A new analysis by clinical data analytics firm Phesi reveals colorectal cancer has the largest available volume of real-world patient data — nearly 6 million patient records across 18,000 cohorts — yet  oncology clinical trials still lag behind in modernization.

The findings, published ahead of ASCO 2025, are based on more than 167 million patient records analyzed via Phesi’s AI-driven Trial Accelerator platform.

Other cancers with significant real-world data volumes include breast (4.8 million), lung (3.0 million), liver (2.3 million) and prostate (2.2 million). Data was also evaluated for the top five causes of cancer deaths as identified by the World Health Organization, unlocking new opportunities for digital twin development and precision trial design.

“Oncology R&D has rapidly advanced in biological understanding — but trial design and operations are not keeping pace,” said Dr Gen Li, CEO of Phesi.

“We now have deeper insight into cancer genetics and therapies, but many trials still suffer from poor patient recruitment and site selection, leading to inefficiencies and increased costs.”

Li argues the sector must shift to AI-powered, data-led ‘precision oncology’, mirroring advances in biomarkers, modalities and patient subtypes. He highlights the example of Keytruda for NSCLC, where genetic targeting has driven meaningful outcomes — but notes 49% of NSCLC trials still do not use biomarker-specific recruitment. Among those that do, 20% of investigators lack experience in lung cancer, undermining effectiveness.

Phesi’s data also shows a stark disparity in investigator site performance: 16% of sites deliver nearly half of patient enrollment, while almost 20% contribute only 3%. This highlights the need for smarter operational planning, using predictive modelling and digital twin simulations from early-phase development onwards.

“Sponsors can no longer afford a scattergun approach to oncology trials,” said Jonathan Peachey, Phesi’s COO.

“With the right data and tools, sponsors can optimize patient targeting, protocol design, site selection and scenario modelling — before a single patient is enrolled. That’s how we accelerate timelines and improve trial success.”

The analysis demonstrates that while sponsors are committed to oncology R&D, true precision requires smarter use of the real-world data now available at scale.