Performance of a Region of Interest–based Algorithm in Diagnosing International Society of Urological Pathology Grade Group ≥2 Prostate Cancer on the MRI-FIRST Database—CAD-FIRST Study

Thibaut Couchoux (1) , Tristan Jaouen (2) , Christelle Melodelima-Gonindard (3) , Pierre Baseilhac (1) , Arthur Branchu (1) , Nicolas Arfi , Richard Aziza (4, 5) , Nicolas Barry Delongchamps (6) , Franck Bladou (7) , Flavie Bratan , Serge Brunelle (8) , Pierre Colin (9) , Jean-Michel Correas (10, 11, 12) , François Cornud (13) , Jean-Luc Descotes (14) , Pascal Eschwege (15) , Gaelle Fiard (14) , Bénédicte Guillaume (16) , Rémi Grange (17) , Nicolas Grenier (18, 19) , Hervé Lang , Frédéric Lefèvre , Bernard Malavaud (20, 5) , Clément Marcelin , Paul Moldovan , Nicolas Mottet , Pierre Mozer , Eric Potiron , Daniel Portalez , Philippe Puech , Raphaele Renard-Penna , Matthieu Roumiguié (20) , Catherine Roy , Marc-Olivier Timsit , Thibault Tricard , Arnauld Villers , Jochen Walz , Sabine Debeer (1) , Adeline Mansuy (1) , Florence Mège-Lechevallier , Myriam Decaussin-Petrucci , Lionel Badet , Marc Colombel , Alain Ruffion , Sébastien Crouzet , Muriel Rabilloud (21, 22, 23, 24, 25) , Rémi Souchon (2) , Olivier Rouvière

1 Hôpital Edouard Herriot [CHU - HCL]
2 LabTAU - Application des ultrasons à la thérapie
3 LECA - Laboratoire d'Ecologie Alpine
4 Service Radiologie et imagerie médicale - Rangueil / Larrey [CHU Toulouse]
5 IUCT Oncopole - UMR 1037 - Institut Universitaire du Cancer de Toulouse - Oncopole
6 INEM - UM 111 (UMR 8253 / U1151) - Institut Necker Enfants-Malades
7 CHU de Bordeaux Pellegrin [Bordeaux]
8 IPC - Institut Paoli-Calmettes
9 EMA - École, mutations, apprentissages
10 Hôpital Necker - Enfants Malades [AP-HP]
11 UPCité - Université Paris Cité
12 LIB - Laboratoire d'Imagerie Biomédicale [Paris]
13 Service Maladies infectieuses et tropicales [AP-HP Hôpital Cochin]
14 Service d'urologie [CHU Grenoble]
15 UNICANCER/ICL - Institut de Cancérologie de Lorraine - Alexis Vautrin [Nancy]
16 BCL, équipe Linguistique de l’énonciation
17 Service de Radiologie [CHU Saint-Etienne]
18 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
19 COMET - COuplages Multiphysiques Et Transferts - LISN
20 Pôle Urologie - Néphrologie - Dialyse - Transplantations - Brûlés - Chirurgie plastique - Explorations fonctionnelles et physiologiques [CHU Toulouse]
21 Département biostatistiques et modélisation pour la santé et l'environnement [LBBE]
22 LBBE - Laboratoire de Biométrie et Biologie Evolutive - UMR 5558
23 CHU Lyon
24 HCL - Hospices Civils de Lyon
25 Service de Biostatistiques [Lyon]

Background and objective: Prostate multiparametric magnetic resonance imaging (MRI) shows high sensitivity for International Society of Urological Pathology grade group (GG) ≥2 cancers. Many artificial intelligence algorithms have shown promising results in diagnosing clinically significant prostate cancer on MRI. To assess a region-of-interest-based machine-learning algorithm aimed at characterising GG ≥2 prostate cancer on multiparametric MRI. Methods: The lesions targeted at biopsy in the MRI-FIRST dataset were retrospectively delineated and assessed using a previously developed algorithm. The Prostate Imaging-Reporting and Data System version 2 (PI-RADSv2) score assigned prospectively before biopsy and the algorithm score calculated retrospectively in the regions of interest were compared for diagnosing GG ≥2 cancer, using the areas under the curve (AUCs), and sensitivities and specificities calculated with predefined thresholds (PIRADSv2 scores ≥3 and ≥4; algorithm scores yielding 90% sensitivity in the training database). Ten predefined biopsy strategies were assessed retrospectively. Key findings and limitations: After excluding 19 patients, we analysed 232 patients imaged on 16 different scanners; 85 had GG ≥2 cancer at biopsy. At patient level, AUCs of the algorithm and PI-RADSv2 were 77% (95% confidence interval [CI]: 70-82) and 80% (CI: 74-85; p = 0.36), respectively. The algorithm's sensitivity and specificity were 86% (CI: 76-93) and 65% (CI: 54-73), respectively. PI-RADSv2 sensitivities and specificities were 95% (CI: 89-100) and 38% (CI: 26-47), and 89% (CI: 79-96) and 47% (CI: 35-57) for thresholds of ≥3 and ≥4, respectively. Using the PI-RADSv2 score to trigger a biopsy would have avoided 26-34% of biopsies while missing 5-11% of GG ≥2 cancers. Combining prostate-specific antigen density, the PI-RADSv2 and algorithm's scores would have avoided 44-47% of biopsies while missing 6-9% of GG ≥2 cancers. Limitations include the retrospective nature of the study and a lack of PI-RADS version 2.1 assessment. Conclusions and clinical implications: The algorithm provided robust results in the multicentre multiscanner MRI-FIRST database and could help select patients for biopsy. Patient summary: An artificial intelligence-based algorithm aimed at diagnosing aggressive cancers on prostate magnetic resonance imaging showed results similar to expert human assessment in a prospectively acquired multicentre test database.

Mots clés: Artificial intelligence; Magnetic resonance imaging; Prostate biopsy; Prostate cancer; Radiomics

Domaines: Statistiques [stat]

Thibaut Couchoux, Tristan Jaouen, Christelle Melodelima-Gonindard, Pierre Baseilhac, Arthur Branchu, et al.. Performance of a Region of Interest–based Algorithm in Diagnosing International Society of Urological Pathology Grade Group ≥2 Prostate Cancer on the MRI-FIRST Database—CAD-FIRST Study. European Urology Oncology, 2024, S2588-9311 (24), pp.00056-7. ⟨10.1016/j.euo.2024.03.003⟩. ⟨hal-04549487⟩