Comprehensive genomic profiling has become routine in modern oncology. Yet despite major advances in sequencing technologies, translating complex molecular data into effective treatment decisions remains a significant challenge.

At the 2026 ASCO Annual Meeting, researchers from Genomate Health will present new findings demonstrating how computational reasoning applied to one of the world’s largest real-world clinicogenomic datasets may help uncover clinically meaningful treatment opportunities that remain underutilized in current practice.

The study, titled “Clinical utility analysis of a computational reasoning system on large-scale tumor genomic data,” analyzed 20,923 patient tumor profiles from the MSK-CHORD real-world clinicogenomic dataset by Memorial Sloan Kettering Cancer Center (MSK), New York, USA, which has been ranked as the number one cancer center in the world multiple times. The dataset integrates large-scale molecular and clinical outcome data across multiple cancer types, providing a uniquely valuable resource for evaluating precision oncology approaches in real-world settings. The goal of the study was to determine whether the Genomate computational platform can identify additional relevant treatment options compared with the state-of-the-art practice at a premier cancer center. 

Presenter: Dóra Lakatos, PhD, Head of Computational Science, Genomate Health
Section: Care Delivery / Models of Care
Abstract Number: 1625
Date & Time: May 30, 2026, 9:00 - 12:00 PM CDT
Location: Hall A, McCormick Place, Chicago

‍

Understanding the full molecular complexity of tumors

Cancer is rarely driven by a single genomic alteration. Tumors typically contain multiple interacting driver alterations that collectively shape treatment response and resistance. In the MSK_CHORD dataset, we identified an average of 4.6 driver alterations per patient, with some tumors harboring as many as 45 distinct drivers. These findings further reinforce the biological complexity of cancer and the limitations of approaches focused only on isolated “actionable biomarkers.”

To address this challenge, researchers applied Digital Drug Assignment (DDA), commercially known as Genomate®, a computational reasoning system designed to evaluate therapies based on the totality of a patient’s molecular profile. Rather than matching one mutation to one drug, DDA integrates functional and pharmacological evidence across the entire genomic context of the tumor and scores molecularly targeted agents according to predicted clinical relevance.

Previous studies have already demonstrated the predictive value of DDA in both clinical trial and real-world settings, including analyses of the SHIVA01 trial and large lung cancer cohorts.

‍

A substantial treatment gap identified across real-world oncology care

Previous research has shown that MTAs assigned a high Genomate score provide increased clinical benefit (NPJ Precis Oncol. 2021 5:59; NPJ Precis Oncol. 2025 9:159). The analysis revealed a striking finding: 75% of tumors in the dataset were associated with at least one high-score molecularly targeted therapy option when analyzed using computational reasoning.

However, among patients with known treatment histories, only 37% actually received a high-score therapy during their treatment course. More specifically, of the first-line targeted therapies administered following NGS testing, only 26.5% were assigned a high score defined as therapies strongly supported by Genomate’s computational reasoning analysis.

This substantial gap highlights a potentially important gap between the treatment opportunities suggested by comprehensive molecular interpretation supported by Genomate® and the therapies ultimately administered in clinical practice. The findings were observed across multiple tumor types, including breast, colorectal, lung, pancreatic, and prostate cancers. These findings underscore the ongoing challenge of translating genomic complexity into actionable treatment decisions.

‍

Implications for precision oncology and drug repurposing

Beyond identifying treatment gaps, the analysis also demonstrated a high prevalence of on-label high-score therapy associations across tumor types, indicating that many patients could still receive therapies already approved within their respective oncologic indications. In addition, a high prevalence of off-label high-score therapy associations was observed across tumor types. Many tumors were associated with therapies already approved in other oncologic indications, suggesting potential opportunities for computationally guided drug repurposing strategies. 

The study adds to a growing body of evidence supporting system-level molecular interpretation approaches in oncology, particularly as genomic datasets continue to expand in scale and complexity.

As precision oncology moves beyond single biomarker paradigms, approaches capable of evaluating the complete molecular context of each patient’s tumor may become increasingly important for supporting more informed and individualized treatment decisions.

Presented at ASCO 2026

This research will be presented at the ASCO Annual Meeting 2026 as a regular poster. It is part of a broader set of studies from the Genomate Health research team exploring how computational reasoning can support precision oncology. Read the full abstract here: https://meetings.asco.org/meetings/2026-asco-annual-meeting/335/17086?presentation=263141

If you are attending ASCO 2026, we invite you to visit our poster presentations and meet the Genomate Health team to learn more about our research. Schedule a meeting with us during the conference to discuss the research and explore potential collaborations.