TY - JOUR
T1 - Predicting response to intravesical BCG in high-risk non-muscle invasive bladder cancer using an artificial intelligence-powered pathology assay
T2 - Development and validation in an international 12 center cohort
AU - Lotan, Yair
AU - Krishna, Viswesh
AU - Abuzeid, Waleed M.
AU - Launer, Bryn
AU - Chang, Sam S.
AU - Krishna, Vrishab
AU - Shingi, Siddhant
AU - Gordetsky, Jennifer B.
AU - Gerald, Thomas
AU - Woldu, Solomon
AU - Shkolyar, Eugene
AU - Hayne, Dickon
AU - Redfern, Andrew
AU - Spalding, Lisa
AU - Stewart, Courtney
AU - Eyzaguirre, Eduardo
AU - Imtiaz, Shamsunnahar
AU - Narayan, Vikram M.
AU - Packiam, Vignesh T.
AU - O'Donnell, Michael A.
AU - Li, Roger
AU - Baekelandt, Loic
AU - Joniau, Steven
AU - Zuiverloon, Tahlita
AU - Fernandez, Mario I.
AU - Schultz, Marcela
AU - Hensley, Patrick J.
AU - Allison, Derek
AU - Taylor, John A.
AU - Hamza, Ameer
AU - Kamat, Ashish
AU - Nimgaonkar, Vivek
AU - Sonawane, Snehal
AU - Miller, Daniel L.
AU - Watson, Drew
AU - Vrabac, Damir
AU - Joshi, Anirudh
AU - Shah, Jay B.
AU - Williams, Stephen B.
N1 - Publisher Copyright:
© 2024 The Author(s).
PY - 2024
Y1 - 2024
N2 - Purpose:There are few markers to identify those likely to recur or progress after treatment with intravesical BCG. We developed and validated artificial intelligence-based histologic assays that extract interpretable features from transurethral resection of bladder tumor digitized pathology images to predict risk of recurrence, progression, development of BCG unresponsive disease, and cystectomy.Materials and Methods:Pre-BCG resection-derived whole-slide images and clinical data were obtained for high-risk non-muscle invasive bladder cancer cases treated with BCG from 12 centers and were analyzed through a segmentation and feature extraction pipeline. Features associated with clinical outcomes were defined and tested on independent development and validation cohorts. Cases were classified into high or low risk for recurrence, progression, BCG unresponsive disease, and cystectomy.Results:944 cases (development:303, validation:641, median follow-up:36 months) representative of the intended use population were included (high-grade Ta:34.1%, high-grade T1:54.8%; carcinoma-in-situ only:11.1%, any carcinoma-in-situ:31.4%). In the validation cohort, "High recurrence risk"cases had inferior high-grade recurrence-free survival versus "Low recurrence risk"cases (HR 2.08, p<0.0001). "High progression risk"patients had poorer progression-free survival (HR 3.87, p<0.001) and higher risk of cystectomy (HR 3.35, p<0.001) than "Low progression risk". Cases harboring the BCG unresponsive disease signature had a shorter time to development of BCG unresponsive disease than cases without the signature (HR 2.31, p<0.0001). AI assays provided predictive information beyond clinicopathologic factors.Conclusions:We developed and validated AI-based histologic assays that identify high-risk non-muscle invasive bladder cancer cases at higher risk of recurrence, progression, BCG unresponsive disease, and cystectomy, potentially aiding clinical decision-making.
AB - Purpose:There are few markers to identify those likely to recur or progress after treatment with intravesical BCG. We developed and validated artificial intelligence-based histologic assays that extract interpretable features from transurethral resection of bladder tumor digitized pathology images to predict risk of recurrence, progression, development of BCG unresponsive disease, and cystectomy.Materials and Methods:Pre-BCG resection-derived whole-slide images and clinical data were obtained for high-risk non-muscle invasive bladder cancer cases treated with BCG from 12 centers and were analyzed through a segmentation and feature extraction pipeline. Features associated with clinical outcomes were defined and tested on independent development and validation cohorts. Cases were classified into high or low risk for recurrence, progression, BCG unresponsive disease, and cystectomy.Results:944 cases (development:303, validation:641, median follow-up:36 months) representative of the intended use population were included (high-grade Ta:34.1%, high-grade T1:54.8%; carcinoma-in-situ only:11.1%, any carcinoma-in-situ:31.4%). In the validation cohort, "High recurrence risk"cases had inferior high-grade recurrence-free survival versus "Low recurrence risk"cases (HR 2.08, p<0.0001). "High progression risk"patients had poorer progression-free survival (HR 3.87, p<0.001) and higher risk of cystectomy (HR 3.35, p<0.001) than "Low progression risk". Cases harboring the BCG unresponsive disease signature had a shorter time to development of BCG unresponsive disease than cases without the signature (HR 2.31, p<0.0001). AI assays provided predictive information beyond clinicopathologic factors.Conclusions:We developed and validated AI-based histologic assays that identify high-risk non-muscle invasive bladder cancer cases at higher risk of recurrence, progression, BCG unresponsive disease, and cystectomy, potentially aiding clinical decision-making.
KW - Artificial intelligence
KW - Bladder cancer
KW - Progression
KW - Recurrence
UR - http://www.scopus.com/inward/record.url?scp=85208256030&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85208256030&partnerID=8YFLogxK
U2 - 10.1097/JU.0000000000004278
DO - 10.1097/JU.0000000000004278
M3 - Article
C2 - 39383345
AN - SCOPUS:85208256030
SN - 0022-5347
JO - Journal of Urology
JF - Journal of Urology
M1 - 10.1097/JU.0000000000004278
ER -