Computed Tomographic Analysis of the Sagittal Orbit-globe Relationship.

Identifying the normal relationship of the orbital rims to the globes is critical in planning surgical correction of craniofacial deformities affecting the orbit. This article illustrates a technical proof of principle and mathematical basis for a computed tomography-based measurement of the sagittal orbit-globe relationship. The technique does not require subject cooperation and is, therefore, optimal for pediatric craniofacial surgical treatment planning and outcome evaluation.

INTRODUCTION

Euophthalmos, the normal relationship of the orbital rims to the globes, is critical for the planning and surgical correction of craniofacial deformities that affect the orbit (e.g., Crouzon and Apert syndrome, achondroplasia, etc.).[1] The establishment of a reproducible facial anthropometric measurement is challenging. When measuring the relationship between the globe and the orbit, surgeons are interested in the distance between landmarks in the anteroposterior plane to answer the question, “How deeply recessed are the patient's eyes?” To answer this question, the distance calculated should not be the distance between two anatomic landmarks in three-dimensional (3D) space, but rather the distance only in the A-P direction. Second, the head and face should be oriented with the facial horizontal plane parallel to the true horizontal plane or in the case of a computed tomography (CT)-based method, the axial plane. Any tilting of the head will alter the distance in anteroposterior dimension from the globe to orbital landmarks. Various devices, such as the orbital anthropometer, can be used to assess the relationship between the globes and peri-orbital soft tissue.[2] While this technique has been employed over the decades, it is difficult to perform in infants and young children, as they do not tolerate the time needed for measurement. In this communication, we describe a CT-based method for measuring the sagittal orbit-globe relationship and provides the mathematical basis for the technique.

MATERIALS AND METHODS

Anatomic landmarks

The anatomic landmarks used in this study have been adapted from traditional anthropometric and cephalometric definitions.[34] They have been modified to be used in a 3D computer environment exhibited by most DICOM viewers. In addition to identifying the nasion (n), there are four peri-orbital soft tissue landmarks for each eye: orbitale superius (os); orbitale inferius (oi); orbitale laterale (ol); anterior cornea (ac). In addition, the bony landmarks Orbitale (Or) and Porion (Po) are used in the reorientation of the axial plane parallel to the frankfort horizontal. landmark and plane definitions are shown in Table 1.

Table 1

Landmark and plane definitions

Image manipulation and calculations

To make the appropriate measurements parallel to the facial horizontal plane, the axial plane is reoriented to coincide with the facial horizontal plane, which we define as the Frankfort Horizontal plane extrapolated to 3D [Table 1 and Figure 1a]. Next, for the measurements to be in the AP direction only, the sagittal plane is defined by choosing two points in the mid-sagittal plane (anterior and posterior nasal spines) [Figure 1b]. These two points comprise a vector lying in the facial horizontal plane, parallel to the mid-sagittal plane. A coronal plane perpendicular to this vector can then be defined and used as a reference point for measurements. Next, the fiducial points described above are digitally placed and the distances between these points and the coronal plane are calculated. Finally, subtracting the distance from the anterior cornea to the coronal plane from all of the landmark-coronal plane distances yields the distance from the anterior cornea to other anatomic landmarks in the A-P direction only, parallel to the facial horizontal [Figure 2]. The mathematical basis for this technique is shown in Table 2.

Figure 1

Defining the coronal plane. First, the axial plane is oriented to the Frankfort horizontal (the plane defined by porion (right arrow) and inferior orbitale (left arrow) (a). Second, two points are chosen in the axial plane parallel to the mid-sagittal plane to define a vector perpendicular to the desired coronal plane (b)

Figure 2

Diagrammatic overview of landmarks and calculations

Table 2

Mathematical basis for calculating the distances from the globe to soft-tissue landmarks which are parallel to the facial horizontal and sagittal planes

Sample calculation

To demonstrate the application of this technique, we applied this to a 13-year-old female who underwent a maxillofacial CT scan following a motor vehicle accident. The scan was acquired without intravenous contrast using helical acquisition with a 320 detector scanner with 0.5 mm collimation (Toshiba Aquilion 320, Toshiba Inc, Tokyo, Japan), with 1 mm bone and soft tissue reconstructions performed in three orthogonal planes. The DICOM viewer used was OsiriX (OsiriX v5.8.5 Foundation, Geneva, Ch).[3] Landmarks were identified by a fellowship trained neuroradiologist. Mathematical calculations were performed in an electronic spreadsheet (Excel 2011, Microsoft Corp, Redmond, WA).

RESULTS

A single DICOM file of a maxillofacial CT belonging to the index patient was imported into the DICOM viewer. The image was reoriented such that the axial plane corresponded to the facial horizontal plane as described above. The technique outlined above was performed, and the resultant measurements are listed in Table 2. Negative numbers correspond to the anterior cornea being posterior to the landmark in comparison. For example, n-ac of − 7.6 means the anterior cornea is 7.6 mm posterior to the soft tissue nasion. The sample calculations were performed on a CT scan from a 13-year-old female who experienced head trauma. The CT scan showed a left retrobulbar hematoma associated with multiple facial and skull base fractures causing mild left eye proptosis. Concordantly, the left anterior cornea is noted to be more anterior relative to adjacent soft-tissue landmarks when compared to the right side [Figure 3 and Table 3].

Figure 3

Computed tomography landmarks placed for an index patient: Right orbitale superius (top arrow), anterior cornea (middle arrow), orbitale inferius (bottom arrow) (a); nasion (arrow) (b); right and left orbitale (arrows) (c)

Table 3

Measurements of the sagittal orbit-globe relationship in sample patient

DISCUSSION

Identifying the normal relationship of the orbital rims to the globes is critical for the planning and surgical correction of craniofacial deformities. This article demonstrates a technical proof of principle of a CT-based method for measuring the sagittal orbit-globe relationship. A single index case demonstrated the utility of this method and the ability to identify asymmetric sagittal orbit-globe relationships.

Fundamental to this technique is the use of mathematical principles to overcome challenges related to head positioning. In a 3D Cartesian coordinate system, any plane in space can be defined mathematically if a single point within the plane and a vector perpendicular to the plane are known. If the perpendicular vector lies in both the facial horizontal and sagittal planes, then the mathematically defined plane perpendicular to this vector is the coronal plane. Once this particular coronal plane is defined, the shortest distance between an anatomic landmark and that coronal plane will be solely in the A-P dimension and parallel to the facial horizontal.

There are limitations to the method demonstrated here. Abnormal anatomy may affect the ability to define the facial horizontal plane using the Frankfort horizontal. In such cases, an alternative method of determining the facial horizontal plane should be used.[56] In addition, The inter and intra-examiner reliability of this technique is unknown. However, the variability of facial landmark placement has been studied and typically is in the realm of 0–2 mm.[7] This technique requires radiation, however many patients with congenital or acquired craniofacial abnormalities are evaluated with CT already and would therefore not require any additional radiation. The mathematical basis within can apply to magnetic resonance imaging (MRI) as well, however, soft tissue landmarks may be more difficult to identify. MRI typically has a lower spatial resolution than CT, and with longer acquisition times, motion can be a limiting factor, in particular for the anterior cornea measurements.

CONCLUSIONS

The CT-based method for measuring the sagittal orbit-globe relationship demonstrated above is a mathematically reliable representation of clinically performed measurements. Clinical measurements are time-consuming, and accordingly are typically only performed on patients with orbital asymmetry, or those with congenital or acquired maxillofacial abnormalities. This virtual technique can be relatively quickly performed, allowing for the creation of a database of normative values, which for practical reasons cannot be easily acquired by conventional techniques. Such normative data when compared to preoperative patient measurements could assist in surgical planning and outcome evaluation.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.