Implementation of an AI-enabled workflow at a community health system was associated with increased biomarker testing completion and more timely testing among patients with non–small cell lung cancer (NSCLC). The findings came from a retrospective and prospective evaluation presented at the 2026 ASCO Quality Care Symposium (Abstract 508).  

“Following implementation of an AI-assisted identification and clinician notification workflow, biomarker testing completion and testing within 30 days improved across NSCLC patient populations, with the largest absolute increase in testing completion observed among patients with unknown-stage disease,” lead study author Yunus Emre Demirhan, MD, of Mercy East Side Communities, and co-authors wrote in their poster.  

“These findings support a potential role for AI-guided workflows in supporting guideline-concordant cancer care delivery, particularly in community hospital settings,” they added.  

The investigators conducted a before-and-after evaluation of an AI-enabled workflow designed to accelerate biomarker testing for adults with stage IB to IV or unknown-stage NSCLC at a community health system. The retrospective portion of the study covered July 1 to December 12, 2024 (n = 471), whereas the prospective portion covered February 12, 2025, to April 10, 2026 (n = 1,115). Biomarker testing included EGFR, ALK, and PD-L1 for patients with early-stage disease, and a multigene panel, PD-L1, and HER2 immunohistochemistry for those with advanced-stage or unknown-stage disease.  

The AI-enabled workflow, Tempus Next, involved nightly ingestion of clinical data, AI-assisted data extraction, clinical curation, identification of testing gaps, and clinician notifications.  

During the prospective portion of the study, 137 patients were identified for review, and clinicians were notified about 120 of them. Of these, 24 subsequently completed biomarker testing, 83 had no documented testing, and 13 were no longer eligible for care.  

These persistent gaps after notification “highlight the need for coordinated follow-up to support completion of clinically appropriate testing,” the study authors noted.  

Retrospective analysis showed that 71.1% of patients completed biomarker testing, compared with 90% after implementation of the AI-enabled workflow. Among patients with early-stage disease, testing completion increased from 68% before implementation to 93.2% afterward. Among those with advanced-stage NSCLC, the rate increased from 82.2% to 91.9%. The largest absolute increase was observed among patients with unknown-stage disease, for whom testing completion increased from 49.6% to 76.8%.  

Comparing the first quarters of 2025 and 2026, the rate of testing completion within 30 days increased significantly among patients with advanced-stage or unknown-stage NSCLC. Multigene panel testing completion increased from 52% in Q1 2025 to 84% in Q1 2026, whereas PD-L1 testing completion increased from 54% to 87%. HER2 immunohistochemistry testing completion increased from 37% to 92% over the same period.  

DISCLOSURES: The study authors reported no conflicts of interest. 

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.