Keywords
skull base - computed tomography - dual energy
Introduction
Due to its complex anatomy and close relationship with osseous structures, skull base
lesions are often evaluated with both computed tomography (CT) and magnetic resonance
imaging (MRI) for diagnostic and preoperative planning purposes. CT can delineate
the osseous anatomy with increased precision in relation to MRI. Additionally, CT
can provide important tips to diagnosis by identifying different patterns of bone
involvement and periosteal reaction or by simply detecting the presence of ossification
or calcification within a lesion.
Review of Literature and Discussion
Slow-growing lesions that do not infiltrate the bone tend to demonstrate smooth cortical
expansion and bone remodeling with preservation the bony cortex, whereas aggressive
tumors or infection typically infiltrate the bone, destroying the adjacent bony cortex
in a permeative pattern.[1] This concept is often used to differentiate between schwannomas and paragangliomas
in the skull base ([Fig. 1]). Well-recognized exceptions to the rule are invasive pituitary adenomas and meningiomas.
Bone invasion in pituitary adenomas does not typically represent a malignant feature
and can be seen in ∼35% of the cases of benign adenomas.[2] Clival invasion is more commonly seen in women, large tumors, and null-cell-type
tumors. These infiltrative adenomas, however, have increased recurrence rates and
complications and therefore their preoperative identification is important for surgical
planning and prognostic purposes.[3]
Fig. 1 Demonstration of two distinctive patterns of bone involvement. Note the permeative
and destructive pattern of bone involvement typically seen in aggressive lesions such
as paragangliomas (A) in relation to a smoothly marginated expansile lesion such as
a schwannoma (B) in the right jugular foramen in these two different patients.
A subset of grade I meningiomas can also invade the bone.[4] These cases tend to be clinically challenging as complete tumor resection becomes
more difficult and therefore the risk of recurrence increases.[5] It has been suggested that the molecular regulators of bone tropism in meningiomas
may depend on their anatomical location as meningiomas of the anterior skull base
show a distinct protein expression pattern compared with spheno-orbital meningiomas.[5] The most typical bony finding in meningiomas, however, is bone hyperostosis ([Fig. 2]). The cause of hyperostosis in meningiomas is controversial. One theory favors focal
vascular disturbances generated by the tumor,[6]
[7]
[8] another suggests osteoblastic stimulation by tumor secreting factors,[9]
[10] and another proposes bone production by the tumor itself.[6]
[8] It is important to recognize, however, that a significant number of patients with
radiologic hyperostosis have been demonstrated to have tumor invasion of the bone.[11]
[12]
Fig. 2 Hyperostosis (white arrowheads) in a cavernous sinus meningioma (black arrowheads).
Hemangiopericytoma is an important diagnostic consideration when bone invasion is
identified in a dural-based mass. These tumors are typically multilobulated, extra-axial
masses, with associated bone erosion. Unlike with meningiomas, however, hyperostosis
and intratumoral calcification are not typically present.[13] Metastasis can also have a similar appearance to hemangiopericytomas and should
also be considered in the appropriate clinical setting.
Another bony change related to meningiomas is the presence of pneumosinus dilatans,
which consists of abnormal expansion of one or more paranasal sinus. This can be another
helpful sign to indicate the presence of a meningioma in the anterior skull base.[14]
CT is also very helpful in identifying patterns of calcification or ossification to
assist in the diagnosis. It is known that ∼90% of craniopharyngiomas calcify, making
this an important diagnostic feature.[15] Additionally, the distinct pattern of chondroid calcification (arc or ringlike calcifications)
in chondrosarcomas can also be useful in pointing to this entity during diagnostic
workup. Chordoma is one of the main differential considerations once chondrosarcomas
are suspected, and it is important to differentiate between arc-whorl intralesional
calcifications seen in chondrosarcomas with fragmented destroyed bone more often seen
in chordomas.[16]
Another use of CT in the skull base is in the identification of bony defects in the
evaluation of cerebral spinal fluid (CSF) leaks. Recent multislice CT scanners can
acquire images with slice thickness as thin as 0.5 to 0.6 mm and can perform multiplanar
reconstructions, providing greater ability to evaluate submillimetric defects. Interactive
multiplanar evaluation (axial, coronal, sagittal, and oblique planes) is important
to identify and correctly describe osseous defects in the evaluation of CSF leaks
([Fig. 3]).[17] The identification of bony defects in these cases is highly sensitive but not definitive
for CSF leak. Stone et al observed that all patients in their 42-patient cohort with
confirmed CSF leak demonstrated bony defects on high-resolution CT. Ten patients with
bony defects demonstrated on CT, however, were not confirmed to have CSF leak.[18]
Fig. 3 Bony defect along the left cribriform plate (arrowhead) in a patient with suspected
cerebrospinal fluid leak.
In addition to providing important characterization of the bony structures, CT can
also provide invaluable information about the relationship of a lesion with the adjacent
vascular structures through computed tomography angiography (CTA).
One major challenge related to vascular imaging in the skull base has been the evaluation
of the cavernous internal carotid artery. This is particular true because the high-density
contrast material within the vessels becomes less conspicuous when surrounded by bone.
Several computerized bone subtraction algorithms have been proposed in the past in
an attempt to overcome this issue. One technique utilizes two imaging acquisitions
(pre- and postcontrast) to subtract the background bone. The major disadvantage of
this technique is the patient's increased radiation exposure.[19] In addition, motion between the two acquisitions can also impact the quality of
the bone subtraction in this technique. Another bone subtraction method uses computerized
imaging processing techniques to differentiate the vessel anatomy from the adjacent
bone by segmenting out only the structure containing the contrast material. This technique
relies on the variation between the densities of different structures as well as few
anatomical landmarks to distinguish between bone and vessel. Unfortunately, overlap
exists between the density of these structures, and the bone subtraction obtained
with this methodology is not consistently reliable. Recent dual-energy technology
can differentiate between contrast material and bone with high precision. This is
possible because the density of the calcium and iodine varies, which causes them to
behave differently depending on the energy applied to X-ray beams (different peak
kilovoltages). Computer algorithms are then applied to the acquired images, allowing
decomposition of few elements and subtraction of the calcium ([Fig. 4]).
Fig. 4 Bone subtraction using dual-energy technique with clear separation between iodine
and calcium.
The evaluation of the petrous, cavernous, and supraclinoid internal carotid artery
is very important in the preoperative planning for tumors that invade the cavernous
sinus. Cavernous sinus meningiomas, for instance, can cause significant narrowing
of the cavernous internal carotid artery. Although this is not easily recognized with
standard multiplanar reconstructions, techniques such as dual-energy bone subtraction
or some postprocessing tools such as curved reformats can be of great assistance.
Even with postprocessing techniques, however, it may be difficult to appreciate smaller
vessels such as posterior communicating arteries that may have been compressed by
the tumor. In these situations, it is often helpful to use thin-section T2-weighted
images to troubleshoot.
Another use of dual-energy technology in the evaluation of skull base pathology relates
to its ability to decrease artifacts that are known to negatively impact image quality,
particularly in the posterior fossa. The petrosal ridge of the temporal bone is the
hardest bone in the human body and is responsible for significant artifact in the
cerebellopontine region. Based on the two polychromatic X-ray beams available in dual-energy
CT, sophisticated reconstruction algorithms can be applied to estimate what a scan
performed with a single monochromatic X-ray beam might have shown. With such approach,
beam-hardening and streak artifacts can be significantly reduced, although often at
the expense of lower signal-to-noise ratio.[20] The same technique can be applied to reduce artifact from external sources or metallic
hardware in the craniocervical junction ([Fig. 5]).[21]
[22]
Fig. 5 Dual-energy acquisition with two different monoenergetic selections (A: 50 keV; B:
100 keV). The 100-keV monoenergetic imaging shows decreased streak artifact from the
suboccipital metallic hardware.
Dual-energy CT can also be used to generate a virtual noncontrasted CT from a contrast
enhanced study by subtracting the iodine material from the image using its material
decomposition capabilities. Yet, there is strong evidence that dual-energy CT acquired
via dual-source technology does not result in increased radiation to patients.[23] In fact, the radiation from dual-energy CTs measured by volume computed tomography
dose index (CTDIvol) have been found to be 12% lower than single-energy CTs.[24]
Additionally, having the ability to generate two scans (contrast-enhanced and virtual
noncontrast) from a single postcontrast acquisition may have further radiation exposure
savings when these two scans are clinically needed. By decomposing the iodine component
from the image, dual-energy CT can also provide maps on which iodine distribution
is color-coded and superimposed on the virtual noncontrast CT, which is thought to
increase visual detection of lesions in the head and neck.[25]
[26]
