Artificial intelligence enters medical imaging and radiation oncology workflows, increasing the need for structured education, digital literacy and safe clinical oversight. A 2026 original publication in European Radiology Experimental maps 29 AI courses across Europe through the Artificial Intelligence in Medical Imaging and Radiation Oncology Education project. Course provision varies by level, content, cost, language and format, leaving uneven access for professionals who need practical AI training.
Course Provision Remains Uneven
The survey ran from September 2024 to January 2025 and collected 215 responses before incomplete and duplicate entries were removed. The final dataset identified 29 unique AI courses for medical imaging and radiation oncology professionals in Europe. Participants came from across Europe, although responses were concentrated in Western Europe. About 60% of respondents were radiographers or radiologists, while about 17% had more technically oriented roles, including medical physicists, engineers, technical physicians or computer scientists.
Universities and other educational institutions offered 15 courses, accounting for 51.7% of identified provision. Industry offered six courses, professional bodies offered five and research institutions or other organisations offered three. Course audiences varied across the medical imaging AI ecosystem. Radiographers were named as an intended audience in 17 courses, medical physicists in 15 and radiologists in 12. Computer scientists, technical physicians, radiation oncologists and engineers were each named in 8 courses.
Most courses had a postgraduate orientation. Twelve courses were aligned with European Qualifications Framework Level 7, while ten were listed as other or non-classified levels. The remainder covered Levels 5, 6 and 8. Standalone courses were most common, with 19 offered independently rather than as part of a wider educational programme. Fourteen courses lasted one week or less, while some used asynchronous e-learning that allowed learners to progress at their own pace.
Online Formats Dominate Delivery
Online delivery was the most common mode of provision. Sixteen courses were designed for fully online delivery through classes or individual e-learning. Eight courses were delivered in person and four used a hybrid approach. The predominance of online learning can support access, particularly when courses are available beyond a single institution. However, short and online formats do not always support practical, hands-on learning. Some online approaches can include interactive learning, simulations and case-based learning, but practical competency development remains a key concern.
Course costs also varied. Twelve courses had no associated fees, representing 41.3% of provision. Fee-paying courses had a mean cost of €2,088 and a median cost of €495, with costs ranging from €171 to €9,628. Duration, delivery format and educational hours varied accordingly. Most courses were accessible beyond the institution or organisation connected with the course. Twenty-two courses were open to external learners, while seven were limited to learners linked to specific programmes or society memberships. Twelve courses did not offer formal academic credits.
English dominated delivery. Twenty-five courses used English, while six were offered in more than one language, including English, Dutch and French. Eleven courses were taught in languages other than English. French appeared in six courses, Dutch in three, Italian in one and German in one. Limited linguistic diversity can restrict access for non-English-speaking professionals and may affect the ability of clinical practitioners to engage with AI training in a meaningful way.
Must Read: Physicians Need More Support to Use AI
Curriculum Gaps Affect AI Readiness
Course content most often focused on clinical AI applications, AI basic principles and AI terminology. Clinical AI applications were included in 25 courses, AI basic principles in 23 and terminology in 20. Other common areas included the impact of AI on the profession, AI implementation, ethical challenges, AI research, regulatory and legal challenges and AI development. Fewer courses covered quality assurance, programming or implementation theories. Quality assurance and programming each appeared in nine courses, while implementation theories appeared in five.
The findings point to a need for broader educational coverage across professional roles, learning levels and practical competencies. Medical imaging and radiation oncology professionals work at the interface between patient care and technological innovation. Their AI education needs differ by seniority, role and responsibility. Foundational AI literacy is needed across professional groups, while advanced competencies are needed for those in leadership, implementation or strategic roles.
Legal and professional expectations also strengthen the case for structured AI education. The European Union AI Act requires explainability of AI-generated outputs and healthcare institutions must enhance employees’ digital literacy and understanding of AI technologies. Understanding AI training and testing is necessary for identifying potential biases or under-representation of specific populations. Without sufficient training, practitioners may struggle to assess AI outputs, recognise AI-generated errors or explain AI outputs to patients. These gaps can affect trust and clinical safety.
Future education also depends on trainers. Academic institutions need to support educators with the knowledge, skills, funding and time required to develop, deploy and evaluate AI educational tools. Training the trainers is part of building a sustainable AI education ecosystem.
AI education for medical imaging and radiation oncology professionals in Europe remains fragmented, despite the identification of 29 courses. Current provision includes online, short, standalone and often postgraduate-level options, with uneven access by geography, language, cost and accreditation. A centralised, searchable database can improve visibility and help professionals select training that matches their needs. Formal curricula, practical competencies, multilingual access and support for educators remain essential for safe, meaningful and equitable AI integration in clinical practice.
Source: European Radiology Experimental
Image Credit: iStock
References:
Decoster R, Erenstein H, Menzinga J et al. (2026) Towards an accessible, centralised, searchable database for AI courses in Europe: the Artificial Intelligence in Medical Imaging and Radiation Oncology Education (AIMIROE) project. Eur Radiol Exp; 10, 80.
