During the 2026 American Society of Clinical Oncology Genitourinary Cancers Symposium (ASCO GU), a major breakthrough drew global attention: the first international external validation of the digital pathology–based AI predictive model MMAI in localized prostate cancer. Oncology Frontier – UroStream conducted an on-site interview with Prof. Anna Wilkins, Honorary Consultant in Clinical Oncology at The Institute of Cancer Research and The Royal Marsden Hospital (UK), who provided an in-depth interpretation of the study’s clinical value, real-world implementation prospects, and the broader advances highlighted at this year’s congress. Oncology Frontier: First, could you elaborate on the core research findings you presented at this ASCO GU congress?
Dr. Anna Wilkins:
At this congress, I presented the first international, independent external validation of the MMAI assay in clinical trial cohorts of localized prostate cancer.
MMAI, or Multimodal Artificial Intelligence pathology model—also known as Machine-Learning Morphometric AI—is a digital pathology–based prognostic model developed by ArteraAI in the United States. It is specifically designed to support precision diagnosis and treatment in prostate cancer, and it represents one of the most extensively validated AI tools in oncology pathology, supported by phase III randomized controlled trial data.
The model was developed to address key limitations in traditional prognostic assessment. Conventional clinical scoring systems often fail to fully capture intratumoral heterogeneity, leading to suboptimal risk stratification. Meanwhile, genomic classifiers, although informative, are associated with high costs, additional tissue requirements, and long turnaround times, limiting their widespread adoption. In contrast, MMAI leverages routinely prepared hematoxylin–eosin (H&E) stained slides, extracting prognostic information directly from standard pathological specimens without the need for additional testing, making it both cost-effective and highly accessible.
The assay requires only standard H&E slides obtained during routine diagnostic workflows. After digital scanning, the images are analyzed by the AI algorithm to generate predictions regarding the risk of disease recurrence and distant metastasis following surgery or radiotherapy.
Our validation study included nearly 1,800 patients with localized prostate cancer from the CHHiP trial, a phase III study comparing conventional versus hypofractionated intensity-modulated radiotherapy. The results demonstrated that, compared with current standard risk stratification systems, MMAI significantly improved predictive accuracy for both recurrence and metastasis, providing high-level evidence to support more precise risk stratification in prostate cancer.
Oncology Frontier: Based on these findings, what impact do you think this assay will have on clinical practice? And what about its safety profile?
Dr. Anna Wilkins:
From a safety perspective, the MMAI assay introduces no additional risk. It relies entirely on existing H&E slides generated during routine biopsy or postoperative pathology. No extra procedures, tissue sampling, or patient burden is required. The process involves only digital scanning and computational analysis, making it inherently safe and non-invasive.
Clinically, the most important value of MMAI lies in its ability to guide individualized treatment decisions, particularly by addressing the long-standing challenge of uncertainty in treatment selection.
For patients classified as low risk, MMAI supports treatment de-escalation. It helps identify individuals who may safely avoid unnecessary androgen deprivation therapy (ADT) or opt for active surveillance, thereby preserving quality of life while maintaining disease control.
Conversely, for high-risk patients, the model identifies those who may benefit from treatment intensification. These patients have a significantly higher risk of recurrence and metastasis, and may require more aggressive strategies, such as combining radiotherapy with systemic therapies, to improve long-term outcomes.
Oncology Frontier: Despite its advantages, what challenges remain for broader global implementation of this technology?
Dr. Anna Wilkins:
This is a very practical question. While MMAI integrates well into routine clinical workflows, its implementation depends heavily on digital pathology infrastructure.
Specifically, widespread adoption requires access to high-quality slide scanners and standardized processes for digital slide preparation. In the UK, digital pathology is expanding but has not yet been universally implemented across all institutions. Globally, disparities are even more pronounced, with significant variability in infrastructure between regions.
Therefore, the most pressing challenge is not the algorithm itself, but the readiness of healthcare systems to support digital pathology at scale.
Oncology Frontier: Beyond MMAI, what other highlights from this ASCO GU meeting should clinicians pay close attention to?
Dr. Anna Wilkins:
Several studies presented at this congress are truly practice-changing. In my view, two advances in muscle-invasive bladder cancer (MIBC) stand out.
First, a new neoadjuvant standard of care is emerging. Data presented at this meeting strongly support the combination of enfortumab vedotin, a Nectin-4–targeting antibody–drug conjugate, with pembrolizumab, a PD-1 inhibitor, as a highly effective neoadjuvant regimen. This combination is rapidly establishing itself as a new standard for MIBC.
Second, the role of genomic biomarkers has been further solidified. Multiple high-quality studies demonstrated that genomic profiling can guide treatment decisions across the entire disease continuum, enabling more precise, individualized management of bladder cancer based on molecular characteristics.
Prof. Anna Wilkins
