New Phesi analysis of 167 million patient records finds colorectal cancer has largest volume of real-world data, but oncology trials still lack modernization

New Phesi analysis of 167 million patient records finds colorectal cancer has largest volume of real-world data, but oncology trials still lack modernization

Phesi CEO urges shift to “precision oncology” and warns that approach to cancer clinical development must evolve at the same speed as our knowledge of biomarkers, modalities and patients is growing 

Boston, US. 15 May, 2025. Phesi, a global provider of patient-centric data analytics, has released a new report ahead of ASCO 2025 analyzing the current state of oncology clinical development. Its latest analysis of 167 million contextualized patient data records via its AI-powered Trial Accelerator™ platform finds that the largest volume of real-world patient data exists on colorectal cancer with almost six million patient records from more than 18,000 cohorts; followed by breast (4.8m), lung (3.0m), liver (2.3m) and prostate cancer (2.2m) (table 1). As part of this new study, Phesi also analyzed the volume of patient data available for the top five causes of cancer death as identified by the World Health Organization (table 2). The availability of such significant volumes of contextualized patient data opens up the potential to generate digital twins in oncology to accelerate and modernize clinical development.

“Oncology R&D is increasingly defined by a more precise understanding of the molecular mechanisms of cancer and of individual patients. But clinical development operations, from planning to implementation, are yet to reflect the same level of precision. Studies suffer from poor patient recruitment, inadequate investigator site selection and poor data quality, increasing costs and cycle times,” said Dr Gen Li, Phesi CEO. “Deeper biological knowledge, combined with new treatments, modalities and combinations offers routes to innovation, but also adds complexity. The fact is, the data haystack is growing as our knowledge of cancer expands, but the needle is shrinking as we understand the critical role patient genetics play in cancer outcomes – as we’ve seen with breakthroughs like Keytruda for non-small-cell lung cancer. Clinical data science coupled with sophisticated digital twins will support development teams to evolve trial planning, all the way from understanding the history of disease to validating control arm data at Phase I and IIA, leading to a higher success rate in Phase IIB and III. The time is now to shift to an AI powered, data-led precision approach to oncology clinical development that leverages the significant volume of data available to the industry.”

Progress made in breast cancer highlights how precise profiling of patients has advanced clinical development. Several biomarkers such as HER2, PD-L1, PIK3CA, BRCa1 and BRCa2 are established, coupled with knowledge of hormone receptors like ER and PR to guide the use of hormone therapies, which are leading to improved patient outcomes. Compared to pancreatic cancer, which is challenging due to asymptomatic early presentation and where only non-specific biomarkers like CA 19-9 are established, and patient outcomes are consequently poor.

Sponsors have access to more data than ever before to fuel oncology R&D; of the top five causes of cancer death (table 2) identified by WHO, Trial Accelerator holds more than 18mn records from over 120,000 investigators – this enables precision in everything from patient selection to investigator site selection. Yet, earlier Phesi studies underscore where opportunities for greater precision continue to be missed. Phesi’s data revealed 51% of NSCLC trials are biomarker specific; deeper analysis of 471 recruiting Phase I NSCLC trials found they targeted more than 20 specific biomarkers, but 20% of investigators lacked a background in lung cancer recruitment. A separate analysis of patient enrollment data from oncology trials revealed a huge disparity between investigator sites; almost one fifth of investigator sites contribute just 3% of patients, and 16% of the best performing trial sites contribute almost half.

“Historically, sponsors have taken a scattergun approach to oncology development that results in costly and poorly performing trials – but with the right data and technology, they can be laser guided,” said Jonathan Peachey, Phesi Chief Operating Officer. “There are four elements to achieving greater precision that sponsors need to optimize – the targeted patient profile, the program, the protocol and the operations plan. Leveraging clinical data science will enable them to optimize these elements, as well as facilitating accurate scenario and prediction modelling before a wet trial ever gets underway, and to develop accurate digital twins. We know that our clients and the clinical development world at large are fully committed to the fight against cancer, but we can also see from the data that sponsors are not yet taking the smartest approach to these trials. Proactive use of AI and clinical data science will solve many of the longstanding issues oncology clinical development faces and most importantly get treatments to patients faster.” 

The full report, Leveraging Data Science to Power to Precision Oncology, is available here: https://info.phesi.com/precision_oncology_report 

Phesi is also launching the latest version of its Digital Patient Profile (DPP) Catalogue at ASCO. The latest catalogue contains 40 DPPs across oncology, metabolic, respiratory, CVS, CNS and inflammation indications. New oncology DPPs include Breast Cancer with PIK3CA and Diffuse large B cell lymphoma (DLBCL). Each DPP provides an accurate view of the patient population for specific disease areas with key demographics including age, sex, comorbidities, outcome measures and concomitant medications.  

Phesi will be at booth number 30148 at ASCO. Book a meeting here: https://info.phesi.com/asco-2025 

Phesi was proud to support Merck in the development of Keytruda for non-small cell lung cancer: https://www.phesi.com/drugs-brought-to-market/.

Phesi Trial Accelerator uses real-world data from more than 90,000 dynamically updated data sources, including observational studies, EHRs, retrospective studies, cohort studies, disease records, pharmacy records, claims data and randomized clinical trial data to provide patient-centric analytics to simulate clinical development and improve decision making. 

–

Table 1. The five cancers with the most available patient data. Source: Phesi Trial Accelerator, May 2025.

Table 2. Top five causes of cancer death as identified by World Health Organization (source: https://www.who.int/news-room/fact-sheets/detail/cancer); showing amount of available data in each indication. Source: Phesi Trial Accelerator, May 2025.