close

Вход

Забыли?

вход по аккаунту

?

911

код для вставкиСкачать
Computer Aided Surgery 3:27–32 (1998)
Clinical Paper
Three-Dimensional Computer Assisted Magnetic
Resonance Imaging for Neurosurgical Planning in
Parasagittal and Parafalcine Central Region Tumors
J. Burtscher,
M.D.,
C. Kremser, Ph.D., M. Seiwald, M.D., A. Obwegeser, M.D., M. Wagner, M.D.,
F. Aichner, M.D., K. Twerdy, M.D., and S. Felber, M.D.
Departments of Neurosurgery (J.B., M.S., A.O., K.T.) and Magnetic Resonance (C.K., M.W., F.A. S.F.),
University of Innsbruck, Innsbruck, Austria
ABSTRACT Usually, conventional magnetic resonance spin echo images (MRI) are sufficient to
establish the diagnosis of intracranial pathology. Planning and executing a neurosurgical procedure
requires the ability of the neurosurgeon to transform these two-dimensional MRI into a threedimensional (3-D) virtual image of the pathology and the surrounding neuronal anatomy. Such
mentally performed transformations after sequential observation of the individual two-dimensional
slices (i.e., MRI and angiography) may be virtual tasks that are very difficult or sometimes impossible
to achieve.
Using 3-D MRI data sets and a semiautomatic computer assisted segmentation technique,
we tried to simulate intraoperative situs-based 3-D MRI reconstructions of parasagittal and parafalcine central region tumors. The MRI reconstructions were integrated into the neurosurgical planning procedure as an additional tool. They proved to be an important adjunct in determining the
distinct anatomy of the intracranial pathology in its relation to the surrounding and overlying brain
and vascular (especially venous) anatomy. With 10 patients with central region parasagittal and
parafalcine tumors, we found that the 3-D MRI reconstructions revealed additional information
compared to conventional cross-sectional images and had an influence on neurosurgical planning
and strategy, improving neurosurgical performance and patient outcome. Comp Aid Surg 3:27–32
(1998) q1998 Wiley-Liss, Inc.
Key words: three-dimensional MRI, magnetic resonance imaging, meningioma surgery
INTRODUCTION
Magnetic resonance imaging (MRI) is the
accepted method of choice for the diagnosis of
central nervous system disorders.2 Present neurosurgical planning often depends on two-dimensional (2-D) information obtained from MRI and
computed tomography (CT) cross-sections.16,17
After sequential reading of a series of 2-D images,
the information has to be mentally transformed
into a virtual three-dimensional (3-D) image of
the complex 3-D anatomy by the neurosurgeon.
Received November 13, 1997; accepted March 14, 1998.
Address correspondence/reprint requests to Dr. Johannes Burtscher, Univ. Klinik für Neurochirurgie, A-6020 Innsbruck,
Austria. E-mail: [email protected]
q1998 Wiley-Liss, Inc.
/ 811c$$7041
06-26-98 12:15:10
igsas
W: IGS
97041
28 Burtscher et al.: 3-D MRI Reconstructions for Neurosurgical Planning
Table 1. MR Technique
Imaging system
sequence
Parameters
TR
TE
TI
a
Matrix
Partitions
3D reconstruction
mode
Workstation
1.5 T Magnetom Vision
(Siemens, Germany)
D MPRAGE
9.7 msec
4 msec
300 msec
127
256 1 256
141
Volume-based surface rendering
Allegro, ISG Technologies (Canada)
These mental transformations are difficult or
sometimes even impossible, and a number of
groups are now integrating 3-D MRI and CT information into the planning of intracranial operations. To approach subcortical and cortical lesions
within the central region safely, exact knowledge
of the gyral and sulcal anatomy of the cerebral
cortex in relation to the tumor and the cortical
veins is essential.4,7,10,12,13
We evaluated the impact of a 3-D display
of the brain, vasculature, and tumor on surgical
decisions during planning and executions of operations for intracranial parasagittal and parafalcine
tumors at the middle one-third of the superior
sagittal sinus and the falx. The 3-D reconstruction
and display is based on 3-D MRI data sets and a
semiautomatic segmentation technique. Tumors
as well as the surrounding or overlying neuronal
and neurovascular anatomy can be interactively
evaluated on the computer screen.
MRI PARAMETERS
All patients were studied on a whole-body superconducting imaging system operating at a field
strength of 1.5 T (Magnetom Vision; Siemens,
Germany) using a circular polarized head coil.
The examination protocol (Table 1) consisted of
a flash 3-D sequence (3-D MPRAGE; see below)
to cover the entire head after intravenous application of 0.1 mmol/kg/BW Gd DTPA (Magnevist,
Schering, Germany).
The fast T1-weighted flash 3-D sequence
is a magnetization prepared rapid gradient echo
(MPRAGE) sequence that uses a 1807 inversion
pulse as preparation pulse. After an inversion
time, TI (after the 1807 pulse), all sampling points
in the slice select direction are acquired. By the
use of the inversion pulse, T1 weighting is
/ 811c$$7041
06-26-98 12:15:10
achieved that provides exact delineation of complex anatomical structures on the MRI.
Thin slices of 1.2 mm, minimal pixel size,
and short TE compensate for intravoxel phase
dispersions due to susceptibility effects at bone/
air to soft tissue interfaces. The resulting images
are nearly isotropic in resolution (0.9 1 0.9 1
11.2 mm); hence secondary reformatted images
have resolution identical to that of the source images. The acquisition time of the MPRAGE sequence is about 5 min. After contrast agent application, tumors with blood – brain barrier disruption and blood vessels at the surface of the brain
are bright and can semiautomatically be segmented from normal brain tissues.
IMAGE PROCESSING
The data were transferred from the MRI scanner
through a network connection to the image processing workstation (Allegro; ISG Technologies,
Ontario, Canada). Segmentation was performed
by a neurosurgeon (J.B.). A volume-rendering
display was obtained by using the initial image
data from the MPRAGE studies.
The structures of interest (tumor, brain, vessels, and skin) are segmented from the gray-level
2-D image and reconstructed into 3-D volumes
and objects. This volume-rendering technique depends on the identification and delineation of regions of interest, which for complex anatomical
objects (e.g., vessels and tumor) often has to be
done by manual contouring. This is a time-consuming and operator-dependent process. Therefore, the operator, who is the neurosurgeon planning the special operative strategy in a given patient, must be trained in the analysis of MRI and
must also be familiar with 3-D image rendering
and postprocessing.
Manual contouring is not necessary in all
patients. By setting a pixel intensity or threshold
range, the workstation aids this process by defining regions with similar intensities, and further
manual editing is not always necessary. In using
this system, the semiautomatic segmentation of
different tissues is done in an interactive fashion
with the operator previewing and, when necessary, editing each section for exact delineation of
anatomical structures. A seed-growing algorithm
that uses threshold ranges to define gray levels
corresponding to specific tissue types is important
for the semiautomatic segmentation technique.
Seed growing enhances threshold segmentation
by allowing the user to place a seed on the struc-
igsas
W: IGS
97041
Burtscher et al.: 3-D MRI Reconstructions for Neurosurgical Planning 29
ture of interest, which then propagates further and
extends to the edge of the structure. The propagation continues until the neighboring pixel is outside the threshold range. In this way, the region
of interest can be quickly and exactly delineated.
Skin and sometimes, owing to the contrast medium enhancement, also tumors and vessels are
readily segmented because of sufficient signal intensity differences. At the brain surface and often
around vessels and tumor, manual editing may be
necessary because of similar gray-scale levels and
threshold values of these neighboring structures.
The delineated regions of interest are reconstructed into a 3-D volume or object. For realtime viewing, surface-shading algorithms are applied to simulate the reflection of an external light
source from the 3-D rendered surfaces.
THREE-DIMENSIONAL DISPLAY
OF INFORMATION
The results of segmentation are rendered as a
surface image showing the precise gyral and sulcal anatomy in relation to tumors and cortical
veins. Deep-seated subcortical lesions as well as
the cortical anatomy can be viewed through transparent bone and cortex using a special feature
of the Allegro software. Other features of the
rendering program allow the operator to assign
colors to each rendered anatomical part or to
zoom in and out on regions of interest from different angles. Additionally, cutaway block views can
be created.
The display of the surface of the head can
be positioned in every direction on the computer
screen, thus allowing for simulation of the expected positioning of the patient’s head intraoperatively. With the head positioned for optimal surgical exposure, craniotomies and cortisectomies
can be simulated. For immediate access, the 3-D
reconstructions were transferred into the operating room.
DATA ANALYSIS
The data analysis was carried out to evaluate the
impact of the additional information generated
by the 3-D MRI reconstructions on neurosurgical
decisions during planning and executions of operations. The neurosurgeons not involved in the
segmentation and reconstruction procedure discussed the anatomic relationships and operative
strategies first relying only on MRI and other
cross sections (e.g., angiography and CT) without
the 3-D MRI information.
/ 811c$$7041
06-26-98 12:15:10
The surgeons were then confronted with a
display of the surface of the head, the underlying
cortical veins, and a projection of the tumor
within the imaging volume and were asked to
give a report of the additional information provided to them by the 3-D MRI reconstructions.
For all patients, the 3-D MRI displays were compared to photos showing the intraoperative anatomical situation.
PATIENTS
The patients for this protocol were selected by a
team of neurosurgeons and neuroradiologists. The
selected group consists of patients with intracranial parasagittal (Fig. 1) and parafalcine (Fig. 2)
tumors, located at the middle one-third of the
superior sagittal sinus and the falx. The pathological diagnosis included ten patients with either parasagittal or parafalcine tumors. The five patients
with parasagittal tumors had meningiomas, and,
among the five patients with parafalcine tumors,
three had meningiomas, one had a glioma, and
one had a metastatic tumor.
RESULTS
We evaluated the impact of a computer assisted
3-D display technique based on 3-D MRI data
sets on surgical planning, operation strategies,
and intraoperative neurosurgical decisions in patients with intracranial parasagittal and parafalcine tumors at the middle one-third of the superior
sagittal sinus and the falx. The neurosurgical staff
reported and analyzed the additional information
provided by the 3-D MRI reconstructions.
The integrated 3-D display of brain, vasculature, and tumor proved to be superior to conventional cross sections alone (e.g., angiography,
MRI, and CT) for preoperative planning. The 3D MRI display improved the surgical decision
during planning and actual execution of the operations by providing improved information about
the optimal sites and extensions of craniotomies
and cortisectomies and the underlying gyral and
sulcal anatomy. Additionally, the delineation of
the distinct anatomy of the tumor in its relation to
the surrounding and overlying brain and vascular
anatomy was significantly improved.
Variable viewing angles, visualization of
gyral and sulcal surface anatomy, selective visualization of tumor and vessels, simulation of surgical exposures, tumor volume measurement, coloring of 3-D objects for better visual delineation,
and 3-D rotational ability were often-used fea-
igsas
W: IGS
97041
30 Burtscher et al.: 3-D MRI Reconstructions for Neurosurgical Planning
Fig. 1. Parasagittal meningioma involving the central sagittal sinus and extending into both hemispheres. A: Posterior
view of the venous system and the meningioma after 3D-MR reconstruction. B: The software window function allowing
for inspection of the tumor, the brain, the cortical vessels, and the superior sagittal sinus after simulation of the
craniotomy. C,D: Integrated display of the brain surface, vessels, and tumor, defining the motor and sensory cortex in
its relation to the tumor borders and to the cortical veins.
tures. The 3-D MRI were compared with the real
intraoperative anatomy. For quality control and
documentation, the intraoperative situation was
photographed. The results correlated well, revealing a good technical and visualization standard
of the integrated 3-D display.
DISCUSSION
In his 1922 Cavendish Lecture, Harvey Cushing8
established the category of parasagittal meningiomas as tumors arising from the wall of the sinus
or its lateral expansion (Fig. 1), whereas falx meningiomas (Fig. 2), completely concealed by the
overlying cerebral cortex, typically do not involve
the sagittal sinus unless they are large or spread
on plaque.9
Both entities, when they are located at the
middle one-third of the superior sagittal sinus or
the falx, are challenging to the neurosurgeons because of their location at the rolandic and motor
/ 811c$$7041
06-26-98 12:15:10
cortex area, as well as the large draining veins
within the surgical field, and of course, in the case
of parasagittal meningiomas, because of tumor
growth within and through the sinus wall.1,5,6,11,18
We believe that, in parasagittal and parafalcine tumor surgery, presurgical angiographic examination is still the method of choice for patency
assessment of the superior sagittal sinus.14,15,19,20
However, the evaluation of the exact tumor location in its relation to the draining cortical veins
and the rolandic and motor cortex area based on
the information of the venous phase of the angiogram (tumor stain) and MRI/CT cross sections is
difficult.
With all of our 10 patients with parasagittal
and parafalcine tumors, we experienced that,
compared to the angiographic examinations as
well as to MRI and CT cross sections, the 3-D
MRI reconstructions revealed additional anatomical information concerning the relationship of the
igsas
W: IGS
97041
Burtscher et al.: 3-D MRI Reconstructions for Neurosurgical Planning 31
Fig. 2. Left central parafalcine lesion. A: View of the brain surface showing the anatomy of the frontoparietal lobes
with the central region on the left. B: Software window function revealing the location and extent of the subcortical
parafalcine tumor at the sensory cortex. Note the big cortical vein in the projection line of the tumor. C,D: A more
lateral view demonstrating the relationship of the tumor to the venous system and to the brain surface. Note the two
skin markers indicating the projection line of the anterior and posterior borders of the tumor on the skin surface from
the surgeon’s point of view.
draining veins to the superior sagittal sinus and
the central region area at the middle one-third of
the superior sagittal sinus and the falx.
In conclusion, volume renderings based on
3-D MRI data sets3 might have an important impact on neurosurgical planning. Surgical performance can be improved and the risk of complications reduced, if the information obtained by the
3-D reconstructions is transferred to the operating
room. The application of 3-D MRI requires close
interaction between neurosurgeons and neuroradiologists. Neurosurgeons who want to integrate
additional information obtained by 3-D volume
renderings into the planning of intracranial operations should be familiar with the principles of
tracing and thresholding techniques. Furthermore,
this raises the need for surgeons to be trained
in the principles of MRI interpretation as well,
/ 811c$$7041
06-26-98 12:15:10
allowing for correct identification of anatomical
structures that have to be segmented.
REFERENCES
1.
2.
3.
4.
igsas
Bonnal J, Brotchi J (1978) Surgery of the superior
sagittal sinus in parasagittal meningiomas. J Neurosurg 48:935.
Brandt-Zavadzki MN, Norman D (1987) Magnetic
Resonance Imaging of the Central Nervous System.
New York: Raven Press.
Brant-Zawatzki MN, Gillan GD, Atkinson DJ, Edalatpour N, Jensen M (1993) Three-dimensional MR
imaging and display of intracranial disease: Improvements with the MP-RAGE sequence and gadolinium. JMRI 3:656–662.
Burtscher J, Voges M, Donauer E, Birbamer G,
Twerdy K, Felber SR (1994) Impact of 3D MRI on
neurosurgery. In Aichner FT, Felber SR, Muller
RN, Rinck PA (eds): Three Dimensional Magnetic
W: IGS
97041
32 Burtscher et al.: 3-D MRI Reconstructions for Neurosurgical Planning
5.
6.
7.
8.
9.
10.
11.
12.
Resonance Imaging. Boston: Blackwell Scientific
Publications, pp 87–94.
Caliauw L (1985) Ergebnisse bei der neurochirurgischen Behandlung parasagittaler Meningiome. Z
Neurochir 46:348–353.
Chan RC, Thompson GB (1984) Morbidity, mortality and quality of life following surgery of intracranial meningiomas: A retrospective study in 257
cases. J Neurosurg 60:52–60.
Ciric IS, Rosenblatt S (1993) Supratentorial craniotomies. In Apuzzo MLJ (ed): Brain Surgery, Complication Avoidance and Management. New York:
Churchill Livingston, pp 51–70.
Cushing H (1922) The meningiomas (dural endotheliomas): Their source and favoured seats of origin. Brain 45:282–316.
Cushing H, Eisenhardt L (1938) Meningiomas:
Their classification, regional behaviour, life history,
and surgical end results. Springfield, IL: Charles C.
Thomas.
Hoehne KH, Bomans M, Pflesser B, Pommert A,
Riemer M, Schiemann T, Tiede U (1994) Perspectives of 3D imaging. In Aichner FT, Felber SR,
Muller RN, Rinck PA (eds): Three-Dimensional
Magnetic Resonance Imaging. Boston: Blackwell
Scientific Publications, pp 27–40.
Hoessly GF, Olivecrona H (1955) Report on 280
cases of verified parasagittal meningioma. J Neurosurg 12:614.
Hu X, Tan KK, Levin DN, Galhotra S, Mullan JF,
/ 811c$$7041
06-26-98 12:15:10
13.
14.
15.
16.
17.
18.
19.
20.
igsas
Hekmatpanah J, Spire JP (1990) Three-dimensional
magnetic resonance imaging of the brain. J Neurosurg 72:433–440.
Kikins R, Gleason PL, Moriarty TM, Moore MR,
Alexander E, Stieg PE, Matsumae M, Lorensen
WE, Cline HE, Black PM, Jolesz FA (1996) Computer-assisted interactive three-dimensional planning for neurosurgical procedures. Neurosurgery
38:640–651.
Marc JA, Schlechter MM (1974) Cortical venous
rerouting in parasagittal meningiomas. Radiology
112:85.
Morris L (1960) Angiography of the superior and
parasagittal sinuses. Br J Radiol 33:606.
Scholz M, Schwechheimer MK, Hardenack M,
Heuser L, Harders AG (1996) MIN-biopsy of brain
tumors: Operative technique and histomorphological results. Min Invas Neurosurg 39:12–16.
Seeger W (1979) Microsurgery of the Brain. Anatomical and Technical Principles I. Wien: SpringerVerlag.
Seeger W (1984) Microsurgery of Cerebral Veins.
Wien: Springer Verlag, p 86.
Waga S, Handa H (1976) Scalp veins as collateral
pathway with parasagittal meningiomas occluding
superior parasagittal sinus. Neuroradiology 11:119.
Yasargil MG, Damur M (1974) Thrombosis of the
cerebral veins and dural sinuses. In Newton TH,
Potts DG (eds): Radiology of the Skull and Brain,
Vol. 2. St. Louis: C.V. Mosby, p 2375.
W: IGS
97041
Документ
Категория
Без категории
Просмотров
13
Размер файла
558 Кб
Теги
911
1/--страниц
Пожаловаться на содержимое документа