713: Al-Driven Digital Avatar for Personalized Prostatectomy Planning- Daniele Amparore

This is one of the entries accepted for adjudication for the 2026 KS Innovation Awards, sponsored by the Vattikuti Foundation. Posting here does not imply being a Finalist.

Here is the Abstract:
Al-Driven Digital Avatar for Personalized Prostatectomy Planning

Presenting Author:
1Daniele Amparore, MD, PhD, FEBU Division of Urology, Department of Surgery INOC – Istituto Nazionale Oncologico Candiolo, IRCCS Candiolo (TO), Italy
Phone: +39 333 3352758
Fax: +39 011 9026244
E-mail: daniele.amparore@unito.it

Contributors:
1A. La Gatta, 2M. Colombo, 3A. Simeri, 1E. Checcucci, 1G. Bignante, 2A. Quarà, 2A. Piana, 2S. De Cillis, 2C. Fiori, 1F. Porpiglia 2F. Piramide, 1Department of Surgery National Oncology Institute of Candiolo (INOC) – IRCCS Candiolo, Turin, Italy 2Department of Urology, San Luigi Gonzaga Hospital, University of Turin, Italy 3Department of Mathematics and Computer Science, University of Calabria, Rende, Cs, Italy

Purpose:
To develop and validate an artificial intelligence model to support patient-specific selection of surgical strategies for optimizing early functional outcomes after robot-assisted radical prostatectomy.

Methods:
We retrospectively analyzed 636 patients undergoing robot-assisted radical prostatectomy. Six combinations of intraoperative variables related to bladder neck management and nerve-sparing techniques were evaluated, with at least 12 cases for each strategy. Machine learning models were trained using preoperative and intraoperative variables to predict trifecta achievement at 3 months, defined as urinary continence, satisfactory erectile function, and prostate-specific antigen less than 0.02 ng/mL. Models were evaluated using stratified 5-fold cross-validation and a 70:30 training/test split. Accuracy, precision, recall, and F1-score were assessed.

Results:
The random forest model showed the best overall performance. On the training set, accuracy, precision, recall, and F1-score were 82.0%, 82.1%, 81.9%, and 81.9%, respectively. On the test set, all four metrics were 69.6%. In the test set, 78.9% of negative and 85.0% of positive cases were correctly classified, while 21.1% and 15.0% were misclassified, respectively.

Conclusions:
An artificial intelligence-based model can integrate patient-specific preoperative and intraoperative information to support selection of surgical strategies during robot-assisted radical prostatectomy. These preliminary findings suggest potential for reducing variability in surgical planning and tailoring intraoperative strategies to improve early trifecta outcomes.

See more at: http://vattikutifoundation.com/

Date
Category
KS Awards, Surgeon, Video Library