Histopathological classification in prostate cancer remains a challenge with high dependence on the expert practitioner. We develop a deep learning (DL) model to identify the most prominent Gleason pattern in a highly curated data cohort and validate it on an independent dataset. The histology images are partitioned in tiles (14,509) and are curated by an expert to identify individual glandular structures with assigned primary Gleason pattern grades. We use transfer learning and fine-tuning approaches to compare several deep neural network architectures that are trained on a corpus of camera images (ImageNet) and tuned with histology examples to be context appropriate for histopathological discrimination with small samples. In our study, the best DL network is able to discriminate cancer grade (GS3/4) from benign with an accuracy of 91%, F1-score of 0.91 and AUC 0.96 in a baseline test (52 patients), while the cancer grade discrimination of the GS3 from GS4 had an accuracy of 68% and AUC of 0.71 (40 patients).
Cancers. 2023 Apr 17*** epublish ***
Ryan Fogarty, Dmitry Goldgof, Lawrence Hall, Alex Lopez, Joseph Johnson, Manoj Gadara, Radka Stoyanova, Sanoj Punnen, Alan Pollack, Julio Pow-Sang, Yoganand Balagurunathan
Department of Machine Learning, H. Lee Moffitt Cancer Center, Tampa, FL 33612, USA., Department of Computer Science and Engineering, University of South Florida, Tampa, FL 33620, USA., Tissue Core Facility, H. Lee Moffitt Cancer Center, Tampa, FL 33612, USA., Analytic Microscopy Core Facility, H. Lee Moffitt Cancer Center, Tampa, FL 33612, USA., Anatomic Pathology Division, H. Lee Moffitt Cancer Center, Tampa, FL 33612, USA., Department of Radiation Oncology, University of Miami Miller School of Medicine, Miami, FL 33136, USA., Desai Sethi Urology Institute, University of Miami Miller School of Medicine, Miami, FL 33136, USA., Genitourinary Cancers, H. Lee Moffitt Cancer Center, Tampa, FL 33612, USA.