ESPU Meeting on Thursday 18, April 2024, 14:50 - 15:50
14:50 - 14:53
S11-1 (OP)
Emre ERDEM 1, Abdulrahman JAFAROV 1, Mustafa Alkan OKTAR 1, Murat Can KARABURUN 1, Aykut AKINCI 1, Tarkan SOYGÜR 2 and Berk BURGU 2
1) Ankara University Faculty of Medicine, Department of Urology, Ankara, TURKEY - 2) Ankara University Faculty of Medicine, Department of Pediatric Urology, Ankara, TURKEY
PURPOSE
In patients with severe hydroureteronephrosis and nocturnal polyuria, night time sleeping adversely affects the drainage of the upper urinary tract since the gravitational force during day time is lacking. This study aimed to investigate the effects of sleeping positions and their influence on renal function in a rat model with hydroureteronephrosis.
MATERIAL AND METHODS
36 Male Sprague-Dawley rats were selected and divided into 3 groups for this experiment. The study involved inducing nocturnal polyuria in the rat models through a standardized partial bladder outlet obstruction procedure [followed by continuous nocturnal monitoring for 2 months.
The investigation focused on 3 forced sleeping positions for 8 hours - supine, prone, and upright. Parameters were recorded including day time and nocturnal urine production, and renal function markers.
Mean creatinine levels were as follows: 0.85±0,02 mg/dL for the supine group, 0.9±0,03 mg/dL for the prone group, 0.6±0,01 mg/dL for the upright group. All kidneys were examined histopathologically for scars and fibrosis after 2 months.
RESULTS
In the upright sleeping group creatinine levels were lower. Preliminary results suggest that sleeping in upright position yielded more favorable outcomes regarding renal function.According to independent t sample test rats adopting the supine and prone positions exhibited significantly decreased urine production and increased renal stress potentially mitigating further renal damage.
CONCLUSIONS
These findings highlight the potential benefits of gravitational force in the drainage of dilated systems. Sleeping at prone or supine positions where drainage of the upper tract is impaired at night, worsens the renal outcome on long run. Since we can not force children sleep at upright position, to reduce the renal damage where drainage of the system is impaired at night, overnight cathetarization can potentially save the kidneys.
14:53 - 14:56
S11-2 (OP)
Tariq Osman ABBAS 1, Sakib MAHMUD 2, Muhammad E. H. CHOWDHURY 2, Santiago VALLASCIANI 3, Saidul KABIR 4, Sreekumar MUTHIYAL 5, Alaa KOKO 5, Ahmed BALLA ABDALLA ALTYEB 6, Abdulrahman ALQAHTANI 7, Amith KHANDAKAR 8 and Sheikh MOHAMMED SHARIFUL ISLAM 9
1) Sidra Medicine, Urology, Doha, QATAR - 2) Qatar University,, Department of Electrical Engineering, Doha, QATAR - 3) Sidra Medicine,, Urology Division, Surgery Department,, Doha, QATAR - 4) University of Dhaka,, Department of Electrical and Electronic Engineering, Dhaka, BANGLADESH - 5) Hamad Medical Corporation, Radiology, Doha, QATAR - 6) Hamad Medical Corporation, Urology, Doha, QATAR - 7) Prince Sattam Bin Abdulaziz University,, 10Department of Biomedical Technology, College of Applied Medical Sciences in Al-Kharj,, Al-Kharj, SAUDI ARABIA - 8) Qatar University,, Department of Electrical Engineering,, Doha, QATAR - 9) Deakin University,, Institute for Physical Activity and Nutrition,, Melbourne, AUSTRALIA
PURPOSE
Kidney ultrasound images are one of the most common methods of monitoring Antenatal or prenatal hydronephrosis (AHN), but grading of this condition is highly subjective and clinicians may select inappropriate therapies or surgical interventions as a result. New approaches are required to differentiate subjects who can facilitate standardization of hydroneprhosis findings in ultrasound.
MATERIAL AND METHODS
An end-to-end deep machine learning framework was developed to sequentially detect ultrasound regions of interest, segment kidneys from 433 US images, and classify AHN severity. We propose the novel Kidney Ultrasound Segmentation Network (KUSNet) for kidney segmentation from ultrasound images, according to the Society of Fetal Urology (SFU) standards. The ground truth kidney masks were generated by two radiologists with more than five years of working experience while the SFU-based annotations for the AHN severity were done by two senior radiologists and three senior urologists. At each stage, the performance of the proposed models was assessed both quantitatively and qualitatively against state-of-the-art networks.
RESULTS
The proposed KUSNet for ultrasound kidney segmentation achieved 97.6% accuracy, 97.4% precision, 97.6% recall or sensitivity, 97.5% f1-score, 95.5% IoU or Jaccard score, and 92.1% Dice score, beating several state-of-the-art networks originally developed for segmenting medical images. On the other hand, the novel PHCNet reached 93.9% accuracy, 93.7% precision, 93.9% recall, 93.8% specificity, and 89.0% f1-score subject-wise when performing multiclass stratification of AHN severity based on the SFU grading system.
CONCLUSIONS
Artificial intelligence-based tools can reliably classify AHN severity to reduce inter- and intra-observer bias, thereby aiding clinicians in the rapid selection of appropriate treatments and surgeries.
14:56 - 14:59
S11-3 (OP)
Roseanne FERREIRA DE FREITAS EUZEBIO 1, Melissa MCGRATH 1 and Luis H. BRAGA 2
1) McMaster University, Surgery-Urology, Hamilton, CANADA - 2) McMaster University - McMaster Children's Hospital, Department of Surgery / Urology, Hamilton, CANADA
PURPOSE
To describe the renal length index (RLI) and evaluate its utility in identifying severe UPJO-like hydronephrosis (HN) suggestive of obstruction.
MATERIAL AND METHODS
A prenatal HN database (2008-23) was reviewed to select patients with unilateral UPJO-like HN. Those with VUR, megaureter, atrophy and other anomalies were excluded. RLI was calculated by using the formula [100%*(Affected Renal Length - Contralateral Renal Length)/Affected Renal Length], based on baseline ultrasound. Data points, baseline renal length(RL), MAG3 t ½ drainage time, curve patterns, and surgical interventions(Pyeloplasty) were collected. Obstruction was defined as a t ½ time >30 min and/or a non-descending drainage curve on MAG3 lasix scan. Fischer's t-test and Wilcoxon-sum were used for statistical analysis.
RESULTS
From 465 patients, 226(48.6%) were SFU 3-4, 162(34.8%) showed obstruction on MAG 3, and 129(27.7%) underwent pyeloplasty. The median RLI was 8.6% (IQR:4.3-16.1%). Patients with obstruction had a notably higher median RLI(16.07%[8.2-24.6%]) compared to non-obstructed counterparts (6.8%[3.4-11.8%],p<0.001). The RLI was threefold higher in the surgical group (18.3% [11.1-25.0%] vs 6.9% [3.4-12.0%], p<0.001). In 93(20%) out of 465 cases, the affected kidney was smaller than the contralateral one. This group had a lower median RLI (5.56% [1.96-11.11%] vs 9.405% [4.92-16.67%],p<0.001) and a decreased obstruction rate (14.0%vs 40.1%,p<0.001) compared to the group with larger affected kidneys. Having an affected RL>contralateral RL increased the likelihood of surgery nearly fourfold (OR 3.9, 95% CI 2.0-7.7).
CONCLUSIONS
Higher RLI values are significantly associated with obstruction on MAG 3 lasix renal scan and, consequently, surgical intervention, particularly in patients whose affected kidneys are larger than the contralateral ones.
15:10 - 15:13
S11-4 (OP)
Lauren ERDMAN 1, Jethro KWONG 2, Ihtisham AHMAD 2, Adree KHONDKER 3, Joana DOS SANTOS 4, Michael CHUA 5, Jin Kyu KIM 3, Armando J LORENZO 5 and Mandy RICKARD 5
1) SickKids, Center for computational medicine, Toronto, CANADA - 2) University of Toronto, Division of Urology, Toronto, CANADA - 3) University of Toronto, Urology, Toronto, CANADA - 4) SickKids, Urolgy, Toronto, CANADA - 5) SickKids, Urology, Toronto, CANADA
PURPOSE
The Hydronephrosis Severity Index (HSI) was developed to use early renal ultrasound images to stratify patients with hydronephrosis (HN) into high, medium, and low risk groups to reduce the invasiveness of investigations for low-risk patients.Herein we present a 2 year follow-up on previously stratified patients to assess longer-term safety of this tool.
MATERIAL AND METHODS
Additional data was collected from the silent trial of the HSI (n = 202). HSI patient stratification was extended to include the original low-risk group (95% sensitive threshold for intervention), and further divided into a medium-risk group (<95% sensitivity, <95% specificity) and a high-risk group (95% specificity).These groups were assessed for future urinary tract infections (UTIs), nuclear scans (NS), and surgery, following HSI calculation.
RESULTS
Among the original 202 silent trial patients, follow-up data was available for 159 patients. Of these 159 patients, 104 were low-risk, 34 were medium-risk, and 21 were high-risk. Within the low-risk group, 1 (1%) had a UTI, 3 (3%) received a NS, 1 (1%) received surgery. In contrast, of the 34 medium-risk patients, 3 (9%) had a UTI, 5 (15%) received a NS, and 3 (9%) received surgery. Finally, among the 21 high-risk patients, 5 (24%) had a UTI, 5 (24%) received a NS, and 15 (71%) received surgery.
CONCLUSIONS
The HSI tool appears to safely and effectively risk-stratify patients into low, medium and high risk levels based on single saggital and transverse renal ultrasound images, up to 2 years later.
15:13 - 15:16
S11-5 (OP)
Bernhard HAID 1, Anna Magdalena BERNSTEINER 2, Franziska RAMESEDER 3 and Josef OSWALD 3
1) Ordensklinikum Linz, Hospital of the Sisters of Charity, Pediatric Urology, Linz, AUSTRIA - 2) Ordensklinikum Linz, Hospital of the Sisters of Charity, Department for Pediatric Urology, Linz, AUSTRIA - 3) Ordensklinikum Linz, Hospital of the Sisters of Charity, Department of Pediatric Urology, Linz, AUSTRIA
PURPOSE
Mercaptoacetlytriglycin (MAG3) scans are used to evaluate children with high-grade hydronephrosis. Based on the time point of administration of furosemide, different protocols are in use. This prospective cohort study was accompanying an institution wide protocol-change from F+20 (furosemide after 20’) to F+0 (simultaneously injection of tracer and furosemide) in children without voluntary bladder control. We aimed at evaluating the impact of the MAG3 protocol used on time to surgery and number of MAG3 scans performed prior to surgery.
MATERIAL AND METHODS
A total of 63 patients evaluated the first time for isolated, unilateral high-grade hydronephrosis (SFU ≥°III) after exclusion of vesicoureteral reflux by a voiding cystourethrography were included. 28 consecutive patients were included before protocol change (F+20), 35 consecutive patients thereafter (F+0). Patients with loss of DRF as indication for surgery were excluded. Demographic data, outcome, ap-diameter as well as time until surgery and number of MAG3 scans prior to surgery were compared between the groups using Mann-Whitney test, Fisher’s exact test and students t-test (Prism 10).
RESULTS
The groups (F+0 vs. F+20) were comparable concerning age (mean 5.1 vs. 5.6 months, p=0.721) and ap-diameter (1.67 vs. 1.68, p=0.95) but not concerning duration of follow-up (median 6.9 vs. 27.7 months, p<0.001). 48.5% (17/35) in the F+0 group and 71.4% (20/28) in the F+20 group underwent pyeloplasty due to persistent obstruction (n=2) in MAG3 scans. Time from 1st MAG3 scan until pyeloplasty was shorter in the F+0 group (7.4 vs. 13.4 months, p<0.001). Children in the F+0 group underwent less MAG3 exams before surgery compared to those in the F+20 group (median 2 vs. 3, p<0.001).
CONCLUSIONS
These preliminary results suggest that in patients with high-grade hydronephrosis use of the F+0 protocol is associated with a shorter time until surgery and less MAG3 exams.
15:16 - 15:19
S11-6 (OP)
Charlotte GROGNET, Arthur LAURIOT DIT PREVOST and Heloise LERISSON
Centre Hospitalier Universitaire de Lille - Hôpital Jeanne de Flandre, Imagerie Pédiatrique, Lille, FRANCE
PURPOSE
MR urography with diuretic injection is an increasingly used imaging modality for the evaluation of pediatric ureteropelvic junction obstruction. It provides anatomical and functional data without irradiation. The aim of the study was to analyze the increase in anteroposterior diameter of renal pelvis after furosemide injection in MR urography to determine a threshold for positive diagnosis. The study also evaluated if this increase could contribute to etiological diagnosis and whether MRI reliably detects crossing vessels.
MATERIAL AND METHODS
70 children who underwent surgery and preoperative MR urography were included in this monocentric retrospective study. The increase in anteroposterior diameter of renal pelvis after furosemide injection in MR urography of the healthy side was compared to the pathological side of a same patient to determine a pathological increase threshold. This increase was also compared according to the etiology (intrinsic vs extrinsic). The association between crossing vessels on MRI and during surgery was also evaluated.
RESULTS
The increase in anteroposterior diameter of renal pelvis was significantly greater on the pathological side compared to the healthy side (p < 0.0001). The optimal threshold for the prediction of pathological character was 6 mm (Sensibility = 68.6%, Specificity = 87.1%). There was no significant difference in increase in anteroposterior diameter of renal pelvis depending on etiology (p = 0.8582). The association between crossing vessels on MRI and during surgery was statistically significant (p < 0.001).
CONCLUSIONS
An increase of more than 6 mm in anteroposterior diameter of renal pelvis after furosemide injection could constitute a MR urography diagnostic criterion for pediatric ureteropelvic junction obstruction.
15:30 - 15:33
S11-7 (OP)
Mathilde GLENISSON 1, Nicolas VINIT 2, Fabienne MARQUANT 3, Perrine PARTY 2, Lilia NEFZI 2, Aline BROCH 2, Caroline ELIE 3, Henri LOTTMANN 2, Nathalie BOTTO 2 and Thomas BLANC 2
1) Hopital Necker Enfants Malades, Department of Pediatric Surgery and Urology, Paris, FRANCE - 2) Hôpital Necker Enfants-Malades, Department of Pediatric Surgery and Urology, Paris, FRANCE - 3) Hôpital Necker Enfants-Malades, Unité de Recherche Clinique/Centre d'investigation Clinique, Paris, FRANCE
PURPOSE
Since its inception in 2016, the use of magnetic-end Double-J ureteral stent (MEDJUS) has rapidly become the drainage of choice in paediatric urology at our institution. However, there are no data specifically evaluating parental satisfaction. This study aims to assess parental satisfaction with the use of MEDJUS.
MATERIAL AND METHODS
The EFUJA study was conducted prospectively (NCT04384939) between 12/2021 and 08/2023. A clinical pathway and prospective research protocol were set up to evaluate the parental satisfaction of the use of MEDJUS. A survey tool specifically addressing satisfaction was applied to all parents of children, immediately following MEDJUS removal (by the nurse) and 7 days following this (by a research assistant).
RESULTS
MEDJUS insertion was attempted in 67 children (35 boys). Median age was 5.9 years (IQR:3.2-10.4). Stent insertion was successful in 65 cases (97%). The two failures occurred during robotic-assisted pyeloplasty with antegrade placement. The MEDJUS was successfully removed in the outpatient clinic in 59 children (92%). The median duration of stent removal was 1 min (1-2).
63 parents responded to the survey immediately after stent removal. They were highly satisfied of the information provided by the surgeons and the nurses before the procedure and during the removal (satisfaction scores 10(10-10)). The median overall experience score was 10(8-10). All parents but one would recommend this clinical pathway if a new JJ stent was needed.
CONCLUSIONS
In this prospective study we confirm that the Black-Star® and its magnetic retrieval device can be safely used in paediatric population and demonstrate excellent parental satisfaction for this care pathway.
15:33 - 15:36
S11-8 (OP)
Yesica QUIROZ MADARRIAGA, Rocio JIMENEZ and Anna BUJONS
Fundacio Puigvert, Paediatric Urology, Barcelona, SPAIN
PURPOSE
The gold standard for treatment of PUJS is still the Anderson-Hynes technique. However, little is known about the repercussions of residual hydronephrosis that may remain after correction of this malformation. The aim of this work is to determine the risk factors that influence the appearance of lithiasis in the long-term follow-up of patients who undergo pyelopasty.
MATERIAL AND METHODS
We retrospectively reviewed 144 medical records of patients undergoing pyeloplasty at a single institution. Finally, 72 patients were included in the analysis, from whom demographic data, preoperative data, surgical approach, postoperative drainage, postoperative data, occurrence of lithiasis, need for surgery and follow-up time were collected. A multivariate analysis with logistic regression was performed, as well as a survival analysis with Kaplan Meyer curves and Cox regression.
RESULTS
79.2% were male, with a mean age of 6.7 years and 51.4% with prenatal diagnosis. The 52.8% were taken to open pyeloplasty and the remaining to robotic approach. The mean renal pelvis diameter (RPD) was 33.8mm (12-66mm). The incidence of lithiasis in 10.4 years of average follow-up (1-32.8 years) was 16.7%, in an average of 6.7 years. In the multivariate analysis only RPD prevailed as a risk factor (OR 1. 17; p<0.023). Regarding the lithiasis-free survival curves, the oldest age group (>10 years) had the worst survival. In the Cox regression of the predictive model, for each additional millimeter in the preoperative DPR, the risk of occurrence of lithiasis is multiplied by 1.2 (p<0.000; 95% CI 1.085 to 1.315).
CONCLUSIONS
Patients with larger preoperative renal pelvis diameters that do not decrease significantly postoperatively have an increased risk of developing lithiasis during follow-up, and those who tend to do so more rapidly are patients who are taken to surgery after 10 years of age.
15:36 - 15:39
S11-9 (OP)
Mariia RIZHKO 1, Mandy RICKARD 2, Joana DOS SANTOS 2, Michael CHUA 2, Armando J LORENZO 2, Anna GOLDENBERG 1 and Lauren ERDMAN 1
1) SickKids, Center for Computational Medicine, Toronto, CANADA - 2) SickKids, Urology, Toronto, CANADA
PURPOSE
Medical image classification with deep learning is challenging, especially in rare conditions when fewer cases are available to learn from. We show that by forcing our algorithm to focus on the kidney in ultrasound images, surgical intervention for hydronephrosis can be predicted accurately when developed using <5% of our original training data set.
MATERIAL AND METHODS
Five datasets were collected from four quaternary care pediatric hospitals in North America. All datasets include sagittal ultrasound images of pediatric kidneys with hydronephrosis. A sub-dataset of convenience from our full set of 2544 ultrasounds, consisting of 66 samples, were segmented to show kidney region. Multiple deep convolutional networks were developed and compared our limited data set model. Our approach used the ResNet18 architecture with a new loss function, combining the binary cross entropy loss with a penalty for class activation map activation outside the area showing the organ of interest.
RESULTS
We achieved comparable performance (>90% area under the receiver operator curve, AUROC) using both the full data set and only 2.6% of the training data required. We evaluated our approach on four additional datasets from other institutions and show that it transfers well (>80% AUROC) across all institutions. We found improved prediction using far fewer samples, as well as improved attention on the kidney, particularly the calyces in our images.
CONCLUSIONS
Our model demonstrates the ability of this new approach to predict just as accurately with significantly less data and improved focus on the organ of interest than required of typical deep learning prediction.