Search

BACKGROUND: Deep learning has been employed using brain functional connectome data for evaluating the risk of cognitive deficits in very preterm infants. Although promising, training these deep learning models typically requires a large amount of labeled data, and labeled medical data are often very difficult and expensive to obtain. OBJECTIVE: This study aimed to develop a self-training deep neural network (DNN) model for early prediction of cognitive deficits at 2 years of corrected age in very preterm infants (gestational age ≤32 weeks) using both labeled and unlabeled brain functional connectome data. MATERIALS AND METHODS: We collected brain functional connectome data from 343 very preterm infants at a mean (standard deviation) postmenstrual age of 42.7 (2.5) weeks, among whom 103 children had a cognitive assessment at 2 years (i.e. labeled data), and the remaining 240 children had not received 2-year assessments at the time this study was conducted (i.e. unlabeled data). To develop a self-training DNN model, we built an initial student model using labeled brain functional connectome data. Then, we applied the trained model as a teacher model to generate pseudo-labels for unlabeled brain functional connectome data. Next, we combined labeled and pseudo-labeled data to train a new student model. We iterated this procedure to obtain the best student model for the early prediction task in very preterm infants. RESULTS: In our cross-validation experiments, the proposed self-training DNN model achieved an accuracy of 71.0%, a specificity of 71.5%, a sensitivity of 70.4% and an area under the curve of 0.75, significantly outperforming transfer learning models through pre-training approaches. CONCLUSION: We report the first self-training prognostic study in very preterm infants, efficiently utilizing a small amount of labeled data with a larger share of unlabeled data to aid the model training. The proposed technique is expected to facilitate deep learning with insufficient training data.

Contrast-enhanced ultrasound (CEUS) is increasingly performed for focal liver lesion evaluation because of its high safety profile and ability to yield a definitive diagnosis in many patients, obviating multiphase CT or MRI. CEUS uses specific ultrasound contrast agents (UCAs) that are exceedingly safe, without the risk of renal or liver toxicity. UCAs are cleared rapidly from the body, allowing for multiple injections during a single examination, with no potential for deposition in the patient. This review highlights the performance of CEUS for liver lesion evaluation and illustrates the imaging appearance of common liver lesions in children and adults.

Background Intestinal ultrasound (IUS) is a noninvasive tool to assess bowel inflammation. There is a paucity of data on its accuracy in pediatric patients. Aim The aim of this study is to evaluate the diagnostic performance of bowel wall thickness (BWT) measured using IUS compared with endoscopic disease activity in children suspected of having inflammatory bowel disease (IBD). Methods We conducted a single-center cross-sectional pilot study of pediatric patients suspected to have previously undiagnosed IBD. Endoscopic inflammation was graded using segmental scores of the Simple Endoscopic Score for Crohn's Disease (SES-CD) and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) and classified as having healthy, mild, or moderate/severe disease activity. Association between BWT and endoscopic severity was assessed using the Kruskal-Wallis test. The diagnostic performance of BWT to detect active disease at endoscopy was evaluated using the area under the receiver operating characteristic curve; sensitivity and specificity were calculated. Results In all, 174 bowel segments in 33 children were assessed by IUS and ileocolonoscopy. An elevated median BWT was associated with increased bowel segment disease severity, classified by the SES-CD (P Conclusion Increasing BWT is associated with increasing endoscopic activity in pediatric IBD. Our study suggests that the optimal BWT cutoff value for detecting active disease may be less than that seen in adults. Additional pediatric studies are needed.

Elevated central venous pressure in those with Fontan circulation causes liver congestion and hepatomegaly. We assessed if liver volume by magnetic resonance imaging (MRI) is associated with adverse cardiovascular outcomes. Retrospective study of 122 patients with Fontan circulation who were10 years old and had a liver MRI with magnetic resonance elastography. Liver volume (ml) was measured by manual segmentation from axial T2-weighted images and was indexed to body surface area. The composite outcome included death, heart transplant, ventricular assist device placement, or nonelective cardiovascular hospitalization. The median age at the time of MRI was 18.9 (interquartile range 15.8 to 25.9) years, and 47% of the patients were women. The mean indexed liver volume was 1,133 ± 180 ml/m

Liver shear stiffness measurement using magnetic resonance elastography (MRE) aids in the noninvasive diagnosis and staging of liver fibrosis. Inadequate breath-holds can lead to inaccurate stiffness estimation and/or failed MRE exams.To prospectively evaluate the performance of compressed sensitivity encoding (C-SENSE) accelerated rapid MRE measurement of liver shear stiffness using displacement wave polarity-inversion motion encoding.Retrospective.Eleven with liver disease and 10 asymptomatic subjects.1.5 T; gradient-recalled-echo (GRE) MRE.All participants underwent: 1) two-dimensional (2D) GRE MRE with inflow saturation using SENSE acceleration factor (R) of 2 (standard of care [SC]); 2) 2D rapid MRE with (RwS); and 3) without (RnS) inflow saturation using C-SENSE R = 3; and 4) spatial three-dimensional (3D) rapid MRE with inflow saturation (R3D) using C-SENSE R = 4; with nominally identical spatial resolution and coverage. Image analyst (D.G., 2 years of experience) drew identical and maximal regions of interest (ROIs) in right hepatic lobe.Linear regression, intra-class correlation coefficients (ICC), Bland-Altman analyses, and the Wilcoxon signed-rank test were used to assess consistency and agreement of liver stiffness measurements for manually drawn identical and maximal ROIs.In 21 participants (37 ± 14 years) with liver stiffness (2.3 ± 0.7 kPa), body mass index (BMI 27 ± 7 kg/mC-SENSE accelerated GRE MRE sequences, using displacement wave polarity-inversion motion encoding, produce equivalent measurements of liver stiffness and have potential clinical benefit in patients with limited breath-holding capacity.1 TECHNICAL EFFICACY: Stage 1.

Introduction: After Fontan operation, decreased venous capacitance and venoconstriction are adaptive mechanisms to maintain venous return and cardiac output. The consequent higher venous pressure may adversely impact end-organ function, exercise capacity and result in worse clinical outcomes. This pilot study evaluated the safety and effect of isosorbide dinitrate (ISDN), a venodilator, on exercise capacity, peripheral venous pressure (PVP), and liver stiffness in patients with Fontan circulation. Methods: In this prospective single-arm trial, 15 individuals with Fontan circulation were evaluated at baseline and after four weeks of therapeutic treatment with ISDN. Primary aims were to assess the safety of ISDN and the effect on maximal exercise (peak oxygen consumption, peak VO2). We also aimed to assess the effect of ISDN on ultrasound-assessed liver stiffness, markers of submaximal exercise, and PVP at rest and peak exercise. Repeated measures t-tests were used to assess change in variables of interest in response to ISDN. Results: Mean age was 23.5 ± 9.2 years (range 11.2-39.0 years), and 10/15 (67%) were male. There was no statistically significant change in peak VO2 (1401±428 to 1428±436 mL/min, p=0.128), but VO2 at the anaerobic threshold increased (1087 + 313 to 1115 + 302 mL/min, p= 0.03). ISDN was also associated with a lower peak exercise PVP (22.5 + 4.5 to 20.6 + 3.0 mmHg, p=0.015). Liver stiffness was lower with ISDN, though the difference was not statistically significant (2.3 + 0.4 to 2.1 + 0.5 m/s, p=0.079). Of the patients completing the trial, mild headache was common (67%), but there were no major adverse events. Conclusions: Treatment with ISDN for 4 weeks is well-tolerated in patients with a Fontan circulation. ISDN is associated with an increase in VO2 at anaerobic threshold, lower peak PVP, and a trend towards lower liver stiffness. Larger, longer studies will be necessary to define the impact of ISDN on clinical outcomes in the Fontan circulation. Clinical Trial Registration: URL: https//clinicaltrials.gov. Unique identifier: NCT04297241

Purpose Define relationships between quantitative magnetic resonance imaging (MRI) metrics and clinical/laboratory data in a pediatric and young adult cohort with autoimmune liver disease (AILD). Materials and methods This prospective, cross-sectional study was institutional review board-approved. Patients enrolled in an institutional AILD registry were divided into groups: (1) autoimmune hepatitis (AIH) or (2) primary sclerosing cholangitis (PSC)/autoimmune sclerosing cholangitis (ASC). Participants underwent serum liver biochemistry testing and research MRI examinations, including 3D magnetic resonance cholangiopancreatography (MRCP), magnetic resonance elastography (MRE), and iron-corrected T1 mapping (cT1). MRCP + and LiverMultiScan (Perspectum Ltd., Oxford, UK) were used to post-process 3D MRCP and cT1 data. Multiple linear regression models were used to assess relationships. Results 58 patients, 35 male, median age 16 years were included; 30 in the AIH group, 28 in the PSC/ASC group. After statistical adjustments for patient age, sex, presence of inflammatory bowel disease (IBD), specific diagnosis (PSC/ASC vs. AIH), and time from diagnosis to MRI examination, left hepatic bile duct maximum diameter was a statistically significant predictor of whole liver mean cT1, cT1 interquartile range (IQR), and MRE liver stiffness (p = 0.01–0.04). Seven laboratory values were significant predictors of whole liver cT1 IQR (p p Conclusions Bile duct diameters and multiple laboratory biomarkers of liver disease are independent predictors of liver stiffness and cT1 IQR in pediatric patients with AILD.

BACKGROUND. CT is the imaging modality of choice to identify lung metastasis. OBJECTIVE. The purpose of this study was to evaluate the performance of reduced-dose CT for the detection of lung nodules in children and young adults with cancer. METHODS. This prospective study enrolled patients 4-21 years old with known or suspected malignancy who were undergoing clinically indicated chest CT. Study participants underwent an additional investigational reduced-dose chest CT examination in the same imaging encounter. Separated deidentified CT examinations were reviewed in blinded fashion by three independent radiologists. One reviewer performed a subsequent secondary review to match nodules between the standard- and reduced-dose examinations. Diagnostic performance was computed for the reduced-dose examinations using the clinical examinations as the reference standard. Intraobserver agreement and interobserver agreement were calculated using Cohen kappa. RESULTS. A total of 78 patients (44 male patients and 34 female patients; mean age, 15.2 ± 3.8 [SD] years) were enrolled. The mean estimated effective dose was 1.8 ± 1.1 mSv for clinical CT and 0.3 ± 0.1 mSv for reduced-dose CT, which is an 83% dose reduction. Forty-five of the 78 (58%) patients had 162 total lung nodules (mean size, 3.4 ± 3.3 mm) detected on the clinical CT examinations. A total of 92% of nodules were visible on reduced-dose CT. The sensitivity and specificity of reduced-dose CT for nodules ranged from 63% to 77% and from 80% to 90%, respectively, across the three reviewers. Intraob-server agreement between clinical CT and reduced-dose CT was moderate to substantial for the presence of nodules (κ = 0.45-0.67) and was good to excellent for the number of nodules (κ = 0.68-0.84) and nodule size (κ = 0.69-0.86). Interobserver agreement for the presence of nodules was moderate for both reduced-dose (κ = 0.53) and clinical (κ = 0.54) CT. A median of one nodule was present on clinical CT in patients with a falsely negative reduced-dose CT examination. CONCLUSION. Reduced-dose CT depicts more than 90% of lung nodules in children and young adults with cancer. Reviewers identified the presence of nodules with moderate sensitivity and high specificity. CLINICAL IMPACT. CT performed at a 0.3-mSv mean effective dose has acceptable diagnostic performance for lung nodule detection in children and young adults and has the potential to reduce patient dose or expand CT utilization (e.g., to replace radiography in screening or monitoring protocols). TRIAL REGISTRATION. ClinicalTrials.gov NCT03681873.

Changes in liver magnetic resonance imaging T1 relaxation times are associated with histologic inflammation and fibrosis.To compare liver T1 measurements obtained using a novel single-breath-hold 3-dimensional (3D) whole-liver T1 estimation method (3D-QALAS) to standard-of-care 2-dimensional (2D) modified Look-Locker (2D-MOLLI) measurements.With institutional review board approval, research magnetic resonance imaging examinations were performed in 19 participants at 1.5 T. T1 relaxometry of the liver was performed using a novel 3D whole-liver T1 estimation method (3D-QALAS) as well as a 2D modified Look-Locker (2D-MOLLI) method. The 3D method covered the entire liver in a single breath hold, whereas 2D imaging was performed at 4 anatomic levels in 4 consecutive breath holds. T1 measurements from parametric maps were obtained by a single operator, and region-of-interest area-weighted mean T1 values were calculated. Pearson correlation (r) was used to assess correlation between T1 estimation methods, and the paired t test and Bland-Altman analysis were used to compare agreement in T1 measurements.In 18 participants (1 participant was excluded from analysis because of respiratory motion artifacts on 3D-QALAS images), 2D-MOLLI and 3D-QALAS mean T1 measurements were strongly correlated (r = 0.95, [95% CI: 0.87-0.98]; P0.0001). 2D-MOLLI T1 values were significantly longer than 3D-QALAS values (647.2 ± 87.3 milliseconds vs. 554.7 ± 75.8 milliseconds; P0.0001) with mean bias = 92.5 milliseconds (95% limits of agreement, 36.8, 148.2 milliseconds).Whole-liver T1 measurements obtained using a novel single-breath-hold 3D T1 estimation method correlate with a standard-of-care multiple consecutive-breath-hold 2D single-slice method but demonstrate systematic bias that should be considered or corrected when used in a clinical or research setting.

BACKGROUND: Patients with autoimmune hepatitis (AIH) and primary or autoimmune sclerosing cholangitis are at nutritional risk; their body composition and has not been extensively studied. We aimed to describe their body composition and identify clinical links. METHODS: Using magnetic resonance imaging (MRI), two reviewers segmented total psoas muscle area (tPMSA), visceral fat area (VFA) and subcutaneous fat area (mm(2)) and measured visceral and subcutaneous thickness (mm). Clinical, laboratory and quality of life (QoL; using PedsQL) data were collected. Sarcopenia was defined as tPMSA ≤5th percentile. Analysis of variance, Wilcoxon rank test and multivariable modelling were performed. A paediatric cohort with non-alcoholic fatty liver disease (NAFLD) was used as a comparator following propensity score matching. RESULTS: Fifty-eight patients with autoimmune liver disease (AILD) (33 [57%] with AIH) were included: median age 16 years (interquartile range [IQR]: 13-18), 33 (57%) male. Median time from diagnosis to MRI was 15 months (IQR: 2-39 months). Two patients (3%) had a BMIz indicative of mild malnutrition. tPMSA was measurable in 52 subjects (90%). Of those, 25 (48%) had sarcopenia. Sarcopenic patients had a lower blood urea nitrogen compared to non-sarcopenic (median [IQR]: 9.5 [8.0, 12.0] vs 11 [10, 14] mg/dL; P = .023). There was no difference in corticosteroid use between groups. The VFA of sarcopenic patients was higher (3156 [2064, 7492]) vs 2084 [688, 3092]) mm(2); P = .005). Patient-reported QoL negatively associated with VFA and general health negatively associated with VFA. Compared with NAFLD, the odds ratio for sarcopenia with AILD was 14.5 (95% confidence interval: 2.3-90.7). CONCLUSION: In autoimmune liver diseases, sarcopenia is highly prevalent, associated with increased visceral fat and QoL.

To compare quantitative biliary measurements obtained with three different magnetic resonance cholangiopancreatography (MRCP) acquisition methods. This retrospective study was IRB-approved. Patients with combinations of clinically indicated 3D FSE MRCP with sensitivity encoding (SENSE), 3D FSE SENSE MRCP with compressed sensing (CS-FSE; acceleration factor 8), and 3D gradient and spin-echo (GRASE) MRCP, acquired between October 2018 and March 2020, were included. The MRCP + Tuning Threshold algorithm (Perspectum Ltd., Oxford, UK) was used to segment 3D biliary models from MRCP data, with multiple metrics quantified from the models. Single measure, two-way, mixed-effects intra-class correlations, Bland–Altman analyses, and Wilcoxon signed-rank tests were used to compare quantitative measurements. From 160 MRCP datasets (25 3D FSE, 67 3D CS-FSE, 68 3D GRASE) in 69 patients, 48 datasets (7 [28%] 3D FSE, 14 [21%] 3D CS-FSE, 27 [40%] 3D GRASE) failed post-processing due to motion artifacts. The remaining 112 MRCP datasets (18 3D FSE, 53 3D CS-FSE, 41 3D GRASE) from 60 patients were included in the analysis. There was good to excellent agreement between 3D FSE and 3D CS-FSE MRCP for diameter of the left and right hepatic ducts, biliary volume, number and length of ducts, and total length of dilations (ICC: 0.83–0.93). The only metrics that exhibited good agreement between 3D FSE and 3D GRASE MRCP were biliary volume (ICC: 0.75) and total number of dilations (ICC: 0.77). 3D CS-FSE MRCP produces comparable biliary diameter metrics and global duct quantification to 3D FSE MRCP at a significantly reduced acquisition time.

Magnetic resonance elastography (MRE) of the liver has emerged as the non-invasive standard for the evaluation of liver fibrosis in chronic liver diseases (CLDs). The utility of MRE in the evaluation of different CLD in both adults and children has been demonstrated in several studies, and MRE has been recommended by several clinical societies. Consequently, the clinical indications for evaluation of CLD with MRE have increased, and MRE is currently used as an add-on test during routine liver MRI studies or as a standalone test. To meet the increasing clinical demand, MRE is being installed in many academic and private practice imaging centers. There is a need for a comprehensive practical guide to help these practices to deliver high-quality liver MRE studies as well as troubleshoot the common issues with MRE to ensure smooth running of the service. This comprehensive clinical practice review summarizes the indications and provides an overview on why to use MRE, technical requirements, system set-up, patient preparation, acquiring the data, and interpretation.

Deep learning Computed Tomography (CT) reconstruction (DLR) algorithms promise to improve image quality but the impact on clinical diagnostic performance remains to be demonstrated. We aimed to compare DLR to standard iterative reconstruction for detection of urolithiasis by unenhanced CT in children and young adults. This was an IRB approved retrospective study involving post-hoc reconstruction of clinically acquired unenhanced abdomen/pelvis CT scans. Images were reconstructed with six different manufacturer-standard DLR algorithms and reformatted in 3 planes (axial, sagittal, and coronal) at 3 mm intervals. De-identified reconstructions were loaded as independent examinations for review by 3 blinded radiologists (R1, R2, R3) tasked with identifying and measuring all stones. Results were compared to the clinical iterative reconstruction images as a reference standard. IntraClass correlation coefficients and kappa (k) statistics were used to quantify agreement. CT data for 14 patients (mean age: 17.3 ± 3.4 years, 5 males and 9 females, weight class: 31-70 kg (n = 6), 71-100 kg (n = 7), > 100 kg (n = 1)) were reconstructed into 84 total exams. 7 patients had urinary tract calculi. Interobserver agreement on the presence of any urinary tract calculus was substantial to almost perfect (k = 0.71–1) for all DLR algorithms. Agreement with the reference standard on number of calculi was excellent (ICC = 0.78–0.96) and agreement on the size of the largest calculus was fair to excellent (ICC = 0.51–0.97) depending on reviewer and DLR algorithm. Deep learning reconstruction of unenhanced CT images allows similar renal stone detectability compared to iterative reconstruction.

Masses and masslike lesions of the pancreas are uncommon in the pediatric population. However, owing to overlapping clinical and imaging features, it can be challenging to differentiate the various causes of pediatric pancreatic masses at initial patient presentation. Clinical data such as patient age, signs and symptoms at presentation, laboratory test results, and potential underlying cancer predisposition syndrome can be helpful when formulating a differential diagnosis. US may be the first imaging study to depict a pancreatic mass in a child, as this examination is frequently performed in children with nonspecific abdominal signs and symptoms because of its wide availability and relatively low cost and the lack of a need for sedation or anesthesia. CT or MRI is typically required for more thorough characterization of the mass and surgical planning. Complete characterization of pancreatic masses includes assessment of vascular involvement, local invasion, and extrapancreatic spread of tumor. The authors provide an up-to-date comprehensive review of the clinical manifestations, histopathologic features, and imaging findings of primary and secondary tumors of the pancreas in children and young adults. Advances in imaging, current prognostic information, and treatment paradigms also are highlighted. Finally, nontumorous masslike lesions of the pediatric pancreas, including vascular malformations, cystic disorders (eg, von Hippel-Lindau syndrome, cystic fibrosis), intrapancreatic accessory spleen, and autoimmune pancreatitis, are discussed.

Elevated magnetic resonance elastography (MRE)-derived liver stiffness may be associated with worse outcomes in people with Fontan circulation. We sought to evaluate the association between liver stiffness and Fontan failure or portal hypertension. Single center cross-sectional retrospective study of people with Fontan circulation who underwent MRE between 2011 and 2020. The cohort was divided into adult (age ≥ 21 years) and pediatric (

To determine if differences between individual sonographers and radiologists performing and interpreting appendix ultrasound affect clinically important outcomes in children with suspected acute appendicitis.Patients18 years of age who presented to our emergency department (ED) with suspected acute appendicitis and underwent an appendix ultrasound were identified. Sonographers who performed fewer than 100 examinations and radiologists who interpreted fewer than 100 examinations during the study period were excluded. Multivariable logistic regression was performed to assess the effect of sonographer, radiologist, clinical variables, and system factors on key clinical outcomes, including hospital admission and appendectomy.In all, 9,283 appendix ultrasounds (mean age, 9.9 ± 4.2 years; 5,400 [58.2%] boys) performed by 31 sonographers (mean number of examinations, 299 ± 140 [range, 115-610]) and interpreted by 31 radiologists (mean number of examinations, 299 ± 157 [range, 101-845]) were included. Mean admission frequency per sonographer was 34.0% ± 3.3% (range, 27.8%-42.6%) and per radiologist was 33.5% ± 3.9% (range, 23.7%-41.6%). Mean appendectomy frequency per sonographer was 20.3% ± 2.6% (range, 14.9%-27.0%) and per radiologist was 20.3% ± 3.1% (range, 15.2%-28.7%). Significant multivariable predictors of hospital admission included temperature (P.0001), white blood cell count (P.0001), male sex (P = .002), imaging performed at the main hospital (versus satellite hospital) (P = .001), abdominal tenderness with ultrasound transducer compression (P.0001), presence of rebound tenderness (P = .001), and presence of acute appendicitis by ultrasound (P.0001), but not sonographer or radiologist. Predictors of appendectomy included weight (P.0001), white blood cell count (P.0001), male sex (P = .0004), abdominal tenderness with ultrasound transducer compression (P.0001), and the presence of acute appendicitis by ultrasound (P.0001), but not sonographer or radiologist.Differences in individual sonographers and radiologists did not predict clinically important outcomes in children undergoing ultrasound in the ED for suspected acute appendicitis.

Repeated administrations of linear gadolinium-based contrast media (GBCM) are associated with T1-weighted (T1-W) signal intensity change in brain structures. The purpose of this study was to compare different brain structures in children after unconfounded, repeated administrations of either a macrocyclic or linear GBCM. We performed a retrospective cohort study, identifying subjects with ≥5 unconfounded administrations of gadoterate meglumine. We matched subjects with repeated administrations of gadopentetate dimeglumine to the gadoterate meglumine arm based on the number of unconfounded GBCM administrations. Two reviewers drew regions of interest on 27 structures in and around the brain. We recorded demographic, modality and study parameters and evaluated them to determine whether they were associated with T1-W signal intensity (SI) changes. Linear mixed effects models evaluated the relationships between the number of GBCM doses and T1-W SI ratio. Finally, we identified differences in the rate of T1-W SI ratio change among individuals using a linear mixed effects model with random slope. We included a total of 52 patients (age range at first MRI: 6.0 months to 17.1 years), 26 in each arm. We detected a significant change in the T1-W SI ratio with repeated administrations of GBCM in one location in the gadoterate meglumine arm and in four locations in the gadopentetate dimeglumine arm. Patient gender and age were not associated with T1-W SI change. Modality vendor, imaging sequence and field strength were variably associated with a systematic difference in the ability to detect a T1-W SI change. Finally, linear mixed effects model with random slope showed that there were individual differences in the slope of SI change at various structures among individuals for both arms. This effect was present in more brain structures in the gadopentetate dimeglumine arm (14 vs. 8). There is a significant change in the T1-W SI ratio over time in multiple brain structures after repeated gadopentetate dimeglumine administrations. This effect was only seen in one ratio after repeated administrations of gadoterate meglumine. There are individual differences in the rate of change of SI ratios over time after repeated administration of gadopentetate dimeglumine and gadoterate meglumine, suggesting that individual differences are present.

INTRODUCTION Pneumatosis intestinalis (PI) is air collections within the wall of the intestine that can be associated with life threatening conditions. Recurrent episodes of PI have been reported; however, little is known about risks for recurrent disease. This study aims to identify predictors of recurrent PI and evaluate long-term outcomes in patients with recurrent PI. METHODS This retrospective case-control study evaluated patients, ages 6 months to 18 years, with imaging evidence of PI over a 30-month period. Images were analyzed by consensus to confirm PI. Recurrent PI was defined as new findings of PI after resolution by imaging and/or at least 1-month interval between episodes of PI. Univariate and multivariable analyses were performed using logistic regression, with significance set to P