Room: Exhibit Hall
Purpose: To determine the frequency of different breast MRI artifacts and rate their clinical impact on diagnosis
Methods: A database of 98 breast MRIs from 12 sites were collected and reviewed by 7 radiologists. Radiologists determined the presence of 15 different artifacts and, if present, rated the severity of each artifact on a 3-point scale. Scale: 1 - minor/little impact on diagnosis, 2 - possible impact on diagnosis, 3 - unacceptable/major impact on diagnosis. Artifacts investigated were: 1.Chemical shift, 2.Aliasing, 3.Motion, 4.Poor fat suppression, 5.Misregistration on subtraction images, 6.Incorrect FOV, 7.Correctable cardiac artifact overlaying anatomy, 8.Truncation, 9.RF leak, 10.Geometric distortion, 11.Other/Undefined/Unsure, 12.No priors uploaded, 13.No tech sheet, 14.Poor positioning, and 15.Poor overall quality.
Results: Artifacts were present in 79 of 98 studies (81%). A total of 156 artifacts were noted as studies with artifacts often included more than a single artifact type. Of the 79 studies with artifacts present, 61 (77%) had severity ratings of 1 (minor/little impact on diagnosis), 11 (14%) had severity ratings of 2(possible impact on diagnosis), and 7 (9%) had severity ratings of 3 (unacceptable/major impact on diagnosis). Patient positioning and poor fat saturation were the most common artifacts encountered (31% and 22% of patients, respectively). Aliasing, motion, misregistration of subtraction images, and incorrect field of view were all present in >10% of patients. Poor fat saturation and incorrect field of view were generally associated with unacceptable severity scores.
Conclusion: Artifacts are common in breast MRI; however, the majority of artifacts have little or no clinical impact. Institutions looking to reduce artifacts should concentrate on fat saturation and field of view as these are often associated with an unacceptable impact on diagnosis.