Evaluating SWAG and Its Validity When Compared to 3D Imagery of Secondarily Grafted Cleft Sites

OBJECTIVETo test validity of 2D Standardized Way to Assess Grafts (SWAG) ratings to assess 3D outcomes of bone grafting (ABG). PATIENTS43 patients (34 UCLP, 9 BCLP) with non-syndromic complete clefts, bone-grafted at mean age 9yrs/3mos, with available post-graft occlusal radiographs and cone beam computed tomography (CBCT) (taken mean 4yrs/9mos post-ABG). MAIN OUTCOME MEASURES2D occlusal radiographs rated twice using SWAG by 6 calibrated raters. 12 scores were averaged and converted to a percentage reflecting bone-fill. Weighted Kappas were assessed for SWAG reliability. 3D cleft-site bone volume was calculated by 1 rater using ITK-SNAP. 13 cleft sites were re-measured by the 'one rater' for 3D reliability using Intraclass Correlation Coefficient (ICC). 2D versus 3D ratings were compared using paired t-test, independent samples t-test, Bland-Altman and Linear Regression. Significance level was P = .5. RESULTS2D reliability was 0.724 (intra-rater) and 0.546 (inter-rater). 3D reliability was 0.986. Bland-Altman plot comparing 2D vs 3D showed for 45 of 47 graft-sites were within 2 SD's. Mean % bone-fill was 64.11% with 2D and 69.06% with 3D (mean difference = 4.95%) that was a non-significant difference in both t-tests. Regression showed a statistically significant relation between the two methods (r2 = 0.46; P = .0001). CONCLUSION2D SWAG systematically and non-significantly underestimated bone-fill. There was a significant correlation between 2D/3D methods. Bland-Altman analysis illustrated the similarity of the two methods. For comparisons of group (cleft treatment Centers') bone grafting outcomes, the 2D method may suffice as a proxy for the 3D method. However, with individual variation up to 40% in 2D estimates of actual 3D volume, 2D SWAG method cannot be used in place of 3D images.