The aims of this study were to determine the sensitivity, specificity, positive and negative predictive values of ultrasonography in detecting zygomaticomaxillary complex fractures, and to highlight factors that may affect the validity of ultrasonography in the diagnosis of zygomaticomaxillary complex fracture. Twenty-one patients with suspected fractures of the zygomaticomaxillary complex presenting at the authors’ hospital were included in this prospective study. All the patients had plain radiographic and computed tomography (CT) investigations. All underwent ultrasonographic examination of the affected region using an ultrasound machine with a 7.5 MHz probe. The different radiologists were not aware of the results of the other two investigations. Statistical significance was inferred at P < 0.05. The validity of ultrasonography varied with fracture sites with a sensitivity of 100% for zygomatic arch fractures, 90% for infraorbital margin fractures and 25% for frontozygomatic suture separation. Specificity was 100% for the three types of fracture. There was no statistically significant difference in the ability of CT scan and ultrasonography to diagnose fractures from various zygomaticomaxillary complex fracture sites ( P = 0.47). Ultrasonography has proved to be a valid tool for the diagnosis of zygomatic arch and displaced infraorbital margin fractures.
The zygomaticomaxillary complex (ZMC) plays a key role in the function and appearance of the facial skeleton. The prominent convex shape of the zygoma gives the contour of the cheek and makes it vulnerable to traumatic injury. The frequency of ZMC fractures is about 45%, and is second only to nasal fractures, which are the most common type of middle face fracture. Following trauma, the alignment of the zygomatic complex is important for facial appearance. Persistent dislocation of the zygomatic arch is a cosmetic problem and it hinders the normal excursion of the coronoid process of the mandible, resulting in limited opening with its attendant functional challenge.
Computed tomography (CT), both coronal and axial planes, is often requested in complex cases to assess blow-out fractures and disruption of the orbital walls because they are considered the gold standard for radiologic diagnosis of ZMC fractures. The drawbacks of CT are the high radiation dose exposure to patients and the potential risk of development of a cataract. Delay in extricating the patient from the machine in an emergency may prove fatal.
Magnetic resonance imaging (MRI) has also been used to assess ZMC fractures. It provides substantially more information about the soft tissue structures than about the underlying bone fractures. Both CT and MRI machines are expensive to buy and operate, and they are not available in many hospitals in Nigeria. With limited access to CT and MRI, standard two-dimensional radiographs continue to be used routinely to diagnose zygomatic bone fractures. These standard views are inadequate for identification of minimal displacement of ZMC fractures due to the superimposition of overlying structures.
Traditionally, ultrasound (US) has been used extensively to image diseases of soft tissues in the abdomen and pelvis. It has been used extensively in the diagnosis of orbital and ocular disease. Until recently, it found little application in oral and maxillofacial surgery including its limited use in the assessment of facial trauma, including ZMC fractures. A few researchers have documented their experience with US in the diagnosis and management of facial trauma. Forrest et al. reported 94% correlation between US and CT findings. Akizuki et al. also reported 92% sensitivity, 100% specificity and 100% positive predictive value compared with CT. McCann et al. documented 85% accuracy between US and conventional radiography.
The purpose of this study was to contribute to ongoing work in the field. It will determine the sensitivity, specificity, positive and negative predictive values of US in detecting ZMC fractures, and highlight factors that may affect the validity of US in the diagnosis of ZMC fracture.
Materials and methods
Twenty-one consecutive patients with fractures to the ZMC, presenting at the authors’ hospital during an 18-month period, were included in this prospective study. Written informed consent was freely obtained from each study participant following a clear explanation of the study objectives and approaches. The procedure and treatment given were also clearly explained to them throughout the study period.
All the patients had plain radiographic investigations carried out by the same radiographer. The standardization of the plain radiographic measurement was done during a pretest period and involved the first author and radiographer, during which criteria for grading the distances between the fractured segments were agreed. For standardization of measurements, repeated radiographs of the same individual were read blindly by the first author and measurements obtained were compared. The difference between the measurements was not significant. The radiographic views included an occipitomental view of the skull (for zygomatic complex fractures) and/or a submentovertical view of the skull (for zygomatic arch fracture).
All the patients had CT investigations using an axial 5 mm thin slice CT (Somatom multislice; Somatom AR. T Siemens, Germany). The CT investigation was performed by a radiologist who was unaware of the results of the plain radiographs.
All the patients had US examination of the affected region carried out by a radiologist using an FF Sonic UF 750 XT (Fukuda Denshi Co. Ltd, Japan) ultrasound machine, generating B mode with a 7.5 MHz curvilinear probe. The US investigation was conducted with the patients in the supine position. A systematic routine of facial scanning was developed, to include vertical and horizontal margins of the orbit to visualize the orbital floor, medial and lateral walls, infraorbital margin and foramen as well as the lateral wall of the maxillary sinus. The zygomatic arch and frontozygomatic suture were also scanned to detect any depression, discontinuity or displacement. The procedure was repeated on the corresponding bone on the other side of the face. The US examination was conducted by the same radiologist to avoid inter-observer variation. This radiologist was different from the one who conducted the CT scan. The radiographer was unaware of the plain radiographs and CT scan diagnoses.
Data analysis was carried out using SPSS version 16.0 (SPSS Inc., Chicago, IL, USA). Analysis included frequencies and cross tabulations. Associations between the discrete variables were tested with Pearson’s χ 2 test. If the frequency of the variables in the cells of the contingency tables was insufficient to meet the assumption of the Pearson’s χ 2 test, the likelihood ratio was used in accordance with standard statistical norms. In addition, wherever tabulated data in the results contained a cell that was zero, the Computer Programme for Epidemiological Analysis was used. The computer analysis programme used PEPI which added 0.0000001 to all the cells that were zero in order to meet one of the assumptions for the use of χ 2 . The test of agreement was done using the kappa coefficient. Statistical significance was inferred at P < 0.05.
The epidemiological methods of validity (sensitivity and specificity) were determined by comparing US findings with those of plain radiographs and also with those of CT investigation ( Fig. 1 ). In the context of this study, sensitivity is defined as the proportion of cases identified by CT scan as having fractures (zygomatic arch, infra-orbital margin and fronto-zygomatic suture separation) and which US also identified as having the fractures. Specificity is defined as the proportion of cases identified by CT scan as not having fractures and which US also identified as not having fractures. The positive predictive value (PPV) is the total proportion of cases identified by US as having fractures, but which the CT scan identified as not having fractures. The negative predictive value (NPV) is the total proportion of cases identified by US as not having fractures but which the CT scan identified as having fractures. Sensitivity, specificity, PPV and NPV were calculated using SPSS.
The patients ages ranged from 12 to 45 years with a mean (±standard deviation) of 26.7 (±7.8) years. Males constituted 81.0% (17/21) of the cases. The male to female ratio was 4.3:1.
Comparison of US and plain radiography findings
The US and plain radiograph findings were compared, with the latter as the standard. The US detected 18 out of the 19 infraorbital margin fractures diagnosed by plain radiographs ( Table 1 ). It showed a false negative in one case. The undetected fracture was not displaced. The agreement between the diagnosis of plain radiograph and US for infraorbital margin fractures was strong ( k = 0.77) and statistically significant ( P < 0.001).
|Infra-orbital margin||19 (67.86%)||18 (66.67%)|
|Frontozygomatic suture||2 (7.14%)||1 (3.70%)|
|Zygomatic arch||7 (25.0%)||8 (29.63%)|
|Total||28 (100%)||27 (100%)|
One (50%) of the two frontozygomatic suture separation cases diagnosed by plain radiographs was detected by US. The agreement ( Table 2 ) between the diagnosis of frontozygomatic suture separation by plain radiographs and US was moderate ( k = 0.64) and statistically significant ( P = 0.001).
|Anatomical site||US||Plain radiographs||Kappa coefficient||P value|
|Fracture present||Fracture absent||Total|