CHALLENGES OF ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING IN DIAGNOSTIC RADIOLOGY.
1Barsha Kumari*, 1David
¹Assistant Professor J P Institute India.
Corresponding Author: Barsha Kumari
DOI: https://doi.org/10.62502/ijmi/bzrd3739
ABSTRACT
Artificial intelligence (AI) represents a ground breaking transformation in contemporary computer systems, playing a pivotal role in the healthcare sector. Its significance is particularly pronounced in the medical field, where it has the potential to revolutionize medical education and applications. However, the clinical integration of AI, especially within the radiology department, presents a distinct set of challenges. The adoption of artificial intelligence and machine learning technologies in routine radiological practices involves numerous tasks and pre-implementation considerations that must be addressed for successful and practical use in daily operations.
Keywords: Artificial intelligence, Education, Radiology
Article Received: Jan-30, 2024 |
Acceptance: Feb-13, 2024 |
Published: March-20, 2024 |
BACKGROUND: Recent medical education and conferences have shown a heightened interest in artificial intelligence (AI) [1]. The radiology department stands out as a major user of AI, incorporating machine learning (ML) and deep learning (DL) techniques, particularly in image analysis [2]. DL, a subset of AI, proves beneficial for quality control, workflow improvement, and reporting [2]. Numerous research studies in recent years have explored various aspects of AI implementation and its benefits in radiology.
Instead of contemplating the replacement of radiologists, the current focus should be on gathering data for radiology references with the ultimate goal of enhancing the clinical efficiency of the department [3]. AI relies on pre-stored data, presenting a challenge in collecting diverse data for different conditions, including normal and abnormal findings [3]. A substantial and varied data sample is crucial for effective implementation.
Efficient data sampling can streamline the reporting system, reduce patient waiting times, and improve overall departmental quality [3]. However, data sampling is not a straightforward task and requires continuous monitoring of technological advancements and emerging possibilities in disease diagnosis [3]. Active participation of radiology professionals, especially radiologists, is essential to address potential errors and challenges related to data, technology, and ethics.
Confidentiality poses a significant challenge for AI and machine learning in clinical practice, emphasizing the importance of safeguarding patient information [3]. The integration of AI can enhance image analysis efficiency, thereby improving the workflow of the radiology department [4]. Radiologists may evolve into data or information managers, highlighting the need for training the new generation of professionals [5].
Cost-effectiveness is a substantial challenge in clinically establishing AI and machine learning in diagnostic radiology [3]. Integrating AI with Picture Archiving and Communication Systems (PACS) and teleradiology systems, though essential, presents practical difficulties and economic challenges [3].
AI and machine learning in medical imaging represent a paradigm shift for traditional radiology researchers. Understanding the technology, processes, limitations, and challenges is crucial for effective implementation in medical imaging practices. Collaboration between radiologists and data scientists, facilitated by initiatives like those organized by RSNA, can address challenges and advance the field using rigorous scientific methodology. This collaborative effort ensures the safe and clinically applicable integration of AI in radiology.
COMPLIANCE WITH ETHICAL STANDARDS
Conflict of Interest None
Statistics and Biometry: No complex statistical methods were necessary for this paper. Informed Consent: None.
Ethical Approval: Not Required.
REFERENCES:
- Kotter, E., Ranschaert, E. (2021) Challenges and solutions for introducing artificial intelligence (AI) in daily clinical workflow. Eur Radiol 31, 5–7 (2021). https://doi.org/10.1007/s00330-020-07148-2
- Moawad AW, Fuentes DT, ElBanan MG, Shalaby AS, Guccione J, Kamel S, Jensen CT, Elsayes (2022) Artificial Intelligence in Diagnostic Radiology: Where Do We Stand, Challenges, and Opportunities. J Comput Assist Tomogr. 2022 Jan-Feb 01;46(1):78-90. doi: 10.1097/RCT.0000000000001247. PMID: 35027520.
- Waller J, O'Connor A, Rafaat E, Amireh A, Dempsey J, Martin C, Umair M. (2022) Applications and challenges of artificial intelligence in diagnostic and interventional radiology. Pol J Radiol. 2022 Feb 25;87: e113-e117. doi: 10.5114/pjr.2022.113531. PMID: 35280945; PMCID: PMC8906183.
- Do HM, Spear LG, Nikpanah M, et al. (2020) Augmented radiologist workflow improves report value and saves time: a potential model for implementation of artificial intelligence. Acad Radiol 2020; 27: 96-105
- Recht MP, Dewey M, Dreyer K et al (2020) Integrating artificial intelligence into the clinical practice of radiology: challenges and recommendations. Eur Radiol. https://doi.org/10.1007/s00330-020- 06672-5