- Email: [email protected]
CT–MR image data fusion for computer-assisted navigated surgery of orbital tumors
European Journal of Radiology 73 (2010) 224–229
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
CT–MR image data fusion for computer-assisted navigated surgery of orbital tumors Stefan Franz Nemec a,∗ , Philipp Peloschek a , Maria Theresa Schmook a , Christian Robert Krestan a , Wolfgang Hauff c , Christian Matula b , Christian Czerny a a b c
Department of Radiology/Division of Neuroradiology and Musculoskeletal Radiology, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria Department of Neurosurgery, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria Department of Ophthalmology, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria
a r t i c l e
i n f o
Article history: Received 27 October 2008 Received in revised form 2 November 2008 Accepted 5 November 2008 Keywords: Computed tomography Magnetic resonance imaging Image fusion Neuronavigation Orbital tumors
a b s t r a c t Purpose: To demonstrate the value of multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) in the preoperative assessment of orbital tumors, and to present, particularly, CT and MR image data fusion for surgical planning and performance in computer-assisted navigated surgery of orbital tumors. Materials and methods: In this retrospective case series, 10 patients with orbital tumors and associated complaints underwent MDCT and MRI of the orbit. MDCT was performed at high resolution, with a bone window level setting in the axial plane. MRI was performed with an axial 3D T1-weighted (w) gradient-echo (GE) contrast-enhanced sequence, in addition to a standard MRI protocol. First, MDCT and MR images were used to diagnose tumorous lesions compared to histology as a standard of reference. Then, the image data sets from CT and 3D T1-w GE sequences were merged on a workstation to create CT–MR fusion images that were used for interventional planning and intraoperative image guidance. The intraoperative accuracy of the navigation unit was measured, defined as the deviation between the same landmark in the navigation image and the patient. Furthermore, the clinical preoperative status was compared to the patients’ postoperative outcome. Results: Radiological and histological diagnosis, which revealed 7 benign and 3 malignant tumors, were concordant in 7 of 10 cases (70%). The CT–MR fusion images supported the surgeon in the preoperative planning and improved the surgical performance. The mean intraoperative accuracy of the navigation unit was 1.35 mm. Postoperatively, orbital complaints showed complete regression in 6 cases, were ameliorated notably in 3 cases, and remained unchanged in 1 case. Conclusion: CT and MRI are essential for the preoperative assessment of orbital tumors. CT–MR image data fusion is an accurate tool for planning the correct surgical procedure, and can improve surgical results in computer-assisted navigated surgery of orbital tumors. © 2008 Published by Elsevier Ireland Ltd.
1. Introduction Tumors of the orbit comprise a number of benign and malignant pathologies that are classified according to the four orbital sections, the intraconal and extraconal compartments, the globe, and the optic nerve [1–4]. Table 1 presents a summary of tumors that can arise from the intra- and extraconal space [4]. The differential diagnosis of some lesions presents a challenge for the preoperative assessment of tumor patients and is extremely important in choosing the correct therapeutic procedure. Therefore, the use of precise radiological techniques plays a crucial role in obtaining an exact
∗ Corresponding author. Tel.: +43 1404007609; fax: +43 1404003777. E-mail address: [email protected] (S.F. Nemec). 0720-048X/$ – see front matter © 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.ejrad.2008.11.003
imaging diagnosis. Multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) are essential modalities with which to achieve a radiological diagnosis and to support surgical planning and performance. MRI provides excellent soft tissue contrast to delineate the anatomy and pathology of the soft tissue, whereas CT, with a high-resolution bone window level setting best evaluates osseus structures [1–5]. The treatment of orbital tumor disease that involves muscular and neurovascular structures requires precise route planning and safe image guidance to provide intraoperative orientation with the greatest accuracy and lowest iatrogenic manipulation. Navigated surgery addresses these concerns with the goal of minimizing surgical morbidity that can be linked to severe postoperative ocular defects or residual tumor. Navigated surgery connects the static visualization of head and neck imaging with the dynamic operative
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
225
Table 1 Intra- and extraconal tumors (according to MF Mafee, shortened version). Intraconal
Extraconal
Common Cavernous hemangioma Optic nerve meningioma Optic nerve glioma Lymphoma Lymphangioma Hemangiopericytoma Rhabdomyosarcoma Metastasis
Capillary hemangioma Cholesterol granulomas Dermoids Lacrimal gland tumors Lymphangioma Peripheral nerve tumors Plasmocytoma Rhabdomyosarcoma
Less common Capillary hemangioma Neurofibroma Schwannoma Lipoma
Fibrous histiocytoma Hemangiosarcoma Hemangiopericytoma Lipoma
procedure [6–14]. The performance of computer-assisted navigated techniques that are used in head and neck surgery have profited significantly from the improved and precise imaging data that have been made possible by the recent advances in computed technology. In this study, we demonstrate the ability of CT and MRI to diagnose orbital tumors, and we present the results of CT–MR image data fusion for computer-assisted surgery of orbital tumors with the help of a navigation system. The protocols that were used are described and the impact of CT, MRI, and CT–MR fusion images in computer-assisted navigated surgery of orbital tumors is discussed. 2. Materials and methods
Fig. 1. A 76-year-old woman with a B-cell lymphoma of the right orbit. Radiologically a pleomorphic adenoma was suspected. (a) The axial CT scan in high-resolution bone window level setting demonstrates an erosion of the lateral orbital wall (arrows). (b) The 3D T1-w sequence delineates a flat extraconal tumor (arrow). The CT scan and the MRI study represent the basis for image fusion.
2.1. Patients The protocol for this study was in accordance with the Declaration of Helsinki. In this retrospective case series, 10 patients (5 women and 5 men with a mean age of 49 of years; range, 25–76 years) were included, after chart review, if the following criteria were fulfilled: (a) clinically suspected benign or malignant orbital tumor with associated complaints; (b) preoperative assessment with standard institutional CT and MRI protocols for orbital tumors; (c) neuronavigated surgery with image guidance based on CT–MR fusion images. A definitive diagnosis was ensured by microscopic examination and corresponding immunhistochemical analysis in each case. Pre-and postoperatively, all patients underwent a profound neuro-opthalmological examination that included testing for visual acuity, visual fields, pupillary responses, ocular motility, palpation of the anterior orbit, and inspection of the external surface of the globe and eyelids. Furthermore, the intraocular pressure was measured and a slit lamp examination was performed. Dilated fun-
duscopy was applied to examine the optic disc and retina. Initially, all 10 patients presented with various complaints, such as unilateral proptosis of the globe, chemosis, diplopia, decreased visual acuity, and ocular pain or pressure sensation. 2.2. Materials MDCT was performed with a 16-row multidetector-unit (Mx 8000 IDT Philips, Eindhoven, the Netherlands) at high resolution with a bone window level setting in a strictly axial plane without inclination of the gantry (120 kV, 200 mAs; FOV ∼ 20 cm; slice thickness 0.8 mm; detector collimation of 4 mm × 0.75 mm). MRI (1.0 T, Gyroscan T10 NT Philips, Eindhoven, the Netherlands) was performed in the axial and coronal planes, with T2-weighted (w) fast-spin-echo sequences [matrix 512 × 512; echo time (TE) 100 ms; repetition time (TR) 3500 ms; slice thickness 4 mm], unenhanced T1-w spin-echo (SE) sequences (matrix 512 × 512; TE 20 ms; TR 550 ms; slice thickness 3 mm), and contrast-enhanced T1-w
Table 2 Patients and pathologies. No.
Age
Sex
Radiology
Location
Histology
Approach
Accuracy (mm)
1 2 3 4 5 6 7 8 9 10
57 47 51 51 25 76 39 47 64 33
M M M F M F F F F M
Lacrimal gland carcinoma Epidermoid Pleomorphic adenoma Cavernous hemangioma Dermoid cyst Pleomorphic adenoma Cavernous hemangioma Pleomorphic adenoma Lacrimal gland carcinoma Pleomorphic adenoma
Extraconal Extraconal Extraconal Intraconal Extraconal Extraconal Intraconal Extraconal Extraconal Extraconal
Adenoidcystic carcinoma Epidermoid Pleomorphic adenoma Cavernous hemangioma Dermoid cyst Primary B-cell lymphoma Cavernous hemangioma Solitary fibrous tumor Adenoidcystic carcinoma Pseudotumor
Supraorbital Lateral Supraorbital Lateral Lateral Supraorbital Lateral Supraorbital Supraorbital Lateral
1.45 1.25 1.5 1.2 1.45 1.55 1.4 1.1 1.3 1.3
226
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
Fig. 2. A 33-year-old man with a pseudotumor of the left orbit. Radiologically a pleomorphic adenoma was suspected. (a) The axial CT scan shows an intact lateral wall (arrows). (b) The 3D T1-w sequence delineates a well-defined extraconal lesion (arrow). (c) Coronal (upper left), sagittal (upper right) and axial (bottom left) navigation image on the workstation monitor. The navigational images are inverted compared with the CT and MRI scan. The navigation images helped the surgeon to approach the lesion (marked by crosshairs). The surgical intervention was documented by a video camera (bottom right).
SE sequences (intravenous application of gadopentetate dimeglumine, Magnevist® ), using a circular polarized head coil. In addition, T1-w SE sequences were obtained, with fat suppression after contrast material administration, in the coronal plane (4 mm slice thickness). MRI was accomplished with a 3D T1-w gradient-echo (GE) contrast-enhanced sequence with a slice thickness of 1 mm in a strictly axial plane without angulation (matrix 256 × 256; TE 6.9 ms; TR 25 ms; measured voxel size 0.90/1.08/2.00 mm; reconstructed voxel size 0.90/0.90/1.00 mm).
Then, image data sets from CT and MRI (3D T1-w GE sequences) were stored on CD ROMs and imported to a workstation (Stealth StationTM , Medtronic USA). A fusion software (Stealth MergeTM , Medtronic USA) was used to search automatically for optimal image-to-image registration between the CT and MRI data sets without manual correction. The CT–MR fusion images were used for preoperative planning and for intraoperative image guidance. For detailed information about the navigation unit, Stealth StationTM , see previous publications [15–17]. In summary, the workstation
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
227
is optical-based and consists of a mobile console with a highresolution monitor, a charge-coupled device (CCD) camera unit, a referencing apparatus with dynamic reference frame (DRF), as well as a Mayfield head-holder and light emitting diodes (LED) pointer that is operated by the surgeon. The camera measures the infrared LED signals and calculates position coordinates. The workstation displays the exact location as a cross-hair on the monitor. The image-to-patient registration to determine corresponding points between the real patient’s anatomy and the navigation images was performed on the basis of at least five to six anatomical surface landmarks to define the coronal, axial, and sagittal plane. Images in the axial, sagittal, and coronal planes, as well as 3D visualizations, were used for image guidance. 2.3. Evaluation Initially, CT and MR images were analysed by a head and neck radiologist (C. C., 20 years experience) to characterize lesions and to differentiate benign and malignant tumors. The radiological diagnosis was compared with the histological standard of reference. Intraoperatively, the accuracy of the navigation unit was measured with a software tool of the workstation by landmark checks in each case. This accuracy was defined as the deviation between the same point in the preoperatively acquired navigation image and the actual patient’s anatomy. Furthermore, the postoperative clinical assessment was reviewed and compared to the preoperative clinical status. 3. Results With the use of CT and MR images, radiological and histological diagnosis was concordant in 7 of 10 cases (70%). Table 2 presents all patients in detail. Pleomorphic adenoma was suspected in one case, and the histopathological examination revealed the presence of an orbital B-cell lymphoma (without other sites of manifestation) (Fig. 1). In another case, a pleomorphic adenoma was suspected, and histopathological examination showed the diagnosis of a rare solitary fibrous tumor. Finally, in a third case, a pleomorphic adenoma was suspected, and histology revealed an orbital pseudotumor (Fig. 2). The intraoperative accuracy of the navigation unit achieved a mean value of 1.35 mm (minimum value 1.1 mm; maximum value 1.55 mm). The navigation system was usable in all cases with no technical difficulties. The fusion images proved to be valuable for surgical planning and supported the surgeon with image guidance during the intervention (Fig. 3). Despite individual differences in surgical approaches, navigation with CT–MR fusion images helped the surgeon to select and to perform the most suitable approach (Fig. 4). Intraoperatively, the image guidance, with the help of fusion images, provided precise visual orientation and early identification of vulnerable neurovascular structures adjacent to the pathology, and therefore, helped to avoid iatrogenic trauma in this series. In all cases, the surgeon was able to perform complete tumor resection, as determined by intraoperative inspection and histology. There were no intraoperative complications. Postoperatively, in 6 of 10 cases orbital/bulbar complaints showed complete regression, in 3 of 10 cases, the complaints were ameliorated notably after surgical intervention, and, in 1 of 10 cases, the complaints remained unchanged. 4. Discussion MDCT and MRI are the methods of choice for the imaging of orbital tumor disease. CT and MR scans are complementary in diagnosing neoplastic disease, each providing different imaging
Fig. 3. A 25-year-old man with a dermoid cyst of the left orbit. (a) The axial CT scan in high-resolution bone window level setting shows an intact lateral orbital wall (arrows). (b) The 3D T1-w sequence delineates a well-defined extraconal lesion (big arrow) with typical intralesional fat (small arrow). (c) The axial fusion image supported the surgeon with precise intraoperative image guidance to approach the lesion. Cross-hairs mark the lesion in its anterior part.
aspects of pathological processes. CT at high resolution, with a bone window level setting, is superior in the visualization of bony involvement. MRI is best for delineating soft tissue and nervous system structures, including the assessment of intracranial tumor extension [1–5]. Orbital tumors, as well as bone and soft tissue changes, were well depicted by CT and MRI with regard to the exact location and relationship of the tumors to the orbital muscles and neurovascular structures. By imaging tumor properties, such as bone erosion and tissue infiltration, both modalities played an important role in characterizing the lesion and helped to delineate benign from malignant pathologies. However, exact diagnosis may be difficult because of
228
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
Fig. 4. A 51-year-old man with a pleomophic adenoma of the right orbit. (a) The 3D T1-w sequence demonstrates a well-defined extraconal tumor (arrow). (b) With the help of the coronal navigational fusion image the surgeon simulated the approach to the tumor (vector). (c) The 3D visualization provided additional information about the tumor location (arrow).
unspecific and overlapping imaging findings, and tumors may be misdiagnosed in cases of unexpected rare tumor entities, as shown in this investigation. Tumors of the orbit are a surgical challenge. The surgical risk, with a considerable rate of morbidity, is related to the close relationship between the tumor and muscular and neurovascular structures. The approach itself contributes to morbidity. On the one hand, invasion of intraorbital structures requires complete tumor resection. On the other hand, the surgeon has to avoid damage to these structures and must focus on their preservation. In this regard,
computer-assisted navigated surgery of orbital tumors offers two main advantages: precise surgical planning and advanced intraoperative orientation, which should ultimately improve the patients’ outcome. Preoperatively, on the basis of standardized approaches, the surgeon plans the surgical vectors, and is able to plan the procedure and the extent of the surgical approach [18]. The precise intraoperative identification of distorted and eroded soft tissue and neurovascular structures during transtumoral dissection without anatomic landmarks is a very important issue of navigated surgery. The navigation system shows the accurate anatomy adjacent to the tumor. Precise intraoperative navigation allows the surgeon to avoid and preserve vital structures, particularly in a complex surgical procedure without real anatomical landmarks for intraoperative orientation. Thus, navigation in orbital surgery helps to reduce surgical manipulation, while at the same time, allowing a more radical tumor resection [6–11]. Furthermore, navigated surgery with CT–MR image data fusion has advantages in preoperative planning and for intraoperative image guidance, as described by other study groups [7,11]. In contrast to computer-assisted surgery with either CT or MRI navigation alone, image data fusion combines and complements the advantages of both CT and MR images, and offers the surgeon more accurate information on the exact geometric and volumetric relationship between the soft tissue structures seen on MRI and the bony structures observed on CT [6,7,10,11]. Although MRI is superior in delineating the lesion itself, the CT information is required to depict the bony structures encountered while approaching the tumor. Otherwise, with CT alone, the depth of the soft tissue pathology cannot be correctly appreciated beyond the margins of the bony defect. The intraoperative accuracy of the navigation unit depends on the quality of the image fusion as well as on the accuracy of imageto-patient registration. Furthermore, the accuracy of the image fusion strongly depends on the preciseness of the obtained CT and MRI data sets [10,11]. The accuracy of the fusion images cannot exceed the slice thickness of the primary CT and MRI data sets. Three-dimensional T1-w GE sequences proved to be very useful for image fusion, due to the slice thickness of 1 mm [19]. The application of high-field MRI (3 TesLa) in image-guided skull base surgery could potentially increase the intraoperative accuracy due to the higher resolution compared to standard-field MRI. Also the use of 64-row MDCT slices may increase the accuracy considering the achievable detector collimation and advanced slice thickness, respectively. However, as yet, there are no studies that have compared field strengths or different MDCT units for this application. The biomechanical features of the orbit, without relevant movements during surgical manipulation, contribute to a high navigational accuracy [6,20]. However, incision of cystic tumors before identifying the tumor borders should be avoided. Critical aspects of navigated orbital tumor surgery should be also considered. The preoperative CT scan is associated with radiation exposure to the eyes. A study that investigated the radiation dose during CT of the sinuses, using similar scan parameters, revealed that the lens doses during high-resolution CT are still below the thresholds of lens opacities and support the role of fine-cut protocols in preoperative assessment [21]. Despite the descriptive study approach and a limited number of patients included in this study, our results demonstrate the benefits of computer-assisted navigated surgery of orbital tumors if performed by an experienced surgeon. Thus, the surgeons must overcome a learning curve to successfully apply this useful and safe technology, which is, however, not a substitute for extensive anatomical knowledge. Although this was a preliminary non-randomized series, our experience with image-guided surgery of the orbit suggests a positive effect on the clinical outcome, due to radical removal of tumorous tissue and conservation of vital orbital structures at the same time.
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
5. Conclusion MDCT and MRI are the modalities of choice for the assessment of orbital tumors. Both modalities play a crucial role in the visualization of benign and malignant tumor properties and are essential for diagnosis. Computer-assisted navigated surgery of orbital tumors with CT–MR fusion images offers advanced preoperative planning and supports the surgeon with precise intraoperative image guidance. References [1] Mueller-Forell W, Pitz S. Orbital pathology. Eur J Radiol 2004;49(2): 105–42. [2] Duvoisin B, Zanella FE, Sievers KW. Imaging of the normal and pathological orbit. Eur Radiol 1998;8(2):175–88. [3] Warner MA, Weber AL, Jakobiec FA. Beningn and malignant tumors of the orbital cavity including the lacrimal gland. Neuroimaging Clin N Am 1996;6(1): 123–42. [4] Mafee MF. Orbit: embryology, anatomy, pathology. In: Som PM, Hugh DC, editors. Head and neck imaging. 4th ed. St. Louis: Mosby; 2003, pp 529–654. [5] Maya MM, Heier LA, Orbital CT. Current use in MR era. Neuroimaging Clin N Am 1998;8(3):651–83. [6] Hejazi N. Frameless image-guided neuronavigation in orbital surgery: practical applications. Neurosurg Rev 2006;29(2):118–22. [7] Leong JL, Batra PS, Citardi MJ. CT–MR image fusion for the management of skull base lesions. Otolaryngol Head Neck Surg 2006;134(5):868–76. [8] Kurtsoy A, Menku A, Tucer B, Oktem IS, Akdemir H. Neuronavigation in skull base tumors. Minim Invasive Neurosurg 2005;48(1):7–12.
229
[9] Rohde V, Spangenberg P, Mayfrank L, Reinges M, Gilsbach JM, Coenen VA. Advanced neuronavigation in skull base tumors and vascular lesions. Minim Invasive Neurosurg 2005;48(1):13–8. [10] Sure U, Alberti O, Petermeyer M, Becker R, Bertalanffy H. Advanced image guided skull base surgery. Surg Neurol 2000;53(6):563–72. [11] Nemec SF, Donat MA, Mehrain S, et al. CT–MR image data fusion for computer assisted navigated neurosurgery of temporal bone tumors. Eur J Radiol 2007;62(2):192–8. [12] Haberland N, Ebmeier K, Hliscs R. Neuronavigation in surgery of intracranial and spinal tumors. J Cancer Res Clin Oncol 2000;126(9):529–41. [13] Gumprecht HK, Widenka DC, Lumenta CB. Brainlab vectorvision neuronavigation system: technology and clinical experience in 131 cases. Neurosurgery 1999;44(1):97–105. [14] Wagner W, Gaab MR, Schroeder HW, Tschiltschke W. Cranial neuronavigation in neurosurgery: assessment of usefulness in relation to type and site of pathology in 284 patients. Minim Invasive Neurosurg 2000;43(3):124–31. [15] Gralla J, Nimsky C, Buchfelder M, Fahlbusch R, Ganslandt O. Frameless stereotactic brain biopsy procedures using Stealth Station: indications, accuracy and results. Zentralbl Neurochir 2003;64(4):166–70. [16] Ganslandt O, Behari S, Gralla J, Fahlbusch R, Nimsky C. Neuronavigation: concept, techniques and applications. Neurol India 2002;50(3):244–55. [17] Wiltfang J, Rupprecht S, Ganslandt O, et al. Intraoperative image-guided surgery of the lateral and anterior skull base in patients with tumors or trauma. Skull Base 2003;13(1):21–9. [18] Day JD, Koos WT, Matula C, Lang J. Color atlas of microsurgical approaches. New York/Stuttgart: Thieme medical publishers; 1997. [19] Czerny C, Rand T, Gstoettner W, Woelfl G, Imhof H, Trattnig S. MR imaging of the inner ear and cerebello-pontine angle: comparison of three-dimensional and two-dimensional sequences. AM J Roentgenol 1998;170(3):791–6. [20] Dorward NL, Alberti O, Velani B, et al. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg 1998;88(4):656–62. [21] Bassim MK, Ebert CS, Sit RC, Senior BA. Radiation dose to the eyes and parotids during CT of the sinuses. Otolaryngol Head Neck Surg 2005;133(4):531–3.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
CT–MR image data fusion for computer-assisted navigated surgery of orbital tumors Stefan Franz Nemec a,∗ , Philipp Peloschek a , Maria Theresa Schmook a , Christian Robert Krestan a , Wolfgang Hauff c , Christian Matula b , Christian Czerny a a b c
Department of Radiology/Division of Neuroradiology and Musculoskeletal Radiology, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria Department of Neurosurgery, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria Department of Ophthalmology, Medical University Vienna, Waehringerguertel 18-20, A-1090 Vienna, Austria
a r t i c l e
i n f o
Article history: Received 27 October 2008 Received in revised form 2 November 2008 Accepted 5 November 2008 Keywords: Computed tomography Magnetic resonance imaging Image fusion Neuronavigation Orbital tumors
a b s t r a c t Purpose: To demonstrate the value of multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) in the preoperative assessment of orbital tumors, and to present, particularly, CT and MR image data fusion for surgical planning and performance in computer-assisted navigated surgery of orbital tumors. Materials and methods: In this retrospective case series, 10 patients with orbital tumors and associated complaints underwent MDCT and MRI of the orbit. MDCT was performed at high resolution, with a bone window level setting in the axial plane. MRI was performed with an axial 3D T1-weighted (w) gradient-echo (GE) contrast-enhanced sequence, in addition to a standard MRI protocol. First, MDCT and MR images were used to diagnose tumorous lesions compared to histology as a standard of reference. Then, the image data sets from CT and 3D T1-w GE sequences were merged on a workstation to create CT–MR fusion images that were used for interventional planning and intraoperative image guidance. The intraoperative accuracy of the navigation unit was measured, defined as the deviation between the same landmark in the navigation image and the patient. Furthermore, the clinical preoperative status was compared to the patients’ postoperative outcome. Results: Radiological and histological diagnosis, which revealed 7 benign and 3 malignant tumors, were concordant in 7 of 10 cases (70%). The CT–MR fusion images supported the surgeon in the preoperative planning and improved the surgical performance. The mean intraoperative accuracy of the navigation unit was 1.35 mm. Postoperatively, orbital complaints showed complete regression in 6 cases, were ameliorated notably in 3 cases, and remained unchanged in 1 case. Conclusion: CT and MRI are essential for the preoperative assessment of orbital tumors. CT–MR image data fusion is an accurate tool for planning the correct surgical procedure, and can improve surgical results in computer-assisted navigated surgery of orbital tumors. © 2008 Published by Elsevier Ireland Ltd.
1. Introduction Tumors of the orbit comprise a number of benign and malignant pathologies that are classified according to the four orbital sections, the intraconal and extraconal compartments, the globe, and the optic nerve [1–4]. Table 1 presents a summary of tumors that can arise from the intra- and extraconal space [4]. The differential diagnosis of some lesions presents a challenge for the preoperative assessment of tumor patients and is extremely important in choosing the correct therapeutic procedure. Therefore, the use of precise radiological techniques plays a crucial role in obtaining an exact
∗ Corresponding author. Tel.: +43 1404007609; fax: +43 1404003777. E-mail address: [email protected] (S.F. Nemec). 0720-048X/$ – see front matter © 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.ejrad.2008.11.003
imaging diagnosis. Multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) are essential modalities with which to achieve a radiological diagnosis and to support surgical planning and performance. MRI provides excellent soft tissue contrast to delineate the anatomy and pathology of the soft tissue, whereas CT, with a high-resolution bone window level setting best evaluates osseus structures [1–5]. The treatment of orbital tumor disease that involves muscular and neurovascular structures requires precise route planning and safe image guidance to provide intraoperative orientation with the greatest accuracy and lowest iatrogenic manipulation. Navigated surgery addresses these concerns with the goal of minimizing surgical morbidity that can be linked to severe postoperative ocular defects or residual tumor. Navigated surgery connects the static visualization of head and neck imaging with the dynamic operative
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
225
Table 1 Intra- and extraconal tumors (according to MF Mafee, shortened version). Intraconal
Extraconal
Common Cavernous hemangioma Optic nerve meningioma Optic nerve glioma Lymphoma Lymphangioma Hemangiopericytoma Rhabdomyosarcoma Metastasis
Capillary hemangioma Cholesterol granulomas Dermoids Lacrimal gland tumors Lymphangioma Peripheral nerve tumors Plasmocytoma Rhabdomyosarcoma
Less common Capillary hemangioma Neurofibroma Schwannoma Lipoma
Fibrous histiocytoma Hemangiosarcoma Hemangiopericytoma Lipoma
procedure [6–14]. The performance of computer-assisted navigated techniques that are used in head and neck surgery have profited significantly from the improved and precise imaging data that have been made possible by the recent advances in computed technology. In this study, we demonstrate the ability of CT and MRI to diagnose orbital tumors, and we present the results of CT–MR image data fusion for computer-assisted surgery of orbital tumors with the help of a navigation system. The protocols that were used are described and the impact of CT, MRI, and CT–MR fusion images in computer-assisted navigated surgery of orbital tumors is discussed. 2. Materials and methods
Fig. 1. A 76-year-old woman with a B-cell lymphoma of the right orbit. Radiologically a pleomorphic adenoma was suspected. (a) The axial CT scan in high-resolution bone window level setting demonstrates an erosion of the lateral orbital wall (arrows). (b) The 3D T1-w sequence delineates a flat extraconal tumor (arrow). The CT scan and the MRI study represent the basis for image fusion.
2.1. Patients The protocol for this study was in accordance with the Declaration of Helsinki. In this retrospective case series, 10 patients (5 women and 5 men with a mean age of 49 of years; range, 25–76 years) were included, after chart review, if the following criteria were fulfilled: (a) clinically suspected benign or malignant orbital tumor with associated complaints; (b) preoperative assessment with standard institutional CT and MRI protocols for orbital tumors; (c) neuronavigated surgery with image guidance based on CT–MR fusion images. A definitive diagnosis was ensured by microscopic examination and corresponding immunhistochemical analysis in each case. Pre-and postoperatively, all patients underwent a profound neuro-opthalmological examination that included testing for visual acuity, visual fields, pupillary responses, ocular motility, palpation of the anterior orbit, and inspection of the external surface of the globe and eyelids. Furthermore, the intraocular pressure was measured and a slit lamp examination was performed. Dilated fun-
duscopy was applied to examine the optic disc and retina. Initially, all 10 patients presented with various complaints, such as unilateral proptosis of the globe, chemosis, diplopia, decreased visual acuity, and ocular pain or pressure sensation. 2.2. Materials MDCT was performed with a 16-row multidetector-unit (Mx 8000 IDT Philips, Eindhoven, the Netherlands) at high resolution with a bone window level setting in a strictly axial plane without inclination of the gantry (120 kV, 200 mAs; FOV ∼ 20 cm; slice thickness 0.8 mm; detector collimation of 4 mm × 0.75 mm). MRI (1.0 T, Gyroscan T10 NT Philips, Eindhoven, the Netherlands) was performed in the axial and coronal planes, with T2-weighted (w) fast-spin-echo sequences [matrix 512 × 512; echo time (TE) 100 ms; repetition time (TR) 3500 ms; slice thickness 4 mm], unenhanced T1-w spin-echo (SE) sequences (matrix 512 × 512; TE 20 ms; TR 550 ms; slice thickness 3 mm), and contrast-enhanced T1-w
Table 2 Patients and pathologies. No.
Age
Sex
Radiology
Location
Histology
Approach
Accuracy (mm)
1 2 3 4 5 6 7 8 9 10
57 47 51 51 25 76 39 47 64 33
M M M F M F F F F M
Lacrimal gland carcinoma Epidermoid Pleomorphic adenoma Cavernous hemangioma Dermoid cyst Pleomorphic adenoma Cavernous hemangioma Pleomorphic adenoma Lacrimal gland carcinoma Pleomorphic adenoma
Extraconal Extraconal Extraconal Intraconal Extraconal Extraconal Intraconal Extraconal Extraconal Extraconal
Adenoidcystic carcinoma Epidermoid Pleomorphic adenoma Cavernous hemangioma Dermoid cyst Primary B-cell lymphoma Cavernous hemangioma Solitary fibrous tumor Adenoidcystic carcinoma Pseudotumor
Supraorbital Lateral Supraorbital Lateral Lateral Supraorbital Lateral Supraorbital Supraorbital Lateral
1.45 1.25 1.5 1.2 1.45 1.55 1.4 1.1 1.3 1.3
226
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
Fig. 2. A 33-year-old man with a pseudotumor of the left orbit. Radiologically a pleomorphic adenoma was suspected. (a) The axial CT scan shows an intact lateral wall (arrows). (b) The 3D T1-w sequence delineates a well-defined extraconal lesion (arrow). (c) Coronal (upper left), sagittal (upper right) and axial (bottom left) navigation image on the workstation monitor. The navigational images are inverted compared with the CT and MRI scan. The navigation images helped the surgeon to approach the lesion (marked by crosshairs). The surgical intervention was documented by a video camera (bottom right).
SE sequences (intravenous application of gadopentetate dimeglumine, Magnevist® ), using a circular polarized head coil. In addition, T1-w SE sequences were obtained, with fat suppression after contrast material administration, in the coronal plane (4 mm slice thickness). MRI was accomplished with a 3D T1-w gradient-echo (GE) contrast-enhanced sequence with a slice thickness of 1 mm in a strictly axial plane without angulation (matrix 256 × 256; TE 6.9 ms; TR 25 ms; measured voxel size 0.90/1.08/2.00 mm; reconstructed voxel size 0.90/0.90/1.00 mm).
Then, image data sets from CT and MRI (3D T1-w GE sequences) were stored on CD ROMs and imported to a workstation (Stealth StationTM , Medtronic USA). A fusion software (Stealth MergeTM , Medtronic USA) was used to search automatically for optimal image-to-image registration between the CT and MRI data sets without manual correction. The CT–MR fusion images were used for preoperative planning and for intraoperative image guidance. For detailed information about the navigation unit, Stealth StationTM , see previous publications [15–17]. In summary, the workstation
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
227
is optical-based and consists of a mobile console with a highresolution monitor, a charge-coupled device (CCD) camera unit, a referencing apparatus with dynamic reference frame (DRF), as well as a Mayfield head-holder and light emitting diodes (LED) pointer that is operated by the surgeon. The camera measures the infrared LED signals and calculates position coordinates. The workstation displays the exact location as a cross-hair on the monitor. The image-to-patient registration to determine corresponding points between the real patient’s anatomy and the navigation images was performed on the basis of at least five to six anatomical surface landmarks to define the coronal, axial, and sagittal plane. Images in the axial, sagittal, and coronal planes, as well as 3D visualizations, were used for image guidance. 2.3. Evaluation Initially, CT and MR images were analysed by a head and neck radiologist (C. C., 20 years experience) to characterize lesions and to differentiate benign and malignant tumors. The radiological diagnosis was compared with the histological standard of reference. Intraoperatively, the accuracy of the navigation unit was measured with a software tool of the workstation by landmark checks in each case. This accuracy was defined as the deviation between the same point in the preoperatively acquired navigation image and the actual patient’s anatomy. Furthermore, the postoperative clinical assessment was reviewed and compared to the preoperative clinical status. 3. Results With the use of CT and MR images, radiological and histological diagnosis was concordant in 7 of 10 cases (70%). Table 2 presents all patients in detail. Pleomorphic adenoma was suspected in one case, and the histopathological examination revealed the presence of an orbital B-cell lymphoma (without other sites of manifestation) (Fig. 1). In another case, a pleomorphic adenoma was suspected, and histopathological examination showed the diagnosis of a rare solitary fibrous tumor. Finally, in a third case, a pleomorphic adenoma was suspected, and histology revealed an orbital pseudotumor (Fig. 2). The intraoperative accuracy of the navigation unit achieved a mean value of 1.35 mm (minimum value 1.1 mm; maximum value 1.55 mm). The navigation system was usable in all cases with no technical difficulties. The fusion images proved to be valuable for surgical planning and supported the surgeon with image guidance during the intervention (Fig. 3). Despite individual differences in surgical approaches, navigation with CT–MR fusion images helped the surgeon to select and to perform the most suitable approach (Fig. 4). Intraoperatively, the image guidance, with the help of fusion images, provided precise visual orientation and early identification of vulnerable neurovascular structures adjacent to the pathology, and therefore, helped to avoid iatrogenic trauma in this series. In all cases, the surgeon was able to perform complete tumor resection, as determined by intraoperative inspection and histology. There were no intraoperative complications. Postoperatively, in 6 of 10 cases orbital/bulbar complaints showed complete regression, in 3 of 10 cases, the complaints were ameliorated notably after surgical intervention, and, in 1 of 10 cases, the complaints remained unchanged. 4. Discussion MDCT and MRI are the methods of choice for the imaging of orbital tumor disease. CT and MR scans are complementary in diagnosing neoplastic disease, each providing different imaging
Fig. 3. A 25-year-old man with a dermoid cyst of the left orbit. (a) The axial CT scan in high-resolution bone window level setting shows an intact lateral orbital wall (arrows). (b) The 3D T1-w sequence delineates a well-defined extraconal lesion (big arrow) with typical intralesional fat (small arrow). (c) The axial fusion image supported the surgeon with precise intraoperative image guidance to approach the lesion. Cross-hairs mark the lesion in its anterior part.
aspects of pathological processes. CT at high resolution, with a bone window level setting, is superior in the visualization of bony involvement. MRI is best for delineating soft tissue and nervous system structures, including the assessment of intracranial tumor extension [1–5]. Orbital tumors, as well as bone and soft tissue changes, were well depicted by CT and MRI with regard to the exact location and relationship of the tumors to the orbital muscles and neurovascular structures. By imaging tumor properties, such as bone erosion and tissue infiltration, both modalities played an important role in characterizing the lesion and helped to delineate benign from malignant pathologies. However, exact diagnosis may be difficult because of
228
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
Fig. 4. A 51-year-old man with a pleomophic adenoma of the right orbit. (a) The 3D T1-w sequence demonstrates a well-defined extraconal tumor (arrow). (b) With the help of the coronal navigational fusion image the surgeon simulated the approach to the tumor (vector). (c) The 3D visualization provided additional information about the tumor location (arrow).
unspecific and overlapping imaging findings, and tumors may be misdiagnosed in cases of unexpected rare tumor entities, as shown in this investigation. Tumors of the orbit are a surgical challenge. The surgical risk, with a considerable rate of morbidity, is related to the close relationship between the tumor and muscular and neurovascular structures. The approach itself contributes to morbidity. On the one hand, invasion of intraorbital structures requires complete tumor resection. On the other hand, the surgeon has to avoid damage to these structures and must focus on their preservation. In this regard,
computer-assisted navigated surgery of orbital tumors offers two main advantages: precise surgical planning and advanced intraoperative orientation, which should ultimately improve the patients’ outcome. Preoperatively, on the basis of standardized approaches, the surgeon plans the surgical vectors, and is able to plan the procedure and the extent of the surgical approach [18]. The precise intraoperative identification of distorted and eroded soft tissue and neurovascular structures during transtumoral dissection without anatomic landmarks is a very important issue of navigated surgery. The navigation system shows the accurate anatomy adjacent to the tumor. Precise intraoperative navigation allows the surgeon to avoid and preserve vital structures, particularly in a complex surgical procedure without real anatomical landmarks for intraoperative orientation. Thus, navigation in orbital surgery helps to reduce surgical manipulation, while at the same time, allowing a more radical tumor resection [6–11]. Furthermore, navigated surgery with CT–MR image data fusion has advantages in preoperative planning and for intraoperative image guidance, as described by other study groups [7,11]. In contrast to computer-assisted surgery with either CT or MRI navigation alone, image data fusion combines and complements the advantages of both CT and MR images, and offers the surgeon more accurate information on the exact geometric and volumetric relationship between the soft tissue structures seen on MRI and the bony structures observed on CT [6,7,10,11]. Although MRI is superior in delineating the lesion itself, the CT information is required to depict the bony structures encountered while approaching the tumor. Otherwise, with CT alone, the depth of the soft tissue pathology cannot be correctly appreciated beyond the margins of the bony defect. The intraoperative accuracy of the navigation unit depends on the quality of the image fusion as well as on the accuracy of imageto-patient registration. Furthermore, the accuracy of the image fusion strongly depends on the preciseness of the obtained CT and MRI data sets [10,11]. The accuracy of the fusion images cannot exceed the slice thickness of the primary CT and MRI data sets. Three-dimensional T1-w GE sequences proved to be very useful for image fusion, due to the slice thickness of 1 mm [19]. The application of high-field MRI (3 TesLa) in image-guided skull base surgery could potentially increase the intraoperative accuracy due to the higher resolution compared to standard-field MRI. Also the use of 64-row MDCT slices may increase the accuracy considering the achievable detector collimation and advanced slice thickness, respectively. However, as yet, there are no studies that have compared field strengths or different MDCT units for this application. The biomechanical features of the orbit, without relevant movements during surgical manipulation, contribute to a high navigational accuracy [6,20]. However, incision of cystic tumors before identifying the tumor borders should be avoided. Critical aspects of navigated orbital tumor surgery should be also considered. The preoperative CT scan is associated with radiation exposure to the eyes. A study that investigated the radiation dose during CT of the sinuses, using similar scan parameters, revealed that the lens doses during high-resolution CT are still below the thresholds of lens opacities and support the role of fine-cut protocols in preoperative assessment [21]. Despite the descriptive study approach and a limited number of patients included in this study, our results demonstrate the benefits of computer-assisted navigated surgery of orbital tumors if performed by an experienced surgeon. Thus, the surgeons must overcome a learning curve to successfully apply this useful and safe technology, which is, however, not a substitute for extensive anatomical knowledge. Although this was a preliminary non-randomized series, our experience with image-guided surgery of the orbit suggests a positive effect on the clinical outcome, due to radical removal of tumorous tissue and conservation of vital orbital structures at the same time.
S.F. Nemec et al. / European Journal of Radiology 73 (2010) 224–229
5. Conclusion MDCT and MRI are the modalities of choice for the assessment of orbital tumors. Both modalities play a crucial role in the visualization of benign and malignant tumor properties and are essential for diagnosis. Computer-assisted navigated surgery of orbital tumors with CT–MR fusion images offers advanced preoperative planning and supports the surgeon with precise intraoperative image guidance. References [1] Mueller-Forell W, Pitz S. Orbital pathology. Eur J Radiol 2004;49(2): 105–42. [2] Duvoisin B, Zanella FE, Sievers KW. Imaging of the normal and pathological orbit. Eur Radiol 1998;8(2):175–88. [3] Warner MA, Weber AL, Jakobiec FA. Beningn and malignant tumors of the orbital cavity including the lacrimal gland. Neuroimaging Clin N Am 1996;6(1): 123–42. [4] Mafee MF. Orbit: embryology, anatomy, pathology. In: Som PM, Hugh DC, editors. Head and neck imaging. 4th ed. St. Louis: Mosby; 2003, pp 529–654. [5] Maya MM, Heier LA, Orbital CT. Current use in MR era. Neuroimaging Clin N Am 1998;8(3):651–83. [6] Hejazi N. Frameless image-guided neuronavigation in orbital surgery: practical applications. Neurosurg Rev 2006;29(2):118–22. [7] Leong JL, Batra PS, Citardi MJ. CT–MR image fusion for the management of skull base lesions. Otolaryngol Head Neck Surg 2006;134(5):868–76. [8] Kurtsoy A, Menku A, Tucer B, Oktem IS, Akdemir H. Neuronavigation in skull base tumors. Minim Invasive Neurosurg 2005;48(1):7–12.
229
[9] Rohde V, Spangenberg P, Mayfrank L, Reinges M, Gilsbach JM, Coenen VA. Advanced neuronavigation in skull base tumors and vascular lesions. Minim Invasive Neurosurg 2005;48(1):13–8. [10] Sure U, Alberti O, Petermeyer M, Becker R, Bertalanffy H. Advanced image guided skull base surgery. Surg Neurol 2000;53(6):563–72. [11] Nemec SF, Donat MA, Mehrain S, et al. CT–MR image data fusion for computer assisted navigated neurosurgery of temporal bone tumors. Eur J Radiol 2007;62(2):192–8. [12] Haberland N, Ebmeier K, Hliscs R. Neuronavigation in surgery of intracranial and spinal tumors. J Cancer Res Clin Oncol 2000;126(9):529–41. [13] Gumprecht HK, Widenka DC, Lumenta CB. Brainlab vectorvision neuronavigation system: technology and clinical experience in 131 cases. Neurosurgery 1999;44(1):97–105. [14] Wagner W, Gaab MR, Schroeder HW, Tschiltschke W. Cranial neuronavigation in neurosurgery: assessment of usefulness in relation to type and site of pathology in 284 patients. Minim Invasive Neurosurg 2000;43(3):124–31. [15] Gralla J, Nimsky C, Buchfelder M, Fahlbusch R, Ganslandt O. Frameless stereotactic brain biopsy procedures using Stealth Station: indications, accuracy and results. Zentralbl Neurochir 2003;64(4):166–70. [16] Ganslandt O, Behari S, Gralla J, Fahlbusch R, Nimsky C. Neuronavigation: concept, techniques and applications. Neurol India 2002;50(3):244–55. [17] Wiltfang J, Rupprecht S, Ganslandt O, et al. Intraoperative image-guided surgery of the lateral and anterior skull base in patients with tumors or trauma. Skull Base 2003;13(1):21–9. [18] Day JD, Koos WT, Matula C, Lang J. Color atlas of microsurgical approaches. New York/Stuttgart: Thieme medical publishers; 1997. [19] Czerny C, Rand T, Gstoettner W, Woelfl G, Imhof H, Trattnig S. MR imaging of the inner ear and cerebello-pontine angle: comparison of three-dimensional and two-dimensional sequences. AM J Roentgenol 1998;170(3):791–6. [20] Dorward NL, Alberti O, Velani B, et al. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg 1998;88(4):656–62. [21] Bassim MK, Ebert CS, Sit RC, Senior BA. Radiation dose to the eyes and parotids during CT of the sinuses. Otolaryngol Head Neck Surg 2005;133(4):531–3.