in care seeking for young children with accidental skull fractures are common.

AIM: We sought to determine the frequency and patterns of delayed medical care seeking for young children with skull fractures. METHODS: We identified accidental skull fractures <4 years old, 2011-2012. Child abuse paediatricians abstracted retrospective data and paediatric radiologists re-reviewed images. 'Delays' were defined as presentation at ≥6 h. 'Minor accidents' included falls <4 feet and low force trauma, while 'major accidents' included higher height falls and major force events. We studied the frequency and duration of care delays, the signs or symptoms leading to care, and the duration of delays after signs or symptoms developed. RESULTS: Two hundred and ten children had accidental skull fractures. Delays were less likely with major accidents (4.9%), than with minor accidents (25.8%) (RR = 0.32 [0.15-0.70]). Children came to care for scalp swelling (STS) (39%), the injury event (36.2%), altered consciousness (15.2%) and vomiting (10.5%). Delayed onset of STS (78.6%) caused most delayed care. Early STS was firm, (17.6%) versus delayed (5.0%), as opposed to soft or fluctuant. CONCLUSION: Delayed care seeking is common for minor, but not major accidental infant and toddler skull fractures. Most followed delayed onset of signs and symptoms. Delayed care seeking alone should not imply child abuse.

Deep medullary vein engorgement and superficial medullary vein engorgement: two patterns of perinatal venous stroke.

Perinatal venous stroke has classically been attributed to cerebral sinovenous thrombosis with resultant congestion or thrombosis of the small veins draining the cerebrum. Advances in brain MRI, in particular susceptibility-weighted imaging, have enabled the visualization of the engorged small intracerebral veins, and the spectrum of perinatal venous stroke has expanded to include isolated congestion or thrombosis of the deep medullary veins and the superficial intracerebral veins. Congestion or thrombosis of the deep medullary veins or the superficial intracerebral veins can result in vasogenic edema, cytotoxic edema or hemorrhage in the territory of disrupted venous flow. Deep medullary vein engorgement and superficial medullary vein engorgement have characteristic findings on MRI and should be differentiated from neonatal hemorrhagic stroke.

Fracture-Associated Bruising and Soft Tissue Swelling in Young Children With Skull Fractures: How Sensitive Are They to Fracture Presence?

OBJECTIVES: The aim of this study was to determine how reliable scalp bruising and soft tissue swelling/cephalohematomas (STS) are for underlying young child skull fractures. METHODS: This was a retrospective clinical and imaging review from 2011 to 2012 of children younger than 4 years with skull fractures from 2 tertiary care hospitals. Imaging was reread by 3 pediatric radiologists. Descriptive statistics were utilized. The retrospective review had institutional review board approval. RESULTS: We identified 218 subjects for review: 210 unintentional and 8 abusive. One hundred forty-three had available 3-dimensional computed tomography reconstructions: 136 unintentional and 7 abused. Two-thirds were younger than 1 year. Twelve subjects had visible scalp bruising, but 73% had clinically and 93% radiologically apparent fracture-associated STS. There was no difference in clinical STS with simple (79%) versus complex (68%) fractures. Nor was there difference in subjects with fractures from minor (77%) versus major (70%) trauma. Unintentionally injured infants did not differ from abused for detectable STS (74% vs 50%). Parietal and frontal bones most frequently sustained fractures and most consistently had associated STS. CONCLUSIONS: Clinically apparent STS is present in approximately three-fourths of children with skull fractures. It may not be important to consistently identify skull fractures in unintentionally injured children. Point-of-care ultrasound may be adequate. For abuse concerns, it is important to identify skull fractures as evidence of cranial impacts and intracranial hemorrhages. The most sensitive, widely available imaging technique, cranial computed tomography scan with 3-dimensional skull reconstruction, should be utilized. Scalp bruising is present in a minority of young children with skull fractures. Its absence does not exclude cranial impact injury.

MRI Vessel Wall Enhancement and Other Imaging Biomarkers in Pediatric Focal Cerebral Arteriopathy-Inflammatory Subtype.

Background and Purpose- Focal cerebral arteriopathy-inflammatory type (FCA-i) is a common cause of pediatric arterial ischemic stroke characterized angiographically by unifocal and unilateral stenosis/irregularity of the large anterior circulation arteries with a presumed inflammatory cause. Arterial vessel wall enhancement (VWE) on vessel wall magnetic resonance imaging is a potential biomarker of inflammation that may improve diagnosis, guide treatment, and predict outcomes in patients with FCA-i. We hypothesized that patients with FCA-i with more severe or extensive VWE would have worse arteriopathy, larger infarcts, worse clinical outcome, and increased risk for infarct progression/recurrence. Methods- Pediatric patients with arterial ischemic stroke, classified as FCA-i, and who underwent vessel wall imaging were retrospectively identified at our institution. Clinical data were reviewed and the Pediatric Stroke Outcome Measure at 1 year was determined as the primary clinical end point. Neuroimaging studies were assessed for infarct size, arteriopathy severity (Focal Cerebral Arteriopathy Severity Score), and VWE. Results- Nine cases of FCA-i with vessel wall imaging were evaluated, and there was a strong correlation between clinical outcome at 1-year with initial infarct volume (Spearman correlation coefficient rho=0.84; P<0.01) and arteriopathy severity (Focal Cerebral Arteriopathy Severity Score; rho=0.85; P<0.01). Patients with infarct progression/recurrence had worse Focal Cerebral Arteriopathy Severity Score at presentation compared with those without progression/recurrence (median [IQR]; 9.0 [8.0-11.8] and 5.0 [4.0-7.0], respectively; P<0.05). On the contrary, measures of VWE were not correlated with arteriopathy severity, infarct size, clinical outcome, or risk of infarct progression/recurrence. Moreover, not all patients with FCA-i demonstrated VWE. Conclusions- VWE may not be a reliable biomarker for the diagnosis or assessment of FCA-i, and future work is needed to assess the utility of vessel wall imaging in pediatric arterial ischemic stroke and FCA-i.

Pathogenic variants in NUBPL result in failure to assemble the matrix arm of complex I and cause a complex leukoencephalopathy with thalamic involvement.

Disorders of the white matter are genetically very heterogeneous including several genes involved in mitochondrial bioenergetics. Diagnosis of the underlying cause is aided by pattern recognition on neuroimaging and by next-generation sequencing. Recently, genetic changes in the complex I assembly factor NUBPL have been characterized by a consistent recognizable pattern of leukoencephalopathy affecting deep white matter including the corpus callosum and cerebellum. Here, we report twin boys with biallelic variants in NUBPL, an unreported c.351 G > A; p.(Met117Ile) and a previously reported pathological variant c. 693 + 1 G > A. Brain magnetic resonance imaging showed abnormal T2 hyperintense signal involving the periventricular white matter, external capsule, corpus callosum, and, prominently, the bilateral thalami. The neuroimaging pattern evolved over 18 months with marked diffuse white matter signal abnormality, volume loss, and new areas of signal abnormality in the cerebellar folia and vermis. Magnetic resonance spectroscopy showed elevated lactate. Functional studies in cultured fibroblasts confirmed pathogenicity of the genetic variants. Complex I activity of the respiratory chain was deficient spectrophotometrically and on blue native gel with in-gel activity staining. There was absent assembly and loss of proteins of the matrix arm of complex I when traced with an antibody to NDUFS2, and incomplete assembly of the membrane arm when traced with an NDUFB6 antibody. There was decreased NUBPL protein on Western blot in patient fibroblasts compared to controls. Compromised NUBPL activity impairs assembly of the matrix arm of complex I and produces a severe, rapidly-progressive leukoencephalopathy with thalamic involvement on MRI, further expanding the neuroimaging phenotype.

Longer upper airway lengths in Robin Sequence: A case-control study using computed tomography.

BACKGROUND: Direct laryngoscopy and intubation are often difficult in children with Robin Sequence. Previous research characterizing anatomic airway differences has focused on parameters influencing airway patency; there is a paucity of data pertaining to intubation trajectories and depth. Such information could impact airway management approaches and decrease the incidence of endotracheal tube malpositioning. AIM: The study goal was to examine whether longitudinal airway parameters pertaining to intubation are different in children with Robin Sequence compared with age-matched controls. METHOD: This case-control study compared patients with RS <4 years of age who had computed tomography scans of the head and neck to age- and sex-matched controls. Measurements were made of the nasopharynx, oropharynx, hypopharynx, tongue, hyoid, and the front teeth to vocal cord, nares to vocal cord, and nasion-basion distances. Statistical analysis was performed using multiple ANCOVA models with the categorical predictor of Robin Sequence vs control and potential covariates including subject height/length, weight, and age. RESULTS: Thirty-three patients with Robin Sequence and 33 control subjects were included. After controlling for subject height/length, mean front teeth to vocal cord distance was 1.2 cm longer (95% CI: 0.9 to 1.6 cm, P < .001) and mean nares to vocal cord distance was 0.8 cm longer (95% CI: 0.4 to 1.2 cm, P < .001) in patients with Robin Sequence than in controls. The tongue was positioned on average 0.5 cm higher (95% CI: 0.3 to 0.8, P < .001) and 0.9 cm more posterior (95% CI: 0.6 to 1.0 cm, P < .001) in cases than in controls. Moreover, in patients with Robin Sequence, the hyoid was positioned on average 0.5 cm more inferiorly (95% CI: 0.2 to 0.8 cm, P < .001) and 0.2 cm more posteriorly (95% CI: 0.1 to 0.4 cm, P < .01) than controls. CONCLUSION: In patients with Robin Sequence under 4 years of age, the mean front teeth to vocal cord distance was found to be 1.2 cm longer while the mean nares to vocal cord distance was found to be 0.8 cm longer controlling for subject length. Clinicians should account for these differences when selecting and placing endotracheal tubes, particularly those with a preformed bend.

Surveillance magnetic resonance imaging for isolated optic pathway gliomas: is gadolinium necessary?

BACKGROUND: Pediatric optic pathway gliomas are typically indolent but have a variable clinical course. Treatment is dictated by symptoms and changes on contrast-enhanced MRI examinations. Gadolinium retention in children has motivated parsimonious use of gadolinium-based contrast agents. OBJECTIVES: To determine surveillance MR factors that motivate changes in tumor-directed therapies and extrapolate cost-efficacy of a non-contrast follow-up protocol. MATERIALS AND METHODS: Using an imaging database search we identified children with isolated optic pathway gliomas and ≥3 follow-up contrast-enhanced MRIs. We reviewed medical records and imaging for: (1) coincident changes on contrast-enhanced MRI and tumor-directed therapy, (2) demographics and duration of follow-up, (3) motivations for intervention, (4) assessment of gadolinium-based contrast agents' utility and (5) health care utilization data. We assessed cost impact in terms of relative value unit (RVU) burden. RESULTS: We included 17 neurofibromatosis type 1 (NF1) and 21 non-NF1 patients who underwent a median 16.9 and 24.3 cumulative contrast-enhanced MR exams over 7.7 years and 8.1 years of follow-up, respectively. Eight children (one with NF1) had intervention based on contrast-enhanced MR findings alone. For these eight, increased tumor size was the only common feature, and it was apparent on non-contrast T2 sequences. For the median patient, a non-contrast follow-up protocol could result in 15.9 (NF1) and 23.3 (non-NF1) fewer gadolinium-based contrast agent administrations, and a 39% lower yearly RVU burden. CONCLUSION: Pediatric patients with isolated optic pathway gliomas undergo a large number of routine contrast-enhanced MR follow-up exams. Gadolinium might not be needed for these exams to inform management decisions. Secondary benefits of a non-contrast follow-up protocol include decreased cost and risk to the patient.

Exome sequencing identifies a recurrent de novo ZSWIM6 mutation associated with acromelic frontonasal dysostosis.

Acromelic frontonasal dysostosis (AFND) is a rare disorder characterized by distinct craniofacial, brain, and limb malformations, including frontonasal dysplasia, interhemispheric lipoma, agenesis of the corpus callosum, tibial hemimelia, preaxial polydactyly of the feet, and intellectual disability. Exome sequencing of one trio and two unrelated probands revealed the same heterozygous variant (c.3487C>T [p. Arg1163Trp]) in a highly conserved protein domain of ZSWIM6; this variant has not been seen in the 1000 Genomes data, dbSNP, or the Exome Sequencing Project. Sanger validation of the three trios confirmed that the variant was de novo and was also present in a fourth isolated proband. In situ hybridization of early zebrafish embryos at 24 hr postfertilization (hpf) demonstrated telencephalic expression of zswim6 and onset of midbrain, hindbrain, and retinal expression at 48 hpf. Immunohistochemistry of later-stage mouse embryos demonstrated tissue-specific expression in the derivatives of all three germ layers. qRT-PCR expression analysis of osteoblast and fibroblast cell lines available from two probands was suggestive of Hedgehog pathway activation, indicating that the ZSWIM6 mutation associated with AFND may lead to the craniofacial, brain and limb malformations through the disruption of Hedgehog signaling.

Prenatal diagnosis of Chudley-McCullough syndrome.

Chudley-McCullough syndrome (CMS) is an autosomal-recessive disorder characterized by a complex brain malformation and profound congenital sensorineural hearing loss. Postnatal brain imaging findings include ventriculomegaly, partial agenesis of corpus callosum, inferior cerebellar dysplasia, arachnoid cysts, and malformations of cortical development including frontal subcortical heterotopia and polymicrogyria. Prenatal diagnosis of CMS is important due to the markedly less severe neurodevelopmental prognosis compared to disorders with similar brain imaging findings. We report prenatal imaging features that help distinguish CMS from other disorders, including slit-like frontal horns, agenesis of the corpus callosum, frontal subcortical heterotopia, arachnoid cysts, and cerebellar dysplasia. © 2016 Wiley Periodicals, Inc.

Near-pair patch generative adversarial network for data augmentation of focal pathology object detection models.

Purpose: The limited volume of medical training data remains one of the leading challenges for machine learning for diagnostic applications. Object detectors that identify and localize pathologies require training with a large volume of labeled images, which are often expensive and time-consuming to curate. To reduce this challenge, we present a method to support distant supervision of object detectors through generation of synthetic pathology-present labeled images. Approach: Our method employs the previously proposed cyclic generative adversarial network (cycleGAN) with two key innovations: (1) use of "near-pair" pathology-present regions and pathology-absent regions from similar locations in the same subject for training and (2) the addition of a realism metric (Fréchet inception distance) to the generator loss term. We trained and tested this method with 2800 fracture-present and 2800 fracture-absent image patches from 704 unique pediatric chest radiographs. The trained model was then used to generate synthetic pathology-present images with exact knowledge of location (labels) of the pathology. These synthetic images provided an augmented training set for an object detector. Results: In an observer study, four pediatric radiologists used a five-point Likert scale indicating the likelihood of a real fracture (1 = definitely not a fracture and 5 = definitely a fracture) to grade a set of real fracture-absent, real fracture-present, and synthetic fracture-present images. The real fracture-absent images scored 1.7±1.0, real fracture-present images 4.1±1.2, and synthetic fracture-present images 2.5±1.2. An object detector model (YOLOv5) trained on a mix of 500 real and 500 synthetic radiographs performed with a recall of 0.57±0.05 and an F2 score of 0.59±0.05. In comparison, when trained on only 500 real radiographs, the recall and F2 score were 0.49±0.06 and 0.53±0.06, respectively. Conclusions: Our proposed method generates visually realistic pathology and that provided improved object detector performance for the task of rib fracture detection.