Temporal Bone - An Imaging Atlas
Von: John I. Lane, Robert J. Witte
Imaging of the temporal bone has recently been advanced with multidetector CT and high-field MR imaging to the point where radiologists and clinicians must familiarize themselves with anatomy that was previously not resolvable on older generation scanners. Most anatomic reference texts rely on photomicrographs of gross temporal bone dissections and low-power microtomed histological sections to identify clinically relevant anatomy. By contrast, this unique temporal bone atlas uses state of the art imaging technology to display middle and inner ear anatomy in multiplanar two- and three-dimensional formats. In addition to in vivo imaging with standard multidetector CT and 3-T MR, the authors have employed CT and MR microscopy techniques to image temporal bone specimens ex vivo, providing anatomic detail not yet attainable in a clinical imaging practice. Also included is a CD that allows the user to scroll through the CT and MR microscopy datasets in three orthogonal planes of section. TOC:Chapter 1. Imaging Technique: Imaging Microscopy.- CT Microscopy.-MR microscopy.- Clinicical Imaging.- Volumetric Multidetector CT.- High Field MR.- Post-processing.- Chapter 2. Anatomy - Middle Ear.- Inner Ear.-Internal Auditory Canal.- Chapter 3. Multiplanar Atlas.- Axial Plane (in the plane of the Lateral Semicircular Canal).- Coronal Plane (perpendicular to the plane of the Lateral Semicircular Canal).-Pöschl Plane (Short axis of the Temporal Bone).- Stenvers Plane (Long axis of the Temporal Bone).- Chapter 4. Advanced Imaging Applications.- Chapter 5. The Temporal Bone Anatomy Tool (CD).
"1 Imaging Technique (p. 1-2)
The gold standard for studying temporal bone anatomy has been histological sectioning by microtome following chemical fi xation and deossifi cation of the temporal bone specimen. This process introduces signifi cant artifacts visible even at low power magnifi cation. Although imaging microscopy cannot compete with histology at higher powers of magnifi cation (i.e., the cellular level), resolution at the 20–100 mm level is achievable, allowing study of the temporal bone without the aforementioned artifacts induced by fi xation and sectioning.
Computed tomography (CT) microscopy has been primarily used in the imaging of small animals in the research laboratory. There are several obvious advantages to imaging microscopy in the study of the temporal bone compared with standard microtomed histological preparations. MicroCT avoids signifi cant tissue destruction and artifacts introduced during the sectioning process such as fractures, soft tissue tears, fl uid, blood or bone dust in the pneumatized spaces, variable section thickness, and wrinkling, which can occur during the mounting process. Specimen preparation was critical to acquiring the best possible imaging dataset.
Despite a careful technique, we encountered small amounts of fl uid and bone dust in the middle ear space. Air–bone interfaces also caused some minor beam-hardening artifact. The temporal bone specimen used in this atlas was harvested from a male cadaver using the block technique, as described by Gulya . The specimen had to be further reduced in size to a maximal diameter of 2.5 cm and to 4 cm in length in order to fi t properly in the microCT scanner.
The specimen was scanned without decalcifi cation or additional fi xation. The microCT scanning technique used to produce the images in this atlas has been previously described in the literature . MicroCT scanning differs from clinical scanning in that the object is rotated between the tube and camera, which are stationary. The specimen is placed on a rotating stage, which turns 360° about its vertical axis in 0.5° increments. At each angle an X-ray exposure is recorded on the charge-coupled device (CCD) camera.
A single acquisition using a 20-mm slice thickness will cover approximately 2 cm of tissue. Our specimen was scanned in two acquisitions, each taking 8 h at 35 KV and 50 mAs. The two volume acquisitions were then combined using the Analyze 3-D voxel registration program, as described by Hanson et al. and Camp et al. [3, 4]. The fi rst image volume was padded with zerovalued voxels to enlarge the volume enough to contain the entire reassembled volume. The region of overlap between the two acquisitions was used to automatically register the second acquisition to the fi rst. Then, an appropriate “crossover” section was determined, and the whole volume assembled. The entire volume dataset consisted of 20 mm cubic voxels. "