Back to search

Marshall University

Artificial intelligence in radiology

Abstract

dc:description.abstract

Introduction: Artificial intelligence (AI) has increasingly transformed radiologic practice by improving diagnostic accuracy, streamlining workflows, and reducing interpretation errors. As AI integration has expanded across imaging modalities, questions have emerged regarding its effectiveness compared to traditional radiologist-only interpretation. Purpose of Study: The purpose of this study has been to evaluate the impact of AI-assisted radiology on diagnostic accuracy, efficiency, and error reduction, while also assessing clinician perceptions of AI as a collaborative tool in imaging analysis. Methodology: This qualitative study has used a systematic review of peer-reviewed literature published between 2015 and 2025, following PRISMA guidelines, combined with an interview with a licensed radiologic technologist. Forty studies were initially identified, and 30 met inclusion criteria related to diagnostic performance, turnaround time, and interpretive reliability. The interview provided experiential insight into the clinical use of AI. Results: The systematic review has shown that AI systems have consistently increased sensitivity and specificity across various imaging domains, including pulmonary, oncologic, neurologic, and musculoskeletal assessments. Multiple studies have demonstrated that AI has detected abnormalities missed during manual review, thereby reducing diagnostic error and supporting quicker clinical intervention. Additionally, AI-assisted workflows have been associated with reduced interpretation times and decreased radiologist workload. The technologist interview supported these findings, noting that AI has improved consistency and accuracy in image review while emphasizing the continued need for human oversight. Discussion: The combined findings have indicated that AI has strengthened radiologic practice by enhancing both diagnostic and operational outcomes. However, the literature and interview have also highlighted the necessity of maintaining clinician involvement to contextualize AI-generated outputs, mitigate algorithmic bias, and ensure patient safety. Ongoing education, regulatory development, and long-term performance monitoring have been identified as essential components of responsible AI integration. Conclusion: This study has demonstrated that AI has served as a valuable adjunct to radiologists by improving diagnostic performance, reducing error rates, and supporting efficient clinical workflows. While AI continues to evolve, radiologists and technologists remain central to ensuring accurate interpretation and optimal patient outcomes.

Degree

thesis:*
Name thesis:degree_name
Healthcare Administration
Level thesis:degree_level
Research Paper
Discipline thesis:degree_discipline
College of Business

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Sweeney, Olivia
Contributors dc:contributor
  • Alberto Coustasse, Dr.PH. MD, MBA, MPH

Subjects

dc:subject × 14

Identifiers

dc:identifier.*
Repository record dc:identifier
https://mds.marshall.edu/etd/2068

Chain of custody

source
Harvested from
Marshall University
Base URL
mds.marshall.edu/do/oai/
Last updated
2026-11-21
Source record
OAI-PMH GetRecord
citation

Sweeney, Olivia. Artificial intelligence in radiology. Research Paper thesis, https://mds.marshall.edu/etd/2068