ESPU-SPU Meeting on Friday 24, September 2021, 13:00 - 13:30
Erman CEYHAN 1, Emin MAMMADOV 1, Sevgen Celik ONDER 2, Hasan Serkan DOGAN 3 and Serdar TEKGUL 3
1) Hacettepe University Faculty of Medicine, Department of Urology, Ankara, TURKEY - 2) Hacettepe University Faculty of Medicine, Department of Pathology, Ankara, TURKEY - 3) Hacettepe University Faculty of Medicine, Department of Urology, Division of Pediatric Urology, Ankara, TURKEY
Malignancy after augmentation cystoplasty (AC) in the long term is reported repetitively up to 5.5%. We assessed the use of urine fluorescent in-situ hybridization (FISH) screening for bladder malignancy after AC in comparison with current methods.
MATERIAL AND METHODS
36/98 patients under follow-up who have completed tenth year after ileal AC included to the study prospectively. 24 (%66.7) patients tested with FISH initially and 28 (77.8%) patients with conventional cytology (CC). Blinded from the cytology results, 32 (%88.9) patients who consented were undergone cystoscopy with random biopsy (native bladder, ileal segment, junction). 2 patients tested with FISH did not consented cystoscopy. This study is registered to the Turkish goverment registry (no:71146310).
Mean follow-up time after AC was 15.4±4.8 years. 2/32 (5.6%) patients diagnosed with adenocarcinoma in cyctoscopic biopsy. FISH analysis of 3/24 (12.5%) patients demonstrated malignant signal. 2 FISH malignant patients were patients who had adenocarcinoma. The third patient’s biopsy was benign and the third year control cystoscopy was normal. 2/3 patients with malignant CC had adenocarcinoma and 1 patient had benign biopsy. The sensitivity and specificity of FISH in our series were 100% and 95% respectively. Whereas the sensitivity and specificity of CC was 100% and 91.6% respectively. The two main morbidities in patients resumed follow-up were recurrent urinary tract infection 8/36 (22.2%) and recurrent bladder stones 3/36 (8.3%).
Despite limited number of patients in this study, FISH is found to be superior to CC with regard to specificity. FISH is a promising tool to select patients for cystoscopy earlier than the tumor progresses to invasive nature.
Maria GRELLA 1, Mathieu PEYCELON 2, Samia LARAQUI 3, Aurelien SCALABRE 4, Alice FAURE 5, Emmanuel SAPIN 6, Valeska BIDAULT 7, Frederic AUBER 8, Diuty SHARMA 9, Luke HARPER 10, Quentin BALLOUHEY 11, Olivier ABBO 12 and Alexis P ARNAUD 3
1) CHU Poitiers, Pediatric surgery department, Poitiers, FRANCE - 2) Indiana University, School of Medicine, Department of Pediatric Urology of Riley Children Hospital, Indianapolis, USA - 3) Univ Rennes, CHU Rennes, General Paediatric Surgery and Paediatric Urology department, Rennes, FRANCE - 4) CHU St Etienne, General Paediatric surgery department, St Etienne, FRANCE - 5) APHM, General Paediatric Surgery and Paediatric Urology department, Marseille, FRANCE - 6) CHU Dijon, General Paediatric surgery department, Dijon, FRANCE - 7) APHP Nord, Reference Center for Rare Diseases (CRMR) Malformations Rares des Voies Urinaires (MARVU); Université de Paris, Department of Pediatric Urology of Robert-Debré University Hospital, Paris, FRANCE - 8) CHU Besançon, General Paediatric surgery department, Besançon, FRANCE - 9) CHU Lille, General Paediatric Surgery and Paediatric Urology department, Lille, FRANCE - 10) CHU Bordeaux, General Paediatric Surgery and Paediatric Urology department, Bordeaux, FRANCE - 11) CHU Limoges, General Paediatric surgery department, Limoges, FRANCE - 12) CHU Toulouse, General Paediatric Surgery and Paediatric Urology department, Toulouse, FRANCE
Urothelial tumor is a rare entity in the pediatric population. Prognosis criteria and long-term follow-up are not standardized. The aim of our study was to describe the treatment and outcomes of a multicentre series.
MATERIAL AND METHODS
International multicentre retrospective study including patients born between 2000 and 2018, treated for urothelial tumor (Ethic review board approval #19.45). Data included patients' characteristics, clinical data, operative notes, pathology reports and outcomes. Statistics: Kaplan Meier survival analysis.
22 patients were included in 12 centres (sex ratio=1.4). Median (min-max) age at first symptoms was 10.8 years (3-17). The most frequent symptom was gross hematuria (45%), but 9 patients (41%) had incidental findings. Two cases of predisposing genetic mutation (1 Lynch, 1 Costello) and 3 familial medical histories of non-urothelial cancer were found. No history of neurogenic bladder nor bladder augmentation were identified. All patients had an endoscopic resection as for first line treatment. Two postoperative complications Clavien 3b were found (1 urethral stenosis, 1 bladder bleeding). All patients had a solitary tumor, without metastasis. Pathological analysis found 9 standard papillomas, 3 inverted papillomas, 3 PUNLMP, 6 LGPUC and 1 invasive urothelial carcinoma. 3 patients required a secondary line surgery for incomplete resection. No adjuvant treatment was needed. After a median follow-up of 11 months (3-44), one patient (Costello syndrome) recurred locally 5months after surgery. The OSR and 1-year EFS were 100% and 90% respectively.
This series underlines the low invasiveness of urothelial tumors in the pediatric population, with a favorable prognosis. However, the small number and the short follow-up must weigh the conclusion. Less than 10% of the patients presented with predisposing factor.
Hriday BHAMBHVANI, Alvaro ZAMORA and Kunj SHETH
Stanford University School of Medicine, Department of Urology, Palo Alto, USA
Machine learning is a promising technique for synthesizing high-dimensional data to supplement clinical decision-making. Currently, there are no reported survival prediction tools for urogenital rhabdomyosarcoma (UG-RMS). Therefore, we sought to develop artificial neural networks (ANNs), a type of machine learning algorithm, and classical Cox proportional hazards (CPH) models to predict both 5-year disease-specific survival (DSS) and overall survival (OS) in pediatric UG-RMS.
MATERIAL AND METHODS
The National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) 18 program database was queried to identify patients under 20 years diagnosed with UG-RMS between 1998 and 2011. ANNs were trained and tested on an 80/20 split of the dataset in a 5-fold cross-validated fashion. Multivariable CPH models were developed in parallel. Variables used for prediction were age, sex, race, primary site, histology, SEER stage, tumor size, surgical excision, and radiation therapy. Receiver operating characteristic curve analysis was conducted, and ANN models were additionally tested for calibration.
277 patients were included. The area under the curve (AUC) for the ANN models was 0.93 for OS and 0.91 for DSS. AUC for the CPH models was 0.82 for OS and 0.84 for DSS. The ANN models were well-calibrated: OS model (slope = 1.02, intercept = -0.06) and DSS model (slope = 0.79, intercept = 0.21). The ANN OS model was deployed in an open access application (https://uorg.shinyapps.io/UroRhabdoSurv/).
ANN models performed better than multivariable CPH models for UG-RMS. These models will need external validation, and investigators can assess the deployed OS model in their own cohorts.