PURPOSE: To evaluate the usefulness of conventional magnetic resonance images (MRI) for arthroscopic surgery in subacromial impingement syndrome of the shoulder, as an alternative to MR arthrography with additional T2 fat saturation images (MRA). MATERIALS AND METHODS: The preoperative MRI of 77 patients (45 females, 32 males) (52 right, 25 left) and MRA of 34 patients (14 females, 20 males) (24 right, 10 left) with subsequent arthroscopic confirmation of subacromial impingement syndrome were reviewed retrospectively. The lesions requiring arthroscopic surgery were 95 subacromial spurs, 101 subacromial bursitis, and 51 full-thickness and 44 partial-thickness tears of the supraspinatus among 111 cases for both studies. A two-by-two table was constructed in order to calculate the sensitivity and specificity of both studies against arthroscopic outcomes. Also we analyzed the false positive and false negative cases of the full-thickness tears individually. RESULTS: The detection rates of subacromial spur and bursitis and full- and partial-thickness tears of the supraspinatus were 91%, 94%, 77%, and 65% in MRI and 93%, 100%, 83%, and 77% in MRA respectively. Their specificities were 33%, 33%, 90%, and 76% in MRI and 50%, 75%, 100%, and 71% in MRA respectively. Eleven false negative cases in regards to MRI resulted in Ellman's grade 3 partial thickness tear (72.7%), mild bursitis (63.6%), greater tuberosity erosion (45.5%), and negative fluid signal of the glenohumeral joint (81.8%). Three false positive cases on the MRI were induced from errors with lower window depth and width on the imagings. Two false negative cases on MRA were induced from the adhesion between Ellman's grade 3 rim-rent tear and the glenohumeral joint cavity. CONCLUSION: Conventional MR images could be used to decide the arthroscopic surgery in subacromial impingement syndrome, as an alternative to MR arthrography with additional T2 fat saturation images.