Clinical application of a dual-surface, four-edge 3D-printed pedicle screw guide in adolescent idiopathic scoliosis surgery: a retrospective comparative study

Accurate pedicle screw placement is essential for safety in scoliosis correction surgery. Conventional freehand techniques are operator-dependent, while navigation-assisted techniques may increase intraoperative imaging requirements and workflow complexity. Patient-specific 3D-printed drill guide templates have been proposed to improve accuracy and reduce imaging use. This retrospective study evaluated the clinical efficacy, safety, and accuracy of a novel dual-surface four-edge 3D-printed drill guide template versus computer navigation-assisted and freehand techniques. A total of 64 patients (24 males, 40 females) undergoing posterior pedicle screw fixation at Fuyang People’s Hospital from January 2017 to January 2024 were included, with 1204 screws placed. In this retrospective, non-randomized cohort, patients were grouped according to the screw placement technique actually used during surgery: Group A (3D-printed guide, n = 22, 430 screws), Group B (navigation-assisted, n = 18, 357 screws), and Group C (freehand, n = 24, 417 screws). Screw accuracy was assessed using the Gertzbein–Robbins system (A–E). The primary metric was the proportion of Grade A screws; Grade A + B was additionally reported as clinically acceptable placement and Grades C–E as clinically relevant breach, together with screw revisions. Baseline characteristics, deformity-correction outcomes, perioperative indicators, intraoperative C-arm fluoroscopy acquisitions, and complications were compared among groups. Baseline characteristics (age, sex, BMI, preoperative Cobb angle, sagittal vertical axis, coronal balance distance, Lenke classification, number of fused segments) were comparable among groups ( P > 0.05). No significant between-group differences were observed in coronal and sagittal offset correction, number of screws, intraoperative blood loss, or length of hospital stay ( P > 0.05). The main curve correction rate was also comparable ( P = 0.608). Group A required significantly fewer recorded intraoperative 2D C-arm fluoroscopy acquisitions than Groups B and C ( P < 0.001). Grade A + B acceptable placement rates were 99.30, 96.36, and 96.64%, and clinically relevant breach (Grades C–E) occurred in 0.70, 3.64, and 3.36% of screws in Groups A, B, and C, respectively; no screw revision was required in any group. In the primary cluster-adjusted GEE analysis, the 3D-printed guide group showed a higher Grade A placement rate than the freehand group (OR = 3.48, 95% CI: 2.67–4.53, Bonferroni-adjusted P < 0.001) and than the navigation-assisted group (OR = 1.59, 95% CI: 1.20–2.10, Bonferroni-adjusted P = 0.004). One postoperative incision infection occurred in Group B, and one case of cerebrospinal fluid leakage occurred in Group C; no other surgery-related complications were reported. In this retrospective cohort, all three techniques were associated with few observed complications and postoperative improvement. The dual-surface, four-edge 3D-printed guide was associated with fewer intraoperative 2D C-arm fluoroscopy acquisitions and high screw-placement accuracy. After accounting for clustering of screws within patients, the comparisons with both freehand and navigation-assisted placement favored the 3D-printed guide; however, these comparative findings should be interpreted cautiously because of the retrospective, non-randomized design and potential temporal confounding. These findings support the clinical feasibility of the guide, but further prospective studies with patient-level and cluster-adjusted analyses are needed to confirm its comparative effectiveness.