Your browser doesn't support javascript.
loading
Predicting Prostate Cancer Molecular Subtype With Deep Learning on Histopathologic Images.
Erak, Eric; Oliveira, Lia DePaula; Mendes, Adrianna A; Dairo, Oluwademilade; Ertunc, Onur; Kulac, Ibrahim; Baena-Del Valle, Javier A; Jones, Tracy; Hicks, Jessica L; Glavaris, Stephanie; Guner, Gunes; Vidal, Igor Damasceno; Markowski, Mark; de la Calle, Claire; Trock, Bruce J; Meena, Avaneesh; Joshi, Uttara; Kondragunta, Chaith; Bonthu, Saikiran; Singhal, Nitin; De Marzo, Angelo M; Lotan, Tamara L.
Afiliación
  • Erak E; Department of Pathology, Johns Hopkins University School of Medicine.
  • Oliveira LD; Department of Pathology, Johns Hopkins University School of Medicine.
  • Mendes AA; Department of Pathology, Johns Hopkins University School of Medicine.
  • Dairo O; Department of Pathology, Johns Hopkins University School of Medicine.
  • Ertunc O; Department of Pathology, Suleyman Demirel University, Turkey.
  • Kulac I; Koç University School of Medicine, Turkey.
  • Baena-Del Valle JA; Fundacion Santa Fe de Bogota University Hospital, Columbia.
  • Jones T; Department of Pathology, Johns Hopkins University School of Medicine.
  • Hicks JL; Department of Pathology, Johns Hopkins University School of Medicine.
  • Glavaris S; Department of Pathology, Johns Hopkins University School of Medicine.
  • Guner G; Hacettepe University, Turkey.
  • Vidal ID; Department of Pathology, University of Alabama School of Medicine, Alabama.
  • Markowski M; Department of Oncology, Johns Hopkins University School of Medicine.
  • de la Calle C; Department of Urology, Johns Hopkins University School of Medicine.
  • Trock BJ; Department of Urology, Johns Hopkins University School of Medicine.
  • Meena A; AIRA Matrix Private Limited, India.
  • Joshi U; AIRA Matrix Private Limited, India.
  • Kondragunta C; AIRA Matrix Private Limited, India.
  • Bonthu S; AIRA Matrix Private Limited, India.
  • Singhal N; AIRA Matrix Private Limited, India.
  • De Marzo AM; Department of Pathology, Johns Hopkins University School of Medicine; Department of Oncology, Johns Hopkins University School of Medicine; Department of Urology, Johns Hopkins University School of Medicine.
  • Lotan TL; Department of Pathology, Johns Hopkins University School of Medicine; Department of Oncology, Johns Hopkins University School of Medicine; Department of Urology, Johns Hopkins University School of Medicine. Electronic address: tlotan1@jhmi.edu.
Mod Pathol ; 36(10): 100247, 2023 Oct.
Article en En | MEDLINE | ID: mdl-37307876
ABSTRACT
Microscopic examination of prostate cancer has failed to reveal a reproducible association between molecular and morphologic features. However, deep-learning algorithms trained on hematoxylin and eosin (H&E)-stained whole slide images (WSI) may outperform the human eye and help to screen for clinically-relevant genomic alterations. We created deep-learning algorithms to identify prostate tumors with underlying ETS-related gene (ERG) fusions or PTEN deletions using the following 4 stages (1) automated tumor identification, (2) feature representation learning, (3) classification, and (4) explainability map generation. A novel transformer-based hierarchical architecture was trained on a single representative WSI of the dominant tumor nodule from a radical prostatectomy (RP) cohort with known ERG/PTEN status (n = 224 and n = 205, respectively). Two distinct vision transformer-based networks were used for feature extraction, and a distinct transformer-based model was used for classification. The ERG algorithm performance was validated across 3 RP cohorts, including 64 WSI from the pretraining cohort (AUC, 0.91) and 248 and 375 WSI from 2 independent RP cohorts (AUC, 0.86 and 0.89, respectively). In addition, we tested the ERG algorithm performance in 2 needle biopsy cohorts comprised of 179 and 148 WSI (AUC, 0.78 and 0.80, respectively). Focusing on cases with homogeneous (clonal) PTEN status, PTEN algorithm performance was assessed using 50 WSI reserved from the pretraining cohort (AUC, 0.81), 201 and 337 WSI from 2 independent RP cohorts (AUC, 0.72 and 0.80, respectively), and 151 WSI from a needle biopsy cohort (AUC, 0.75). For explainability, the PTEN algorithm was also applied to 19 WSI with heterogeneous (subclonal) PTEN loss, where the percentage tumor area with predicted PTEN loss correlated with that based on immunohistochemistry (r = 0.58, P = .0097). These deep-learning algorithms to predict ERG/PTEN status prove that H&E images can be used to screen for underlying genomic alterations in prostate cancer.
Palabras clave

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Tipo de estudio: Prognostic_studies / Risk_factors_studies Idioma: En Revista: Mod Pathol Asunto de la revista: PATOLOGIA Año: 2023 Tipo del documento: Article

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Tipo de estudio: Prognostic_studies / Risk_factors_studies Idioma: En Revista: Mod Pathol Asunto de la revista: PATOLOGIA Año: 2023 Tipo del documento: Article
...