Development and Validation of a Deep Learning Method to Predict Cerebral Palsy From Spontaneous Movements in Infants at High Risk.

Importance: Early identification of cerebral palsy (CP) is important for early intervention, yet expert-based assessments do not permit widespread use, and conventional machine learning alternatives lack validity. Objective: To develop and assess the external validity of a novel deep learning-based method to predict CP based on videos of infants' spontaneous movements at 9 to 18 weeks' corrected age. Design, Setting, and Participants: This prognostic study of a deep learning-based method to predict CP at a corrected age of 12 to 89 months involved 557 infants with a high risk of perinatal brain injury who were enrolled in previous studies conducted at 13 hospitals in Belgium, India, Norway, and the US between September 10, 2001, and October 25, 2018. Analysis was performed between February 11, 2020, and September 23, 2021. Included infants had available video recorded during the fidgety movement period from 9 to 18 weeks' corrected age, available classifications of fidgety movements ascertained by the general movement assessment (GMA) tool, and available data on CP status at 12 months' corrected age or older. A total of 418 infants (75.0%) were randomly assigned to the model development (training and internal validation) sample, and 139 (25.0%) were randomly assigned to the external validation sample (1 test set). Exposure: Video recording of spontaneous movements. Main Outcomes and Measures: The primary outcome was prediction of CP. Deep learning-based prediction of CP was performed automatically from a single video. Secondary outcomes included prediction of associated functional level and CP subtype. Sensitivity, specificity, positive and negative predictive values, and accuracy were assessed. Results: Among 557 infants (310 [55.7%] male), the median (IQR) corrected age was 12 (11-13) weeks at assessment, and 84 infants (15.1%) were diagnosed with CP at a mean (SD) age of 3.4 (1.7) years. Data on race and ethnicity were not reported because previous studies (from which the infant samples were derived) used different study protocols with inconsistent collection of these data. On external validation, the deep learning-based CP prediction method had sensitivity of 71.4% (95% CI, 47.8%-88.7%), specificity of 94.1% (95% CI, 88.2%-97.6%), positive predictive value of 68.2% (95% CI, 45.1%-86.1%), and negative predictive value of 94.9% (95% CI, 89.2%-98.1%). In comparison, the GMA tool had sensitivity of 70.0% (95% CI, 45.7%-88.1%), specificity of 88.7% (95% CI, 81.5%-93.8%), positive predictive value of 51.9% (95% CI, 32.0%-71.3%), and negative predictive value of 94.4% (95% CI, 88.3%-97.9%). The deep learning method achieved higher accuracy than the conventional machine learning method (90.6% [95% CI, 84.5%-94.9%] vs 72.7% [95% CI, 64.5%-79.9%]; P < .001), but no significant improvement in accuracy was observed compared with the GMA tool (85.9%; 95% CI, 78.9%-91.3%; P = .11). The deep learning prediction model had higher sensitivity among infants with nonambulatory CP (100%; 95% CI, 63.1%-100%) vs ambulatory CP (58.3%; 95% CI, 27.7%-84.8%; P = .02) and spastic bilateral CP (92.3%; 95% CI, 64.0%-99.8%) vs spastic unilateral CP (42.9%; 95% CI, 9.9%-81.6%; P < .001). Conclusions and Relevance: In this prognostic study, a deep learning-based method for predicting CP at 9 to 18 weeks' corrected age had predictive accuracy on external validation, which suggests possible avenues for using deep learning-based software to provide objective early detection of CP in clinical settings.

Uveitis in Juvenile Idiopathic Arthritis: 18-Year Outcome in the Population-based Nordic Cohort Study.

PURPOSE: To assess the long-term outcome of uveitis in juvenile idiopathic arthritis (JIA). DESIGN: Population-based, multicenter, prospective JIA cohort, with a cross-sectional assessment of JIA-associated uveitis (JIA-U) 18 years after the onset of JIA. PARTICIPANTS: A total of 434 patients with JIA, of whom 96 had uveitis, from defined geographic areas of Denmark, Finland, Norway, and Sweden. METHODS: Patients with onset of JIA between January 1997 and June 2000 were prospectively followed for 18 years. Pediatric rheumatologists and ophthalmologists collected clinical and laboratory data. MAIN OUTCOME MEASURES: Cumulative incidence of uveitis and clinical characteristics, JIA and uveitis disease activity, ocular complications, visual outcome, and risk factors associated with the development of uveitis-related complications. RESULTS: Uveitis developed in 96 (22.1%) of 434 patients with JIA. In 12 patients (2.8%), uveitis was diagnosed between 8 and 18 years of follow-up. Systemic immunosuppressive medication was more common among patients with uveitis (47/96 [49.0%]) compared with patients without uveitis (78/338 [23.1%]). Active uveitis was present in 19 of 78 patients (24.4%) at the 18-year visit. Ocular complications occurred in 31 of 80 patients (38.8%). Short duration between the onset of JIA and the diagnosis of uveitis was a risk factor for developing ocular complications (odds ratio [OR], 1.4; 95% confidence interval [CI], 1.1-1.8). Patients with a diagnosis of uveitis before the onset of JIA all developed cataract and had an OR for development of glaucoma of 31.5 (95% CI, 3.6-274). Presence of antinuclear antibodies (ANAs) was also a risk factor for developing 1 or more ocular complications (OR, 3.0; 95% CI, 1.2-7.7). Decreased visual acuity (VA) <6/12 was found in 12 of 135 eyes (8.9%) with uveitis, and 4 of 80 patients (5.0%) with JIA-U had binocular decreased VA <6/12. CONCLUSIONS: Our results suggest that uveitis screening should start immediately when the diagnosis of JIA is suspected or confirmed and be continued for more than 8 years after the diagnosis of JIA. Timely systemic immunosuppressive treatment in patients with a high risk of developing ocular complications must be considered early in the disease course to gain rapid control of ocular inflammation.

Salivary Oral Microbiome of Children With Juvenile Idiopathic Arthritis: A Norwegian Cross-Sectional Study.

Background: The oral microbiota has been connected to the pathogenesis of rheumatoid arthritis through activation of mucosal immunity. The objective of this study was to characterize the salivary oral microbiome associated with juvenile idiopathic arthritis (JIA), and correlate it with the disease activity including gingival inflammation. Methods: Fifty-nine patients with JIA (mean age, 12.6 ± 2.7 years) and 34 healthy controls (HC; mean age 12.3 ± 3.0 years) were consecutively recruited in this Norwegian cross-sectional study. Information about demographics, disease activity, medication history, frequency of tooth brushing and a modified version of the gingival bleeding index (GBI) and the simplified oral hygiene index (OHI-S) was obtained. Microbiome profiling of saliva samples was performed by sequencing of the V1-V3 region of the 16S rRNA gene, coupled with a species-level taxonomy assignment algorithm; QIIME, LEfSe and R-package for Spearman correlation matrix were used for downstream analysis. Results: There were no significant differences between JIA and HC in alpha- and beta-diversity. However, differential abundance analysis revealed several taxa to be associated with JIA: TM7-G1, Solobacterium and Mogibacterium at the genus level; and Leptotrichia oral taxon 417, TM7-G1 oral taxon 352 and Capnocytophaga oral taxon 864 among others, at the species level. Haemophilus species, Leptotrichia oral taxon 223, and Bacillus subtilis, were associated with healthy controls. Gemella morbillorum, Leptotrichia sp. oral taxon 498 and Alloprevotella oral taxon 914 correlated positively with the composite juvenile arthritis 10-joint disease activity score (JADAS10), while Campylobacter oral taxon 44 among others, correlated with the number of active joints. Of all microbial markers identified, only Bacillus subtilis and Campylobacter oral taxon 44 maintained false discovery rate (FDR) < 0.1. Conclusions: In this exploratory study of salivary oral microbiome we found similar alpha- and beta-diversity among children with JIA and healthy. Several taxa associated with chronic inflammation were found to be associated with JIA and disease activity, which warrants further investigation.

Efficacy and safety of intraarticular corticosteroid injections in adolescents with juvenile idiopathic arthritis in the temporomandibular joint: a Norwegian 2-year prospective multicenter pilot study.

BACKGROUND: Intraarticular corticosteroids (IACs) have been used to treat temporomandibular joint (TMJ) arthritis. However, prospective clinical studies with magnetic resonance imaging (MRI) scoring are lacking. The aim of this study was to examine efficacy and safety of a single IAC in the TMJ in adolescents with juvenile idiopathic arthritis (JIA) in a clinical setting. METHODS: In this Norwegian prospective multicenter pilot study 15 patients with JIA (mostly persistent oligoarthritis or RF negative polyarthritis categories) and a clinically and MRI-verified diagnosis of TMJ arthritis were treated with IACs and followed for 2 years. Demographics, systemic medication, general disease activity and outcome measures were recorded including a pain-index score and maximal incisal opening (MIO). Inflammation and bone damage scores were assessed, using two recently published MRI scoring systems with masked radiological evaluation. RESULTS: Among the 15 patients, 13 received a single IAC (5 bilateral), and 2 repeated IACs once unilaterally. Thus, the total number of IACs was 22. Median age was 15 years and the majority had an age not thought of as critical regarding mandibular growth retardation due to steroid injection. During the 2-year observation period systemic medication with disease modifying antirheumatic drugs (DMARDs) including biologics was initiated or adjusted in 10/15 (67%) patients. At the 2-months study visit after injection we observed a minimal improvement in MIO from median 44 (1st, 3rd quartiles; 36, 48) mm to 45 (43, 47) mm, p = 0.045 and decreased MRI mean additive inflammatory score from 4.4 ± 1.8 standard deviations (SD) to 3.4 ± 2.0, p = 0.040. From baseline to the 2-months follow-up pain improved in 6/11 patients but pain scores were not significantly improved. MRI-assessed damage increased in two patients with repeated IACs, and decreased in 3 patients but most of the patients were stable over the 2-year follow-up. Intra-rater repeatability of the MRI scoring system domains varied from poor to excellent. CONCLUSIONS: In this pilot study of predominately single IACs to the TMJ in combination with systemic treatment we observed improvement in MRI-assessed inflammation, mostly stable condylar bone conditions and minimal clinical improvement in adolescents with JIA and TMJ arthritis. No severe side effects were seen.

The Predictive Accuracy of the General Movement Assessment for Cerebral Palsy: A Prospective, Observational Study of High-Risk Infants in a Clinical Follow-Up Setting.

BACKGROUND: Early prediction of cerebral palsy (CP) using the General Movement Assessment (GMA) during the fidgety movements (FM) period has been recommended as standard of care in high-risk infants. The aim of this study was to determine the accuracy of GMA, alone or in combination with neonatal imaging, in predicting cerebral palsy (CP). METHODS: Infants with increased risk of perinatal brain injury were prospectively enrolled from 2009-2014 in this multi-center, observational study. FM were classified by two certified GMA observers blinded to the clinical history. Abnormal GMA was defined as absent or sporadic FM. CP-status was determined by clinicians unaware of GMA results. RESULTS: Of 450 infants enrolled, 405 had scorable video and follow-up data until at least 18-24 months. CP was confirmed in 42 (10.4%) children at mean age 3 years 1 month. Sensitivity, specificity, positive and negative predictive values, and accuracy of absent/sporadic FM for CP were 76.2, 82.4, 33.3, 96.8, and 81.7%, respectively. Only three (8.1%) of 37 infants with sporadic FM developed CP. The highest accuracy (95.3%) was achieved by a combination of absent FM and abnormal neonatal imaging. CONCLUSION: In infants with a broad range of neonatal risk factors, accuracy of early CP prediction was lower for GMA than previously reported but increased when combined with neonatal imaging. Sporadic FM did not predict CP in this study.

Participation in school and physical education in juvenile idiopathic arthritis in a Nordic long-term cohort study.

BACKGROUND: The aim of the study was to describe school attendance and participation in physical education in school among children with juvenile idiopathic arthritis (JIA). METHODS: Consecutive cases of JIA from defined geographical areas of Finland, Sweden and Norway with disease onset in 1997 to 2000 were followed for 8 years in a multi-center cohort study, aimed to be as close to population-based as possible. Clinical characteristics and information on school attendance and participation in physical education (PE) were registered. RESULTS: Participation in school and in PE was lowest initially and increased during the disease course. Eight years after disease onset 228/274 (83.2%) of the children reported no school absence due to JIA, while 16.8% reported absence during the last 2 months due to JIA. Full participation in PE was reported by 194/242 (80.2%), partly by 16.9%, and none by 2.9%. Lowest participation in PE was found among children with ERA and the undifferentiated categories. Absence in school and PE was associated with higher disease activity measures at the 8-year visit. School absence > 1 day at baseline predicted use of disease-modifying anti-rheumatic drugs, including biologics (DMARDs) (OR 1.2 (1.1-1.5)), and non-remission off medication (OR 1.4 (1.1-1.7) 8 years after disease onset. CONCLUSION: School absence at baseline predicted adverse long-term outcome. In children and adolescents with JIA participation in school activities is mostly high after 8 years of disease. For the minority with low participation, special attention is warranted to promote their full potential of social interaction and improve long-term outcome.

The Norwegian version of the Juvenile Arthritis Multidimensional Assessment Report (JAMAR).

The Juvenile Arthritis Multidimensional Assessment Report (JAMAR) is a new parent/patient-reported outcome measure that enables a thorough assessment of the disease status in children with juvenile idiopathic arthritis (JIA). We report the results of the cross-cultural adaptation and validation of the parent and patient versions of the JAMAR in the Norwegian language. The reading comprehension of the questionnaire was tested in 10 JIA parents and patients. Each participating centre was asked to collect demographic and clinical data and the JAMAR in 100 consecutive JIA patients or all consecutive patients seen in a 6-month period and to administer the JAMAR to 100 healthy children and their parents. The statistical validation phase explored descriptive statistics and the psychometric issues of the JAMAR: the 3 Likert assumptions, floor/ceiling effects, internal consistency, Cronbach's alpha, interscale correlations, test-retest reliability, and construct validity (convergent and discriminant validity). A total of 301 JIA patients (3.3% systemic, 41.2% oligoarticular, 25.9% RF negative polyarthritis, and 29.6% other categories) and 74 healthy children were enrolled in three centres. The JAMAR components discriminated well healthy subjects from JIA patients. All JAMAR components revealed good psychometric performances. In conclusion, the Norwegian version of the JAMAR is a valid tool for the assessment of children with JIA and is suitable for use both in routine clinical practice and clinical research.

Computer-based video analysis identifies infants with absence of fidgety movements.

BackgroundAbsence of fidgety movements (FMs) at 3 months' corrected age is a strong predictor of cerebral palsy (CP) in high-risk infants. This study evaluates the association between computer-based video analysis and the temporal organization of FMs assessed with the General Movement Assessment (GMA).MethodsInfants were eligible for this prospective cohort study if referred to a high-risk follow-up program in a participating hospital. Video recordings taken at 10-15 weeks post term age were used for GMA and computer-based analysis. The variation of the spatial center of motion, derived from differences between subsequent video frames, was used for quantitative analysis.ResultsOf 241 recordings from 150 infants, 48 (24.1%) were classified with absence of FMs or sporadic FMs using the GMA. The variation of the spatial center of motion (CSD) during a recording was significantly lower in infants with normal (0.320; 95% confidence interval (CI) 0.309, 0.330) vs. absence of or sporadic (0.380; 95% CI 0.361, 0.398) FMs (P<0.001). A triage model with CSD thresholds chosen for sensitivity of 90% and specificity of 80% gave a 40% referral rate for GMA.ConclusionQuantitative video analysis during the FMs' period can be used to triage infants at high risk of CP to early intervention or observational GMA.

High prevalence of abnormal motor repertoire at 3 months corrected age in extremely preterm infants.

AIMS: To compare early motor repertoire between extremely preterm and term-born infants. An association between the motor repertoire and gestational age and birth weight was explored in extremely preterm infants without severe ultrasound abnormalities. METHODS: In a multicentre study, the early motor repertoire of 82 infants born extremely preterm (ELGAN:<28 weeks) and/or with extremely low birth weight (ELBW:<1000 g) and 87 term-born infants were assessed by the "Assessment of Motor Repertoire - 2 to 5 Months" (AMR) which is part of Prechtl's "General Movement Assessment", at 12 weeks post-term age. Fidgety movements were classified as normal if present and abnormal if absent, sporadic or exaggerated. Concurrent motor repertoire was classified as normal if smooth and fluent and abnormal if monotonous, stiff, jerky and/or predominantly fast or slow. RESULTS: Eight-teen ELBW/ELGAN infants had abnormal fidgety movements (8 absent, 7 sporadic and 3 exaggerated fidgety movements) compared with 2 control infants (OR:12.0; 95%CI:2.7-53.4) and 46 ELBW/ELGAN infants had abnormal concurrent motor repertoire compared with 17 control infants (OR:5.3; 95%CI:2.6-10.5). Almost all detailed aspects of the AMR differed between the groups. Results were the same when three infants with severe ultrasound abnormalities were excluded. In the remaining ELBW/ELGAN infants, there was no association between motor repertoire and gestational age or birth weight. CONCLUSION: ELBW/ELGAN infants had poorer quality of early motor repertoire than term-born infants.The findings were not explained by severe abnormalities on neonatal ultrasound scans and were not correlated to the degree of prematurity. The consequences of these abnormal movement patterns remain to be seen in future follow-up studies.

Effects of High Intensity Interval Training on Pregnant Rats, and the Placenta, Heart and Liver of Their Fetuses.

OBJECTIVE: To investigate the effects of high intensity interval training (HIIT) on the maternal heart, fetuses and placentas of pregnant rats. METHODS: Female Sprague-Dawley rats were randomly assigned to HIIT or sedentary control groups. The HIIT group was trained for 6 weeks with 10 bouts of high intensity uphill running on a treadmill for four minutes (at 85-90% of maximal oxygen consumption) for five days/week. After three weeks of HIIT, rats were mated. After six weeks (gestational day 20 in pregnant rats), echocardiography was performed to evaluate maternal cardiac function. Real-time PCR was performed for the quantification of gene expression, and oxidative stress and total antioxidant capacity was assessed in the tissue samples. RESULTS: Maternal heart weight and systolic function were not affected by HIIT or pregnancy. In the maternal heart, expression of 11 of 22 genes related to cardiac remodeling was influenced by pregnancy but none by HIIT. Litter size, fetal weight and placental weight were not affected by HIIT. Total antioxidant capacity, malondialdehyde content, peroxidase and superoxide dismutase activity measured in the placenta, fetal heart and liver were not influenced by HIIT. HIIT reduced the expression of eNOS (p = 0.03), hypoxia-inducible factor 1α (p = 0.04) and glutathione peroxidase 4.2 (p = 0.02) in the fetal liver and increased the expression of vascular endothelial growth factor-ß (p = 0.014), superoxide dismutase 1 (p = 0.001) and tissue inhibitor of metallopeptidase 3 (p = 0.049) in the fetal heart. CONCLUSIONS: Maternal cardiac function and gene expression was not affected by HIIT. Although HIIT did not affect fetal growth, level of oxidative stress and total antioxidant capacity in the fetal tissues, some genes related to oxidative stress were altered in the fetal heart and liver indicating that protective mechanisms may be activated.

Coronary flow reserve in pregnant rats with increased left ventricular afterload.

BACKGROUND: Coronary flow reserve (CFR) is used as a measure of coronary endothelial function. We investigated the effect of increased afterload on CFR of pregnant and non-pregnant rats. METHODS: Afterload increase in Wister rats (both pregnant and non-pregnant) was achieved by the infusion of angiotensin II (Ang II) for ∼10 days or by subjecting them to transverse aortic constriction (TAC) for ∼14 days. Control groups were infused with 0.9% NaCl or had sham surgery, respectively. In pregnant rats, the experiments were performed close to term gestation. Doppler velocity waveforms of the left main coronary artery were recorded using a high resolution ultrasound imaging system (Vevo 770, VisualSonics, Canada) at baseline while the animals were anesthetized with 1.5% inhaled isoflurane, and during maximal coronary dilatation obtained by the inhalation of 3.5% of isoflurane. CFR was calculated as the ratio between the peak coronary flow velocities (CFRpeak) and the velocity-time integrals (CFRVTI) recorded at hyperemia and at baseline. RESULTS: CFR could be calculated in 60 of 75 (80%) animals. There were no differences in CFR between intervention and control groups irrespective of whether afterload was increased by Ang II or TAC. In the TAC-study CFRpeak (1.54±0.07 vs 1.85±0.17; p = 0.03) was decreased in pregnant compared to non-pregnant shams. When sham animals from both studies were pooled together both CFRpeak (1.42±0.07 vs 1.86±0.16; p = 0.005) as well as CFRVTI (1.45±0.07 vs 1.78±0.12; p = 0.03) were significantly lower in pregnant rats compared to non-pregnant. CONCLUSIONS: CFR can be measured non-invasively in rats using Doppler echocardiography and high concentrations of inhaled isoflurane as a coronary vasodilator. In pregnant rats, CFR is reduced close to term. CFR is not affected by increased left ventricular afterload caused by chronic Ang II infusion or TAC.

Effect of transverse aortic constriction on cardiac structure, function and gene expression in pregnant rats.

BACKGROUND: There is an increased risk of heart failure and pulmonary edema in pregnancies complicated by hypertensive disorders. However, in a previous study we found that pregnancy protects against fibrosis and preserves angiogenesis in a rat model of angiotensin II induced cardiac hypertrophy. In this study we test the hypothesis that pregnancy protects against negative effects of increased afterload. METHODS: Pregnant (gestational day 5.5-8.5) and non-pregnant Wistar rats were randomized to transverse aortic constriction (TAC) or sham surgery. After 14.2 ± 0.14 days echocardiography was performed. Aortic blood pressure and left ventricular (LV) pressure-volume loops were obtained using a conductance catheter. LV collagen content and cardiomyocyte circumference were measured. Myocardial gene expression was assessed by real-time polymerase chain reaction. RESULTS: Heart weight was increased by TAC (p<0.001) but not by pregnancy. Cardiac myocyte circumference was larger in pregnant compared to non-pregnant rats independent of TAC (p = 0.01), however TAC per se did not affect this parameter. Collagen content in LV myocardium was not affected by pregnancy or TAC. TAC increased stroke work more in pregnant rats (34.1 ± 2.4 vs 17.5 ± 2.4 mmHg/mL, p<0.001) than in non-pregnant (28.2 ± 1.7 vs 20.9 ± 1.5 mmHg/mL, p = 0.06). However, it did not lead to overt heart failure in any group. In pregnant rats, α-MHC gene expression was reduced by TAC. Increased in the expression of ß-MHC gene was higher in pregnant (5-fold) compared to non-pregnant rats (2-fold) after TAC (p = 0.001). Nine out of the 19 genes related to cardiac remodeling were affected by pregnancy independent of TAC. CONCLUSIONS: This study did not support the hypothesis that pregnancy is cardioprotective against the negative effects of increased afterload. Some differences in cardiac structure, function and gene expression between pregnant and non-pregnant rats following TAC indicated that afterload increase is less tolerated in pregnancy.