You have accessJournal of UrologyKidney Cancer: Localized: Surgical Therapy II (PD16)1 Sep 2021PD16-04 A PILOT STUDY INVESTIGATING THE FEASIBILITY OF USING A FULLY AUTOMATIC SOFTWARE TO ASSESS THE RENAL AND PADUA SCORE Mathieu Carlier, Fabien Lareyre, Cédric Adam, Marion Carrier, Matthieu Durand, and Juliette Raffort Mathieu CarlierMathieu Carlier More articles by this author , Fabien LareyreFabien Lareyre More articles by this author , Cédric AdamCédric Adam More articles by this author , Marion CarrierMarion Carrier More articles by this author , Matthieu DurandMatthieu Durand More articles by this author , and Juliette RaffortJuliette Raffort More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000001998.04AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Minimally invasive approaches have demonstrated their interest to improve the peri-operative outcomes during partial nephrectomy for renal cell carcinoma (RCC). Several image-based morphometric scoring systems such as the RENAL and the PADUA scores have been developed to evaluate the complexity of the procedure. Artificial intelligence (AI) has brought new insights in medical imaging analysis, offering the opportunity to perform complex tasks and to develop pattern recognition. We aimed to develop a new imaging software based on AI-techniques to enable an automatic detection and a 3D visualization of RCC from CT scans. This pilot study addressed the feasibility to use it to evaluate the anatomical features included in the RENAL and the PADUA scores. METHODS: The study included patients with RCC who underwent robot-assisted nephrectomy at the University Hospital of Nice between January 2015 and December 2018. Diagnosis of RCC was based on CTA and confirmed by post-operative histopathological examination. Clinical characteristics were collected using manual and electronic medical records. Imaging data were extracted in DICOM format. The pipeline to enable an automatic visualization of RCC consisted in 4 sequential steps: image pre-processing, detection of the kidney using the active contour method, segmentation of the renal vascularization using a hybrid approach combing expert system with deep learning and 3D-viusalization of the tumor kidney. RENAL and PADUA scores were assessed by a trained operator on a semi-automatic commercialized software to serve as a ground truth. Concordance of the staging obtained with the automatic software were blindly assessed by the same operator. RESULTS: In total, 41 CTA from patients with RCC were analysed, with a median age of 66 years (57-70) and 80.5% of men. The median PADUA score assessed with the commercialized software was 9 (7-11) and the renal score was 8 (5.5-9). The automatic pipeline enabled to automatically detect the tumoral kidney and provided a 3D-visualization in all cases, with a computational time less than 20 seconds. Compared to the ground truth, concordances for staging the anatomical features of the RENAL scores were respectively: 87.8% for radius, 85.4% for exophytic rate, 82.9% for location to the polar lines and 92.7% for the antero-posterior location. For the PADUA scores, concordances were 90.2% for tumor size, 85.4% for exophytic rate, 87.8% for polar location and 100% for renal rim. CONCLUSIONS: The automatic software allowed to accurately detect and provide a 3D-visualization of the tumoral kidney in all cases, with a good concordance with the ground truth to assess the RENAL and PADUA scores. By adding further algorithms, this software could be trained to automatically calculate morphometric scores, saving time and improving reproducibility for clinicians. Source of Funding: None © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e279-e279 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Mathieu Carlier More articles by this author Fabien Lareyre More articles by this author Cédric Adam More articles by this author Marion Carrier More articles by this author Matthieu Durand More articles by this author Juliette Raffort More articles by this author Expand All Advertisement Loading ...