Significant variability exists in preoperative planning software measures of glenoid morphology for shoulder arthroplasty

Abstract

Background & HypothesisWe sought to assess the reliability of 4 different shoulder arthroplasty 3-dimensional preoperative planning programs. Comparison was also made to manual measurements conducted by 2 fellowship-trained musculoskeletal radiologists. We hypothesized that there would be significant variation in measurements of glenoid anatomy affected by glenoid deformity. MethodsA retrospective review of computed tomography (CT) scans of patients undergoing shoulder arthroplasty was undertaken. A total of 76 computed tomographies were analyzed for glenoid version and inclination by 4 templating software systems (VIP, Blueprint, TrueSight, ExactechGPS). Inter-rater reliability was assessed via intra-class correlation coefficient (ICC). For those shoulders with glenohumeral arthritis (58/76), ICC was also calculated when sub-grouping by modified Walch classification. Lin's concordance correlation coefficient was calculated for each system with 2 musculoskeletal-trained radiologists’ measurements. ResultsMeasurements of glenoid version and inclination differed between at least 2 programs by 5º-10º in 75% and 92% of glenoids respectively, and by >10º in 18% and 45% respectively. ICC was excellent for version but only moderate for inclination. ICC was highest among Walch A glenoids for both version (near excellent) and inclination (good), and lowest among Walch D for version (near poor) and Walch B for inclination (moderate). When measuring version, VIP had the highest concordance with manual measurement; Blueprint had the lowest. For inclination Blueprint had the highest concordance; ExactechGPS had the lowest. Discussion & ConclusionDespite overall high reliability for measures of glenoid version between 4 frequently utilized shoulder arthroplasty templating softwares, this reliability is significantly affected by glenoid deformity. The programs were overall less reliable when measuring inclination, and a similar trend of decreasing reliability with increasing glenoid deformity emerged that was not statistically significant. Concordance with manual measurement is also variable. Further research is needed to understand how this variability should be accounted for during shoulder arthroplasty preoperative planning. Level of EvidenceLevel III; Retrospective Comparative Study