Background Generative artificial intelligence (AI) facilitates the development of digital twins, which enable virtual representations of real patients to explore, predict and simulate patient health trajectories, ultimately aiding treatment selection and clinical trial design, among other applications. Recent advances in forecasting utilizing generative AI, in particular large language models (LLMs), highlights untapped potential to overcome real-world data (RWD) challenges such as missingness, noise and limited sample sizes, thus empowering the next generation of AI algorithms in healthcare.

Methods We developed the Digital Twin - Generative Pretrained Transformer (DT-GPT) model, which leverages biomedical LLMs using rich electronic health record (EHR) data. Our method eliminates the need for data imputation and normalization, enables forecasting of clinical variables, and prediction exploration via a chatbot interface. We analyzed the method’s performance on RWD from both a long-term US nationwide non-small cell lung cancer (NSCLC) dataset and a short-term intensive care unit (MIMIC-IV) dataset.

Findings DT-GPT surpassed state-of-the-art machine learning methods in patient trajectory forecasting on mean absolute error (MAE) for both the long-term (3.4% MAE improvement) and the short-term (1.3% MAE improvement) datasets. Additionally, DT-GPT was capable of preserving cross-correlations of clinical variables (average R2 of 0.98), and handling data missingness as well as noise. Finally, we discovered the ability of DT-GPT both to provide insights into a forecast’s rationale and to perform zero-shot forecasting on variables not used during the fine-tuning, outperforming even fully trained, leading task-specific machine learning models on 14 clinical variables.

Interpretation DT-GPT demonstrates that LLMs can serve as a robust medical forecasting platform, empowering digital twins that are able to virtually replicate patient characteristics beyond their training data. We envision that LLM-based digital twins will enable a variety of use cases, including clinical trial simulations, treatment selection and adverse event mitigation.

M Bordukova, N Makarov, R Rodriguez-Esteban and F Schmich are allemployees of F. Hoffmann-La Roche. MP Menden is a former employee of AstraZeneca. MP Menden collaborates and is financially supported by GSK, F. Hoffmann-La Roche and AstraZeneca. The authors have no other relevant affiliations or financial involvement with any organization orentity with a financial interest in or financial conflict with the subjectmatter or materials discussed in the manuscript apart from those disclosed.

This study was funded by F. Hoffmann-La Roche, Helmholtz Association, Munich School for Data Science - MUDS, and European Union's Horizon 2020 Research and Innovation Programme.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Flatiron Health data are available for research via Institutional review board (IRB) approval of a master study protocol with waiver of informed consent (IRB # RWE-001, "The Flatiron Health Real-World Evidence Parent Protocol", Tracking # FLI1-18-044 by the Copernicus Group IRB), obtained prior to study conduct, which covers the data from all sites represented. The MIMIC-IV dataset is publicly and openly available. For MIMIC-IV, the collection of patient information and creation of the research resource was reviewed by the Institutional Review Board at the Beth Israel Deaconess Medical Center, who granted a waiver of informed consent and approved the data sharing initiative.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

The Flatiron Health datasets are available upon request for the specific purpose of replicating results at PublicationsDataAccess@flatiron.com. The MIMIC-IV dataset is publicly available.