AI at ASCO: Updates in Precision, Research, and Prevention

This year’s American Society of Clinical Oncology (ASCO) Annual Meeting featured an Artificial Intelligence (AI) Track to showcase how digital technologies are transforming the landscape of cancer care by offering new opportunities for personalized treatments, accelerated research, and improved risk stratification. In this month’s post, we’ll dig into exciting presentations that highlight the impact of AI across the oncology spectrum, from optimizing prostate cancer therapy to streamlining research and paving the way for targeted screening programs.

ArteraAI’s Predictive Power for Treatment De-Escalation

One of the most exciting advancements comes from Artera (previously featured on our podcast), whose team presented validation data at ASCO for their ArteraAI Prostate Test. While treatments like abiraterone, an androgen receptor pathway inhibitor (ARPI), combined with radiation and long-term androgen deprivation therapy (ADT) have significantly improved outcomes for high-risk patients, concerns about side effects, follow-up, and cost can lead to hesitation in its usage.

Artera’s multimodal AI (MMAI) platform is designed to provide personalized prognostic and predictive insights for high-risk prostate cancer patients. The MMAI model analyzed the STAMPEDE trial patient cohort and revealed that only a quarter of high-risk, non-metastatic prostate cancer patients derive meaningful benefit from abiraterone intensification. This finding suggests that up to 75% of such patients can be spared from additional cost and toxicity by de-escalating treatment without compromising survival.

This commercially available test is already used to guide treatment decisions for lower-risk patients (who can be followed with active surveillance) and intermediate-risk patients (predicting benefit from short-term ADT). The latest enhancement for high-risk patients further solidifies ArteraAI’s utility, offering a pathway to optimize patient outcomes while reducing the burden of overtreatment.

Accelerating Research With ASCOmind

Anybody who has been to ASCO knows that the sheer volume of people, posters, presentations, and ideas can be overwhelming—even for a large venue like McCormick Place! ASCOmind is an innovative AI platform developed by the IMO Health team, designed to deliver instant abstract insights. ASCOmind comprises six collaborative AI agents—Pre-processor, Categorizer, MetadataExtractor, Analyzer, Visualizer, and ProtocolMaker—built on a GPT4o-based system.

Tested on multiple myeloma (MM) abstracts from the 2024 meeting, ASCOmind demonstrated efficiency and accuracy. While it would typically take a human reviewer over an hour to manually abstract 51 predefined data elements from an abstract, ASCOmind completed the same task in under 5 minutes per article, with visual summaries generated in under 10 minutes. The system was able to accurately categorize studies, identify patient populations, treatment types, and outcomes, with minimal human correction needed.

By automating and accelerating the analysis of vast amounts of research data, ASCOmind can support quicker decision-making and help clinicians more efficiently extract actionable insights from the burgeoning body of oncology literature.

AI for Cancer Risk Stratification

Beyond treatment and research, AI is also proving to be instrumental in the realm of early detection. Traditional cancer screening programs often face limitations due to difficulty of consistent follow-up, high costs, and potential for false positives, along with many common cancers not having a guideline-approved screening test. Researchers from Rambam Health Care Campus and Technion in Israel showcased the XGBoost machine learning model’s application for 10-year cancer risk stratification using routine health data.

The team developed two XGBoost models. Model 1, a baseline, used only age and gender. The more advanced Model 2 incorporated 53 health indicators, including demographics, lifestyle factors (such as smoking, alcohol, and physical activity), 26 laboratory parameters (eg, complete blood counts, biochemistry, lipids), 9 medication categories, and 7 disease categories including routine periodic health check data, matched with national cancer registry records. Model 2 significantly outperformed Model 1 (area under the curve of 0.799 vs 0.706). Against a population baseline risk of 7.3%, Model 2 identified:

• A low-risk group (bottom 50%) with only a 1.9% 10-year cancer risk
• A high-risk group (90-99th percentile) with a 25.1% risk (3.5-fold increase)
• A very high-risk group (top 1%) with 74.4% risk (10-fold increase)

The most important predictive features included age, monocyte count, albumin, total bilirubin, and low-density lipoprotein (LDL) cholesterol. This bidirectional stratification holds immense promise, suggesting a future where we can reduce screening for low-risk populations while intensifying detection for high risk groups.

Conclusion

In summary, the 2025 ASCO Annual Meeting’s dedicated AI Track made it clear that AI is rapidly reshaping oncology—from personalizing prostate cancer treatment and accelerating research workflows to revolutionizing cancer risk stratification. Tools like ArteraAI are enabling more precise, less toxic care by identifying which patients truly benefit from intensive therapies, while platforms like ASCOmind are making it possible to synthesize and visualize complex research data in minutes rather than hours. Meanwhile, advanced risk models such as the XGBoost approach are paving the way for smarter, more targeted cancer screening strategies.

Collectively, these innovations highlight AI’s growing role in making cancer care more efficient, equitable, and effective, offering hope for better outcomes and a more personalized approach to prevention and treatment.

About Dr. Haque

Waqas Haque, MD, MPH, is a Hematology/Oncology Fellow at the University of Chicago. He recently completed his Internal Medicine Residency at New York University (NYU) in a Clinical Investigator Track. Dr. Haque has balanced his patient care work with a variety of research projects. During his fellowship training at the University of Chicago, he plans to further his work in innovative clinical trial design, value-based care delivery to cancer patients, and clinical investigation.