Principal Investigators:
Tamara Lotan, MD (Johns Hopkins University)
Angelo De Marzo, MD, PhD (Johns Hopkins University)

Co-Investigators: Corrine Joshu, PhD (Johns Hopkins University), Bruce Trock, PhD (Johns Hopkins University), Misop Han, MD, (Johns Hopkins University)

Description:

• The Gleason grading system has been a standard grading system for prognosticating prostate cancer by pathologists for decades, and is a manual determination made of how abnormal the tumor area looks under a microscope. However, standard pathology procedures for grading prostate cancer are time consuming and imperfect.
• Digital pathology is revolutionizing the way pathologists diagnose and grade malignancies.
• Lotan and team are developing artificial intelligence (AI)-based digital pathology technologies to improve the precision and reproducibility of Gleason grading, and to more accurately predict clinical outcomes and molecular tumor characteristics.
• In this project, Dr. Lotan and team will use samples from prostate cancer biopsy and radical prostatectomy cohorts to improve and validate AI algorithms for prediction of Gleason grade, disease progression, metastasis, and recurrence in prostate cancer.
• Whether AI algorithms offer improved prediction of clinical outcomes compared with classical pathology grading systems, and when combined with clinical-pathological variables, will be determined.
• AI algorithms will also be developed that can identify prostate cancer molecular subclasses with specific tumor mutations.
• If successful, this project will result in the development of validated AI systems that offer more accurate and rapid prostate cancer prognostication, which will improve treatment selection and clinical outcomes for patients.

What this means to patients: Digital pathology and artificial intelligence technologies offer the opportunity for predicting oncologic outcomes and tumor molecular characteristics more accurately than the human eye. In this project, Dr. Lotan and team will develop and validate AI systems with high accuracy for predicting prostate cancer aggressiveness, tumor molecular characteristics, and clinical outcomes, which will enable clinicians to choose more appropriate treatment strategies for patients.