Quantitative Decision Reliability of AI-Accelerated Computational Modeling and Simulation for Cardiovascular Devices

The Food and Drug Administration's Center for Devices and Radiological Health seeks a single contractor to develop quantitative methods for evaluating the reliability of AI-accelerated computational modeling and simulation in cardiovascular device regulatory submissions, with proposals due August 7 and questions due July 31. This is a sole-source procurement with no set-aside.

Work is performed at the contractor's facility with no FDA facility access required and no travel authorized; bi-weekly virtual meetings with FDA staff are mandatory. The contract is a 12-month Firm Fixed Price award to the lowest-priced technically acceptable offeror. The scope comprises four primary tasks: a literature review of surrogate modeling methods documenting at least 10 publications with surrogate model accuracy data; computational fluid dynamics development and evaluation using the FDA nozzle benchmark or justified alternative, including generation of at least 100 simulation points and three or more surrogate models (including neural network-based approaches); decision reliability assessment through mathematical procedures combining surrogate approximation error, numerical error, and model-comparator discrepancy, implemented in a Python Jupyter notebook; and preparation of a draft journal publication. Deliverables are due at 3, 9, and 12 months.

The contractor must demonstrate expertise in cardiovascular computational fluid dynamics with peer-reviewed publications, medical device regulatory science experience, surrogate modeling expertise, and understanding of regulatory frameworks. NAICS code is 541715, Research and Development in the Physical, Engineering, and Life Sciences. The contract falls under FAR Part 12 for commercial services. Quotes must be submitted via email to Telisha.Wilson@fda.hhs.gov with subject line "ALP-133983" and remain valid through September 30, 2026.