Abstract
Developmental dysplasia of the hip is a condition in which the acetabulum provides insufficient coverage of the femoral head in the hip joint. This configuration gives poor biomechanical load distribution, with increased stress at the superior aspect of the joint surfaces, and can often lead to degenerative arthritis. Morphologically, the poor coverage may be due to an acetabulum that is too shallow or oriented in valgus.
The dysplastic deformity can be treated surgically with a group of similar procedures, often labeled periacetabular osteotomies or rotational acetabular osteotomies. Each involves separating the acetabulum from the pelvis and fixating the fragment back to the pelvis in an orientation with increased coverage of the femoral head. This redistributes the biomechanical loads relative to acetabulum.
Bone remodeling at the level of trabeculae is an accepted concept under research; however, it is unclear whether the hip undergoes gross morphology changes in response to changes in biomechanical loading. An understanding of the degree to which this remodeling occurs (if at all) may have an impact on surgical planning.
In this retrospective study, computed tomography (CT) scans of 13 patients (2 male, 11 female, 40 ± 9 years of age) undergoing unilateral periacetabular osteotomies were examined; scans were taken both pre-operatively and at least a year post-operatively with an in-plane resolution of 0.55 mm and a slice thickness of 1.25 mm. Scans were segmented to produce triangulated meshes for the proximal femurs and the pelvis. These scans were manually processed to isolate the articular portions of the femoral heads and acetabulums, respectively; the fovea, acetabular fossa, any osteophytes and any segmentation artifacts were excluded.
Post-operative meshes were registered to their pre-operative counterparts for both the femoral head and the acetabulum, for both the operative and non-operative hips, using the iterative closest point (ICP) algorithm to 20 iterations. To account for differences in defining the edges of the articular surfaces in the manual isolation, metrics were only calculated using points that were within 0.3 mm of a normal from the opposing mesh. With the resulting matched data, nearest neighbour distances were calculated to form the remodeling metrics. Select spurious datapoints were removed manually.
For the operative femoral heads, the registered post-operative points were 0.24±0.53 mm outside of the pre-operative points. The maximum deviation was on average 1.94 mm with worst-case of 2.99 mm; the minimum deviation was −0.62 mm with worst-case of −2.06 mm. Positive numbers indicate the post-operative points are ‘outside’ of the pre-operative points – that is, farther from the head centre. The non-operative femoral heads have similar deviation values, 0.21±0.46 mm outside, with maximum and minimum deviation averaging to 1.24 mm and −0.74 mm respectively, with worst cases of 2.99mm and −1.80mm.
For the operative acetabulums, the post-operative deviations were −0.08±0.43mm. The maximum and minimum deviations averaged to 0.62mm and −0.82mm, with worst cases of 2.14mm and −1.51mm across the set. Again, the non-operative acetabulums were very similar; post-operative deviations were −0.02±0.43mm, maximum and minimum deviations averaged to 1.24mm and −0.65mm, with worst cases of 1.97mm and −2.00mm.
These quantitative measurements were reflected in manual examination of the meshes; generally speaking, there were small deviations with no overarching patterns across the anatomy.
All metrics were very similar across the same anatomy (that is, femoral head or acetabulum) regardless of whether the hip operative or non-operative. Femurs tended to ‘grow’ slightly post-operatively, but by less than a half voxel in size. Given that the CT voxels are large compared to the measured deviations, it is possible the results may be sensitive to the manual segmentations used as source data.
Manual examination of the deviations indicated a few potential trends. Seven operative and eleven non-operative acetabulums had a small patch of positive deviation (1mm to 1.5mm) in the anterosuperior aspect. This can be seen in the plot as the yellow-red area near the top right of the leftmost rendering. Other high-deviation areas included the superior aspect of the acetabulum (both positive and negative) and the superior aspect of the femoral head (generally positive).
The edges of the mesh were often a source of high deviation. This is likely an artifact of over-inclusion the manual isolation of the articular surfaces, as joint surfaces become non-articular as they move away from the joint interface.
Overall, the superior and anterosuperior aspects of the acetabulum and the superior aspect of the femoral head showed some indication of systemic changes; further study may clarify whether these data represent consistent anatomical changes. However, as the magnitude of the deviations between pre- and post-operative scans are on or below the order of the CT voxel size, we conclude that (in the absence of other strongly compelling evidence) periacetabular osteotomies for adults should be planned without the expectation of gross remodeling of the articular surfaces.
