Three-dimensional-fast Imaging Employing Steady-State Acquisition and T2-weighted Fast Spin-Echo Magnetic Resonance Sequences on Visualization of Cranial Nerves III-XII
Cheng Yu-Shu,Zhou Zheng-Rong,Peng Wei-Jun,Tang Feng
DOI: https://doi.org/10.1097/00029330-200802010-00020
2008-01-01
Abstract:Because of the small diameter and complex anatomic course of the cranial nerves except for the optic nerve, trigeminal nerve, facial nerve, and cochlear and vestibular nerve, other cranial nerves are difficult to be visualized in magnetic resonance imaging (MRI) scanning with conventional thickness (5-10 mm). With the rapid development of MRI technology high spatial resolution, three-dimensional and two-dimensional magnetic resonance imaging technologies have been used in recent years in the observation of normal and abnormal cranial nerves, including three-dimensional constructive interference in steady state (3D CISS) sequence, threedimensional magnetization-prepared rapid gradient-echo (3D MP-RAGE) sequence, three-dimensional fast inflow with steady-state precession (3D FISP) sequence, and some fast spin-echo (FSE) sequences.1-6 Threedimensional fast imaging employing steady-state acquisition (3D-FIESTA) is a fast three-dimensional steady-state imaging sequence with high spatial resolution and contrast between the organizational structures. And this sequence was reportedly used in the study on the diseases in the cerebellopontine angle; inner ear and posterior fossa tumors.7,8 However, the reports about the value of 3D-FIESTA sequence for the visualization of normal cranial nerves are still rare. This study aimed to evaluate the sensitivity of 3D-FIESTA MR sequence and 2D FSE T2-weighted imaging (T2WI) for the visualization of cranial nerves (CN) III-XII in their cisternal courses. METHODS Volunteers Twenty-five healthy volunteers (11 men and 14 women) were included in this study. Their ages ranged from 22 to 46 years (mean age 29.1 years). MRI All of the 25 volunteers underwent MRI using a dualgradient 1.5 T MRI system (Signal Infinity Twinspeed, GE Medical System, USA) with 3D FIESTA and 2D FSE sequences. 3D-FIESTA sequence TR/TE = 5.7/1.4 ms; flip angle = 65° field of view (FOV) = 180 mm × 180 mm; slab thickness = 66-76 mm; matrix =512 × 512; number of slabs = 1; effective slice thickness = 1 mm; number of acquisitions =2-3; scan time = 3-4.5 minutes. 2D FSE T2WI sequence TR/TE = 4000 ms/78-83 ms; flip angle =180°; field of view = 180 mm × 180 mm; echo train length (ETL) =13-14; slice thickness = 2 mm; inter-slice spacing =0.2 mm; number of acquisitions =2, scan time =6-7 minutes. Image analysis Analysis of the data set obtained with the 3D-FIESTA sequence was based on the original axial 3D images with an effective slice thickness of 1 mm. The visualizations of cranial nerves III-XII in their cisternal courses were described as “identified with certainty”, “probably identified”, or “not identified”. Two radiologists evaluated independently the visualization rates of the cranial nerves III-XII in their cisternal courses with the two MRI sequences of 3D-FIESTA and 2D FSE T2WI. RESULTS Oculomotor nerve (CN III) The rates of visualization of CN III using 3D FIESTA and 2D FSE were 100% and 68%, respectively. CN III was identified with certainty in 80% by 3D FIESTA sequence (Figure 1A) and 22% by 2D FSE sequence (Figure 1B), "probably identified” in 20% by 3D FIESTA sequence and 46% by 2D FSE sequence, and “not identified” in 32% by 2D FSE sequence.Figure 1.: Bilateral oculomotor nerves identified with 3D-FIESTA sequence and 2D-FSE sequence. (A) Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Bilateral oculomotor nerves were identified with certainty (white arrow). (B) Axial image of 2D FSE sequence with a slice thickness of 2 mm. Bilateral oculomotor nerves were also identified with certainty (white arrow). The same volunteer as Figure 1A.Figure 2. Bilateral trochlear nerves identified with 3D-FIESTA sequence and 2D-FSE sequence. (A) Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Bilateral trochlear nerves were identified with certainty (white arrow). (B) Axial image of 2D FSE sequence with a slice thickness of 2 mm. Bilateral trochlear nerves were not identified. The same volunteer as Figure 2A.Figure 3. Left trigeminal nerve identified with 3D-FIESTA sequence and 2D-FSE sequence. (A) Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Left side of V1 (thick white arrow), V2 (thin white arrow) and V3 (black arrow) were identified with certainty. (B) Axial image of 2D FSE sequence with a slice thickness of 2 mm. Left side of V1 (thick white arrow), V2 (black arrow) were probably identified and V3 (thin white arrow) was identified with certainty. The same volunteer as Figure 3A.Figure 4. Bilateral abducens nerves identified with 3D-FIESTA sequence and 2D-FSE sequence. (A) Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Bilateral abducens nerves were identified with certainty (white arrow). (B) Axial image of 2D-FSE sequence with a slice thickness of 2 mm. Bilateral abducens nerves were probably identified (white arrow). The same volunteer as Figure 4A.Trochlear nerve (CN IV) We succeeded in identifying CN IV in 38%, including the part “probably identified” in 22%, “identified with certainty” in 16% (Figure 2A), and “not identified” in 62% by 3D FIESTA sequence, whereas none of CN IV was identified by 2D FSE sequence (Figure 2B). Trigeminal nerve (CN V) CN V was identified with certainty in 100% both by the 3D FIESTA and 2D FSE. By using 3D FIESTA we identified CN V1 and CN V2 in 100%, CN V3 in 82% (Figure 3A). By the 2D FSE we identified CN V1 in 76%, CN V2 in 62%, and CN V3 in 50% (Figure 3B). Abducens nerve (CN VI) Using the 3D FIESTA we succeeded in identifying CN VI in 100% (Figure 4A), including the part “probably identified” in 18%, “identified with certainty” in 82%. Using the 2D FSE we identified CN VI in 52% (Figure 4B), including the part “probably identified” in 42%, "identified with certainty” in 10%, “not identified” in 48%. Facial (CN VII) and vestibulocochlear nerves (CN VIII) We identified CN VII and CN VIII with certainty in 100% (Figure 5) both by the 3D FIESTA and 2D FSE sequences.Figure 5.: Bilateral facial and vestibulocochlear nerves identified with 3D-FIESTA sequence and 2D-FSE sequence. Axial images of 3D-FIESTA sequence (A) with an effective slice thickness of 1 mm and 2D-FSE sequence (B) with a slice thickness of 2 mm. Bilateral facial and vestibulocochlear nerves were identified with certainty (white arrow) using the two sequences.Figure 6. Bilateral complex of the glossopharyngeal, vagus, and accessorius nerves identified with 3D-FIESTA sequence and 2D-FSE sequence. (A). Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Bilateral complex of the glossopharyngeal, vagus, and accessorius nerves were identified with certainty (white arrow). (B) Axial image of 2D-FSE sequence with a slice thickness of 2 mm. Right side of the complex of the glossopharyngeal, vagus, and accessorius nerves was identified (white arrow). The same volunteer as Figure 6A.Figure 7. Right hypoglossal nerve identified with 3D-FIESTA sequence and 2D-FSE sequence. (A) Axial image of 3D-FIESTA sequence with an effective slice thickness of 1 mm. Right side of hypoglossal nerve was identified with certainty (white arrow). (B) Axial image of 2D-FSE sequence with a slice thickness of 2 mm. Bilateral hypoglossal nerves were not identified. The same volunteer as Figure 7A.Complex of the glossopharyngeal, vagus, and accessorius nerves (CN IX-XI) Using the 3D FIESTA, we succeeded in identifying this lower nerve complex in 100%, including the part "probably identified” in 12%, “identified with certainty" in 88% (Figure 6A). Using the 2D FSE, we identified the lower nerve complex in 84%, including the part "probably identified” in 46%, “identified with certainty" in 38% (19/50) (Figure 6B), “not identified” in 16%. Hypoglossal nerve (CN XII) Using the 3D FIESTA sequence, we succeeded in identifying this lower nerve in 66% including the part "probably identified” in 52%, “identified with certainty" in 14% (Figure 7A), and “not identified” in 34%. However, none of CN XII was identified using the 2D FSE sequence (Figure 7B). DISCUSSION Most cranial nerves are slim with complex anatomic courses and difficult to be identified using routine MR sequences with 5-10 mm thickness except for CN II, V, VII and VIII. In recent years, some 3D MR sequences have been used in the study of cranial nerves, but the report about the use of 3D FIESTA sequence and T2-weighted FSE sequence with 2-mm scanning thickness to visualize the normal nerves of CN III-XII in their cisternal courses has not been found. In this study we analyzed the value of these two sequences on the visualization of the normal nerves of CN III-XII in their cisternal courses and found that most of the nerves of CN III-XII in their cisternal courses could be reliably assessed when using two sequences. Our results suggest that the 3D-FIESTA sequence was more superior to increasing the rates of visualization of the nerves III, IV, V1, V2, V3, VI, IX-XI and XII than the T2-weighted FSE sequence. However, the identification of the cranial nerves was related to many factors including spatial resolution of image, diameter and course of the cranial nerve and the signal contrast between the nerve and surrounding structures. High spatial resolution of MRI is the precondition for the visualization of the cranial nerves. 3D FIESTA MR sequence is a three-dimensional fast steady-state sequence, by which the image acquired has high spatial resolution and good contrast of different structures and may provide more subtle anatomic details. Cisternal portion of the cranial nerves can be well visualized when using the 3D sequence because of the natural contrast between the nerves and surrounding cerebrospinal fluid (CSF). In this study most of the nerves of CN III-XII in their cisternal courses could be reliably assessed when using the 3D-FIESTA sequence. And the 3D-FIESTA sequence is more superior to increasing the rates of visualization of the nerves III, IV, V1, V2, V3, VI, IX-XI and XII than that of the T2-weighted FSE sequence. The results may be related to the good contrast between the nerves and surrounding CSF in the 3D FIESTA sequence and different effective thickness of 1 mm in the 3D-FIESTA sequence and 2 mm in the T2-weighted FSE sequence. So we think that the spatial resolution is a very important factor for identification of the cranial nerves. Another effective way to improve spatial resolution of images is reducing the field of view (FOV). In this study the FOV was comparatively large to 18 cm × 18 cm. In our following study we reduced the FOV to 16 cm × 16 cm and 14 cm × 14 cm, with which the spatial resolution of images was improved in some degrees. So we may try the best to reduce the FOV in order to improve the spatial resolution of the images to find the slim cranial nerves. Although the 3D-FIESTA sequence is superior to displaying the cranial nerves in their cisternal courses, the signal-to-noise ratio (SNR) of the images using 3D-FIESTA is rather lower than that using routine sequences with 5 mm or 10 mm slice thickness. Then, in the diagnosis of common intracranial diseases such as cerebral infarction and demyelination, we should use the 3D-FIESTA combined with routine spin-echo (SE) or FSE sequences instead of that using 3D-FIESTA sequences only. Yet, in case of cranial nerve impaired or other cranial nerve disorders including inflammation, primary neural tumors such as schwannomas and neurofibromas, vessel compression or contact on cranial nerves, and perineural spread of head and neck malignancies, 3D-FIESTA is necessary and important to make a correct diagnosis as a complementary sequence. In addition, another advantage of the 3D-FIESTA sequence lies in the shorter acquisition time than that of routine SE or FSE sequences. Thus, the use of the 3D-FIESTA sequence whether combined with other routine sequences or not may add less scan time and help to keep the patients immobility easily. Multi plane reconstruction of 3D imaging is helpful to increase the scope of visualization of cranial nerves and makes the lesions of cranial nerves more intuitionistic to benefit the treatment. Most cranial nerves could be identified in the cisternal segments by 3D-FIESTA and 2D FSE sequences owing to the natural contrast with the cerebrospinal fluid. As the extracranial cranial nerves lose the natural contrast of CSF and the branches of nerves are smaller with their courses more complex, the extracranial normal nerves are difficult to be identified except for the optic nerve with good contrast of orbital fat. At the same conditions of MR scanning the differences in the identification rates of various cranial nerves are related to the anatomic characteristics, especially to the nerve diameter. Trigeminal nerve is the largest one among the cranial nerves III-XII and the trunks of the trigeminal nerve can be identified using 3D-FIESTA and 2D FSE sequences in this study. Because the trochlear nerve is the smallest of all the cranial nerves with a diameter of 0.4 mm (range 0.2-0.6 mm) only,9 the identification rates of the trochlear nerve were the lowest in the cranial nerves III-XII, with a rate of 38% by 3D-FIESTA and 0 by T2-weighted 2D FSE sequences respectively. Although the abducens nerve is relatively small with an average diameter of 2.2 mm (range 1.0-2.9 mm),10 the identification rate of CN VI was 100% by 3D-FIESTA in this study. However, using T2-weighted 2D FSE sequence, the identification rate of CN VI was only 52% for a possible reason of the relatively low spatial resolution of the FSE sequence. The diameter of hypoglossal nerve trunk was only 1.3-1.6 mm,11 and the identification rates of CN XII were relatively low: 66% by 3D FIESTA sequence and 0 by T2-weighted 2D FSE sequence. In this study the identification rates of CN VII and VIII were all 100% by the two different sequences because of the two reasons of large diameters (CN VII: 1.8 mm, ranged 1-2 mm, CN VIII: 3 mm, ranged 2-5 mm9) and the transverse courses paralleled to the scanning planes. For the similar reason the oculomotor nerve and the lower nerve complex (CN IX-XI) are easily identified by the two sequences. In the peripheral nerve system, inflammation, primary or secondary tumors and vascular oppression lesions might lead to damages to the functions of cranial nerves. On radiology early detection and accurate evaluation of cranial nerve lesions may be helpful for the clinical diagnosis, treatment and prognosis assessment.12-14 For the cranial nerves are slim with complex anatomic course, appropriate imaging modalities are essential to the detection of cranial nerve lesions. The images obtained with 3D-FIESTA sequence highlight the contrast of cranial nerves and CSF and are useful to display the subtle anatomic details. Another advantage of 3D FIESTA sequence is a rather shorter time of the scanning. As a complement to conventional scanning sequences, 3D-FIESTA sequence is helpful to increase the visualization of normal and abnormal cranial nerves. In conclusion, using the 3D-FIESTA sequence and 2D T2-weighted FSE sequences with 2-mm slice thickness most normal cranial nerves in their cisternal courses could be identified. Furthermore, it is better to identify the slim cranial nerves using 3D FIESTA sequence rather than that by 2D FSE sequences with 2 mm slice thickness.