Impact of MRI integration in CT-based planning on the accuracy of patient-specific 3D-printed shelf implant placement

Abstract

Introduction: Hip dysplasia, characterized by an insufficient acetabular coverage of the femoral head, increases hip joint stress and predisposes to degenerative changes. A novel 3D-printed, patient-specific extracapsular shelf implant was developed to increase femoral head coverage. Accurate implant placement is crucial. This cadaveric study compared CT-only surgical planning with combined CT- and MRI-based planning to evaluate whether MRI integration improves positioning accuracy. Methods: Two cohorts of non‑dysplastic cadaveric hips were studied. In cohort 1 (five hips), implant design was based solely on CT imaging. In cohort 2 (four hips), both CT and MRI datasets were used, incorporating capsular soft‑tissue anatomy. Postoperative implant positioning was compared with preoperative plans using point-cloud analyses, clockface coverage graphs, and Dice coefficients. Acceptable placement was defined as: median Euclidean distance < 5 mm, median angular deviation < 5°, and Dice > 0.75 respectively. Results: In the CT‑only cohort, three of five implants failed one or more accuracy thresholds, with Euclidean distances up to 8.5 mm, coverage deviations up to 6°, and Dice coefficients as low as 0.37. CT‑only designed implants consistently tilted away from the acetabular rim, reflecting underestimation of hip capsule thickness and insertion height. In the CT+MRI cohort, all four implants met the <5 mm and <5° deviation thresholds, and three achieved Dice ≥0.75. No consistent deviation patterns were observed. Conclusion: Combined CT and MRI planning improved implant positioning accuracy by better accounting for variations in hip capsule morphology. MRI integration demonstrated superior performance over CT‑only planning for patient‑specific shelf implant placement.