Research Presentation Session: Paediatrics

RPS 312 - Foetal and paediatric thoraco-abdominal imaging

February 28, 11:30 - 12:30 CET

7 min
Fractal analysis of perfusion in neuroblastoma using MRI predicts volumetric tumour response and outcome after induction chemotherapy
Florian Michallek, Berlin / Germany
Author Block: F. Michallek, M. Dewey, A. Eggert, T. M. Thole-Kliesch; Berlin/DE
Purpose: Angiogenesis is a hallmark of cancer, however, evidence on the role of tumour perfusion assessment in paediatric patients with neuroblastoma is sparse, although contrast-enhanced magnetic resonance imaging (MRI) is routinely used in neuroblastoma patients. Fractal analysis allows to characterise perfusion patterns on routine contrast-enhanced MRI, which has potential as a non-invasive imaging biomarker for predicting treatment response and outcome by providing information on vascular changes during induction chemotherapy.
Methods or Background: In a national multi-centre setting (30 centres), we retrospectively performed fractal analysis of contrast-enhanced, fat-saturated, T1-weighted MRI. Patients underwent MRI at the timepoint of diagnosis (TP1) and prior to surgical tumour resection (TP2). At both timepoints, fractal analysis was performed to quantify complexity of the tumour perfusion pattern by fractal dimension (FD). At TP1, we investigated, whether fractal analysis allows to predict local tumour response. At TP2, we evaluated prognostic implications of fractal analysis for predicting event-free-survival (EFS). Additionally, we applied fractal analysis to CD34-immunohistochemistry in a subset of patients to correlate imaging findings with underlying vascular structure. We adhered to the STARD guideline.
Results or Findings: We included n=73 patients in separate discovery (n=36; single-centre) and validation cohorts (n=37; 29 centres). At TP1, fractal analysis predicted relative volumetric tumour response with mean bias -0.05 (limits of agreement: -0.35–0.25) using Bland-Altman analysis. Areas with low FD at TP1 showed high tumour response, while high FD-areas tended to persist. At TP2, fractal analysis separated patients with and without favourable EFS in Kaplan-Meier-Analysis (p=0.0019) in tumours lacking MYCN-amplification. In CD34-immunohistochemistry, FD increased significantly after induction chemotherapy (FD-TP1: 1.23±0.09, FD-TP2: 1.44±0.07, p<0.001).
Conclusion: In a retrospective multi-centre setting, fractal analysis of perfusion predicted volumetric response and outcome after induction chemotherapy in patients with neuroblastoma.
Limitations: Prospective validation is lacking making this a limitation of the study.
Funding for this study: No funding was obtained for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Neuroblastoma Registry 2016 German Clinical Trial Register with the DRKS-ID: DRKS00023442; Ethics committee number: 16-432.
7 min
Feasibility and diagnostic accuracy of early postoperative MRI after resection of neuroblastic tumours
Maryanna Chaika, Tuebingen / Germany
Author Block: M. Chaika, M. Esser, P. Krumm, R. Guglielmi, S. Gassenmaier, C. Urla, S. Warmann, J. F. Schäfer; Tübingen/DE
Purpose: In standard MRI, it may be problematic to differentiate residual tumour (RT) from scar, reactive changes, or recurrence after three months. The aim of this study was to evaluate a short MRI protocol performed in the early postoperative period.
Methods or Background: The inclusion criteria were: histologically confirmed neuroblastic tumour, resection by the reference surgery in our centre, early postoperative MRI and MRI preoperatively, and standardised MRI Protocol. The MRI protocol included the following sequences: T1w vibe Dixon before and after contrast, T2w with fat saturation and DWI with the calculation of the ADC map. The analysis was performed by 3 independent readers (resident/adults radiologist/board certified pediatric radiologist) using a 4-points Likert-scale . A multimodal reference standard was determined by: F/U imaging, consensus tumour-board, and consensus between senior radiologist and surgeon.
Results or Findings: Thirty-nine patients with a median of 46 age months (5-177). MRI was performed in the mean 8(+/- 5) days after surgery. RT was found in 13 patients by MRI and confirmed by the reference standard, with a typical location at the mesenteric root and retrocrural. 4 RTs were expected by the surgeons with a median volume of 19 ml (1-34 ml), and nine tumours were unexpected with a median volume of 1 ml (0.25-7). Sensitivity, specificity, and accuracy (reader) were 77,54, and 70 %(Resident), 81,85, and 82 %(Adult Radiologist), 92,92, and 92 %(Pediatric Radiologist). Reading the postoperative MRI alone, the diagnostic performance of the paediatric radiologist were 88,74, and 84%.
Conclusion: Early MRI protocol is feasible for determination of residual tumour. Experience in pediatric imaging is crucial to achieving high diagnostic precision. Reading the preoperative MRI improves diagnostic accuracy.
Limitations: Small sample size. The results are preliminary and need to be confirmed by follow-up studies.
Funding for this study: No funding from any specific grant was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The local institutional review board (Ethics Committee at the Medical Faculty of the Eberhard Karls University, Tuebingen, Germany) approved this retrospective, monocentric study.
7 min
Neonatal pulmonary MRI at term-equivalent age for defining bronchopulmonary dysplasia severity and its relationship with short-term respiratory outcomes
Gianluca Folco, Imperia / Italy
Author Block: A. M. Munari1, G. Folco1, C. B. Monti1, F. Rizzetto1, N. A. R. Panarisi2, S. Zirpoli1; 1Milan/IT, 2Assago/IT
Purpose: This prospective study aims to demonstrate that neonatal pulmonary magnetic resonance imaging (MRI), using commercially available sequences, can assess lung parenchymal injury associated with bronchopulmonary dysplasia (BPD) and may predict short-term respiratory outcomes.
Methods or Background: Pulmonary MRI, using turbo spin echo and gradient echo sequences during natural sleep, was performed in premature infants who underwent routine brain MRI at term-equivalent age. Lung MRIs were scored using the modified Ochiai scoring system and a new proposed score, the Bronchopulmonary Dysplasia Magnetic Resonance Index (BPDMRI), based on four hyperaeration (hyperexpansion, emphysema, cysts, mosaic pattern) and four parenchymal variables (triangular subpleural opacities, fibrous stripes and bands, edema and atelectasis, distortion of bronchovascular bundles). Both scores were correlated with two short-term respiratory outcomes: the length of mechanical ventilation and the length of any respiratory support until discharge.
Results or Findings: The study included 25 premature infants (9 with BPD, 36%). Pulmonary MRI showed a wide variety in appearance of pulmonary parenchyma among BPD patients and revealed significant structural differences across the range of MRI scores. Both the modified Ochiai score and the BPDMRI resulted positively correlated to the length of mechanical ventilation (r=0.828, p<0.001; r=0.900, p<0.001) and the length of any respiratory support until discharge (r=0,897, p<0.001; r=0.953, p<0.001). In the severe disease group, the length of respiratory support showed a smaller IQR when related with the BPDMRI rather than when related with the clinical definition.
Conclusion: Pulmonary MRI identified lung parenchymal abnormalities associated with BPD, showing a strong correlation with the length of mechanical ventilation and the length of any respiratory support until discharge.
Limitations: Single centre, retrospective study is the limitation of this study.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was ethically approved.
7 min
Applicability of the O-RADS scoring system in children and adolescents: results of a preliminary retrospective monocentric study
Elisa Mercanzin, Bergamo / Italy
Author Block: E. Mercanzin1, P. Clauser2, V. Rizzo3, S. Sironi1, A. Hojreh2; 1Bergamo/IT, 2Vienna/AT, 3Maglie/IT
Purpose: The study aims to assess the accuracy of the Ovarian-Adnexal Reporting & Data System (O-RADS) classification on ultrasound (US) and magnetic resonance imaging (MRI) in paediatric patients.
Methods or Background: In this single-centre retrospective study, all US and MRI studies performed between 2005 and 2022 for suspected ovarian lesions in patients < 18 years were included. Histopathology was considered as standard of reference. Two readers in consensus evaluated imaging characteristics as defined by the O-RADS (features, size, wall thickness, ascites). Each lesion was then assigned an O-RADS category. Chi-square test was used, sensitivity, specificity, and accuracy were calculated using a cutoff >O-RADS 3, and compared.
Results or Findings: Included in this study were 95 patients (mean age 12,6, range 0-18). Histology found 45 non-neoplastic cystic masses (47.4%), 47 benign neoplasms (49.5%) and 3 malignant neoplasms (3.2%). MRI was available in 80 cases and US in 72 cases. No significant difference in the US features was found between malignant and benign lesions (p>0.2). On MRI a significant difference was found between benign and malignant masses in terms of lesion diameter (p=0.02) and enhancement (p=0.025). MRI correctly classified all malignant cases (3/3, 100%), while US only 1/2 (50%). Specificity was comparable between modalities (respectively 89% vs 88%, p<0.0001). Accuracy was higher for MRI (87% vs 90%, p<0.0001).
Conclusion: Our results showed that the MRI O-RADS demonstrated a high diagnostic accuracy and can be applied in paediatric patients. In this population, a higher specificity would be desirable to reduce false positives and unnecessary follow up or surgery. A lower accuracy was found for US.
Limitations: This was a retrospective study, its limitations are the retrospective evaluation of the US images, low number of malignant lesions.
Funding for this study: No research funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics committee: Medical University of Vienna 2057/2017 approved this study.
7 min
The distal colostogram for anorectal malformation: considerations in low- and middle-income countries
Miriam Leiderer, Passau / Germany
Author Block: M. Leiderer; Passau/DE
Purpose: The purpose of this study was to evaluate imaging challenges of anorectal malformations (ARM) in older male children and those with prior loop colostomy.
Methods or Background: Treatment of ARM in low- and middle-income countries suffers from limited resources in paediatric surgery. Children undergo colostomy in the newborn period, and definitive treatment is delayed. Consequently, children are older than the recommended age for definitive reconstruction at time of surgery. A loop colostomy adds additional technical challenges. We reviewed all colostograms obtained for patients with ARM presenting for definitive repair at a specialised hospital in Uganda, between April and August 2023.
Results or Findings: Twenty-three patients were included, twelve male, of which seven with suspicion of rectourinary fistula. Three of these were up to 12 m (5 m, 9 m, 12 m), in all of which the fistula tract was visualised. No fistula tract was visualised in older patients (22 m, 3 y, 4y , 7 y), however, in three of these the fistula origin was seen as a small anterior tapering of the rectal pouch. In one patient with no sign of fistula the absence of a fistula was confirmed during surgery. Fourteen patients (61%) had a fecaloma in the distal pouch.
Conclusion: For all the patients with suspected genitourinary fistula older than 12 m the fistula did not opacify, an age group never expressly reported previously. Possible explanations are occlusion by accumulated feces and mucus or through colonic distention; and ability of older children to react to abdominal pressure by tensing the striated perirectal muscular complex. Fecaloma because of loop colostomy was common, leading to chronic rectal distension associated with a poorer functional outcome and increased infection risk.
Limitations: The small number of cases warrants further research with a larger case number. Also, no follow-up is available yet to assess functional outcome, making these the limitations of this study.
Funding for this study: No funding was received.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: The requirement for informed consent was waived for this retrospective study.
7 min
Evaluation of perfusion impairment in Kawasaki disease using fully quantitative cardiovascular magnetic resonance myocardial perfusion: correlation with left ventricular remodelling
Zhongqin Zhou, Chengdu / China
Author Block: Z. Zhou, L-y. Wen, S. Azhe, L. Hu, Y-K. Guo; Chengdu/CN
Purpose: This study aims to assess perfusion impairment in Kawasaki disease (KD) using fully quantitative cardiovascular magnetic resonance (CMR) myocardial perfusion and correlates it with left ventricular (LV) remodeling.
Methods or Background: Eight-seven children (60 males, 7.49±2.15 years) diagnosed with KD and 33 controls (22 males, 8.22±2.81 years) were enrolled to complete CMR. A dual-bolus protocol was used to acquire perfusion images. Myocardial blood flow (MBF) was acquired through post processing and corrected for the heart rate-blood pressure product (MBF corrected [MBFcor]). The diameter of coronary artery in patients with KD was standardized by Z score. Z score≥5 represented mid-giant coronary aneurysm.
Results or Findings: MBFcor was lower in patients with KD than that of controls (2.23±0.50 vs 2.48±0.67 mL/g/min, p=0.031), especially in patients at acute phase (2.08±0.54 vs 2.48±0.67 mL/g/min, p=0.026). The MBFcor in patients with Z score≥5 decreased when compared with the other patients with KD and controls (2.04±0.54 vs 2.34±0.44 vs 2.48±0.57 mL/g/min, p=0.005). MBFcor was correlated with Z score in patients with KD (r=-0.305, p=0.004). Multivariate analysis revealed that age, acute phase and Z score was correlated with MBFcor (β=-0.238, p=0.019; β=-0.217, p=0.031; β=-0.290, p=0.005, respectively). Radial peak strain (PS), the absolute value of circumference PS, LV mass index and LV remodeling index was correlated with MBFcor (r=0.233, p=0.035; r=0.259, p=0.019; r=-0.268, p=0.012; r=-0.391, p<0.001, respectively). Multivariate analysis revealed that acute phase and MBFcor was associated with LV remodeling index (β=0.290, p=0.003; β=-0.235, p=0.025, respectively)
Conclusion: Fully quantitative CMR myocardial perfusion can evaluate myocardial perfusion impairment in children with KD. The perfusion impairment in KD is related to acute disease and coronary artery dilation and may cause abnormalities in LV function and LV remodeling.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Medical Ethics Committee of our hospital.

This session will not be streamed, nor will it be available on-demand!