Comparison of Direct 3D Scans and Plaster Casts in Digital Ankle-Foot Orthosis Workflows

Background: Digital ankle-foot orthosis (AFO) workflows depend on accurate lower limb geometry acquisition. Direct 3D scanning on the lower limb may reduce plaster use, but its geometrical agreement with orthotist plaster casting remains insufficiently investigated. Objective: This study compared posterior lower limb surfaces captured by direct 3D scanning with surfaces obtained from orthotist plaster casts digitised using the same scanner. Methods: Seven healthy paediatric participants underwent bilateral direct scanning and bilateral plaster casting, giving 14 limb comparisons. Cast meshes were digitised, trimmed to reduce boundary effect, globally registered to the direct scans, and segmented into posterior foot, ankle, calf and knee proxy regions. Cast-to-scan distances were then calculated. Results: The ankle region showed the greatest residual mismatch, with a weighted mean distance of 3.49 mm (95% CI 3.16-3.81 mm). It also had the highest proportions of vertices exceeding 2 mm (74.8%) and 5 mm (21.2%), and was the highest-error region in 13 of 14 limbs. The knee region had the lowest weighted mean distance (2.28 mm, 95% CI 1.92-2.60 mm) but showed the largest upper-tail spread. Conclusions: These preliminary findings indicate direct scanning and plaster casting did not produce completely interchangeable posterior limb surfaces in this cohort. The ankle was the most consistently discrepant AFO-relevant region, indicating that digital AFO workflows should include local geometric quality control rather than relying only on whole-limb agreement.