Researchers at the Icahn School of Medicine at Mount Sinai have developed an advanced artificial intelligence (AI)-driven tool to improve the management and prognosis of prostate cancer.
The tool, PATHOMIQ_PRAD, designed for patients with intermediate-risk prostate cancer, uses deep learning to extract morphological features from datasets derived from biopsy or surgical hematoxylin- and eosin-stained whole-slide images. It aims to identify those at higher risk of rapid disease progression and provide more timely, accurate predictions for earlier interventions and more targeted, personalized, treatment plans.
“About 60 percent of patients in the intermediate-risk group don’t have a clear treatment plan, and around 30 to 50 percent see their cancer progress after the first round of therapy. We’re finding that some of these patients are at higher risk for rapid progression, so identifying them early is critical,” says co-corresponding author Ash Tewari, MBBS, MCh, FRCS (Hon.), DSc (Hon.), Professor and Chair, Milton and Carroll Petrie Department of Urology, Mount Sinai Health System and Director of the Center of Excellence for Prostate Cancer at the Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai. “We developed this tool to analyze samples from biopsies or surgeries, providing a clearer understanding of which patients may require more aggressive treatment earlier to improve their outcomes. PATHOMIQ_PRAD has the potential to become a routine part of clinical decision-making.”
The figure on the left provides a schematic overview of the Pathomiq_PRAD AI pipeline, including the pre-processing steps and deep learning algorithms used for morphology quantification and risk-score prediction. Figure on right shows a cluster map of different prostate cancer patterns detected by the AI tool PATHOMIQ_PRAD from stained tissue images. The boxes highlight close-up views of various tissue structures in the cancer and surrounding area. Nair, et al. (2024). A novel artificial intelligence–powered tool for precise risk stratification of prostate cancer progression in patients with clinical intermediate risk. European Urology: https://doi.org/10.1016/j.eururo.2024.07.013
PATHOMIQ_PRAD scores range from 0 to 1, with higher scores indicating high-risk features. The study analyzed large datasets to classify patients into high- and low-risk groups using pre-determined PATHOMIQ_PRAD clinical cut offs of 0.45 for BCR and 0.55 for metastasis. These limits were based on factors such as the chances of cancer returning or spreading. The study found that the tool outperformed existing benchmark cancer outcomes over the next five years compared to other current tools.
“One key advantage of the tool is its ability to analyze specific regions of tissue that may hold clues to previously undiscovered drivers of prostate cancer progression. This insight could potentially lead to advancements in understanding racial disparities in prostate cancer outcomes, helping us explore why certain populations face more aggressive disease,” says co-corresponding author Sujit S. Nair, PhD, Assistant Professor of Urology and Director of GU Immunotherapy Research in the Department of Urology at the Icahn School of Medicine at Mount Sinai. “These developments are exciting, and we are working toward further validation studies.”
Details on the findings were reported in the September 2024 online issue of European Urology. The paper is titled “A Novel Artificial Intelligence-Powered Tool for Precise Risk Stratification of Prostate Cancer Progression in Patients With Clinical Intermediate Risk.”
The researchers worked in collaboration with PATHOMIQ INC., based in Cupertino, California, to develop PATHOMIQ_PRAD. Next, the researchers plan to conduct large-scale clinical validation studies with a more diverse patient population. They also aim to pursue regulatory approval to develop PATHOMIQ_PRAD as a Lab Developed Test, allowing its use in CLIA-certified labs. Additionally, the Icahn School of Mount Sinai at Mount Sinai and PATHOMIQ teams are integrating the tool with advanced genomic profiling techniques, such as spatial transcriptomics and mass cytometry, to enhance understanding of the biology behind the regions identified by PATHOMIQ_PRAD.
“Our study pioneers an innovative design, and we’re thrilled about the AI tool’s potential to transform risk stratification for prostate cancer patients, paving the way for future advances,” says Rachel Brody, MD, PhD, Professor of Pathology, Molecular and Cell-Based Medicine at the Icahn School of Medicine at Mount Sinai and a co-author of the study.
“One key advantage of the tool is its ability to analyze specific regions of tissue that may hold clues to previously undiscovered drivers of prostate cancer progression,” says co-corresponding author Sujit S. Nair, PhD, Assistant Professor of Urology and Director of GU Immunotherapy Research in the Department of Urology at the Icahn School of Medicine at Mount Sinai.
The study highlights the clinical potential of this innovative AI tool, which uses digitized whole-slide images to transform prostate cancer.
“By analyzing various tissue types—epithelial, stromal, and immune cells—it generates a detailed score for each patient, predicting outcomes and offering a powerful new way to guide treatment decisions,” says co-corresponding author Dimple Chakravarty, PhD, Assistant Professor of Urology at the Icahn School of Medicine at Mount Sinai. “Ours is the first AI tool designed specifically for intermediate-risk prostate cancer patients that is both scalable and generalizable. It can be used for risk stratification from biopsy and surgery specimens. It’s affordable, quick, and adaptable for use in various health care settings.”