유방촬영술 정도관리를 위한 현장조사, 팬톰검사 및 임상영상 평가

Abstract

Purpose: The purpose of this study was to establish a quality standard for mammographic equipment in Korea and to eventually improve mammographic quality in clinics and hospitals throughout Korea by educating technicians and clinic personnel. Materials and Methods: For the phantom test and on site assessment, we visited 37 sites and examined 43 sets of mammographic equipment. Items that were examined include phantom test, radiation dose measurement, developer assessment, etc. The phantom images were assessed visually and by optical density measurements. For the clinical image assessment, clinical images from 371 sites were examined following the new Korean standard for clinical image evaluation. The items examined include labeling, positioning, contrast, exposure, artifacts, collimation among others. Results: Quality standard of mammographic equipment was satisfied in all equipment during on site visits. Average mean glandular dose was 114.9 mRad. All phantom image test scores were over 10 points (average, 10.8 points). However, optical density measurements were below 1.2 in 9 sets of equipment (20.9%). Clinical image evaluation revealed appropriate image quality in 83.5%, while images from non-radiologist clinics were adequate in 74.6% (91/122), which was the lowest score of any group. Images were satisfactory in 59.0% (219/371) based on evaluation by specialists following the new Korean standard for clinical image evaluation. Satisfactory images had a mean score of 81.7 (1 S.D.=8.9) and unsatisfactory images had a mean score of 61.9 (1 S.D=11). The correlation coefficient between the two observers was 0.93 (p<0.01) in 49 consecutive cases. Conclusion: The results of the phantom tests suggest that optical density measurements should be performed as part of a new quality standard for mammographic equipment. The new clinical evaluation criteria that was used in this study can be implemented with some modifications for future mammography quality control by the Korean government.