Henkjan Huisman

PhD opportunity at Radboud University Medical Center in Nijmegen, the Netherlands, within the COMPASS4MD HORIZON-funded project. The project focuses on synthetic imaging data for medical AI regulation : when synthetic radiology images can provide valid evidence for the pre-market clinical evaluation of medical AI. The PhD candidate will combine technical research with stakeholder interviews and consensus development to help define the evidence required for regulatory use. This is a four-year, full-time PhD position (36 hours/week), based in Nijmegen , with a starting date of 1 February 2027 . Joint supervision is mentioned with Michail Klontzas (University of Crete), Henkjan Huisman (Radboud University Medical Center), and Merel Huisman as PI. The post welcomes candidates with an MSc in AI, data science, engineering, medical informatics, or a related field , plus an interest in regulatory science . Applications must be submitted through the official application portal only . The deadline is 22 October 2026 .

Radboud University Medical Center (Radboudumc) is recruiting a PhD candidate for a 4-year project on synthetic imaging data, medical imaging AI, radiology, and medical AI regulation . The project asks when synthetic medical images can be used as evidence for the development and regulatory evaluation of diagnostic AI systems, including questions of representativeness, fidelity, privacy, bias, and scientific validity. The PhD is part of COMPASS4MD , a Horizon Europe consortium focused on new methodologies for evaluating clinical evidence for medical devices. Your research will combine technical analysis of synthetic imaging datasets and generation methods with regulatory science and research methodology. You will study how synthetic data are used in research and regulatory submissions, assess their quality for different purposes, and help define minimum evidence standards for regulatory use. Research topics include: comparison of synthetic versus real clinical images; metrics for fidelity and clinical validity; interactions between privacy protection, bias, and representativeness; use of synthetic data for training augmentation versus external validation; evidence requirements across detection, segmentation, classification, and quantification tasks; and special issues for rare and ultra-rare diseases. The technical work will involve image quality and distributional analyses, downstream AI performance evaluation, bias and representativeness assessment, and privacy-related analyses. You will also conduct a structured literature review, interview stakeholders such as regulators, notified bodies, and AI developers, and contribute to a Delphi study. The project aims to produce evidence standards for the regulatory submission of diagnostic medical AI using synthetic imaging data. You will be based at Radboudumc in the Department of Medical Imaging and embedded in the Diagnostic Image Analysis Group (DIAG). The project is jointly supervised by Dr Merel Huisman at Radboudumc and Dr Michail Klontzas at the University of Crete, with Prof. Henkjan Huisman as promotor. Collaboration with the University of Crete and other international partners is a key part of the role. Applicants should hold an MSc in Artificial Intelligence, Data Science, Computer Science, Biomedical Engineering, Technical Medicine, Medical Image Analysis, Medical Informatics, or a closely related field. A background in health law or technology regulation is also welcome if paired with empirical research experience and strong quantitative skills. Desired experience includes Python, machine learning, medical imaging or computer vision, synthetic image generation, image analysis, statistics, and an interest in MDR and the AI Act. English proficiency at C1–C2 level is required. The position is a paid PhD employment contract with salary scale 10A (€3,217–€4,077). The preferred start date is 1 February 2027. The vacancy closes on 22 October 2026 . Apply via the vacancy page on the Radboudumc careers website.