CORR Insights^®: Sex-specific Differences of the Infraacetabular Corridor: A Biomorphometric CT-based Analysis on a Database of 523 Pelves

Where Are We Now?

In the current study by Gras and colleagues, the authors used biomorphometric analysis of 523 pelves to determine the orientation and diameter of the infraacetabular corridor. The authors found that a minimum 5 mm corridor was present in 93% of pelves. This work may result in improved surgical techniques for the placement of infraacetabular screws, which can be challenging, and which may help improve the surgical treatment in patients with certain patterns of acetabular fracture. The strength of this paper is the large quantity of scans used and the objective technique used to determine the size and location of the corridor.

The use of large anatomic datasets is becoming much more common in orthopedics and has moved from research to development. For example, the SOMA™ database (Stryker Orthopaedics, Mahwah, NJ, USA) consists of more than 12,500 bones [3]. Stryker has used this database for determining the fit of fracture fixation plates and joint replacements in an attempt to improve the design of these devices. The ADaM™ service (Materialise NV, Leuven, Belgium) utilizes advanced statistical methods to analyze 3-D patient anatomy.

However, it is important to understand the details of the models used. In this paper, the pelvis models only include the periosteal surface. The cortex was not modeled. If there is variation in cortical thickness, for some patients, there may not be room for the infraacetabular screw even if the diameter of the infraacetabular corridor were 5 mm as measured to the periosteal surface. In addition, a majority of the pelvis scans (70%) were taken from CT scans from patients who required a CT angiography. These patients did not have a periacetabular fracture. A malreduced fracture may further reduce the corridor width.

Where Do We Need To Go?

These issues can be addressed, and the techniques developed in this paper can be applied to many interesting clinical problems in orthopaedics. However, the anatomic models used must have the detail and clinical relevance to the question being asked. In this paper, it would have been better to use a pelvis model that included the endosteal surface. This can be done by segmenting the scan based on Hounsfield unit with a threshold of around 600 Hounsfield units. This value has been shown to be the best match of the cortico-cancellous interface and will likely better model the true size of the corridor [1]. The scans should also be from patients with a periacetabular fracture.

In a similar fashion, if an anatomic database is used to design an implant for a patient population with osteoarthritis (OA), then the database should contain anatomy from patients with OA. The anatomy of patients with OA is not the same as patients without OA [2, 4].

How Do We Get There?

As with all computer models, the key step is validation. In-vitro simulation of this model is complex. Ideally, a clinical trial should be done to validate the results of this model. A preoperative CT scan is commonly done to plan these surgeries so the method used in this paper can be used for preoperative planning. Next, the prospective determination of screw size and trajectory can be compared to the actual surgical procedure. This is the best way to assure the essential factors were included in the model. Modeling the most-relevant variables is necessary to provide clinically useful information that can guide the treatment of real world conditions.