COTA launches new real-world data platform to accelerate cancer research and advance generative AI models in oncology

Automation of electronic health record data abstraction in COTA’s new large-scale dataset product Vista will advance real-world evidence in oncology and pave the way for deployment of trusted generative AI in cancer care.  

Oncology real-world data and analytics company COTA has just announced the launch of Vista, a large-scale automated electronic health record (EHR) dataset. Cota developed Vista using automation to increase the pace, and therefore scale, of EHR data abstraction in order both to advance real-world evidence (RWE) insights in oncology and to pave the way for development of trusted generative AI in cancer care. 

“Real-world data that tells a more complete story of a patient has accelerated our understanding of cancer and the development of lifesaving and life-extending therapies,” said Miruna Sasu, CEO of COTA. “Generative AI has even more potential to advance treatment options for millions of people if we can build trusted large language models for oncology. That means first curating and validating data at a fundamentally different pace and scale. That is what we are doing with Vista.” 

Vista combines automated data abstraction techniques and machine learning algorithms with medical expert oversight to glean clinically relevant information from EHRs at scale. The resultant information offers improved insights into the holistic journeys of populations of people with cancer. Vista is intended to power both retrospective analyses and patient matching into clinical trials. 

While many clinical datasets remain restricted by stringent criteria to include only the most straightforward cancer patients, Vista incorporates both oncological and non-oncological co-morbidity information about patients with any cancer diagnosis. This provides the most authentic and complete picture possible of real-world cancer populations and factors that impact individuals’ health journeys through cancer treatment. 

COTA proffer Vista as a new standalone dataset product but it also a precursor to their existing more targeted datasets, Vantage and Focus: combined use of these enables researchers to transition seamlessly from broad hypothesis generation to deep clinical studies of specific patient cohorts. 

COTA states that: “unreliable data is the single largest barrier to unlocking generative AI as a breakthrough tool to transform cancer therapy for millions of patients.” As Vista sources data directly from EHRs with COTA oncologists overseeing the data to ensure the information is clinically precise and relevant, it provides the transparency that is critical to powering trustworthy AI tools.