Much discussion surrounding artificial intelligence (AI) has been in relation to the role of the radiologist. However, radiographers – often the bridge between doctors and patients – appear to be a much-overlooked professional in the ongoing AI debate, say some observers.

Delegates attending today’s joint session on AI and the radiography profession, hosted by the European Federation of Radiographer Societies (EFRS) and the International Society of Radiographers & Radiological Technologists (ISRRT), will hear from radiography representatives about the various related issues challenging the profession.

The impact of AI and machine automation on radiographers’ role and responsibilities requires more attention, as was highlighted at ECR 2019, according to Maryann Hardy, PhD, professor of radiography and imaging practice research at the University of Bradford’s Faculty of Health Studies, U.K. A panel of AI radiology experts at last year’s congress struggled to answer how AI could affect radiographers, finally suggesting that there might be a greater opportunity for patient interaction or radiographer reporting of images with AI support, she said. This suggestion flies in the face of existing arguments from radiologists that such a move could undermine professional autonomy and expert decision-making.

“Importantly, the response also demonstrated a lack of consideration of what the impact of AI might be on image acquisition processes and the profession responsible for image production, workflow, and patient care during the acquisition process,” Hardy told ECR Today.

The delivery of successful radiology services is dependent on team-working between radiographers, radiologists, nurses, and other service support staff. Professionals should understand the actual and potential impact of AI across the service and not just consider it through their own professional lens, she noted.

‘Optimisation’ risks

AI promises ‘workflow optimisation’ by offering efficiencies in examination time and dose reduction as well as workforce costs and numbers, continued Hardy. Rapid advances in computer and imaging technologies, plus the associated increasing demand for imaging, have meant that few younger radiographers have had the opportunity to gain high levels of expertise to work across multiple modalities. Junior radiographers, in particular, do not have the knowledge, skills, and confidence to modify pre-set image acquisition protocols in response to patient need and clinical question, she remarked.

As a result, radiographers risk being deskilled in their core competencies of deciding the best imaging technique and exposure parameters to use. At the same time, many decision-making processes related to vetting, sequence selection, and image and dose optimisation are automatic, without the appropriate scrutiny as would be expected if the decision was made by a radiographer, she said.

In the rush for throughput and financial efficiency, systems of machine governance, responsibility, and ethical practices that ensure patients remain the central service focus, may not have been thoroughly considered. Radiographers, therefore, must engage with AI and decide on how machine decision-making should be audited. Consequently, questions must be addressed, such as who is responsible for the machine’s decision, how the level of machine decision-making is communicated to the patient and, importantly, who is responsible if an error occurs, Hardy explained.

Training requirements

With this in mind, radiographer education – with its changing technological emphasis and the need for new competencies, including AI fundamentals and ethical technology – will form another element of her presentation.

In the U.K., for example, the Topol Review of 2019 (‘Preparing the healthcare workforce to deliver the digital future’) has argued that all healthcare professional pre-registration courses should have a computer science component by 2025 to prepare the workforce for the digital future, while all health professionals should update their skills to ensure they can use, understand, and challenge digital systems and AI in the delivery of care.

But Hardy questions where the expertise is to deliver such education and more importantly, the readiness of healthcare organisations to future-proof its delivery processes and workforce, particularly given the challenge of increasing demand and staff shortages. She believes the need for radiography education in AI implementation has reached a high level of urgency.

“With the rise of machine learning, the days of the traditional radiographer are limited, and radiographers must evolve to become experts in these new technologies or face extinction,” noted Hardy. “While changes in education can enable evolution, are clinical centres ready to employ the computerised imaging science expert that will be the radiographer of the future? Or will resistance to adopting new radiographer roles and responsibilities be the death knell of radiography as a graduate profession?”

Melissa Jackowski, EdD, chair of the American Society of Radiologic Technologists (ASRT), emphasises that undergoing change and experiencing technological innovation is not a new experience for radiographers. During this transition, more than ever, radiologists and medical imaging and radiation therapy professionals will need to understand each other’s role and stand together to advocate for their profession and their patients, she noted.

At today’s session, she will be presenting findings from a 2019 national research study conducted by the ASRT regarding professionals’ overall comfort with technology, their understanding of AI and machine learning (ML), use in departments, the effects of AI/ML, and how they believe these technologies will change the role of the radiological technologist.

The study found that AI is already having an impact on radiographers’ daily practice. In general, when asked “How often do you use AI/ML/automated features on your equipment?”, 18% of respondents chose “all of the time” (90–100% of the time), nearly 35% chose “most of the time” (51–89% of the time), and only 10% chose “rarely” (1–15% of the time).

ASRT survey question from 2019 showed that while most medical imaging technologists and radiation therapists believe artificial intelligence, machine learning and automated features will have a beneficial or neutral effect on performing exams or therapy, there is some concern that it could negatively affect patient interaction protocols (provided by Melissa Jackowski, EdD).

Survey respondents were inclined to see AI as having a beneficial impact on considerations such as safety and quality, but they expressed concern that it could alter the more human aspects of the profession, such as patient interaction and creativity, which she states, must not be lost during the incorporation of AI into radiography practice. Furthermore, respondents showed no widespread consensus that AI would adversely affect their professional prospects.

“I believe the key message is that medical imaging and radiation therapy professionals, and radiologists, need to be leaders in the development of use-case scenarios related to their respective roles and implementation of this and any technological changes in our field. This is part of our role as competent practitioners and patient advocates,” Jackowski said.

Joint Session of the EFRS and the ISRRT
, Friday, July 17, 17:00–18:00
Artificial intelligence and the radiographer profession

  • Chairpersons’ introduction
    Jonathan McNulty; Dublin/IE
    Donna Newman; Fargo, ND/US
  • Exploring the current landscape and evidence-base relating to artificial intelligence (AI) and the radiographer profession
    Nicholas Hans Woznitza; London/UK
  • Ethical considerations in AI
    Adrian Brady; Cork/IE
  • Considering AI in our education and training programmes
    Maryann Hardy; Bradford/UK
  • Horizon scanning: the future of AI and the radiographer profession
    Melissa Jackowski; Angier, NC/US
  • Live Q&A: What steps can be taken to better prepare radiographers for AI?


Ongena YP, Haan M, Yakar D, Kwee TC (2019) Patients’ views on the implementation of artificial intelligence in radiology: development and validation of a standardized questionnaire. Eur Radiol. doi: 10.1007/s00330-019-06486-0:

Wong SH, Al-Hasani H, Alam Z, Alam A (2019) Artificial intelligence in radiology: how will we be affected? Eur Radiol. 29(1):141-143:

Pinto Dos Santos D, Giese D, Brodehl S et al (2019) Medical students’ attitude towards artificial intelligence: a multicentre survey. Eur Radiol. 29(4):1640-1646:

Geis JR, Brady A, Wu CC et al (2019) Ethics of artificial intelligence in radiology: summary of the joint European and North American multisociety statement. Insights Imaging. 10(1):101:

Pesapane F, Volonté C, Codari M, Sardanelli F (2019) Artificial intelligence as a medical device in radiology: ethical and regulatory issues in Europe and the United States. Insights Imaging. 9(5):745-753: