Patient-specific 3D-printed shelf implant for the treatment of hip dysplasia tested in an experimental animal pilot in canines

The concept of a novel patient-specific 3D-printed shelf implant should be evaluated in a relevant large animal model with hip dysplasia. Therefore, three dogs with radiographic bilateral hip dysplasia and a positive subluxation test underwent unilateral acetabular augmentation with a 3D-printed dog-specific titanium implant. The contralateral side served as control. The implants were designed on CT-based pelvic bone segmentations and extended the dysplastic acetabular rim to increase the weight bearing surface without impairing the range of motion. Outcome was assessed by clinical observation, manual subluxation testing, radiography, CT, and gait analysis from 6 weeks preoperatively until termination at 26 weeks postoperatively. Thereafter, all hip joints underwent histopathological examination. The implantation and recovery from surgery was uneventful. Clinical subluxation tests at the intervention side became negative. Imaging showed medialization of the femoral head at the intervention side and the mean (range) CE-angle increased from 94° (84°–99°) preoperative to 119° (117°–120°) postoperative. Gait analysis parameters returned to pre-operative levels after an average follow-up of 6 weeks. Histology showed a thickened synovial capsule between the implant and the femoral head without any evidence of additional damage to the articular cartilage compared to the control side. The surgical implantation of the 3D shelf was safe and feasible. The patient-specific 3D-printed shelf implants restored the femoral head coverage and stability of dysplastic hips without complications. The presented approach holds promise to treat residual hip dysplasia justifying future veterinary clinical trials to establish clinical effectiveness in a larger cohort to prepare for translation to human clinic.


Standardized CT-scan protocol
Under a general anesthesia protocol (Appendix 1.5.1) the dogs were positioned in dorsal recumbency with extended and slightly internally rotated femora, consistent with hip dysplasia (HD) position I for radiographic canine HD screening programs (Fig. S1). In the ventrodorsal HD position I radiograph the pelvis is exactly symmetrical, left and right femur are parallel and the patella is positioned midline on the distal femur. The CT scans were made using a 64 slice CT scanner (Siemens Somatom Definition AS, Siemens Healthcare) with the following standardized parameters: 120 kV, 250 mas, 1000 ms tube rotation time, 0.6 mm slice thickness, 0.35 spiral pitch factor, 512 × 512 pixel matrix. Reconstructions were made in transverse and sagittal planes using soft tissue and bone reconstruction kernels and images were reviewed in soft tissue/bone settings (window length 50, width 300, and window length 600, width 3000, respectively).

Anesthesia and analgesia protocols a. Protocol during imaging:
For CT and radiographs anesthesia was provided by intravenous dexmedetomidine (10 μg/kg), induction by intravenous propofol (1-2 mg/kg) and maintenance by inhalation of 1-1.5% isoflurane. After imaging, anesthesia was antagonized with atipamezole depending on the administered dose of dexmedetomidine.

b. Protocol during implantation:
For the implantation of the personalized 3D implant anesthesia was provided by preoperative intravenous administration of dexmedetomidine (10 μg/kg), induced by intravenous propofol (1-2 mg/kg) and maintained by intravenous propofol (100 mg/kg/hr) and inhalation of 1-1.5% isoflurane. A morphine injection (0.1 mg/kg) diluted with levobupivacaine (1ml/5kg) was administered in the epidural space between L7 and S1 on top of the cauda equina. After postoperative imaging, the anesthesia was antagonized with atipamezole depending on the administered dose of dexmedetomidine. At induction dogs received antibiotic prophylaxis consisting of cefazoline 20 mg/kg intravenously, repeated every 90 minutes during surgery.
Postoperative analgesia was provided by intravenous buprenorphine for 24 hours (20 μg/kg, one day, three times daily) and subcutaneous carprofen (4 mg/kg). Thereafter analgesia was continued by oral administration of carprofen (2 mg/kg, 7 days, twice a day) and tramadol (Tramagetic once Daily 200 mg, 3-7 days on indication, twice a day. Postoperative antibiotic prophylaxis consisted of oral administration of amoxycillin / clavulanic acid (Synulox; 12.5 mg/kg PO, 7 days, twice a day).

Surgical approach
All implantations were performed by one surgical team (a board certified veterinary surgeon and a medical doctor) to prevent variations in placement and screw insertion torque.
The skin incision was centred at the level of the greater trochanter and was placed over the cranial border of the shaft of the femur. Distally, it extended one third the length of the femur; proximally, it curved slightly cranially to end just short of the dorsal midline. The skin margins were undermined and retracted. An incision was made through the superficial leaf of the fasciae latae, along the cranial border of the biceps femoris muscle. The biceps femoris muscle was retracted caudally to allow incision in the deep leaf of the fasciae latae to free the insertion of the tensor fasciae latae muscle. The incision continued proximally through intermuscular septum between the cranial border of the superficial gluteal muscle and the tensor fasciae latae muscle. The fasciae latae and the attached tensor fasciae latae muscle were retracted cranially and the biceps caudally.
Blunt dissection and separation along the neck of the femur with the fingertip allowed visualization of a triangle bounded dorsally by the middle and deep gluteal muscles, laterally by the vastus lateralis muscle, and medially by the rectus femoris muscle. The insertion of the rectus femoris muscle was exposed by tenotomy of a portion of the deep gluteal tendon close to the trochanter, leaving enough tendon on the bone to allow suturing. The deep gluteal muscle was split proximally, parallel to its fibres, and the pedicle was allowed to retract exposing the cranial and dorsal rim of the acetabulum. Using a periosteal elevator, the deep gluteal muscle was elevated from the bone cranially and dorsally to the acetabulum to free the acetabular rim for the acetabular rim extension implant. The joint capsule was covered by areolar tissue, which was carefully cleared away by blunt dissection without opening the joint. The ilium bone was exposed ventrally, just cranial to the rectus femoris muscle to allow the implant to curve around the ventral border of the ilium. The implant was fitted to its designated site at the rim of the acetabulum and after rechecking its proper placement, a 3.5 mm drill guide was attached to the middle screw hole and the first hole was drilled with the 2.8 mm drill under continuous irrigation with saline solution. Next, the depth was measured using a depth gauge and double-checked with the pre-operative planning. The

Histology (Methods)
After six months of follow-up, all dogs were euthanized. Each hip joint was harvested using an electric multipurpose saw (Bosch PSA 700, Gerlingen, Germany)(Main manuscript: Figure 6A The intervention side demonstrated a normal volume of cartilage with smooth surface and all zones of the acetabular and femoral cartilage were intact. The corresponding Safranin O / Fast Green appeared also to be normal. There were no abnormalities observed in the tide mark, nor subchondral changes.
Synovial lining was composed of 1-2 layers of cells, while villous hyperplasia and cell infiltrates were absent. Synovium was negative for proteoglycan and collagen type II.

Video 1.
In silico 3D model of the hind limb of a dog showing hip instability due to a dysplastic acetabular rim followed by the virtual implantation of a 3D printed shelf implant. After implantation the hip stability improves without limiting the hip's range of motion.

Video 2.
Video demonstrating the preoperative Ortolani test in dog #1. A distinct audible and palpable click is observed indicating reduction of the femoral head in the acetabulum marking a positive Ortolani test.

Video 3.
Video demonstrating the postoperative Ortolani test in dog #1. No audible or palpable click is observed indicating a stable hip joint that does not allow subluxation of the femoral head marking a negative Ortolani test.