Volume 29, Issue 165, November 2025

Artificial Intelligence in Medical Imaging: Current Applications, Limitations, and Future Perspectives - Narrative Review

Mateusz Mierniczek^1♦, Maria Mierniczek², Aleksandra Mierniczek³, Sonia Mojzyk⁴, Agnieszka Partyka⁵

¹St. Raphael’s Voivodeship Specialist Hospital in Czerwona Góra, Czerwona Góra 10, 26-060 Chęciny, Poland
²St. Raphael’s Voivodeship Specialist Hospital in Czerwona Góra, Czerwona Góra 10, 26-060 Chęciny, Poland
³Ministry of Interior and Administration Hospital in Gdansk, Kartuska 4/6, 80-104 Gdańsk, Poland
⁴7th Navy Hospital, Polanki 117, 80-305 Gdańsk, Poland
⁵Murcki Hospital, Sokołowskiego 2, 40-749 Katowice, Poland

^♦Corresponding author
Mateusz Mierniczek; St. Raphael’s Voivodeship Specialist Hospital in Czerwona Góra, Czerwona Góra 10, 26-060 Chęciny, Poland

ABSTRACT

Artificial intelligence (AI) has developed from a theory to a technology that has become practical in current healthcare, especially in image diagnosis. This is a review of where current usage of this technology is, its limitations at present, and its possible futures. Our findings indicate that AI systems are already augmenting radiological practice across various imaging modalities, including X-ray, CT, and MRI. These tools show real promise in speeding up diagnosis by detecting subtle or early signs that might be missed during standard screenings, particularly lung and breast cancers, and by analyzing heart and brain images. However, many unanswered questions remain regarding the reliability of AI tools in routine clinical practice, despite their impressive technical performance. Model bias, a lack of varied, high-quality data, and difficult moral conundrums related to AI use are some of the main obstacles. We observe that clearer legal and regulatory frameworks, as well as greater transparency, are increasingly required, often referred to as "explainable AI". Looking ahead, the field is evolving. The next generation of AI may involve multimodal and foundation models that integrate imaging data with other clinical information. The use of AI is focused on supporting, not replacing, radiologists and on analyzing medical images.

Keywords: Medical imaging, Artificial intelligence, Diagnostic imaging, Radiomics

Medical Science, 2025, 29, e219ms3735

PDF

DOI: https://doi.org/10.54905/disssi.v29i165.e219ms3735

Published: 29 November 2025