학술논문
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.
Document Type
Academic Journal
Author
Couchoux T; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Jaouen T; LabTau, INSERM Unit 1032, Lyon, France.; Melodelima-Gonindard C; Laboratoire d'écologie Alpine, CNRS, UMR 5553, Grenoble, France; Université Grenoble Alpes, Grenoble, France.; Baseilhac P; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Branchu A; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Arfi N; Department of Urology, Hôpital Saint Joseph Saint Luc, Lyon, France.; Aziza R; Department of Radiology, Institut Universitaire du Cancer de Toulouse, Toulouse, France.; Barry Delongchamps N; Department of Urology, Hôpital Cochin, Assistance Publique-Hôpitaux de Paris, Paris, France.; Bladou F; Department of Urology, Centre Hospitalier Universitaire de Bordeaux, Bordeaux, France.; Bratan F; Department of Diagnostic and Interventional Imaging, Hôpital Saint Joseph Saint Luc, Lyon, France.; Brunelle S; Department of Radiology and Medical Imaging, Institut Paoli-Calmettes Cancer Center, Marseille, France.; Colin P; Department of Urology, Hôpital privé La Louvrière, Lille, France.; Correas JM; Department of Radiology, Hôpital Necker, Assistance Publique-Hôpitaux de Paris, Paris, France.; Cornud F; Department of Radiology, Hôpital Cochin, Assistance Publique-Hôpitaux de Paris, Paris, France.; Descotes JL; Université Grenoble Alpes, Grenoble, France; Department of Urology, Centre Hospitalier Universitaire de Grenoble, Grenoble, France.; Eschwege P; Department of Urology, Centre Hospitalier Régional et Universitaire de Nancy, Vandoeuvre, France.; Fiard G; Université Grenoble Alpes, Grenoble, France; Department of Urology, Centre Hospitalier Universitaire de Grenoble, Grenoble, France.; Guillaume B; Department of Radiology, Centre Hospitalier Universitaire de Grenoble, Université Grenoble Apes, Grenoble, France.; Grange R; Department of Radiology, University Hospital of Saint-Etienne, Saint-Priest-en-Jarez, France.; Grenier N; Department of Radiology, Centre Hospitalier Universitaire de Bordeaux, Hôpital Pellegrin, Bordeaux, France.; Lang H; Department of Urology, Centre Hospitalier Universitaire de Strasbourg, Nouvel Hôpital Civil, Strasbourg, France.; Lefèvre F; Department of Radiology, Centre Hospitalier Régional et Universitaire de Nancy, Vandoeuvre, France.; Malavaud B; Department of Urology, Institut Universitaire du Cancer de Toulouse, Toulouse, France.; Marcelin C; Department of Radiology, Centre Hospitalier Universitaire de Bordeaux, Hôpital Pellegrin, Bordeaux, France.; Moldovan PC; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Mottet N; Department of Urology, University Hospital of Saint-Etienne, Saint-Priest-en-Jarez, France.; Mozer P; Department of Urology, Hôpital Pitié-Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, France.; Potiron E; Clinique Urologique de Nantes, Saint-Herblain, France.; Portalez D; Department of Radiology, Institut Universitaire du Cancer de Toulouse, Toulouse, France.; Puech P; Department of Radiology, Centre Hospitalier Régional et Universitaire de Lille, Lille, France.; Renard-Penna R; Department of Radiology, Hôpital Pitié-Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, France; GRC no 5, ONCOTYPE-URO, Sorbonne Universités, Paris, France.; Roumiguié M; Department of Urology, Toulouse-Rangueil University Hospital, Toulouse France.; Roy C; Department of Radiology B, Centre Hospitalier Universitaire de Strasbourg, Nouvel Hôpital Civil, Strasbourg, France.; Timsit MO; Department of Urology, Hôpital Européen Georges Pompidou, Assistance Publique-Hôpitaux de Paris, Paris, France.; Tricard T; Department of Urology, Centre Hospitalier Universitaire de Strasbourg, Nouvel Hôpital Civil, Strasbourg, France.; Villers A; Department of Urology, Univ. Lille, CHU Lille, Lille, France.; Walz J; Department of Urology, Institut Paoli-Calmettes Cancer Center, Marseille, France.; Debeer S; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Mansuy A; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France.; Mège-Lechevallier F; Department of Pathology, Hospices Civils de Lyon, Pierre-Bénite, France.; Decaussin-Petrucci M; Department of Pathology, Hospices Civils de Lyon, Pierre-Bénite, France.; Badet L; Department of Urology, University Hospital of Saint-Etienne, Saint-Priest-en-Jarez, France; Department of Urology, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France; Université Lyon 1, Université de Lyon, Lyon, France.; Colombel M; Department of Urology, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France; Université Lyon 1, Université de Lyon, Lyon, France.; Ruffion A; Université Lyon 1, Université de Lyon, Lyon, France; Department of Urology, Centre Hospitalier Lyon Sud, Hospices Cibvils de Lyon, Pierre-Bénite, France.; Crouzet S; LabTau, INSERM Unit 1032, Lyon, France; Department of Urology, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France; Université Lyon 1, Université de Lyon, Lyon, France.; Rabilloud M; Université Lyon 1, Université de Lyon, Lyon, France; Pôle Santé Publique, Service de Biostatistique et Bioinformatique, Hospices Civils de Lyon, Lyon, France; CNRS, UMR 5558, Laboratoire de Biométrie et Biologie Évolutive, Équipe Biostatistique-Santé, Villeurbanne, France.; Souchon R; LabTau, INSERM Unit 1032, Lyon, France.; Rouvière O; Department of Urinary and Vascular Imaging, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France; LabTau, INSERM Unit 1032, Lyon, France; Université Lyon 1, Université de Lyon, Lyon, France. Electronic address: olivier.rouviere@netcourrier.com.
Source
Publisher: Elsevier B.V Country of Publication: Netherlands NLM ID: 101724904 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 2588-9311 (Electronic) Linking ISSN: 25889311 NLM ISO Abbreviation: Eur Urol Oncol Subsets: MEDLINE
Subject
Language
English
Abstract
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.
(Copyright © 2024 The Author(s). Published by Elsevier B.V. All rights reserved.)
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.
(Copyright © 2024 The Author(s). Published by Elsevier B.V. All rights reserved.)