Three-dimensional magnetic resonance imaging in skull base lesions

Three-dimensional magnetic resonance imaging in skull base lesions

Am I Otolarvngol 12:139-145.1991 Three-Dimensional Magnetic Resonance Imaging in Skull Base Lesions G. GREVERS, MD, J. ASSAL, MD, T. VOGL, MD, AND C...

7MB Sizes 0 Downloads 128 Views

Am I Otolarvngol 12:139-145.1991

Three-Dimensional Magnetic Resonance Imaging in Skull Base Lesions G. GREVERS, MD, J. ASSAL, MD, T. VOGL, MD, AND C. WILIMZIG, MD

Three-dimensional (3D) magnetic resonance (MR) imaging is a new digital technique developed 2 years ago by a multidisciplinary group of head and neck surgeons, clinical radiologists, and mathematicians at the University of Munich. In this study, the clinical value of this method, which has been improved significantly during the last 9 months, is evaluated in lesions of the skull base. Our results indicate that 3D reconstruction based on two-dimensional (2D) MR images reveals topographic details of interesting structures. In addition, this method offers new possibilities for the preoperative planning of tumor resection, particularly in lesions close to the skull base. However, this imaging technique will have to be improved before it achieves widespread clinical use. AM J OTOLARYNGOL 12:139-145. Copyright 0 1991 by W.B. Saunders Company Key words: skull base surgery, three-dimensional imaging, magnetic resonance imaging.

proved. We are now able to generate 3D images of acceptable quality. Recent reports by our group have already stated that the skull base would be the main area of interest for 3D imaging of the head and neck.” This advanced technique offers an additional imaging possibility for preoperative and radiotherapy planning. In this paper, the clinical value of our improved 3D MR imaging method is assessed in volunteers and in patients with different tumors near the skull base.

The rapid development of computed tomography (CT] and magnetic resonance (MR) imaging in recent years has significantly expanded the possibilities for diagnostic imaging of head and neck lesions. Further progress is expected from threedimensional (3D) imaging in this area. However, in the head and neck, the clinical value of 3D imaging is controversial, mainly due to the fact that 3D reconstruction techniques demonstrated to date are of poor clinical value. All of these former methods were based on CT. In 1988, a multidisciplinary group of head and neck surgeons, radiologists, and mathematicians was convened to develop a 3D technique based on two-dimensional (ZD) MR images. This method was called “ray tracing mode.“* The topographic relationship of an anatomic structure, its surrounding tissue, and the surface of the head can be demonstrated by this technique. During the first year, the quality of the 3D images was inadequate for clinical use because the resolution of the different tissues was not satisfactory. Recently, however, the technique has been im-

MATERIALS

AND METHODS

Four healthy volunteers and 16 patients with head and neck lesions were examined with 3D MR imaging. Magnetic resonance images of the volunteers were obtained to optimize the parameters of the 3D sequences

and to investigate the 3D reconstruction method generated by a separate display, using a special imaging unit. After primary standard 2D MR imaging, all patients underwent a 3D MR examination in a 1.5 T Magnetom (Siemens, Erlangen, Germany) using a 25-cm diameter headcoil. The patients were examined before and after application of the paramagnetic contrast medium Gadolinium (Gd)-DTPA. In 80% of the tumors, there was a

Received February 1, 1991, from the Departments of Otorhinolaryngology, Head and Neck Surgery, and Radiology, University of Munich, Munich, Germany. Accepted for publication April 6. 1991. Address correspondence and reprint requests to G. Grevers, MD, Department of Otorhinolaryngology. Head and Neck Surgery, University of Munich, Marchioninistr. 15. D-~CICICI Miinchen 70, Germany. Copyright 0 1991 by W.B. Saunders Company 0196-0709/91/1203-0011$5.00/O

139

significant increase in signal intensity. Thus, many tumors contrasted with the surrounding tissue. A further improvement was achieved by a new fast turbo-flash 3D sequence, which shortens the time of measurement, thus reducing motion artefacts. The following parameters proved to be optimal: TR, 10 milliseconds; TE, 4 milliseconds: flip, 15’; 128 slices; and a resultant slice thickness of 1.4 mm. The measuring time for the 3D sequence was 11:53 minutes per patient; the total examination time was approximately 1 hour.

3D MR IMAGING IN SKULL BASE LESIONS

140

The region of interest is then determined to define the size of the window to be cut into the head model. An additional scout view of the frontal orientation in a transverse slice is helpful in determining the depth and angle of intersection. The location and size of the window must be evaluated individually for each case, depending on the kind of lesion.

4 p

‘. .

RESULTS

GV “B

_

__

-

YE

Figure 1. Ray tracing method. Contiguous MR slices define a 3D lattice of data that is detected by a succession of rays descending through the data cube from a point of view outside the original data set. The grey values of one of the rays that traces each of the slices are demonstrated in the curve beyond the image. GV, grey value; R, ray; VB, volume beginning; VE, volume end; P, point of view. (Reprinted with permission.‘]

Three-Dimensional Reconstruction Mode. For 3D reconstruction, 128 contiguous MR slices define a 3D lattice of data detected by a succession of rays descending through the data cube from a point of view outside the original data set. This ray tracing mode (Fig 1) calculates the surface of the patient’s head.

Figure 2. A patient with a squamous cell carcinoma of the right maxillary sinus. A 3D surface reconstruction (A) based on a rapid turboflash sequence measurement is compared with a preoperative picture of the patient (B). The tumor (t) is visualized in a sagittal 3D orientation (C). An additional scout view (s) shows the sagittal orientation in a transverse slice. (D) Coronal reconstructed 3D image windowed in the region of interest. The delineation of the tumor (t) is visible as well as its infiltration and extension. The corresponding 2D image (E) Shows a coronal Tl-weighted spin echo image (TR = 700, TE = 22). The tissue contrast between tumor (t) and orbital fat (0) is superior compared with panel D.

Three-dimensional MR imaging enables us to visualize a lesion through an individual window in relation to the surrounding tissue and, simultaneously, in topographic relation to the surface of the head. Additionally, the structure of interest can be observed from any desired direction. A patient with a squamous cell carcinoma of the right maxillary sinus was examined using turboflash sequence imaging (Fig ZA). The result of the 3D surface reconstruction obtained was compared with a preoperative picture of the patient (Fig ZB). Progress in the quality of computed surface reconstruction becomes apparent by observing the almost identical physiognomy. Figure 2C shows the tumor and the surrounding structures by the window technique. In Figs 2D and E, a frontal Tl-weighted 2D slice is compared with the corresponding frontal 3D image. The 3D pictures in Figs 2C and D are sufficient to visualize the tumor as well as its infiltration, extension,

GRJNERS ET AL

Figure 2.

(Con’t)

and relationship to the surrounding tissue. One can appreciate the destruction of the anterior and posterior walls of the right maxillary sinus as well as the bony defect of the inferior orbital rim caused by tumor infiltration. An aggressive fibroma located at the left temporal bone is demonstrated in Fig 3A. By using the window technique, the internal carotid artery and the vertebral artery can be seen (Fig 3B). One can notice a nonhomogeneous, enhancing, infiltrating tumor mass in the retromaxillary region and close to the carotid artery. Another patient with an ossifying fibroma underwent MR examination with Tl-weighted and 3D sequence measurement. The lesion and the in-

flammatory reaction of the meninges are visible in the 3D images (Fig 4). Typically, a central bandlike region of signal hypointensity can be observed. A profound contrast enhancement due to the concomitant meningeal reaction can be detected at the cranial border of the tumor [Fig 4A). In the infratemporal region, the tumor spreads toward the lateral pterygoid muscle and the parotid gland. The superficial temporal region is involved, displaying a significant enhancement. The tumor extension toward the inferolateral border of the orbital apex can be noticed in another section, located 15 mm anterior (Fig 4B). The coronal 3D images (Fig 4) provide an excellent delineation of the tumor extension.

142

3D MR IMAGING

Figure

Figure

2.

2.

(Con’t)

(Con’t)

IN SKULL

BASE LESIONS

GREVERSETAL

Figure 2.

(Con?)

DISCUSSION

The diagnostic value of 3D reconstruction is controversial. Many clinicians and radiologists still believe that the third dimension does not supply sufficient additional information compared with 2D MR or CT. On the other hand, reconstruction methods, mainly based on CT, have been established for routine clinical use in other medical fields, such as orthopedic surgery (hipjoints), neurosurgery (brain tumors], and craniofacial surgery (craniofacial dysmorphologies).“‘” The 3D technique based on CT is mentioned in the otorhinolaryngologic literature in relation to the preoperative planning of craniofacial tumor resection7 The new 3D MR reconstruction technique developed by our group 2 years agol*’ has yet to provide satisfactory results in the evaluation of

head and neck lesions. The clinical value of this technique, however, can be increased by improving the quality and resolution of the 3D images. Our results using the improved 3D MR in various lesions of the skull base have provided new aspects of diagnostic value in this delicate area. We performed 3D examinations before and after the application of the paramagnetic contrast medium Gd-DTPA, achieving a better evaluation of lesion and vascularization. The duration of the complete 3D MR imaging examination was approximately 1 hour. Our findings in different lesions of the head and neck indicate that 3D MR imaging is useful mainly in lesions close to the skull base. In these areas, the technique has significant advantages concerning topography and the evaluation of surrounding structures. Because of the high operative risk as well as the sophisticated anatomy in this area, all diagnostic modalities should be used prior to skull base surgery. In

Figure 3. (A) A patient with an aggressive fibroma (f) located underneath the temporal bone (coronal 3D I the Turn& (f) and-its relationship to the carotid artery (arrows) and the skull base. Arrow heads: vertebral artery.

GREVERS

ET AL

Figure 4. Coronal 3D images demonstrate an ossifying fibroma after administration of Gd-DTPA in window technique. Both the band-like tumor (t) and the concomitant inflammatory reaction (i) of the meninges are visible in the 3D image (A). (B) Another section of the tumor (t) located 15 mm anterior. The tumor (t) extends toward the orbital apex (0).

other regions of the head and neck, 3D MR did not show advantages when compared with normal 2D MR. The diagnostic value of 3D MR is limited even in the skull base. To achieve a technique that is ready for routine clinical use, the quality of our 2D data must be improved. By further development of hardware and software, the loss of information caused by transferring the 128 2D slices into a special 3D display can be reduced. A multidisciplinary approach involving clinicians and radiologists is necessary to achieve optimal results. References 1. Grevers G. Wilimzig C, Vogl T. et al: Eine neue Methode zur 3D-Rekonstruktion im Kopf-Hals-Bereich. Laryngo-RhinoOtologie 1990; 69:187-190

2. Grevers G, Vogl T, Wilimzig C: 3D-Reconstructions of head and neck lesions, in Proceedings of the XIV World Congress of Otorhinolaryngology, Head and Neck Surgery, Madrid, Spain, September 10 through 15,1989. Amsterdam, The Netherlands, Kugler & Ghedini Publications (in press) 3. Herman GT: Three-dimensional scanner. J Comput Assist Tomogr

imaging on a CT or MR 1988; 12:450-458

4. Marsh JL, Vannier MW: Three-dimensional ing from CT scans for the study of craniofacial J Craniofac Genet Dev Biol 1989; 9:61-75

surface imagdysmorphology.

5. Schad LR, Boesecke R. Schlegel W, et al: Three dimensional image correlation of CT, MR. and PET studies in radiotherapy treatment planning of brain tumors. J Comput Assist Tomogr 1987; 11:948-954 6. Takahashi H. Takagi dimensional reconstruction window and its membrane. 101,5:517-521

A, Sando I: Computer-aided threeand measurement of the round Otolaryngol Head Neck Surg 1989:

7. Zinreich SJ, Mattox DE, Johns ME, et al: 3-D CT for cranial facial and laryngeal surgery. Laryngoscope 1988; 98:1212-1219