Ethics approval
Animal handling was in accordance with the European Directive for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes (86/609/EU). The experiments were approved by the National Central Committee for Experiments on Animals (CCD) and a maximum of three experimental dogs could be used to evaluate safety (AVD1080020173505) after which a new clinical trial should be started to investigate effectiveness in symptomatic dog patients. The working protocol (WP3505-01-1) was further supervised by the local Animal Welfare Body and followed the ARRIVE guidelines.
Study design
Prior to the implantation of the personalized 3D-printed implant (T = 0) and during the 6 months follow up period, clinical observation, manual subluxation testing, imaging, gait analysis were conducted. Upon termination of the study, histology of the hip joints was performed (Table 1).
Animals
In this pilot study three female mongrel dogs (Marshall, North Rose, New York) with natural occurring, radiographically confirmed, asymptomatic bilateral hip dysplasia were included. The mean (range) age of the dogs was 25 (24–25) months and the mean body weight was 26 (24–29) kg. The hip with the worst dysplastic parameters15 based on radiological examination and manual subluxation (Ortolani16) testing (Fig. 1A–D) was chosen as the intervention side for the 3D shelf implant (N = 3) and the contralateral hip served as control (N = 3) (Table 2). All subluxation tests were performed under general anesthesia by two board-certified veterinary surgeons who were blinded for each other’s results. The three dogs were housed in a group enclosure with cage enrichment and were put on an ad libitum diet. Furthermore, the dogs were housed with a regular 24-h day-night rhythm and were allowed in an outdoor pen at least twice daily.
Laxity due to hip dysplasia is confirmed based on clinical examination (A-D) and is counteracted by implantation of the 3D-printed shelf implant (E). (A) The limb is in neutral flexion and in an adducted position, and force is applied toward the dorsum of the dog along the femoral axis (red arrows). (B) This force causes dorsal subluxation in a hip with joint laxity due to hip dysplasia. (C) During the Ortolani (reduction) test, the limb is slowly abducted (blue arrow) while force on the femur (red arrows) is maintained. (D) A positive Ortolani sign is evident when a click is heard or palpated as the subluxated femoral head reduces into the acetabulum (green arrows)17. (E) Introduction of the shelf implant ideally stabilizes the joint by reinforcing the hip capsule and labrum as a weight bearing and stabilizing surface (purple arrows). In close-up the internal 2 mm offset of the implant is visible that allows the capsule attachment to remain unaffected (F).
The intervention/imaging
At the initiation of the study (− 6 weeks), a CT-scan with a standardized protocol (Appendix 1) was made of the entire pelvic area and femora (120 kV, 250 mas, 0.6 mm slice thickness). The CT scans were semi-automatically segmented using imaging processing software, Mimics Medical 21.0 (Materialise, Leuven, Belgium). Standardized bone threshold values (HU 226—upper boundary) were used to guide the semi-automatic CT-based anatomical model. This model was saved and transferred using Stereolithography (STL file) to design the 3D shelf implant.
Implant design
The patient-specific 3D-printed acetabular shelf implants were designed, by the primary author using Freeform Plus software (Geomagic, 3D Systems, Leuven, Belgium), as described prior by Willemsen et al.11 (Figs. 2 and 3). The implants consisted of two subsections; the ‘rim extension part’ and the ‘implant-bone interface or attachment part’. For the rim extension part a 20°–30° increase in CE-angle was pursued and the effect on the range of motion was monitored by performing an in silico range of motion (ROM) simulation, for each individual hip. Thereafter, the outcomes were reviewed with a board-certified surgeon and the design was altered if clinically needed (Figs. 1E, 3) (Video 1). The rim of the acetabulum received an offset of 2 mm not to interfere with the attachment of the joint capsule on the acetabular rim and to allow the hip capsule to be interposed between the implant and the cartilage of the femoral head (Figs. 1F, 3).
Rendering of a canine pelvis with a 3D-designed acetabular rim implant for the left dysplastic hip. Orientation: left is cranial, top is dorsal.
Digital rendering of the implant designed for dog #1. (A) The external implant surface with the clockface positions (green arrows) on the rendered implant. (B) The internal implant surface shows the internal offset (X) that allows the capsule attachment to remain unaffected and the 70% porous inner shell (Y) allowing bone ingrowth for osseous integration and secondary implant fixation.
The implant-bone interface part was also designed patient specific to be able to incorporate 5 locking screws and an additional ilium flange for ease in positioning and for additional stability. Thereafter, the implant bone interface was designed with a porous (70%, 1 mm sized Dode-Medium unit cell) inner shell to optimize bone ingrowth, osseointegration and secondary implant fixation (Fig. 3). Locking screw holes were planned in the implant in such a way that the screw trajectory remained sufficiently distant from the acetabulum but at the same time purchasing the maximal possible bone stock for the preferred screw length. The screws were placed bi-cortical and generally not parallel to each other.
The implants were manufactured from medical grade titanium alloy Ti-6Al-4V ELI grade 23 by direct metal printing using a ProX DMP320 machine (3D Systems, Leuven, Belgium). Postprocessing included hot-isostatic-pressing, polishing, screw wiretapping and a standard intermediate cleaning step (incl. ultrasonic cleaning and automated cleaning) by the implant manufacturer. Additionally, final (manual) cleaning and autoclave sterilization was performed by our in-house sterilization facility.
Orthogonal radiographs and CT of the pelvis and hips were made at − 6, 0, + 6, + 12, + 26 weeks from the implantation (Table 1) and parameters such as the center-edge (CE)-angle18 were assessed by a board-certified veterinary radiologist. Subsequently, the CT-scans were uploaded into image analysis software Mimics (Medical v20, Materialise, Leuven, Belgium) to calculate the percentage of femoral head coverage by using multiplanar reconstruction19. The acetabular coverage was measured in + 20° posterior pelvic tilt in relation to the cranial–caudal axis to simulate the functional standing posture of a dog19. Additionally, the accuracy of the placement was analyzed by rigidly overlaying the preplanning with the postoperative 3D models with an iterative closest point (ICP) algorithm20 and subsequently calculating the average implant transformation matrix in mm21.
The surgeries were performed by a board-certified veterinary surgeon under a standardized general anesthesia protocol (Appendix 2b) and consisted of a cranio-dorsal approach (Appendix 3) to the hip joint leaving the joint capsule intact22. The implant was fitted to the bone and positioned over the hip joint capsule and was fixated with five 3.5 mm locking screws (DePuys Synthes, Raynham, Massachusetts, USA). Full weight bearing was allowed directly postoperatively and reintroduction of the dog into the study group from an individual cage was done 24 h postoperatively. Due to the surgical nature of the intervention only blinding occurred during histological examination.
Outcomes
General health assessments, orthopedic examinations, and subjective locomotion evaluations were performed weekly during the whole experiment. The subluxation (Ortolani) tests of the hips were assessed under general anesthesia at − 6, 0, 6, 12, 26 weeks (Fig. 1) (Table 1).
Gait analysis was performed using a standardized gait protocol (Appendix 4) using a force plate23,24 for objective evaluation of vertical (Fz) ground reaction forces (N/kg) measuring differences between the intervention and control limb and the distribution ratio between front-limb and hind-limb loading before surgery at − 1 (baseline) and after surgical intervention at 1, 2, 4, 8, 12, 16, 20, and 26 weeks (Table 1).
The dogs were followed for 26 weeks to allow enough time for initial surgical recovery, secondary implant fixation and to asses tissues changes to the joint capsule or cartilage after implant intervention. At final follow‐up, the dogs were euthanized (Appendix 2c). Each hip joint was harvested and macroscopically evaluated before histological examination was performed on the capsule and femoral and acetabular cartilage of the decalcified joints using a standardized staining protocol for Hematoxylin and Eosin (HE) staining, and Safranin O/Fast Green staining19 (Appendix 5).
