EVALUATION OF INTER-RATER AGREEMENT IN 1.5T MAGNETIC RESONANCE BRAIN IMAGING ARTIFACTS REDUCTION BEFORE AND AFTER IMAGE QUALITY OPTIMIZATION”AND PERCEPTION OF RADIOLOGY PROFESSIONALS ON IT. A RELIABILITY STUDY

Abstract

Background: Magnetic Resonance Imaging (MRI) at 1.5 Tesla is widely used in clinical practice for brain imaging because of its accessibility and reliable diagnostic performance. However, image artifacts such as motion, susceptibility, and flow-related distortions often compromise diagnostic accuracy and interpretation⁶ . These artifacts can mask pathology or create misleading appearances, thereby reducing confidence in clinical decision-making. Several strategies have been proposed to optimize MRI image quality, including adjustments in acquisition protocols, use of artifact suppression techniques, and application of advanced postprocessing
tools⁷ . Despite these efforts, assessment of image quality remains largely subjective, with potential variability among different raters⁸ .
Methods: A comparative study was conducted on 41 patients undergoing routine brain MRI at 1.5T. Conventional images (pre intervention) were compared with optimized images (postintervention), incorporating protocol adjustments. Two independent skilled and senior radiologists, blinded to acquisition status, assessed image overall image quality using a 5-point Likert scale. Inter-rater agreement was measured with Cohen’s kappa test9
Results: A total of 53 patients underwent brain MRI at 1.5T during the study period. Of these, 12 cases were excluded according to the predefined exclusion criteria. The final analysis was therefore conducted on 41 patients, which met the sample size requirement for this exploratory study. Where male candidates are 35 and females were 18 only (randomly selected) In the analysis of 41 brain MRI cases at 1.5T, artifacts were initially identified in 17% of scans by Rater 1 (R1) and in 10% by Rater 2 (R2). Following the implementation of image quality optimization strategies, a substantial reduction in artifact prevalence was observed. Specifically, 86% of the artifact-affected cases demonstrated improvement according to R1, while 75% showed improvement according to R2. These findings suggest that the applied strategies were effective in mitigating artifacts and enhancing overall image quality. Also, A cross-sectional survey was conducted among radiologists, radiographers, and radiology technologists working in various healthcare settings. The survey included questions assessing demographic information, professional background, knowledge of image quality optimization applications in radiology, attitudes towards image quality optimization integration, perceived benefits and challenges, and overall awareness of image quality optimization advancements in the field. The preliminary analysis indicated that while a majority of radiology professionals recognize the potential benefits of image quality optimization in enhancing diagnostic accuracy, there is a significant variation in the level of knowledge and understanding of image quality optimization applications. The perception of image quality optimization was generally positive, with respondents acknowledging its role in improving patient outcomes radiological procedures. However, there was a notable demand for more comprehensive educational programs to bridge the knowledge gap.
Conclusion: Image quality optimization strategies effectively reduced artifacts in 1.5T brain MRI, with inter-rater agreement fairly observed. While improvements were consistent, subjective variability highlights the need for integrating objective measures in future evaluations.