Overcoming the Challenges in the Development and Implementation of Artificial Intelligence in Radiology: A Comprehensive Review of Solutions Beyond Supervised Learning

Abstract

Artificial intelligence (AI) in radiology is a rapidly developing field with several prospective clinical studies demonstrating its benefits in clinical practice. In 2022, the Korean Society of Radiology held a forum to discuss the challenges and drawbacks in AI development and implementation. Various barriers hinder the successful application and widespread adoption of AI in radiology, such as limited annotated data, data privacy and security, data heterogeneity, imbalanced data, model interpretability, overfitting, and integration with clinical workflows. In this review, some of the various possible solutions to these challenges are presented and discussed; these include training with longitudinal and multimodal datasets, dense training with multitask learning and multimodal learning, self-supervised contrastive learning, various image modifications and syntheses using generative models, explainable AI, causal learning, federated learning with large data models, and digital twins.

The Promise of Explainable AI in Digital Health for Precision Medicine: A Systematic Review

“…Deep-learning models offer more exact and efficient diagnosis for diseases requiring analysis of medical images (i.e., cancer, dementia), compared with human experts [ 72 ]. Explainable artificial-intelligence approaches to deep-learning models of medical images often include some form of visual explanation highlighting the image segments the model used to make the diagnosis [ 73 , 74 ]. Deep learning can also reduce drug discovery costs by efficiently screening for potential candidates, reducing time compared with traditional methods [ 29 ].…”

Artificial Intelligence in Breast Imaging: Opportunities, Challenges, and Legal–Ethical Considerations

“…Overcoming these hurdles requires a concerted effort from researchers, developers, policymakers, and stakeholders to ensure that AI is harnessed effectively and responsibly in diverse applications. 32 , 50 , 51 …”

Updated Primer on Generative Artificial Intelligence and Large Language Models in Medical Imaging for Medical Professionals

“…Sixth, the interpretability and control of these models remain challenging, particularly in complex architectures such as GANs. The opaque nature of these models impedes their fine-tuning, understanding, and decision-making processes, which limits their practical applications in medical research and clinical settings [ 11 ]. Hallucinations, characterized by the generation of unrealistic or spurious outputs, challenge the credibility and reliability of generated content [ 140 141 ].…”

“…Medical imaging has emerged as a prominent area that can be explored using generative models. GANs and diffusion models have been used in studies focusing on image reconstruction and quality enhancement [ 11 ]. The paramount importance of maintaining privacy in medical research has also led to the use of synthetic data.…”