BACKGROUND: Scapular notching can happen at diverse location depending on implant design or operative technique, therefore, it is easily misdiagnosed. Thus, this study purposed to suggest a method helpful to assess scapular notching. METHODS: The subjects were 73 cases of reverse shoulder arthroplasty (RSA) for cuff tear arthropathy during the period from May 2009 to April 2014 and followed-up for over a year. There was medialized RSA in 22 cases, bone increased offset RSA (BIO-RSA) in 36 cases, and metal increased offset RSA (metal-RSA) in 15 cases. Scapular notching was not determined by bone defect at the inferior of glenosphere as Sirveaux's classification, but scapular notching at the site where the rotational route of the polyethylene of humeral implant met the scapular neck were examined. The results were compared with conventional method. RESULTS: By conventional method, scapular notching was observed in 10 cases (45.5%) in medialized RSA, 12 cases (33.3%) in BIO-RSA, and none in metal-RSA. By new method, it was observed in 9 cases (40.9%) in medialized RSA, 10 cases (27.8%) in BIO-RSA, and none of metal-RSA. The site of scapular notching was apart from glenoshpere in 18 cases, and at inferior of glenosphere in 1 case. Absorption of bone graft was observed in 4 (11.1%) out of 36 cases of BIO-RSA. CONCLUSIONS: It is hard to distinguish scapular notching from absorption of bone graft in BIO-RSA, and bone absorption at the lateral lower end of glenoid in medialized RSA. Thus, it is considered useful to assess scapular notching at the site where the rotational route of the polyethylene insert meets scapular neck.