Artificial intelligence-based detection of paediatric appendicular skeletal fractures: performance and limitations for common fracture types and locations

BACKGROUND

Research into artificial intelligence (AI)-based fracture detection in children is scarce and has disregarded the detection of indirect fracture signs and dislocations.

OBJECTIVE

To assess the diagnostic accuracy of an existing AI-tool for the detection of fractures, indirect fracture signs, and dislocations.

MATERIALS AND METHODS

An AI software, BoneView (Gleamer, Paris, France), was assessed for diagnostic accuracy of fracture detection using paediatric radiology consensus diagnoses as reference. Radiographs from a single emergency department were enrolled retrospectively going back from December 2021, limited to 1,000 radiographs per body part. Enrolment criteria were as follows: suspected fractures of the forearm, lower leg, or elbow; age 0-18 years; and radiographs in at least two projections.

RESULTS

Lower leg radiographs showed 607 fractures. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were high (87.5%, 87.5%, 98.3%, 98.3%, respectively). Detection rate was low for toddler's fractures, trampoline fractures, and proximal tibial Salter-Harris-II fractures. Forearm radiographs showed 1,137 fractures. Sensitivity, specificity, PPV, and NPV were high (92.9%, 98.1%, 98.4%, 91.7%, respectively). Radial and ulnar bowing fractures were not reliably detected (one out of 11 radial bowing fractures and zero out of seven ulnar bowing fractures were correctly detected). Detection rate was low for styloid process avulsions, proximal radial buckle, and complete olecranon fractures. Elbow radiographs showed 517 fractures. Sensitivity and NPV were moderate (80.5%, 84.7%, respectively). Specificity and PPV were high (94.9%, 93.3%, respectively). For joint effusion, sensitivity, specificity, PPV, and NPV were moderate (85.1%, 85.7%, 89.5%, 80%, respectively). For elbow dislocations, sensitivity and PPV were low (65.8%, 50%, respectively). Specificity and NPV were high (97.7%, 98.8%, respectively).

CONCLUSIONS

The diagnostic performance of BoneView is promising for forearm and lower leg fractures. However, improvement is mandatory before clinicians can rely solely on AI-based paediatric fracture detection using this software.

Parameters of glucose metabolism and the aging brain: a magnetization transfer imaging study of brain macro- and micro-structure in older adults without diabetes

Given the concurrent, escalating epidemic of diabetes mellitus and neurodegenerative diseases, two age-related disorders, we aimed to understand the relation between parameters of glucose metabolism and indices of pathology in the aging brain. From the Leiden Longevity Study, 132 participants (mean age 66 years) underwent a 2-h oral glucose tolerance test to assess glucose tolerance (fasted and area under the curve (AUC) glucose), insulin sensitivity (fasted and AUC insulin and homeostatic model assessment of insulin sensitivity (HOMA-IS)) and insulin secretion (insulinogenic index). 3-T brain MRI was used to detect macro-structural damage (atrophy, white matter hyper-intensities, infarcts and/or micro-bleeds) and magnetization transfer imaging (MTI) to detect loss of micro-structural homogeneity that remains otherwise invisible on conventional MRI. Macro-structurally, higher fasted glucose was significantly associated with white matter atrophy (P = 0.028). Micro-structurally, decreased magnetization transfer ratio (MTR) peak height in gray matter was associated with higher fasted insulin (P = 0.010), AUCinsulin (P = 0.001), insulinogenic index (P = 0.008) and lower HOMA-IS index (P < 0.001). Similar significant associations were found for white matter. Thus, while higher glucose was associated with macro-structural damage, impaired insulin action was associated more strongly with reduced micro-structural brain parenchymal homogeneity. These findings offer some insight into the association between different parameters of glucose metabolism (impairment of which is characteristic of diabetes mellitus) and brain aging.

Visceral adipose tissue is associated with microstructural brain tissue damage

OBJECTIVE

Obesity has been associated with microstructural brain tissue damage. Different fat compartments demonstrate different metabolic and endocrine behaviors. The aim was to investigate the individual associations between abdominal visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) and microstructural integrity in the brain.

METHODS

This study comprised 243 subjects aged 65.4 ± 6.7 years. The associations between abdominal VAT and SAT, assessed by CT, and magnetization transfer imaging markers of brain microstructure for gray and white matter were analyzed and adjusted for confounding factors.

RESULTS

VAT was associated with normalized MTR peak height in gray (β -0.216) and white matter (β -0.240) (both P < 0.01) after adjustment for confounding factors. After adjustment for sex, age, and descent, SAT was associated with normalized MTR peak height in gray and white matter, but not after additional correction for BMI, hypertension, current smoking, statin use, and type 2 diabetes (respectively, β -0.055 and β 0.035, both P > 0.05). Stepwise linear regression analysis showed that only VAT was associated with normalized MTR peak height in gray and white matter (both P < 0.001).

CONCLUSIONS

Our data indicate that increased abdominal VAT rather than SAT is associated with microstructural brain tissue damage in elderly individuals.

An in vivo study on brain microstructure in biological and chronological ageing

This study aimed to investigate whether magnetization transfer imaging (MTI) parameters of cortical gray and white matter and subcortical gray matter structures differ between subjects enriched for human familial longevity and control subjects to provide a thorough description of the brain phenotype of familial longevity. Moreover, we aimed to describe cerebral ageing effects on MTI parameters in an elderly cohort. All subjects were included from the Leiden Longevity Study and underwent 3 Tesla MTI of the brain. In total, 183 offspring of nonagenarian siblings, who are enriched for familial factors of longevity, were contrasted with 163 environmentally and age-matched controls. No differences in cortical and subcortical gray matter and white matter MTI parameters were found between offspring and control subjects using histogram-based and voxel-wise analyses. Cortical gray matter and white matter MTI parameters decreased with increasing chronological age (all p < 0.001). Decrease of white matter magnetization transfer ratio (MTR) was homogeneous throughout the whole mean white matter skeleton except for parts of the callosal splenium and partly the posterior limb of the internal capsule and superior region of the corona radiata (p < 0.05). Mean MTR of subcortical gray matter structures decreased with increasing age (p amygdala, caudate nucleus and putamen < 0.001; p pallidum = 0.001, p thalamus = 0.002). In conclusion, the brain phenotype of human familial longevity is - at a mean age of 66 years - not characterized by preserved macromolecular brain tissue integrity.

Preserved white matter integrity is a marker of familial longevity

OBJECTIVE

Brain tissue integrity is highly heritable, and its decline is a common phenomenon of ageing. This study aimed to determine whether the phenotype of familial longevity is marked by a relative preservation of brain tissue microstructure.

METHODS

Participants were enrolled in the Leiden Longevity Study. In total, 185 middle-aged to elderly offspring of nonagenarian siblings, who were enriched for familial factors of longevity, were contrasted with 171 environment- and age-matched controls. All subjects underwent 3T whole brain magnetic resonance diffusion tensor imaging.

RESULTS

Voxel-wise analysis revealed widespread age-related decrease of white matter fractional anisotropy and increases of axial, radial, and mean diffusivity (all p < 0.003). Offspring showed higher mean white matter fractional anisotropy (mean [standard error]: offspring, 0.3232 [0.0009]; controls, 0.3212 [0.0009]; p = 0.04) compared to control subjects independent of cardiovascular risk factors. When differences in white matter diffusion parameters between offspring and control subjects were assessed voxel-wise, offspring showed higher white matter fractional anisotropy and lower white matter radial diffusivity predominantly in the callosal genu and body (both p < 0.003). With the effect of chronological age on white matter microstructure taken into account, offspring can be considered 4.5 years "biologically younger" compared to control subjects with regard to white matter integrity.

INTERPRETATION

Both middle-aged to elderly offspring of nonagenarian siblings and control subjects show common age-related decline of white matter integrity, but it is less marked in the callosal genu and body in the offspring. This corresponds to a biological age benefit of 4.5 years of the offspring as compared to the control subjects.

Microstructural brain tissue damage in metabolic syndrome

OBJECTIVE

We investigated the association between metabolic syndrome risk factors and brain tissue integrity, as assessed by magnetic resonance imaging.

RESEARCH DESIGN AND METHODS

From the Leiden Longevity Study, which is a community-based study of long-lived subjects, their offspring, and partners thereof, 130 subjects (61 men; mean age 66 years) were included. A metabolic syndrome score was computed by summing the individual number of components according to the Adult Treatment Panel III criteria. We performed linear and logistic regression analysis and used standardized β-values to assess the association between metabolic syndrome and brain macrostructure (brain volume and white matter lesion load, lacunar infarcts, and cerebral microbleeds) and microstructure (mean magnetization transfer ratio [MTR], MTR histogram peak height, fractional anisotropy, and mean diffusivity [MD]). Linear and stepwise regression analysis was performed to identify the individual contribution of one metabolic syndrome parameter adjusting for the four other parameters. Models were adjusted for age, sex, and relation to long-lived family.

RESULTS

Brain macrostructure was not associated with metabolic syndrome. In contrast, metabolic syndrome was associated with decreased gray (β = -0.3 P = 0.001) and white matter peak height (β = -0.3, P = 0.002) and increased gray matter MD (β = 0.2, P = 0.01, P = 0.01). Serum HDL cholesterol (β = 0.22, P = 0.012), triglycerides (β =-0.25, P = 0.002), BMI (β =-0.2, P = 0.014), and diastolic blood pressure (β = -0.17, P = 0.047, and β = -0.23, P = 0.009, for gray and white matter, respectively) were independent factors in these changes in brain microstructure.

CONCLUSIONS

In early manifest metabolic syndrome, brain tissue decline can be detected. Serum HDL cholesterol, triglycerides, BMI, and diastolic blood pressure were independent factors in brain tissue integrity.

Increased functional connectivity and brain atrophy in elderly with subjective memory complaints

Subjective memory complaints (SMC) are common among elderly. Although subtle changes in memory functioning can hardly be determined using neuropsychological evaluation, neuroimaging studies indicate regionally smaller brain structures in elderly with SMC. Imaging of resting-state functional connectivity is sensitive to detect changes in neurodegenerative diseases, but is currently underexplored in SMC. Here, we investigate resting-state functional connectivity and brain structure in SMC. We analyzed magnetic resonance imaging data of 25 elderly with SMC and 29 age-matched controls (mean age of 71 years). Voxel-based morphometry and volume measurements of subcortical structures were employed on the structural scans using FSL. The dual regression method was used to analyze voxel-wise functional connectivity in relation to eight well-characterized resting-state networks. Group differences were studied with two-sample t-tests (p<0.05, Family-Wise Error corrected). In addition to gray matter volume reductions (hippocampus, anterior cingulate cortex (ACC), medial prefrontal cortex, cuneus, precuneus, and precentral gyrus), elderly with SMC showed increased functional connectivity in the default mode network (hippocampus, thalamus, posterior cingulate cortex (PCC), cuneus, precuneus, and superior temporal gyrus) and the medial visual network (ACC, PCC, cuneus, and precuneus). This study is the first which demonstrates that, in addition to smaller regional brain volumes, increases in functional connectivity are present in elderly with SMC. This suggests that self-reported SMC is a reflection of objective alterations in brain function. Furthermore, our results indicate that functional imaging, in addition to structural imaging, can be a useful tool to objectively determine a difference in brain integrity in SMC.

Brain tissue volumes in familial longevity: the Leiden Longevity Study

Atrophy is one of the major age-related changes in the brain. The absence of brain atrophy in elderly individuals reflects deceleration in the process of biological aging. Moreover, results from human twin studies suggest a large genetic influence on the variance of human brain tissue volumes. To investigate the association of brain volumes with exceptional longevity, we tested whether middle-aged to elderly offspring of nonagenarian siblings have larger brain volumes than their spouses using magnetic resonance imaging. No differences in whole brain, gray matter and white matter volume were found. These brain volumes were associated with chronological age in offspring and control subjects (all P < 0.001). Left amygdalar volume of the offspring was larger (P = 0.03) compared with control subjects [mean volume offspring (cm3) (95% confidence interval, CI) = 1.39 (1.36-1.42), mean volume control subjects (cm3) (95% CI) = 1.32 (1.29-1.35)]. Association of left amygdalar volume with familial longevity was particularly pronounced when offspring with the oldest long-lived parent were compared with control subjects (P = 0.01). Amygdalar volumes were not associated with chronological age in both groups. Our findings suggest that the observed association of a larger left amygdalar volume with familial longevity is not caused by a relative preservation of the left amygdala during the course of aging but most likely a result of early development caused by a genetic familial trait.

Lower susceptibility to cerebral small vessel disease in human familial longevity: the Leiden Longevity Study

BACKGROUND AND PURPOSE

On MRI, cerebral white matter lesions, lacunar infarcts, and cerebral microbleeds are common imaging correlates of cerebral small vessel damage in apparently healthy elderly individuals. We investigated whether middle-aged to elderly offspring of nonagenarian siblings, who are predisposed to become long-lived as well, have a lower prevalence of white matter lesions, lacunar infarcts, and cerebral microbleeds than control subjects.

METHODS

All subjects were from the Leiden Longevity Study. In this study, middle-aged to elderly offspring of nonagenarian siblings, who are predisposed to become long-lived as well, were contrasted to their spouses. Cerebral small vessel disease was assessed using 3-T MRI.

RESULTS

Offspring were less likely to have severe periventricular frontal caps (odds ratio [OR], 0.3; 95% confidence interval [CI], 0.1-1.1; P trend=0.01) and severe periventricular bands (OR, 0.4; 95% CI, 0.2-0.8; P trend=0.02). Moreover, offspring were less likely to have frontal (OR, 0.4; 95% CI, 0.2-0.9; P trend=0.05), parietal (OR, 0.4; 95% CI, 0.1-0.9; P trend=0.001), temporal (OR, 0.3; 95% CI, 0.1-0.8]; P trend=0.004), and occipital subcortical white matter lesions (OR, 0.3; 95% CI, 0.1-0.6; P trend=0.001). Prevalence of lacunar infarcts also was lower in offspring (OR, 0.3; 95% CI, 0.1-1.1; P=0.07). Prevalence of microbleeds was not significantly different in offspring and control subjects.

CONCLUSIONS

Exceptional familial longevity is associated with a lower susceptibility to white matter lesions and lacunar infarcts, but not cerebral microbleeds.