Mimicking clinical trials with synthetic acute myeloid leukemia patients using generative artificial intelligence

Abstract

Clinical research relies on high-quality patient data, however, obtaining big data sets is costly and access to existing data is often hindered by privacy and regulatory concerns. Synthetic data generation holds the promise of effectively bypassing these boundaries allowing for simplified data accessibility and the prospect of synthetic control cohorts. We employed two different methodologies of generative artificial intelligence – CTAB-GAN+ and normalizing flows (NFlow) – to synthesize patient data derived from 1606 patients with acute myeloid leukemia, a heterogeneous hematological malignancy, that were treated within four multicenter clinical trials. Both generative models accurately captured distributions of demographic, laboratory, molecular and cytogenetic variables, as well as patient outcomes yielding high performance scores regarding fidelity and usability of both synthetic cohorts (n = 1606 each). Survival analysis demonstrated close resemblance of survival curves between original and synthetic cohorts. Inter-variable relationships were preserved in univariable outcome analysis enabling explorative analysis in our synthetic data. Additionally, training sample privacy is safeguarded mitigating possible patient re-identification, which we quantified using Hamming distances. We provide not only a proof-of-concept for synthetic data generation in multimodal clinical data for rare diseases, but also full public access to synthetic data sets to foster further research.

Critical Appraisal and Future Challenges of Artificial Intelligence and Anticancer Drug Development

“…The second approach to developing in silico clinical trials is to retrospectively use big data and artificial intelligence algorithms to identify patients eligible for clinical trials and achieving this by aggregating and structuring data from patients' medical records [20,21]. A concrete illustration of this strategy comes from the study by Eckardt and coworkers [22]. The authors demonstrated the feasibility of two different technologies of generative AI to develop synthetic clinical trial data for patients with acute myeloid leukemia and that closely mimic disease biology and clinical behaviors.…”

“…Additionally, machine learning models reach a high level of complexity and may constitute impenetrable "black boxes" with the output difficult to interpret. There is also a risk for an unavoidable widening gap between how fast technologies like AI are devised and how quickly regulatory agencies may establish legal rules for safe implementation [22].…”