Alberto Martini1,2, Akriti Gupta1, Shivaram Cumarasamy1, Sara C Lewis3, Kenneth G Haines4, Alberto Briganti2, Francesco Montorsi2, Ashutosh K Tewari1. 1. Department of Urology, Icahn School of Medicine at Mount Sinai, New York City, New York, USA. 2. Department of Urology, Vita-Salute San Raffaele University, Milan, Italy. 3. Department of Radiology, Icahn School of Medicine at Mount Sinai, New York City, New York, USA. 4. Department of Pathology, Icahn School of Medicine at Mount Sinai, New York City, New York, USA.
Abstract
OBJECTIVES: To create a model that predicts side-specific seminal vesicle invasion using clinical, biopsy and multiparametric magnetic resonance imaging data. METHODS: We analyzed data from 544 patients who underwent robot-assisted radical prostatectomy at a single institution. To develop a side-specific predictive model, we ultimately considered four variables: prostate-specific antigen, highest ipsilateral biopsy Gleason grade, highest ipsilateral percentage core involvement and seminal vesicle invasion on multiparametric magnetic resonance imaging. A binary multivariable logistic regression model was fitted to predict seminal vesicle invasion. A nomogram was then built based on the coefficients of the resulting logit function. The leave-one-out cross validation method was used for internal validation, and the decision curve analysis for the evaluation of the net clinical benefit. RESULTS: We relied on 804 side-specific cases after excluding negative biopsy observations (n = 284). Seminal vesicle invasion was reported on multiparametric magnetic resonance imaging in 41 (5%) cases, and on final pathology in 64 (8%) cases. All variables in the model emerged as predictors of seminal vesicle invasion (all P ≤ 0.001) and were subsequently considered to build a nomogram. The area under the curve of multiparametric magnetic resonance imaging alone in predicting seminal vesicle invasion was 59.1%; whereas one of the clinical variables only was 85.1%. The area under the curve of the nomogram resulting from their combination was 86.5%. After internal validation, this resulted in 84.7%. The model achieved good calibration and the decision curve analysis showed its clinical benefit, especially when compared with relying only on multiparametric magnetic resonance imaging prediction of seminal vesicle invasion. CONCLUSIONS: A nomogram based on clinical and multiparametric magnetic resonance imaging data can predict seminal vesicle invasion and serve as a tool to urologists for surgical planning.
OBJECTIVES: To create a model that predicts side-specific seminal vesicle invasion using clinical, biopsy and multiparametric magnetic resonance imaging data. METHODS: We analyzed data from 544 patients who underwent robot-assisted radical prostatectomy at a single institution. To develop a side-specific predictive model, we ultimately considered four variables: prostate-specific antigen, highest ipsilateral biopsy Gleason grade, highest ipsilateral percentage core involvement and seminal vesicle invasion on multiparametric magnetic resonance imaging. A binary multivariable logistic regression model was fitted to predict seminal vesicle invasion. A nomogram was then built based on the coefficients of the resulting logit function. The leave-one-out cross validation method was used for internal validation, and the decision curve analysis for the evaluation of the net clinical benefit. RESULTS: We relied on 804 side-specific cases after excluding negative biopsy observations (n = 284). Seminal vesicle invasion was reported on multiparametric magnetic resonance imaging in 41 (5%) cases, and on final pathology in 64 (8%) cases. All variables in the model emerged as predictors of seminal vesicle invasion (all P ≤ 0.001) and were subsequently considered to build a nomogram. The area under the curve of multiparametric magnetic resonance imaging alone in predicting seminal vesicle invasion was 59.1%; whereas one of the clinical variables only was 85.1%. The area under the curve of the nomogram resulting from their combination was 86.5%. After internal validation, this resulted in 84.7%. The model achieved good calibration and the decision curve analysis showed its clinical benefit, especially when compared with relying only on multiparametric magnetic resonance imaging prediction of seminal vesicle invasion. CONCLUSIONS: A nomogram based on clinical and multiparametric magnetic resonance imaging data can predict seminal vesicle invasion and serve as a tool to urologists for surgical planning.
Authors: Rocco S Flammia; Benedikt Hoeh; Gabriele Sorce; Francesco Chierigo; Lukas Hohenhorst; Zhen Tian; Jordan A Goyal; Costantino Leonardo; Alberto Briganti; Markus Graefen; Carlo Terrone; Fred Saad; Shahrokh F Shariat; Francesco Montorsi; Felix K H Chun; Michele Gallucci; Pierre I Karakiewicz Journal: Prostate Date: 2022-04-11 Impact factor: 4.012