Is the circle perfect? Radiographic validation of the best fit circle in nonarthritic shoulders

Anatomic reconstruction of the humeral articular segment is a recognized goal of implant designs for anatomic shoulder arthroplasty. The "best fit circle" (BFC) is used to determine the quality of proximal humeral reconstruction after anatomic shoulder arthroplasty. The purpose of this study was to validate the BFC using plain radiographs in a nonarthritic cohort.

Two hundred consecutive patients evaluated for nonarthritic shoulder conditions with true anterior-posterior radiographs were identified. 110 (55 percent) were male. The mean age was 56 ± 14 years. A native circle (NC) was determined from the humeral bone articular contour. The BFC was determined according to Youderian et al. Radius of curvature (ROC), center offset (CO), and articular segment thickness (AST) were measured. BFC and NC measurements were compared using 2 tailed paired Wilcoxon rank-sum tests. Pearson's correlation coefficients were used to evaluate associations between the BFC and NC measurements.

The NC ROC was 0.21 ± 0.81 mm greater than the BFC ROC (P < .001) and was greater than the BFC ROC in 124 cases (62 percent). The difference was greater than 2 mm in 6 cases (3 percent). The NC CO was 0.09 ± 1.21 mm medial to the BFC CO (P = .30) and was medial to the BFC CO in 110 cases (55 percent). The CO difference was greater than 2 mm in 19 cases (9.5 percent). The mean NC AST was 0.81 ± 1.27 mm greater than BFC AST (P < .001) and greater than the BFC AST in 150 cases (75 percent). The difference in AST was greater than 2 mm in 46 cases (23 percent). The NC AST:ROC ratio was significantly greater than the BFC AST:ROC (0.79 vs. 0.75; P < .001), and greater in 140 cases (70 percent). The NC AST:ROC was greater than 0.80 in 72 cases (36 percent) compared to 39 cases (19.5 percent) for the BFC. There was a significant difference in the position of the CO between males and females.

This study validates the BFC as a surrogate of the normal proximal humeral bony articular anatomy in most shoulders. On average, the NC is significantly larger than the BFC. In most cases the differences are small and likely clinically irrelevant. Nevertheless, there were outliers and using the BFC to determine the size of a prosthetic humeral head implant could risk error in these cases.