Purpose <p>This study aimed to investigate the effect of 3D rapid prototyping(3DRP) curve-specific scoliosis model in pedicle screws placement, and contrast it with standard free-hand techniques.</p> Methods <p>A retrospective review on scoliosis patients operated on by a single surgeon from 2014 to 2018 identified 48 patients with curve-specific models manufactured by 3DRP and used intra-operatively for guidance of instrumentation. They were age and gender-matched to 48 scoliosis patients in whom pedicle screws were placed following the standard free-hand technique. Screw position was determined on CT by a grading system as Grade 0 (no violation), Grade 1 (&lt; 2&#xa0;mm perforation), Grade 2 (&lt; 4&#xa0;mm perforation), or Grade 3 (&gt; 4&#xa0;mm perforation).</p> Results <p>1485 screws (686 in the 3DRP group vs. 799 in the control group) were analyzed. Patients in the 3DRP group had higher Cobb angles and more challenging deformity. Although the overall percentages of critical perforations (those with risk of complications) were comparable between the two groups, the distribution of screw perforation were different. Screws in the 3DRP group were less likely to be critical both medially and laterally, and more likely to be critical anteriorly. Furthermore, laterally and anteriorly, the difference between them were largely due to a difference in Grade 2 perforation, whereas medially, the difference resulted from a higher percentage of both Grade 2 and Grade 3 perforation.</p> Conclusion <p>3DRP scoliosis model represents an affordable and accessible approach to reduce medial pedicle screw perforation with high risk of complications in the thoracic and lumbar spine.</p>