Key points are not available for this paper at this time.
You have accessJournal of UrologyBladder Cancer: Invasive VI (MP77)1 May 2024MP77-12 ARTIFICIAL INTELLIGENCE-BASED MORPHOLOGIC MODEL FOR PREDICTION OF IMMUNOTHERAPY RESPONSE IN LOCALLY ADVANCED AND METASTATIC UROTHELIAL CARCINOMA OF THE BLADDER Ross S. Liao, Monica Nair, Parag Jain, Chensu Xie, Hassan Muhammad, Wei Huang, Hirak Basu, George Wilding, Rajat Roy, Claudia Marcela Diaz-Montero, Tae Hyun Hwang, Scott Dawsey, Jane Nguyen, Eric Klein, Shilpa Gupta, and Omar Mian Ross S. LiaoRoss S. Liao , Monica NairMonica Nair , Parag JainParag Jain , Chensu XieChensu Xie , Hassan MuhammadHassan Muhammad , Wei HuangWei Huang , Hirak BasuHirak Basu , George WildingGeorge Wilding , Rajat RoyRajat Roy , Claudia Marcela Diaz-MonteroClaudia Marcela Diaz-Montero , Tae Hyun HwangTae Hyun Hwang , Scott DawseyScott Dawsey , Jane NguyenJane Nguyen , Eric KleinEric Klein , Shilpa GuptaShilpa Gupta , and Omar MianOmar Mian View All Author Informationhttps://doi.org/10.1097/01.JU.0001009404.49693.12.12AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Immune checkpoint inhibitors (ICI) are key therapeutic options for loco-regionally advanced and metastatic urothelial bladder cancer (UBC). However, only some patients have an objective response, and adverse events warrant careful selection of patients who will elicit the most benefit. Artificial intelligence (AI) technology has the ability to identify clinically-relevant morphologic patterns from digitized pathology slides not routinely recognized on pathologic review to predict response to ICI. METHODS: We analyzed H&E whole slide images (WSI) from transurethral resection of bladder tumor (TURBT) from 116 patients with locally advanced and metastatic UBC at a single institution from 2015-2020. Adjacent multiplex immunohistochemistry (IHC)-stained specimens were analyzed for myeloid and lymphoid markers. 20 patients with complete/partial response or stable disease after ICI were deemed "responders". 20 patients with progressive clinical/radiographic disease despite ICI were deemed "non-responders". WSI were divided into small image patches termed image tiles. Tiles were then processed by multiple AI encoder models to extract morphologic features which were combined by an aggregation model to represent WSI morphology and used to predict response. AUC was used to evaluate performance of response prediction. RESULTS: Our AI model had an AUC of 0.708 for classifying patients (n=40) into responder and non-responder cohorts. Among the top 50% (n=20) with high probability of ICI response predicted by our model, 75% were clinical responders. Among the bottom 50% (n=20) who had high risk of no response to ICI by our model, 75% showed clinical non-response. Multiplex IHC was performed using adjacent sections which yielded deeper mechanistic insights into the immunologic milieu driving predictive morphologic patterns. CONCLUSIONS: By applying an innovative AI-based morphologic analysis to TURBT WSIs, we generated a computational biomarker capable of predicting patient response to immunotherapy for patients with UBC. This approach holds clinical utility for therapeutic patient selection and merits further investigation. Download PPT Source of Funding: Department of Defense, PathomIQ Inc © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e1257 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Ross S. Liao More articles by this author Monica Nair More articles by this author Parag Jain More articles by this author Chensu Xie More articles by this author Hassan Muhammad More articles by this author Wei Huang More articles by this author Hirak Basu More articles by this author George Wilding More articles by this author Rajat Roy More articles by this author Claudia Marcela Diaz-Montero More articles by this author Tae Hyun Hwang More articles by this author Scott Dawsey More articles by this author Jane Nguyen More articles by this author Eric Klein More articles by this author Shilpa Gupta More articles by this author Omar Mian More articles by this author Expand All Advertisement PDF downloadLoading ...
Liao et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: