Mp37-10 Exploration and Construction of Puncture-Free and Function-Preserving Prostate Cancer Diagnosis and Treatment System

You have accessJournal of UrologyProstate Cancer: Localized: Surgical Therapy I (MP37)1 May 2024MP37-10 EXPLORATION AND CONSTRUCTION OF PUNCTURE-FREE AND FUNCTION-PRESERVING PROSTATE CANCER DIAGNOSIS AND TREATMENT SYSTEM Zengjun Wang Zengjun WangZengjun Wang View All Author Informationhttps://doi.org/10.1097/01.JU.0001008948.02935.01.10AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: To explore the accuracy, safety and feasibility of puncture-free radical prostatectomy with preservation of urinary control and ejaculatory function in highly suspicious prostate cancer patients based on artificial intelligence and imaging histology technology. METHODS: We used artificial intelligence deep learning (AI-DL) to construct a focal tumor prediction model for the suspected prostate cancer population , and determined the presence and location of non-primary foci of prostate cancer through imaging histology and machine learning models. RESULTS: We accumulated 125 prostate cancer patients with a single, clearly confined lesion on MRI, and after robot-assisted radical prostatectomy for prostate cancer, their resected specimens were completely and adequately sampled for pathologic testing. It was found that 56 of them were confirmed to have a single lesion by postoperative pathology, while 69 were confirmed to have multifocal or diffuse prostate cancer, i.e., the presence of cancerous foci that were not visualized on prostate MRI, by postoperative pathology. We divided all cases into two groups based on the presence of foci that did not show up on MRI by postoperative pathology, and randomly divided them into a training set and a validation set in a ratio of 3:1. The area under the ROC curve of the training set was 0.8654 and the area under the ROC curve of the validation set was 0.8908, and the sensitivity and specificity of the validation set for diagnosing the presence of an opacified cancerous lesion were 0.8235 and 0.8571, respectively. CONCLUSIONS: Our model has higher sensitivity and specificity for diagnosing prostate cancer, which can be treated directly with minimally invasive robotic-assisted urethral vesicoureteral prostatectomy (Madigan Prostatectomy) without preoperative prostate puncture biopsy, avoiding the missed diagnosis and misdiagnosis of biopsy as well as the pain of poor prognosis of traditional surgery, such as postoperative complications, for patients. Source of Funding: No funding information available © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e606 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Zengjun Wang More articles by this author Expand All Advertisement PDF downloadLoading ...