Retrospective Analysis of Presymptomatic Treatment In Sturge-Weber Syndrome.

Background: Ninety percent of infants with Sturge-Weber syndrome (SWS) brain involvement have seizure onset before 2 years of age; this is associated with worse neurologic outcome. Presymptomatic treatment before seizure onset may delay seizure onset and improve outcome, as has been shown in other conditions with a high-risk of developing epilepsy such as tuberous sclerosis complex. Electroencephalogram (EEG) may be a biomarker to predict seizure onset. This retrospective clinical data analysis aims to assess impact of presymptomatic treatment in SWS. Methods: This two-centered, IRB-approved, retrospective study analyzed records from patients with SWS brain involvement. Clinical data recorded included demographics, age of seizure onset (if present), brain involvement extent (unilateral versus bilateral), port-wine birthmark (PWB) extent, family history of seizure, presymptomatic treatment if received, neuroscore, and anti-seizure medication. EEG reports prior to seizure onset were analyzed. Results: Ninety-two patients were included (48 females), and 32 received presymptomatic treatment outside of a formal protocol (5 aspirin, 16 aspirin and levetiracetam; 9 aspirin and oxcarbazepine, 2 valproic acid). Presymptomatically-treated patients were more likely to be seizure-free at 2 years (15 of 32; 47% versus 7 of 60; 12%; p<.001). A greater percentage of presymptomatically-treated patients had bilateral brain involvement (38% treated versus 17% untreated; p=.026). Median hemiparesis neuroscore at 2 years was better in presymptomatically-treated patients. In EEG reports prior to seizure onset, the presence of slowing, epileptiform discharges, or EEG-identified seizures was associated with seizure onset by 2 (p=.001). Conclusion: Presymptomatic treatment is a promising approach to children diagnosed with SWS prior to seizure onset. Further study is needed, including prospective drug trials, long-term neuropsychological outcome, and prospective EEG analysis to assess this approach and determine biomarkers for presymptomatic treatment.

Autism-associated brain differences can be observed in utero using MRI.

Developmental changes that occur before birth are thought to be associated with the development of autism spectrum disorders. Identifying anatomical predictors of early brain development may contribute to our understanding of the neurobiology of autism spectrum disorders and allow for earlier and more effective identification and treatment of autism spectrum disorders. In this study, we used retrospective clinical brain magnetic resonance imaging data from fetuses who were diagnosed with autism spectrum disorders later in life (prospective autism spectrum disorders) in order to identify the earliest magnetic resonance imaging-based regional volumetric biomarkers. Our results showed that magnetic resonance imaging-based autism spectrum disorder biomarkers can be found as early as in the fetal period and suggested that the increased volume of the insular cortex may be the most promising magnetic resonance imaging-based fetal biomarker for the future emergence of autism spectrum disorders, along with some additional, potentially useful changes in regional volumes and hemispheric asymmetries.

Association Between Long-Term Exposure to Ambient Air Pollution and the Risk of Mild Cognitive Impairment in a Chinese Urban Area: A Case-Control Study.

Background: As a prodromal stage of dementia, significant emphasis has been placed on the identification of modifiable risks of mild cognitive impairment (MCI). Research has indicated a correlation between exposure to air pollution and cognitive function in older adults. However, few studies have examined such an association among the MCI population inChina. Objective: We aimed to explore the association between air pollution exposure and MCI risk from the Hubei Memory and Aging Cohort Study. Methods: We measured four pollutants from 2015 to 2018, 3 years before the cognitive assessment of the participants. Logistic regression models were employed to calculate odds ratios (ORs) to assess the relationship between air pollutants and MCI risk. Results: Among 4,205 older participants, the adjusted ORs of MCI risk for the highest quartile of PM2.5, PM10, O3, and SO2 were 1.90 (1.39, 2.62), 1.77 (1.28, 2.47), 0.56 (0.42, 0.75), and 1.18 (0.87, 1.61) respectively, compared with the lowest quartile. Stratified analyses indicated that such associations were found in both males and females, but were more significant in older participants. Conclusions: Our findings are consistent with the growing evidence suggesting that air pollution increases the risk of mild cognitive decline, which has considerable guiding significance for early intervention of dementia in the older population. Further studies in other populations and broader geographical areas are warranted to validate these findings.

Noninvasive Gamma Sensory Stimulation May Reduce White Matter and Myelin Loss in Alzheimer's Disease.

BACKGROUND: Patients with Alzheimer's disease (AD) demonstrate progressive white matter atrophy and myelin loss. Restoring myelin content or preventing demyelination has been suggested as a therapeutic approach for AD. OBJECTIVE: Herein, we investigate the effects of non-invasive, combined visual and auditory gamma-sensory stimulation on white matter atrophy and myelin content loss in patients with AD. METHODS: In this study, we used the magnetic resonance imaging (MRI) data from the OVERTURE study (NCT03556280), a randomized, controlled, clinical trial in which active treatment participants received daily, non-invasive, combined visual and auditory, 40 Hz stimulation for six months. A subset of OVERTURE participants who meet the inclusion criteria for detailed white matter (N = 38) and myelin content (N = 36) assessments are included in the analysis. White matter volume assessments were performed using T1-weighted MRI, and myelin content assessments were performed using T1-weighted/T2-weighted MRI. Treatment effects on white matter atrophy and myelin content loss were assessed. RESULTS: Combined visual and auditory gamma-sensory stimulation treatment is associated with reduced total and regional white matter atrophy and myelin content loss in active treatment participants compared to sham treatment participants. Across white matter structures evaluated, the most significant changes were observed in the entorhinal region. CONCLUSIONS: The study results suggest that combined visual and auditory gamma-sensory stimulation may modulate neuronal network function in AD in part by reducing white matter atrophy and myelin content loss. Furthermore, the entorhinal region MRI outcomes may have significant implications for early disease intervention, considering the crucial afferent connections to the hippocampus and entorhinal cortex.

Human-to-monkey transfer learning identifies the frontal white matter as a key determinant for predicting monkey brain age.

The application of artificial intelligence (AI) to summarize a whole-brain magnetic resonance image (MRI) into an effective "brain age" metric can provide a holistic, individualized, and objective view of how the brain interacts with various factors (e.g., genetics and lifestyle) during aging. Brain age predictions using deep learning (DL) have been widely used to quantify the developmental status of human brains, but their wider application to serve biomedical purposes is under criticism for requiring large samples and complicated interpretability. Animal models, i.e., rhesus monkeys, have offered a unique lens to understand the human brain - being a species in which aging patterns are similar, for which environmental and lifestyle factors are more readily controlled. However, applying DL methods in animal models suffers from data insufficiency as the availability of animal brain MRIs is limited compared to many thousands of human MRIs. We showed that transfer learning can mitigate the sample size problem, where transferring the pre-trained AI models from 8,859 human brain MRIs improved monkey brain age estimation accuracy and stability. The highest accuracy and stability occurred when transferring the 3D ResNet [mean absolute error (MAE) = 1.83 years] and the 2D global-local transformer (MAE = 1.92 years) models. Our models identified the frontal white matter as the most important feature for monkey brain age predictions, which is consistent with previous histological findings. This first DL-based, anatomically interpretable, and adaptive brain age estimator could broaden the application of AI techniques to various animal or disease samples and widen opportunities for research in non-human primate brains across the lifespan.

Linking maternal disrupted interaction and infant limbic volumes: The role of infant cortisol output.

Despite a large animal literature documenting the role of low maternal nurturance and elevated glucocorticoid production on offspring limbic development, these pathways have not yet been assessed during human infancy. Informed by animal models, the present study examined whether 1) maternal disrupted interaction is related to infant cortisol levels, 2) infant cortisol levels are associated with infant limbic volumes, and 3) infant cortisol levels mediate associations between maternal disrupted interaction and infant limbic volumes. Participants included 57 mother-infant dyads. Infant saliva was measured at one time point before and two time points after the Still-Face Paradigm (SFP) at age 4 months. Five aspects of maternal disrupted interaction were coded during the SFP reunion episode. Between 4 and 25 months (M age = 11.74 months, SD = 6.12), under natural sleep, infants completed an MRI. Amygdala and hippocampal volumes were calculated via automated segmentation. Results indicated that 1) maternal disrupted interaction, and specifically disoriented interaction, with the infant was associated with higher infant salivary cortisol (AUCg) levels during the SFP, 2) higher infant AUCg was related to enlarged bilateral amygdala and hippocampal volumes, and 3) infant AUCg mediated the relation between maternal disrupted interaction and infant amygdala and hippocampal volumes. Findings are consistent with controlled animal studies and provide evidence of a link between increased cortisol levels and enlarged limbic volumes in human infants. Results further suggest that established interventions to decrease maternal disrupted interaction could impact both infant cortisol levels and infant limbic volumes.

Challenges of implementing computer-aided diagnostic models for neuroimages in a clinical setting.

Advances in artificial intelligence have cultivated a strong interest in developing and validating the clinical utilities of computer-aided diagnostic models. Machine learning for diagnostic neuroimaging has often been applied to detect psychological and neurological disorders, typically on small-scale datasets or data collected in a research setting. With the collection and collation of an ever-growing number of public datasets that researchers can freely access, much work has been done in adapting machine learning models to classify these neuroimages by diseases such as Alzheimer's, ADHD, autism, bipolar disorder, and so on. These studies often come with the promise of being implemented clinically, but despite intense interest in this topic in the laboratory, limited progress has been made in clinical implementation. In this review, we analyze challenges specific to the clinical implementation of diagnostic AI models for neuroimaging data, looking at the differences between laboratory and clinical settings, the inherent limitations of diagnostic AI, and the different incentives and skill sets between research institutions, technology companies, and hospitals. These complexities need to be recognized in the translation of diagnostic AI for neuroimaging from the laboratory to the clinic.

Negative versus withdrawn maternal behavior: Differential associations with infant gray and white matter during the first 2 years of life.

Distinct neural effects of threat versus deprivation emerge by childhood, but little data are available in infancy. Withdrawn versus negative parenting may represent dimensionalized indices of early deprivation versus early threat, but no studies have assessed neural correlates of withdrawn versus negative parenting in infancy. The objective of this study was to separately assess the links of maternal withdrawal and maternal negative/inappropriate interaction with infant gray matter volume (GMV), white matter volume (WMV), amygdala, and hippocampal volume. Participants included 57 mother-infant dyads. Withdrawn and negative/inappropriate aspects of maternal behavior were coded from the Still-Face Paradigm at four months infant age. Between 4 and 24 months (M age = 12.28 months, SD = 5.99), during natural sleep, infants completed an MRI using a 3.0 T Siemens scanner. GMV, WMV, amygdala, and hippocampal volumes were extracted via automated segmentation. Diffusion weighted imaging volumetric data were also generated for major white matter tracts. Maternal withdrawal was associated with lower infant GMV. Negative/inappropriate interaction was associated with lower overall WMV. Age did not moderate these effects. Maternal withdrawal was further associated with reduced right hippocampal volume at older ages. Exploratory analyses of white matter tracts found that negative/inappropriate maternal behavior was specifically associated with reduced volume in the ventral language network. Results suggest that quality of day-to-day parenting is related to infant brain volumes during the first two years of life, with distinct aspects of interaction associated with distinct neural effects.

Experimental study on the dissipation performance of supersaturated total dissolved gas in microbubble treatment.

The production of total dissolved gas (TDG) supersaturation resulting from dam discharges has been identified as a causative factor for gas bubble disease (GBD) or mass mortality in fish. In this study, the mitigation solution for fish refuge in supersaturated TDG water was explored by using microbubbles generated by aeration to enhance supersaturated TDG dissipation. The effects of various aeration factors (aeration intensity, water depth, and aerator size) on the dissipation processes of supersaturated TDG were quantitatively investigated through a series of tests conducted in a static aeration column. The results indicated that the dissipation rates of supersaturated TDG increased as a power function with the factors of aeration intensity and aerator size and decreased as a power function with increasing water depth. A universal prediction model for the dissipation rate of supersaturated TDG in the aeration system was developed based on the dimensional analysis of the comprehensive elements, and the parameters in the model were determined using experimental data. The outcomes of this study can furnish an important theoretical foundation and scientific guidance for the utilization of aeration as a measure to alleviate the adverse impacts of supersaturated TDG on fish.

BOston Neonatal Brain Injury Dataset for Hypoxic Ischemic Encephalopathy (BONBID-HIE): Part I. MRI and Manual Lesion Annotation.

Hypoxic ischemic encephalopathy (HIE) is a brain injury that occurs in 1 ~ 5/1000 term neonates. Accurate identification and segmentation of HIE-related lesions in neonatal brain magnetic resonance images (MRIs) is the first step toward predicting prognosis, identifying high-risk patients, and evaluating treatment effects. It will lead to a more accurate estimation of prognosis, a better understanding of neurological symptoms, and a timely prediction of response to therapy. We release the first public dataset containing neonatal brain diffusion MRI and expert annotation of lesions from 133 patients diagnosed with HIE. HIE-related lesions in brain MRI are often diffuse (i.e., multi-focal), and small (over half the patients in our data having lesions occupying <1% of brain volume). Segmentation for HIE MRI data is remarkably different from, and arguably more challenging than, other segmentation tasks such as brain tumors with focal and relatively large lesions. We hope that this dataset can help fuel the development of MRI lesion segmentation methods for HIE and small diffuse lesions in general.

Diagnosis and treatment of transnasal endoscopic optic canal decompression for traumatic optic neuropathy.

Objective: To investigate the clinical efficacy and prognostic factors of transnasal endoscopic optic decompression in the treatment of traumatic optic neuropathy (TON). Methods: A retrospective analysis was performed on 13 TON patients in The Seventh Affiliated Hospital of Sun Yat-sen University and Shenzhen Eye Hospital in Shenzhen City (China) from June 2020 to April 2022. These patients had received transnasal endoscopic optic decompression, and hormonal and neurotrophic drugs were given after surgery. Visual acuity (VA) improvement was used as the criterion to judge clinical efficacy. Results: In a total of 13 patients, 13 injured eyes (12 men and 1 woman; mean age, 28.0 ± 11.8 years) received transnasal endoscopic optic decompression. After surgery, nine patients had improved VA, whereas four patients failed to show any improvement, resulting in a total effective rate of 69.2%. Of the six patients with no light perception preoperatively, three had effective results after the operation, giving an effective rate of 50.0%. Of the seven patients with residual light sensation preoperatively, six had effective results after the operation, giving an effective rate of 85.7%. Of the 10 patients operated on within 7 days after injury, seven had effective results, giving an effective rate of 70%. Of the three patients injured and operated on after 7 days, two had effective results, giving an effective rate of 66.7%. Conclusion: Transnasal endoscopic optic nerve decompression is an effective treatment method for TON. The presence of residual light perception and the timing of surgery within 7 days are crucial to the prognosis.

Maternal Childhood Abuse Versus Neglect Associated with Differential Patterns of Infant Brain Development.

Severity of maternal childhood maltreatment has been associated with lower infant grey matter volume and amygdala volume during the first two years of life. A developing literature argues that effects of threat (abuse) and of deprivation (neglect) should be assessed separately because these distinct aspects of adversity may have different impacts on developmental outcomes. However, distinct effects of threat versus deprivation have not been assessed in relation to intergenerational effects of child maltreatment. The objective of this study was to separately assess the links of maternal childhood abuse and neglect with infant grey matter volume (GMV), white matter volume (WMV), amygdala and hippocampal volume. Participants included 57 mother-infant dyads. Mothers were assessed for childhood abuse and neglect using the Adverse Childhood Experiences (ACE) questionnaire in a sample enriched for childhood maltreatment. Between 4 and 24 months (M age = 12.28 months, SD = 5.99), under natural sleep, infants completed an MRI using a 3.0 T Siemens scanner. GMV, WMV, amygdala and hippocampal volumes were extracted via automated segmentation. Maternal history of neglect, but not abuse, was associated with lower infant GMV. Maternal history of abuse, but not neglect, interacted with age such that abuse was associated with smaller infant amygdala volume at older ages. Results are consistent with a threat versus deprivation framework, in which threat impacts limbic regions central to the stress response, whereas deprivation impacts areas more central to cognitive function. Further studies are needed to identify mechanisms contributing to these differential intergenerational associations of threat versus deprivation.

Prevalence and risk factors for subjective cognitive decline and the correlation with objective cognition among community-dwelling older adults in China: Results from the Hubei memory and aging cohort study.

INTRODUCTION: The prevalence and risk factors for subjective cognitive decline (SCD) and its correlation with objective cognition decline (OCD) among community-dwelling older adults is inconsistent. METHODS: Older adults underwent neuropsychological and clinical evaluations to reach a consensus on diagnoses. RESULTS: This study included 7486 adults without mild cognitive impairment and dementia (mean age: 71.35 years [standard deviation = 5.40]). The sex-, age-, and residence-adjusted SCD prevalence was 58.33% overall (95% confidence interval: 58.29% to 58.37%), with higher rates of 61.25% and 59.87% in rural and female subgroups, respectively. SCD global and OCD language, SCD memory and OCD global, SCD and OCD memory, and SCD and OCD language were negatively correlated in fully adjusted models. Seven health and lifestyle factors were associated with an increased risk for SCD. DISCUSSION: SCD affected 58.33% of older adults and may indicate concurrent OCD, which should prompt the initiation of preventative intervention for dementia. HIGHLIGHTS: SCD affects 58.33% of older adults in China. SCD may indicate concurrent objective cognitive decline. Difficulty finding words and memory impairments may indicate a risk for AD. The presence of SCD may prompt preventative treatment initiation of MCI or dementia. Social network factors may be initial targets for the early prevention of SCD.

Segmentation ability map: Interpret deep features for medical image segmentation.

Deep convolutional neural networks (CNNs) have been widely used for medical image segmentation. In most studies, only the output layer is exploited to compute the final segmentation results and the hidden representations of the deep learned features have not been well understood. In this paper, we propose a prototype segmentation (ProtoSeg) method to compute a binary segmentation map based on deep features. We measure the segmentation abilities of the features by computing the Dice between the feature segmentation map and ground-truth, named as the segmentation ability score (SA score for short). The corresponding SA score can quantify the segmentation abilities of deep features in different layers and units to understand the deep neural networks for segmentation. In addition, our method can provide a mean SA score which can give a performance estimation of the output on the test images without ground-truth. Finally, we use the proposed ProtoSeg method to compute the segmentation map directly on input images to further understand the segmentation ability of each input image. Results are presented on segmenting tumors in brain MRI, lesions in skin images, COVID-related abnormality in CT images, prostate segmentation in abdominal MRI, and pancreatic mass segmentation in CT images. Our method can provide new insights for interpreting and explainable AI systems for medical image segmentation. Our code is available on: https://github.com/shengfly/ProtoSeg.

Maternal Childhood Maltreatment Is Associated With Lower Infant Gray Matter Volume and Amygdala Volume During the First Two Years of Life.

Background: Childhood maltreatment affects approximately 25% of the world's population. Importantly, the children of mothers who have been maltreated are at increased risk of behavioral problems. Thus, one important priority is to identify child neurobiological processes associated with maternal childhood maltreatment (MCM) that might contribute to such intergenerational transmission. This study assessed the impact of MCM on infant gray and white matter volumes and infant amygdala and hippocampal volumes during the first 2 years of life. Methods: Fifty-seven mothers with 4-month-old infants were assessed for MCM, using both the brief Adverse Childhood Experiences screening questionnaire and the more detailed Maltreatment and Abuse Chronology of Exposure scale. A total of 58% had experienced childhood maltreatment. Between 4 and 24 months (age in months: mean = 12.28, SD = 5.99), under natural sleep, infants completed a magnetic resonance imaging scan using a 3T Siemens scanner. Total brain volume, gray matter volume, white matter volume, and amygdala and hippocampal volumes were extracted via automated segmentation. Results: MCM on the Adverse Childhood Experiences and Maltreatment and Abuse Chronology of Exposure scales were associated with lower infant total brain volume and gray matter volume, with no moderation by infant age. However, infant age moderated the association between MCM and right amygdala volume, such that MCM was associated with lower volume at older ages. Conclusions: MCM is associated with alterations in infant brain volumes, calling for further identification of the prenatal and postnatal mechanisms contributing to such intergenerational transmission. Furthermore, the brief Adverse Childhood Experiences questionnaire predicted these alterations, suggesting the potential utility of early screening for infant risk.

Patterns of Neural Functional Connectivity in Infants at Familial Risk of Developmental Dyslexia.

Importance: Developmental dyslexia is a heritable learning disability affecting 7% to 10% of the general population and can have detrimental impacts on mental health and vocational potential. Individuals with dyslexia show altered functional organization of the language and reading neural networks; however, it remains unknown how early in life these neural network alterations might emerge. Objective: To determine whether the early emergence of large-scale neural functional connectivity (FC) underlying long-term language and reading development is altered in infants with a familial history of dyslexia (FHD). Design, Setting, and Participants: This cohort study included infants recruited at Boston Children's Hospital between May 2011 and February 2019. Participants underwent structural and resting-state functional magnetic resonance imaging in the Department of Radiology at Boston Children's Hospital. Infants with FHD were matched with infants without FHD based on age and sex. Data were analyzed from April 2019 to June 2021. Exposures: FHD was defined as having at least 1 first-degree relative with a dyslexia diagnosis or documented reading difficulties. Main Outcomes and Measures: Whole-brain FC patterns associated with 20 predefined cerebral regions important for long-term language and reading development were computed for each infant. Multivariate pattern analyses were applied to identify specific FC patterns that differentiated between infants with vs without FHD. For classification performance estimates, 99% CIs were calculated as the classification accuracy minus chance level. Results: A total of 98 infants (mean [SD] age, 8.5 [2.3] months; 51 [52.0%] girls) were analyzed, including 35 infants with FHD and 63 infants without FHD. Multivariate pattern analyses identified distinct FC patterns between infants with vs without FHD in the left fusiform gyrus (classification accuracy, 0.55 [99% CI, 0.046-0.062]; corrected P < .001; Cohen d = 0.76). Connections linking left fusiform gyrus to regions in the frontal and parietal language and attention networks were among the paths with the highest contributions to the classification performance. Conclusions and Relevance: These findings suggest that on the group level, FHD was associated with an early onset of atypical FC of regions important for subsequent word form recognition during reading acquisition. Longitudinal studies linking the atypical functional network and school-age reading abilities will be essential to further elucidate the ontogenetic mechanisms underlying the development of dyslexia.

A Role for Data Science in Precision Nutrition and Early Brain Development.

Multimodal brain magnetic resonance imaging (MRI) can provide biomarkers of early influences on neurodevelopment such as nutrition, environmental and genetic factors. As the exposure to early influences can be separated from neurodevelopmental outcomes by many months or years, MRI markers can serve as an important intermediate outcome in multivariate analyses of neurodevelopmental determinants. Key to the success of such work are recent advances in data science as well as the growth of relevant data resources. Multimodal MRI assessment of neurodevelopment can be supplemented with other biomarkers of neurodevelopment such as electroencephalograms, magnetoencephalogram, and non-imaging biomarkers. This review focuses on how maternal nutrition impacts infant brain development, with three purposes: (1) to summarize the current knowledge about how nutrition in stages of pregnancy and breastfeeding impact infant brain development; (2) to discuss multimodal MRI and other measures of early neurodevelopment; and (3) to discuss potential opportunities for data science and artificial intelligence to advance precision nutrition. We hope this review can facilitate the collaborative march toward precision nutrition during pregnancy and the first year of life.

Increased Breastfeeding Proportion Is Associated with Improved Gross Motor Skills at 3-5 Years of Age: A Pilot Study.

Breastmilk provides key nutrients and bio-active factors that contribute to infant neurodevelopment. Optimizing maternal nutrition could provide further benefit to psychomotor outcomes. Our observational cohort pilot study aims to determine if breastfeeding extent and breastmilk nutrients correlate with psychomotor outcomes at school age. The breastfeeding proportion at 3 months of age and neurodevelopmental outcomes at 3-5 years of age were recorded for 33 typically developing newborns born after uncomplicated pregnancies. The association between categorical breastfeeding proportion and neurodevelopmental outcome scores was determined for the cohort using a Spearman correlation with and without the inclusion of parental factors. Vitamin E and carotenoid levels were determined in breastmilk samples from 14 of the mothers. After the inclusion of parental education and income as covariates, motor skill scores positively correlated with breastmilk contents of α-tocopherol (Spearman coefficient 0.88, p-value = 0.02), translutein (0.98, p-value = 0.0007), total lutein (0.92, p-value = 0.01), and zeaxanthin (0.93, p-value = 0.0068). Problem solving skills negatively correlated with the levels of the RSR enantiomer of α-tocopherol (-0.86, p-value = 0.03). Overall, higher exposure to breastfeeding was associated with improved gross motor and problem-solving skills at 3-5 years of age. The potential of α-tocopherol, lutein, and zeaxanthin intake to provide neurodevelopmental benefit is worthy of further investigation.

Deep learning of birth-related infant clavicle fractures: a potential virtual consultant for fracture dating.

BACKGROUND: In infant abuse investigations, dating of skeletal injuries from radiographs is desirable to reach a clear timeline of traumatic events. Prior studies have used infant birth-related clavicle fractures as a surrogate to develop a framework for dating of abuse-related fractures. OBJECTIVE: To develop and train a deep learning algorithm that can accurately date infant birth-related clavicle fractures. MATERIALS AND METHODS: We modified a deep learning model initially designed for face-age estimation to date infant clavicle fractures. We conducted a computerized search of imaging reports and other medical records at a tertiary children's hospital to identify radiographs of birth-related clavicle fracture in infants ≤ 3 months old (July 2003 to March 2021). We used the resultant database for model training, validation and testing. We evaluated the performance of the deep learning model via a four-fold cross-validation procedure, and calculated accuracy metrics: mean absolute error (MAE), root mean square error (RMSE), intraclass correlation coefficient (ICC) and cumulative score. RESULTS: The curated database consisted of 416 clavicle radiographs from 213 infants. Average chronological age (equivalent to fracture age) at time of imaging was 24 days. This model estimated the ages of the clavicle fractures with MAE of 4.2 days, RMSE of 6.3 days and ICC of 0.919. On average, 83.7% of the fracture age estimates were accurate to within 7 days of the ground truth. CONCLUSION: Our deep learning study provides encouraging results for radiographic dating of infant clavicle fractures. With further development and validation, this model might serve as a virtual consultant to radiologists estimating fracture ages in cases of suspected infant abuse.