Establishment of clinical safety in monitor diagnosis of breast cancer from the viewpoint of medical image perception

点击翻译按钮获取中文摘要

(1)Interobserver agreement in quality control of mammography was evaluated using a breast phantom. Eleven radiologists interpreted a breast phantom image (CIRS Model X) by four viewing methods : (1)original screen-film (SF), (2)soft-copy reading of the digitized film image, (3)hard-copy reading of computed radiography using imaging plate (FCR), and (4)soft-copy reading of FCR. 11 radiologists judged each item of the phantom using a three-point rating scale. With regard to "mass" depiction, kappa value of FCR was significantly lower than that of SF (p<0.05). There was no significant difference in kappa value for detecting "microcalcifications". However, score of "microcalcifications" by FCR hard-copy reading was significantly lower than that by the other three viewing methods (p<0.05). Viewing methods could affect the judgment of the phantom image, and special attention to viewing conditions are recommended for quality control of FCR mammogram.(2)Significance of internal echoes at interpretation of breast US image was evaluated 100 breast masses (50 benign,50 malignant) were interpreted by 6 radiologists and 20 radiological technologists using a four-point rating scale. Double reading was done in 50 cases. US images were modified to make the internal echoes invisible by painting with black color using the Adobe Photoshop. ROC analysis showed a significant difference in diagnostic performance between the original US images and modified US images (p<0.05). Kappa value of posterior echoes was highest whereas that of halo showed the lowest value in both the original and modified images. Reproducibility in halo and gland surface was fair, and that in posterior echoes and the D/W was good. The internal echoes are important for image perception of breast lesions. Lack of internal echoes should degrade the diagnostic performance of breast US.