Prostate cancer therapy personalization via multi-modal deep learning on randomized phase III clinical trials

Abstract

Prostate cancer is the most frequent cancer in men and a leading cause of cancer death. Determining a patient’s optimal therapy is a challenge, where oncologists must select a therapy with the highest likelihood of success and the lowest likelihood of toxicity. International standards for prognostication rely on non-specific and semi-quantitative tools, commonly leading to over- and under-treatment. Tissue-based molecular biomarkers have attempted to address this, but most have limited validation in prospective randomized trials and expensive processing costs, posing substantial barriers to widespread adoption. There remains a significant need for accurate and scalable tools to support therapy personalization. Here we demonstrate prostate cancer therapy personalization by predicting long-term, clinically relevant outcomes using a multimodal deep learning architecture and train models using clinical data and digital histopathology from prostate biopsies. We train and validate models using five phase III randomized trials conducted across hundreds of clinical centers. Histopathological data was available for 5654 of 7764 randomized patients (71%) with a median follow-up of 11.4 years. Compared to the most common risk-stratification tool—risk groups developed by the National Cancer Center Network (NCCN)—our models have superior discriminatory performance across all endpoints, ranging from 9.2% to 14.6% relative improvement in a held-out validation set. This artificial intelligence-based tool improves prognostication over standard tools and allows oncologists to computationally predict the likeliest outcomes of specific patients to determine optimal treatment. Outfitted with digital scanners and internet access, any clinic could offer such capabilities, enabling global access to therapy personalization.

Prostate Cancer: Advances in Radiation Oncology, Molecular Biology, and Future Treatment Strategies

“…The data involved pathology samples from 5,654 trial patients with high and sufficient quality digital histopathology image data. The results confirmed that artificial intelligence model did outperform traditional clinical risk stratification for predicting outcome, therefore improvements in personalization strategies will help identify patients who could benefit from augmented therapy and potentially titrate therapy for those with favorable features (35).…”

“…Prolaris, Decipher, and Oncotype genomic profiling testing are available to patients to help define molecular signaling that may suggest a different disease process than implied by traditional biomarkers and tools used to assign risk. In the future, next generation sequencing may be used to complement more traditional biomarkers defined on immunohistochemical staining including markers for neuroendocrine expression (33)(34)(35).…”

Evaluation of Patients for Radiotherapy for Prostate Adenocarcinoma

“…Additionally, while the usage of artificial intelligence is still an NCCN category IIB recommendation at this time for aiding with risk stratification, new studies suggest it will have an increasing role in cancer therapy precision [8,71].…”

2022 European Society for Medical Oncology: Meeting highlights

“…Genomic classifiers and artificial intelligence-based tools using digital histopathology may provide additional guidance. 5,6 PRESTO focused on patients that had progressed biochemically despite maximal local therapy to the pelvis, including salvage radiotherapy. 7 In this setting, intermittent ADT is a standard of care (SoC).…”