Abstract
BACKGROUND: Preoperative CT-based planning is established in shoulder arthroplasty surgery. Automated planning software has become available to assist the surgeon and may increase reliability and efficiency. This study aims to evaluate the reliability of an automated 3-dimensional (3D) planning software package (Blueprint™ v2.1.5, Wright Medical Ltd) in the assessment of the arthritic shoulder against manual multiplanar measurement (MM). METHODS: 74 CT studies acquired for preoperative shoulder arthroplasty planning were reviewed on two occasions by four different evaluators, taking manual measurement (MM) of glenoid version and inclination adjusted with multiplanar reformation and adhering to modified Freidman and Maurier methods. 15 scans were not processed by Blueprint due to incompatible scanning protocols or severe scapular dysmorphia. 59 Blueprint measures were compared with the manual data. RESULTS: Version: Intra-observer reliability of glenoid version MM was excellent (mean ICC 0.92). Inter-observer reliability between all four readers was good (ICC 0.89). A Bland-Altman analysis of Blueprint versus MM for version measurements demonstrated a mean pair difference of -5.77 (95% CI -7.25 to 4.29). Inclination: Intra-rater and inter-rater reliabilities were good (ICC 0.85 and 0.80 respectively). Blueprint and MM values for inclination followed a more convergent pattern than for version. Bland-Altman analysis for inclination did not show substantial bias, with a mean pair difference of 1.4 (95% CI -0.1 to 2.9). CONCLUSION: Manual preoperative planning for shoulder arthroplasty is time consuming and requires experience. Automated 3D planning offers a consistent tool to assist the surgeon, notwithstanding intra-operative anatomical and technical variation, and margin of error. Surgeons should as ever be mindful of the specifics of a given automated program and our data quantified a bias for retroversion which may be important for measures close to the thresholds for augmentation or customised implants.