In our latest interview, we spoke to Tugba Akinci D’Antonoli, a radiology resident at Cantonal Hospital Baselland and a researcher at the University of Basel, Switzerland. D’Antonoli is a member of the 2023–2025 trainee editorial board for Radiology: Artificial Intelligence, a scientific editorial board member for European Radiology and Diagnostic and Interventional Radiology and is also a member of the Young Club Committee in the European Society of Medical Imaging Informatics.
What is your background/experience with artificial intelligence and what first attracted you to the topic?
Reflecting on my journey into AI, I’ve come to realize that my interest emerged through a series of seemingly minor coincidences, as is often the case for many people. However, if I had to pinpoint a defining moment, it would be during my first radiology residency (as I am currently undergoing my second residency training in a different country—if you’re curious why, just look up “obtaining non-EU medical diploma recognition in an EU country”). During that time, I took the initiative to bridge communication gaps between vendors and our hospital’s IT department, working collaboratively to troubleshoot system issues. Although I didn’t realize it at the time, this experience laid the foundation for my understanding of radiology workflows and their integration into hospital systems. Later in my career, I was involved in establishing a radiology department from scratch and ensuring its operations met safety and efficiency standards. These experiences sparked my interest in imaging informatics, giving me a unique edge early in my career.
Later, during my ESOR fellowships, my involvement in international collaborations opened the door to new intellectual horizons. I developed a keen interest in statistics, scientific methodology, and artificial intelligence, along with a strong desire to further expand my knowledge. Today I am still learning, whether through exploring new research papers or keeping up with the latest developments, and I’m grateful to continue this journey of growth and discovery.
What are the biggest challenges to AI adoption in clinical practice?
The biggest challenges to AI adoption in clinical practice are twofold: return on investment (ROI) and optimizing human-machine interactions.
I consider myself fortunate to have experienced the evolution of radiology—from manual processes to automation. I still remember the era of recording the dictation of a report, which were then transcribed by staff. Today, dictation is automated, freeing up staff for other tasks. However, this transition wasn’t smooth; early dictation systems were often inaccurate, requiring extensive proofreading, which at first seemed more tedious than the previous system. This highlights a key hurdle: humans need time to trust and adapt to new technologies.
Another major challenge is justifying the ROI. Many radiology departments operate without automation, and convincing hospital administrators to invest in AI tools can be difficult. There’s ongoing debate about whether these tools provide significant clinical benefits, such as improving patient outcomes or saving lives, versus merely offering marginal time savings. To drive adoption, we need strong evidence demonstrating their reliability and tangible benefits in clinical settings. Human factors—such as trust, usability, and integration—cannot be overlooked. AI must be proven as a reliable tool for healthcare professionals before it can be widely implemented.
Give us an example (or an educated guess) of what you think AI will be able to do in 3 years. What about in 10 years?
Predicting the future is inherently risky—consider Geoff Hinton’s famous 2016 prediction that radiologists would soon be obsolete. Yet, here we are, still integral to clinical practice after eight years.
Therefore, my cautious prediction is that over the next three years we will see significant advancements in non-interpretive tasks. For instance, large language models could become standard in reporting systems, helping radiologists correct typos, draft impressions from findings, and highlight overlooked details; some organizations are already piloting these tools. This will streamline workflows and enhance report quality.
Looking a decade ahead, I anticipate automation will extend to many tasks we currently handle manually. Regardless of the specific advancements, AI will enable us to spend more time collaborating with other healthcare professionals and engaging with patients, as Eric Topol aptly describes in Deep Medicine. The goal isn’t to replace radiologists, but to empower them with tools that enhance accuracy, efficiency, and patient care.
As a radiologist, what is your advice for younger colleagues wanting to dive into the topic of Artificial Intelligence, Machine/Deep Learning, and/or Radiomics? What are your tricks for staying up to date in this fast-evolving field?
To my younger colleagues, I would say to stay curious and embrace technology. Stay curious and believe in your ability to learn new things—mastery comes with time and persistence. Technology will inevitably shape your future, so resist the urge to shy away from it.
I strongly recommend joining medical imaging informatics societies, both local and international, such as EuSoMII or SIIM. These communities provide opportunities to connect with like-minded individuals, collaborate on research, and grow professionally. Another valuable tip is to engage in peer review. Reviewing papers keeps you at the forefront of new developments and sharpens your critical thinking, which can directly benefit your own research.
Finally, never stop learning. AI, machine learning, and radiomics are evolving rapidly, but staying connected through publications, conferences, and professional networks will ensure you remain well-informed and prepared for the future.
