Replicating clinical trial results through AI: Providence and Microsoft see possibilities through revolutionary new tool

Providence and Microsoft Research have developed an AI framework that could revolutionize clinical trials and bring new treatments into play quicker than before

RENTON, Wash. [Oct. 31, 2025] — Following a three-year study that assessed deidentified data from cancer patients across the Providence family of organizations, researchers from Providence and Microsoft developed TRIALSCOPE — a new AI-powered framework designed to both simulate and validate clinical trial outcomes using real-world data, enabling researchers to reproduce results of large, historical clinical trials just from observational patient data. This approach doesn’t replace validation but offers a way to reduce early risk and optimize trial planning before enrolling patients.

“The results are very promising,” said Carlo Bifulco, M.D., medical director of Cancer Genomics and Precision Oncology, Providence Cancer Institute. “And actually, the technology has moved even beyond what was possible when we designed this trial in 2022. The AI tools keep improving so we can simulate trial results from the patients’ data.”

Such advancements are especially significant given that conventional clinical trials are often lengthy and limited in scope, with years spent recruiting enough participants.

Typically, a new medication or treatment undergoes clinical trials and then approvals for up to 15 years before it’s available. Clinical trials usually only include a relatively small number of patients and often take years to recruit sufficient participation.

TRIALSCOPE has the potential to take years off this process by finding patients based on data in their medical records.

“This de-risks clinical trials by using real-world data from patients who have already received treatments, allowing researchers to generate insights without exposing new patients to new medication,” said Brian Piening, Ph.D., director of research for Providence Genomics and co-author of the study. “While the smaller, simulated datasets still require careful validation, TRIALSCOPE’s potential is invaluable, giving researchers a powerful new framework to help reduce the need for large initial participant pools and accelerating the path to more effective studies.”

“By applying cutting-edge modeling to real-world data, TRIALSCOPE opens new possibilities for exploring trial outcomes and advancing precision medicine,” said Hoifung Poon, general manager of Real-World Evidence at Microsoft Research. “This approach helps researchers ask bigger questions and uncover patterns that traditional methods often miss.”

Results from the study, which were recently featured in NEJM AI, a New England Journal of Medicine publication, supported that TRIALSCOPE could reproduce the results of lung and pancreatic cancer clinical trials based on data in the patients’ electronic health records (e.g. diagnoses, treatments), rather than relying on data collected specifically for a clinical trial.

Dr. Bifulco noted that leveraging historical patient data not only streamlines the process of testing new hypotheses but also makes the development of studies for new medications significantly more efficient — from initial planning through final assessment.

One key application is to find new successful treatment strategies by mining compassionate-use data, where individual patients gain access to experimental therapies when all other options have failed. Because these are typically one-off treatments, it can be difficult to extrapolate results to all cancer patients without a framework like TRIALSCOPE.

In addition, TRIALSCOPE has cost benefits as it can review the same amount of data that it would take an entire team of people to comb through — and at a much quicker pace.

Given the success of TRIALSCOPE’s initial testing, Dr. Bifulco sees potential for using the framework to design future clinical trials.

“We hope TRIALSCOPE can help us develop new cancer treatments that save lives,” said Bill Wright, Ph.D., chief research officer, Providence. “As it continues to evolve, TRIALSCOPE and similar AI tools will help accelerate research and usher in a new era of truly personalized precision medicine.”