Patient-specific desktop 3D-printed guides for pelvic tumour resection surgery: a precision study on cadavers.

PURPOSE: 3D-printed patient-specific instruments have become a useful tool to improve accuracy in pelvic tumour resections. However, their correct placement can be challenging in some regions due to the morphology of the bone, so it is essential to be aware of the possible placement errors in each region. In this study, we characterize these errors in common pelvic osteotomies.
METHODS: We conducted an experiment with 9 cadaveric specimens, for which we acquired a pre-operative computed tomography scan. Small PSIs were designed for each case following a realistic surgical approach for four regions of the pelvis: iliac crest (C), supra-acetabular (S), ischial (I), and pubic (P). Final surgical placement was based on a post-operative scan. The resulting positions were compared with pre-operative planning, obtaining translations, rotations, and maximum osteotomy deviations in a local reference frame defined based on the bone's morphology.
RESULTS: Mean translations and rotations in the direction of the osteotomy plane were as follows: C = 5.3 mm, 6.7°; S = 1.8 mm, 5.1°; I = 1.5 mm, 3.4°; P = 1.8 mm, 3.5°. Mean translations in the remaining axes were below 2 mm. Maximum osteotomy deviations (75% of cases) were below 11.8 mm in C (7.8 mm for half-length), 7.8 mm in S (5.5 mm for half-length), 5.5 mm in I, and 3.7 mm in P.
CONCLUSION: We have characterized placement errors for small PSIs in four regions of the pelvis. Our results show high errors in C and S PSIs in the direction of the resection plane's normal, and thus large osteotomy deviations. Deviations in short osteotomies in S, I and P and placement errors in the remaining directions were low. The PSIs used in this study are biocompatible and can be produced with a desktop 3D printer, thus minimizing manufacturing cost.