Deep generative model of the distal tibial classic metaphyseal lesion in infants: assessment of synthetic images.

Background: The classic metaphyseal lesion (CML) is a distinctive fracture highly specific to infant abuse. To increase the size and diversity of the training CML database for automated deep-learning detection of this fracture, we developed a mask conditional diffusion model (MaC-DM) to generate synthetic images with and without CMLs. Purpose: To objectively and subjectively assess the synthetic radiographic images with and without CMLs generated by MaC-DM. Materials and Methods: For retrospective testing, we randomly chose 100 real images (50 normals and 50 with CMLs; 39 infants, male = 22, female = 17; mean age = 4.1 months; SD = 3.1 months) from an existing distal tibia dataset (177 normal, 73 with CMLs), and generated 100 synthetic distal tibia images via MaC-DM (50 normals and 50 with CMLs). These test images were shown to 3 blinded radiologists. In the first session, radiologists determined if the images were normal or had CMLs. In the second session, they determined if the images were real or synthetic. We analyzed the radiologists' interpretations and employed t-distributed stochastic neighbor embedding technique to analyze the data distribution of the test images. Results: When presented with the 200 images (100 synthetic, 100 with CMLs), radiologists reliably and accurately diagnosed CMLs (kappa = 0.90, 95% CI = [0.88-0.92]; accuracy = 92%, 95% CI = [89-97]). However, they were inaccurate in differentiating between real and synthetic images (kappa = 0.05, 95% CI = [0.03-0.07]; accuracy = 53%, 95% CI = [49-59]). The t-distributed stochastic neighbor embedding analysis showed substantial differences in the data distribution between normal images and those with CMLs (area under the curve = 0.996, 95% CI = [0.992-1.000], P < .01), but minor differences between real and synthetic images (area under the curve = 0.566, 95% CI = [0.486-0.647], P = .11). Conclusion: Radiologists accurately diagnosed images with distal tibial CMLs but were unable to distinguish real from synthetically generated ones, indicating that our generative model could synthesize realistic images. Thus, MaC-DM holds promise as an effective strategy for data augmentation in training machine-learning models for diagnosis of distal tibial CMLs.

Dating birth-related clavicular fractures: pediatric radiologists versus artificial intelligence.

BACKGROUND: Fracture dating from skeletal surveys is crucial in the diagnosis and investigation of infant abuse. However, this task is challenging because of the subjective nature of the radiologic interpretation and the lack of ground truth. Researchers have used birth-related clavicle fractures as a surrogate to study the radiographic pattern of healing; however, they did not elucidate the accuracy performance of the radiologists in dating fractures. OBJECTIVE: To determine the accuracy of radiologists in dating birth-related clavicle fractures and compare their performance to that achieved by computer algorithm. MATERIALS AND METHODS: We used a previously assembled birth-related clavicle fracture database consisting of 416 anteroposterior clavicle radiographs as the study cohort. The average and standard deviation of the fracture age within this database were 24 days and 18 days, respectively. Three blinded radiologists independently estimated the ages of the clavicle fractures depicted in the radiographs within the database. We compared these estimation results to those made by a recently published deep-learning (DL) model conducted with the identical infant cohort. We calculated standard error metrics to compare the accuracy performances of the radiologists and the computer model. RESULTS: The intra- and inter-reader agreements of the fracture age estimates by the radiologists were moderate to good. The radiologists estimated the fracture ages with a mean absolute error (MAE) of 6.1-7.1 days, and standard deviation of the absolute error of 6.3-8.3 days. The accuracy performances of the three radiologists were not significantly different from one another. In comparison, the DL model estimated the age of clavicle fractures with an MAE of 4.2 days, significantly lower than all of the radiologists (P < 0.001). CONCLUSION: Three experienced pediatric radiologists dated clavicular fractures with moderate-good intra- and inter-reader agreements. The correlations between the radiologists' estimates and the ground truth were moderate to good. The fracture ages assigned by the DL model showed superior correlation with the ground truth compared to radiologists' dating estimates.

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.

Machine learning to identify distal tibial classic metaphyseal lesions of infant abuse: a pilot study.

BACKGROUND: The classic metaphyseal lesion (CML) is an injury that is highly specific for infant abuse, and the distal tibia is one of the most common sites of occurrence. A machine learning tool that identifies distal tibial CMLs on infant skeletal surveys could assist radiologists in the diagnosis of infant abuse. OBJECTIVE: To develop and evaluate a machine learning-based classification algorithm to identify distal tibial CMLs on skeletal surveys performed for suspected infant abuse. MATERIALS AND METHODS: We reviewed medical records of infants (≤1 year old) who had skeletal surveys for suspected abuse at a large tertiary children's hospital over the past 13 years to identify those at low and high risk for abuse. Normal distal tibial radiographs from the low-risk group formed the normal study cohort; radiographs with distal tibial CMLs from the high-risk group formed the abnormal study cohort. We used these two cohorts to train a machine learning algorithm to classify distal tibial radiographs as normal or abnormal. We systematically evaluated this algorithm using a fivefold cross-validation procedure and statistically analyzed the results. RESULTS: The normal study cohort consisted of 177 radiographs from 89 infants, and the abnormal study cohort consisted of 73 radiographs from 35 infants. Our machine learning algorithm showed an overall performance accuracy of 93% and Kappa of 0.84. The overall sensitivity and specificity of the model were 88% and 96%, respectively. CONCLUSION: Our developed machine learning model shows encouraging results as an automated tool to identify CMLs of the distal tibia on skeletal surveys performed for suspected infant abuse.

Temporal Pattern of Radiographic Findings of Costochondral Junction Rib Fractures on Serial Skeletal Surveys in Suspected Infant Abuse.

BACKGROUND. Costochondral junction (CCJ) rib fractures pose a challenge in the radiographic detection and dating of infant abuse. OBJECTIVE. The purpose of this article was to assess the temporal pattern of radio-graphic findings of CCJ fractures on a serial skeletal survey (SS). METHODS. Reports of SSs performed for suspected infant abuse were reviewed to identify those reporting a CCJ fracture. Study inclusion required undergoing initial and approximately 2-week follow-up SSs that included anteroposterior and bilateral oblique radiographs of the reported CCJ rib fracture. Two pediatric radiologists retrospectively classified fractures in terms of the primary injury pattern (bucket-handle: visible crescentic fracture line; corner: visible triangular fracture line; other) and secondary healing pattern (growth disturbance; sclerosis; subperiosteal new bone formation [SPNBF]). Discrepant readings were resolved by consensus. RESULTS. The final cohort included 26 infants with 81 CCJ fractures. On initial SS, 59% (48/81) of fractures showed a primary pattern, most commonly a bucket-handle pattern (46%; 37/81); 6% (5/81) showed a primary pattern on follow-up SS (p < .001). On initial SS, the prevalence of a secondary pattern was 89% (72/81), most commonly a growth disturbance (85%; 69/81), followed by sclerosis (57%; 46/81); 80% (65/81) showed a secondary pattern on follow-up SS (p = .12). Overall prevalence of SPNBF was 28% (23/81). Addition of bilateral oblique views on initial SS resulted in a significant increase relative to the anteroposterior views alone in the detection of primary and secondary patterns by 15% (p = .04) and 30% (p < .001), respectively. CONCLUSION. A bucket-handle appearance is the most common primary pattern of fracture. The significantly lower prevalence of a primary pattern on follow-up vs initial SSs suggests that the CCJ fracture line is usually visible for only approximately 2 weeks. A growth disturbance of the rib end is the most common secondary pattern, followed by bony sclerosis, consistent with a healing injury. SPNBF is uncommon. Most CCJ fractures are in a healing phase at initial diagnosis. The signs of repair commonly remain visible on 2-week follow-up. The increased diagnostic yield of oblique views provides support to the inclusion of these projections in routine SS protocols. CLINICAL IMPACT. The findings will help radiologists improve the diagnosis and dating of CCJ rib fractures.

Subperiosteal new bone formation with the distal tibial classic metaphyseal lesion: prevalence on radiographic skeletal surveys.

BACKGROUND: The classic metaphyseal lesion (CML) is a strong indicator of infant abuse, and the distal tibia is one of the most common sites for this injury. OBJECTIVE: To determine the prevalence of subperiosteal new bone formation accompanying the distal tibial CMLs identified on infant skeletal surveys. MATERIALS AND METHODS: Skeletal surveys performed for suspected infant abuse (2005-2017) were reviewed. Inclusion criteria were 1) anteroposterior (AP) and lateral radiographs of a distal tibial CML from the initial survey, 2) AP radiograph from the 2-week follow-up survey, 3) additional fractures, 4) child protection team consults and 5) mandated report filing for suspected abuse. We identified 22 distal tibial CMLs from 16 infants. Radiographs of these lesions were shown on the picture archiving and communication system to two blinded pediatric radiologists. Readers indicated the presence/absence of subperiosteal new bone formation on individual and combinations of images. RESULTS: Inter-reader agreements were fair (kappa=0.47). The prevalence of subperiosteal new bone formation on initial AP radiograph was 34%. Significant increases in the prevalence were found with the addition of follow-up AP radiograph (57%; P<0.001), initial lateral radiograph (57%; P=0.002) and follow-up AP plus initial lateral radiographs (71%; P<0.001). Statistically significant increases in prevalence were also noted when the third view was added to the other two views (increase of 14%; P=0.024). CONCLUSION: Even when skeletal surveys include initial AP, lateral and follow-up AP radiographs of the tibia, nearly one-third of distal tibial CMLs will fail to demonstrate subperiosteal new bone formation.

Lower-extremity growth patterns and skeletal maturation in children with unilateral fibular hemimelia.

BACKGROUND: Fibular hemimelia is the most common congenital long-bone deficiency. It is usually unilateral and results in a limb-length discrepancy. The literature generally subscribes to the concept of constant inhibition, a process by which limb-length ratios between the shorter and longer extremity remain constant throughout growth, but scientific data supporting this concept are sparse. Additionally, recent literature suggests that these children have abnormal skeletal maturation. OBJECTIVE: To elucidate the lower-extremity long-bone growth patterns and skeletal maturation of children with unilateral fibular hemimelia. MATERIALS AND METHODS: We reviewed medical records of children with unilateral fibular hemimelia seen at a large pediatric hospital over a 17-year period. Inclusion criteria were: at least two scanograms prior to any shortening/lengthening procedure, and no other congenital or acquired disorders. We collected the study cohort's femoral and tibial lengths (scanogram reports), plotted them against patient chronological ages and compared them to published growth standards. When these children's bone ages (Greulich and Pyle) were available, we plotted them against the children's chronological ages. RESULTS: Twenty-three children were included (total=115 scanograms). At least 1 bone-age assessment was performed in 19 children (total=84 bone ages). All bone growth curves were within normal growth standards for the femur and tibia. Length ratios between shorter and longer limbs remained constant. Skeletal maturation was within two standard deviations of normal in 90% of bone ages. CONCLUSION: Lower-extremity long bones of children with unilateral fibular hemimelia have relatively normal growth curves, supporting and confirming the concept of constant inhibition. Most children show normal skeletal maturation.

The distal tibial classic metaphyseal lesion: medial versus lateral cortical injury.

BACKGROUND: The distal tibia is a common location for the classic metaphyseal lesion (CML). Prior radiologic-pathologic studies have suggested a tendency for medial, as opposed to lateral, cortical injury with the CML, but there has been no formal study of the geographic distribution of this strong indicator of abuse. OBJECTIVE: This study compares medial versus lateral cortical involvement of distal tibial CMLs in a clinical cohort of infants with suspected abuse. MATERIALS AND METHODS: Reports of 1,020 skeletal surveys performed for suspected abuse (July 2005-June 2016) were reviewed. Twenty-six distal tibial CMLs (14 unilateral, 6 bilateral) with anteroposterior (AP) and lateral projections on the initial skeletal survey and at least an AP view on the follow-up survey were identified in 20 infants. Two blinded pediatric radiologists determined if the medial and/or lateral margins of the distal tibial metaphysis were involved by the CML. RESULTS: Average interreader absolute agreement and kappa scores were 0.69-0.90 and 0.45-0.72, respectively. Average intrareader absolute agreement and kappa scores were 0.65-0.88 and 0.44-0.57, respectively. Analyses showed that the distal tibial CML almost always involved the medial cortical margin (reader 1=89%, reader 2=88%, pooled=89%) and the fracture infrequently involved the lateral cortical margin (reader 1=12%, reader 2=38%, pooled=26%). The percentage point difference between fracture involvement in medial and lateral margins was statistically significant from zero (P<0.001). CONCLUSION: The distal tibial CML is most often encountered medially; lateral involvement is uncommon. This observation should help guide the radiologic diagnosis and could have implications for understanding the biomechanics of this distinctive injury.

Biomechanics of the classic metaphyseal lesion: finite element analysis.

BACKGROUND: The classic metaphyseal lesion (CML) is strongly associated with infant abuse, but the biomechanics responsible for this injury have not been rigorously studied. Radiologic and CT-pathological correlates show that the distal tibial CML always involves the cortex near the subperiosteal bone collar, with variable extension of the fracture into the medullary cavity. Therefore, it is reasonable to assume that the primary site of bone failure is cortical, rather than intramedullary. OBJECTIVE: This study focuses on the strain patterns generated from finite element modeling to identify loading scenarios and regions of the cortex that are susceptible to bone failure. MATERIALS AND METHODS: A geometric model was constructed from a normal 3-month-old infant's distal tibia and fibula. The model's boundary conditions were set to mimic forceful manipulation of the ankle with eight load modalities (tension, compression, internal rotation, external rotation, dorsiflexion, plantar flexion, valgus bending and varus bending). RESULTS: For all modalities except internal and external rotation, simulations showed increased cortical strains near the subperiosteal bone collar. Tension generated the largest magnitude of cortical strain (24%) that was uniformly distributed near the subperiosteal bone collar. Compression generated the same distribution of strain but to a lesser magnitude overall (15%). Dorsiflexion and plantar flexion generated high (22%) and moderate (14%) localized cortical strains, respectively, near the subperiosteal bone collar. Lower cortical strains resulted from valgus bending, varus bending, internal rotation and external rotation (8-10%). The highest valgus and varus bending cortical strains occurred medially. CONCLUSION: These simulations suggest that the likelihood of the initial cortical bone failure of the CML is higher along the margin of the subperiosteal bone collar when the ankle is under tension, compression, valgus bending, varus bending, dorsiflexion and plantar flexion, but not under internal and external rotation. Focal cortical strains along the medial margins of the subperiosteal bone collar with varus and valgus bending may explain the known tendency for focal distal tibial CMLs to occur medially. Further research is needed to determine the threshold of applied forces required to produce this strong indicator of infant abuse.

Infant lower extremity long bone growth rates: comparison of contemporary with early 20th century data using mixed effect models.

OBJECTIVES: Modern reference standards for long bone growth are lacking for infants (≤1 year). This study develops a quantitative framework to characterize lower extremity bone shaft growth during infancy based on radiographic images, and compares it to data from the 1930s. METHODS: Femoral, tibial, and fibular shaft length measurements (diaphysis plus metaphysis) were collected retrospectively from 70 infants on initial and follow-up skeletal surveys performed for suspected abuse (7/2005-2/2013). These serial skeletal survey data (SSSD) were compared to the Denver Child Research Council data (DCRCD), a 1930's longitudinal dataset from 80 infants. Mixed effects regression models were developed to estimate growth trajectories from these data. Growth trajectories and short-term (≤2 months) growth rates were compared. RESULTS: Statistically distinct models described the contemporary (SSSD) and historic (DCRCD) datasets; however, there was substantial overlap (77-90%) between their confidence bands for the three measured bones. Based on developed models, the average long bone shafts of the DCRCD are shorter at birth than SSSD (femur: 77.0 vs. 82.3 mm; tibia: 64.4 vs. 68.2 mm; fibula: 61.0 vs. 64.4 mm), but the DCRCD long bone growth rates are faster than SSSD (femur: 0.21 vs. 0.17 mm/day; tibia: 0.16 vs. 0.14 mm/day; fibula: 0.15 vs. 0.14 mm/day). Short-term growth rates of these bones decreased with age. The effect of sexual dimorphism on long bone growth during infancy was non-significant. CONCLUSION: This study provides reference standards for long bone growth rates during the dynamic period of infancy that may aid clinical assessment, and also inform research studies of disorders associated with altered skeletal growth.

Growth recovery lines are more common in infants at high vs. low risk for abuse.

BACKGROUND: Growth recovery lines, also known as growth arrest lines, are transverse radiodense metaphyseal bands that develop due to a temporary arrest of endochondral ossification caused by local or systemic insults. OBJECTIVE: To determine if growth recovery lines are more common in infants at high risk versus low risk for abuse. MATERIALS AND METHODS: Reports of American College of Radiology compliant skeletal surveys (1999-2013) were reviewed with clinical records. Infants at low risk for abuse had a skull fracture without significant intracranial injury, history of a fall and clinical determination of low risk (child protection team/social work assessment). Infants at high risk had significant intracranial injury, retinal hemorrhages, other skeletal injuries and clinical determination of high risk. There were 52 low-risk infants (mean: 4.7 months, range: 0.4-12 months) and 21 high-risk infants (mean: 4.2 months, range: 0.8-9.1 months). Two blinded radiologists independently evaluated the skeletal survey radiographs of the knees/lower legs for the presence of at least one growth recovery line. RESULTS: When growth recovery lines are scored as probably present or definitely present, their prevalence in the low-risk group was 38% (standard deviation [SD] = 8%; reader 1 = 17/52, reader 2 = 23/52) vs. 71% (SD = 7%; reader 1 = 16/21, reader 2 = 14/21) in the high-risk group (P < 0.001; odds ratio 4.0, 95% CI: 1.7-9.5). CONCLUSION: Growth recovery lines are encountered at a significantly higher rate in infants at high risk vs. low risk for abuse. This suggests that abused infants are prone to a temporary disturbance in endochondral ossification as a result of episodic physiological stresses.

Bone age assessment practices in infants and older children among Society for Pediatric Radiology members.

BACKGROUND: Numerous bone age estimation techniques exist, but little is known about what methods radiologists use in clinical practice. OBJECTIVE: To determine which methods pediatric radiologists use to assess bone age in children, and their confidence in these methods. MATERIALS AND METHODS: Society for Pediatric Radiology (SPR) members were invited to complete an online survey regarding bone age assessment. Respondents were asked to identify the methods used and their confidence with their technique for the following groups: Infants (<1 year old), 1- to 3-year-olds and 3- to 18-year-olds. RESULTS: Of the 937 SPR members invited, 441 responded (47%). For infants, 70% of respondents use the hand/wrist method of Greulich and Pyle, 27% use a hemiskeleton method (e.g., Sontag or Elgenmark), and 14.4% use the knee method of Pyle and Hoerr. Of these respondents, 34% were not confident with their technique. For 1- to 3-year-olds, 86% used Greulich and Pyle, and 19% used a hemiskeleton method; 21% were not confident with their technique in this age group. For 3- to 18-year-olds, 97% used Greulich and Pyle, and only 6% of respondents were not confident with their technique in this category. A logistic regression analysis demonstrated that the chronological age of the patient had the greatest impact on reader confidence, with the odds ratios for confidence being 4 times greater in the 3- to 18-year-olds category compared to the younger groups. CONCLUSION: For children older than 3 years, the majority of pediatric radiologists are very confident in their use of Greulich and Pyle for bone age assessment. However a variety of methodologies are used when assessing bone age in infants and younger children, and pediatric radiologists are less confident assessing bone age in these children. This survey highlights the need for a consensus protocol on bone age assessment of younger children and infants that provides readers with a higher degree of confidence.

Posteromedial subtalar coalitions: prevalence and associated morphological alterations of the sustentaculum tali.

BACKGROUND: A variant form of subtalar coalition isolated to the posterior sustentaculum has been previously described, though its prevalence is not known and its relationship to the middle facet has not been characterized. OBJECTIVE: To determine the prevalence and morphological alterations of isolated posteromedial subtalar coalitions. MATERIALS AND METHODS: Computed tomography (CT) images of the foot or ankle performed from 2004 to 2012 were reviewed and 97 patients (mean age: 13.3+/-2.9 years; range: 9-19 years) with subtalar coalition were identified. In 41 (42%) of these, the condition was bilateral, yielding a total of 138 subtalar coalitions. In the 33 patients where CT demonstrated isolated posteromedial subtalar coalitions, multiplanar reformats along the long axis of the sustentaculum tali were generated, from which the anteroposterior dimensions of the sustentaculum tali and middle facet were measured. Posterior sustentaculum measurements defining the posterior extension of the sustentaculum beyond the middle facet were directly measured by two radiologists. Ratios of middle facet to posterior sustentaculum measurements were calculated. Thirty-three patients undergoing CT for ankle fracture served as controls. RESULTS: Ninety-seven of 138 coalitions (70.2%) affected the middle facet and 2/138 (1.4%) involved the posterior facet. There were 39 (28.2%) posteromedial subtalar coalitions in 33 patients. Mean AP measurements of the middle facet and posterior sustentaculum in patients with posteromedial subtalar coalitions were 12.6 mm and 18.2 mm, respectively, compared to 16.6 mm and 9.2 mm in controls (P<0.001). Mean middle facet/posterior sustenaculum (MF/PS) ratios were 0.74 for posteromedial subtalar coalitions vs. 1.92 for controls (P<0.001). CONCLUSION: Posteromedial subtalar coalitions comprise more than one-quarter of subtalar coalitions, and are associated with an intact, but shorter, middle facet and longer sustentaculum tali. This observation may aid in accurate diagnosis and management of this relatively common disorder.

Prevalence of Capitellar Osteochondritis Dissecans in Children With Chronic Radial Head Subluxation and Dislocation.

OBJECTIVE: An abnormal radiocapitellar articulation may play a role in the development of capitellar osteochondritis dissecans (OCD). The objective of this study is to determine the prevalence of capitellar OCD in pediatric patients with chronic radial head subluxation (RHS) or radial head dislocation (RHD). MATERIALS AND METHODS: We searched the radiology reports generated at a large pediatric hospital between 2004 and 2012, to identify patients 8-18 years old with chronic RHS or RHD. Patients with underlying bony abnormalities that could obscure, be confused with, or increase the risk of OCD of the elbow were excluded. Two pediatric radiologists assessed images of the elbow, grading radiocapitellar alignment as follows: normal, subluxed, dislocated, or intermittently subluxed and dislocated. Readers also reviewed images to detect capitellar OCD. The differences in the prevalence of OCD noted in elbows in each category of radiocapitellar alignment were assessed. RESULTS: A total of 118 cases of chronic RHS, RHD, or both were detected in 100 patients. The radiocapitellar alignment of these cases was defined as follows: 53-58% were dislocated, 26-31% were subluxed, 14-15% were intermittently subluxed and dislocated, and less than 1% had normal alignment. There were 14 cases of capitellar OCD (11.9%). One elbow with OCD was categorized by both readers as dislocated, with the remaining 13 elbows (93%) deemed by the two readers to be either subluxed or intermittently subluxed and dislocated. The percentage of elbows with RHS that were found to have capitellar OCD was 32.3-33.3%, compared with 1.4-1.6% of elbows with RHD. CONCLUSION: The prevalence of capitellar OCD is greater in elbows with RHS than in those with RHD. This finding suggests that abnormal radiocapitellar mechanics are a factor in the development of OCD. In children with RHS and new or worsening pain, capitellar OCD should be considered, and imaging strategies should be directed toward early identification of this important and treatable condition.

Infant bone age estimation based on fibular shaft length: model development and clinical validation.

BACKGROUND: Bone age in infants (<1 year old) is generally estimated using hand/wrist or knee radiographs, or by counting ossification centers. The accuracy and reproducibility of these techniques are largely unknown. OBJECTIVE: To develop and validate an infant bone age estimation technique using fibular shaft length and compare it to conventional methods. MATERIALS AND METHODS: We retrospectively reviewed negative skeletal surveys of 247 term-born low-risk-of-abuse infants (no persistent child protection team concerns) from July 2005 to February 2013, and randomized them into two datasets: (1) model development (n = 123) and (2) model testing (n = 124). Three pediatric radiologists measured all fibular shaft lengths. An ordinary linear regression model was fitted to dataset 1, and the model was evaluated using dataset 2. Readers also estimated infant bone ages in dataset 2 using (1) the hemiskeleton method of Sontag, (2) the hemiskeleton method of Elgenmark, (3) the hand/wrist atlas of Greulich and Pyle, and (4) the knee atlas of Pyle and Hoerr. For validation, we selected lower-extremity radiographs of 114 normal infants with no suspicion of abuse. Readers measured the fibulas and also estimated bone ages using the knee atlas. Bone age estimates from the proposed method were compared to the other methods. RESULTS: The proposed method outperformed all other methods in accuracy and reproducibility. Its accuracy was similar for the testing and validating datasets, with root-mean-square error of 36 days and 37 days; mean absolute error of 28 days and 31 days; and error variability of 22 days and 20 days, respectively. CONCLUSION: This study provides strong support for an infant bone age estimation technique based on fibular shaft length as a more accurate alternative to conventional methods.

Absence of rickets in infants with fatal abusive head trauma and classic metaphyseal lesions.

PURPOSE: To determine if rickets is present in cases of infant homicide with classic metaphyseal lesions (CMLs) and other skeletal injuries. MATERIALS AND METHODS: This study was exempt from the institutional human subjects board review because all infants were deceased. An archival review (1984-2012) was performed of the radiologic and histopathologic findings of 46 consecutive infant fatalities referred from the state medical examiner's office for the evaluation of possible child abuse. Thirty infants with distal femoral histologic material were identified. Additional inclusion criteria were as follows: (a) The medical examiner determined that the infant had sustained a head injury and that the manner of death was a homicide, (b) at least one CML was evident at skeletal survey, (c) CMLs were confirmed at autopsy, and (d) non-CML fractures were also present. Nine infants (mean age, 3.9 months; age range, 1-9 months) were identified. Two pediatric radiologists independently reviewed the skeletal surveys for rachitic changes at the wrists and knees. A bone and soft tissue pathologist reviewed the distal femoral histologic slices for rickets. RESULTS: There were no radiographic or pathologic features of rickets in the cohort. CONCLUSION: The findings provide no support for the view that the CML is due to rickets. Rather, they strengthen a robust literature that states that the CML is a traumatic injury commonly encountered in physically abused infants.

The yield of high-detail radiographic skeletal surveys in suspected infant abuse.

BACKGROUND: Skeletal surveys are routinely performed in cases of suspected child abuse, but there are limited data regarding the yield of high-detail skeletal surveys in infants. OBJECTIVE: To determine the diagnostic yield of high-detail radiographic skeletal surveys in suspected infant abuse. MATERIAL AND METHODS: We reviewed the high-detail American College of Radiology standardized skeletal surveys performed for suspected abuse in 567 infants (median: 4.4 months, SD 3.47; range: 4 days-12 months) at a large urban children's hospital between 2005 and 2013. Skeletal survey images, radiology reports and medical records were reviewed. A skeletal survey was considered positive when it showed at least one unsuspected fracture. RESULTS: In 313 of 567 infants (55%), 1,029 definite fractures were found. Twenty-one percent (119/567) of the patients had a positive skeletal survey with a total of 789 (77%) unsuspected fractures. Long-bone fractures were the most common injuries, present in 145 children (26%). The skull was the site of fracture in 138 infants (24%); rib cage in 77 (14%), clavicle in 24 (4.2%) and uncommon fractures (including spine, scapula, hands and feet and pelvis) were noted in 26 infants (4.6%). Of the 425 infants with neuroimaging, 154 (36%) had intracranial injury. No significant correlation between positive skeletal survey and associated intracranial injury was found. Scapular fractures and complex skull fractures showed a statistically significant correlation with intracranial injury (P = 0.029, P = 0.007, respectively). CONCLUSION: Previously unsuspected fractures are noted on skeletal surveys in 20% of cases of suspected infant abuse. These data may be helpful in the design and optimization of global skeletal imaging in this vulnerable population.

Pediatric CT dose reduction for suspected appendicitis: a practice quality improvement project using artificial Gaussian noise--part 1, computer simulations.

OBJECTIVE: The purpose of this study was to develop a departmental practice quality improvement project to systematically reduce CT doses for the evaluation of suspected pediatric appendicitis by introducing computer-generated gaussian noise. MATERIALS AND METHODS: Two hundred MDCT abdominopelvic examinations of patients younger than 20 years performed with girth-based scanning parameters for suspected appendicitis were reviewed. Two judges selected 45 examinations in which the diagnosis of appendicitis was excluded (14, appendix not visualized; 31, normal appendix visualized). Gaussian noise was introduced into axial image series, creating five additional series acquired at 25-76% of the original dose. Two readers reviewed 270 image series for appendix visualization (4-point Likert scale and arrow localization). Volume CT dose index (CTDIvol) and size-specific dose estimate (SSDE) were calculated by use of patient girth. Confidence ratings and localization accuracy were analyzed with mixed models and nonparametric bootstrap analysis at a 0.05 significance level. RESULTS: The mean baseline SSDE for the 45 patients was 16 mGy (95% CI, 12-20 mGy), and the corresponding CTDIvol was 10 mGy (95% CI, 4-16 mGy). Changes in correct appendix localization frequencies were minor. There was no substantial trend with decreasing simulated dose level (p = 0.46). Confidence ratings decreased with increasing dose reduction (p = 0.007). The average decreases were -0.27 for the 25% simulated dose (p = 0.01), -0.17 for 33% (p = 0.03), and -0.03 for 43% (p = 0.65). CONCLUSION: Pediatric abdominal MDCT can be performed with 43% of the original dose (SSDE, 7 mGy; CTDIvol, 4.3 mGy) without substantially affecting visualization of a normal appendix.

Imaging of skeletal injuries associated with abusive head trauma.

Skeletal injuries are commonly encountered in infants and young children with abusive head trauma. Although certain patterns of intracranial injury suggest abuse, none are diagnostic. Therefore demonstration of associated unsuspected skeletal injuries has important implications, particularly when highly specific fractures are present. Skull fractures are commonly associated with intracranial injury, but no fracture pattern is indicative of physical abuse. Other skeletal injuries including classic metaphyseal lesions and rib, spine and scapular fractures are strong predictors of abusive head trauma in infants with intracranial injury. It is mandatory to perform rigorous skeletal surveys in infants and young children with clinical and neuroimaging findings concerning for abusive head trauma.

Retroclival collections associated with abusive head trauma in children.

Retroclival collections are rare lesions reported almost exclusively in children and strongly associated with trauma. We examine the incidence and imaging characteristics of retroclival collections in young children with abusive head trauma. We conducted a database search to identify children with abusive head trauma ≤ 3 years of age with brain imaging performed between 2007 and 2013. Clinical data and brain images of 65 children were analyzed. Retroclival collections were identified in 21 of 65 (32%) children. Ten (48%) were subdural, 3 (14%) epidural, 2 (10%) both, and 6 (28%) indeterminate. Only 8 of 21 retroclival collections were identifiable on CT and most were low or intermediate in attenuation. Eighteen of 21 retroclival collections were identifiable on MRI: 3 followed cerebral spinal fluid in signal intensity and 15 were bloody/proteinaceous. Additionally, 2 retroclival collections demonstrated a fluid-fluid level and 2 enhanced in the 5 children who received contrast material. Sagittal T1-weighted images, sagittal fluid-sensitive sequences, and axial FLAIR (fluid-attenuated inversion recovery) images showed the retroclival collections best. Retroclival collections were significantly correlated with supratentorial and posterior fossa subdural hematomas and were not statistically correlated with skull fracture or parenchymal brain injury. Retroclival collections, previously considered rare lesions strongly associated with accidental injury, were commonly identified in this cohort of children with abusive head trauma, suggesting that retroclival collections are an important component of the imaging spectrum in abusive head trauma. Retroclival collections were better demonstrated on MRI than CT, were commonly identified in conjunction with intracranial subdural hematomas, and were not significantly correlated with the severity of brain injury or with skull fractures.