Intraoperative navigated 3-dimensional ultrasound angiography in tumor surgery

Surgical Neurology 66 (2006) 581 – 592 www.surgicalneurology-online.com Technique-Imaging Intraoperative navigated 3-dimensional ultrasound angiography in tumor surgery Ola M. Rygh, MDa,b,d,4, Toril A. Nagelhus Hernes, PhDb,c,d, Frank Lindseth, PhDc,d, Tormod Selbekk, MScc,d, Tomm Brostrup Mqller, PhD, MDa,d, Geirmund Unsgaard, PhD, MDa,b,d a Department of Neurosurgery, St. Olav University Hospital, 7006 Trondheim, Norway b The Norwegian University of Science and Technology, 7006 Trondheim, Norway c SINTEF Health Research, 7491 Trondheim, Norway d National Centre for 3D ultrasound in Surgery, St. Olav University Hospital, 7006 Trondheim, Norway Received 20 January 2006; accepted 23 May 2006 Abstract Background: Avoiding damage to blood vessels is often the concern of the neurosurgeon during tumor surgery. Using angiographic image data in neuronavigation may be useful in cases where vascular anatomy is of special interest. Since 2003, we have routinely used 3D ultrasound angiography in tumor surgery, and between January 2003 and May 2005, 62 patients with different tumors have been operated using intraoperative 3D ultrasound angiography in neuronavigation. Methods: An ultrasound-based neuronavigation system was used. In addition to 3D ultrasound tissue image data, 3D ultrasound angiography (power Doppler) image data were acquired at different stages of the operation. The value and role of navigated 3D ultrasound angiography as judged by the surgeon were recorded. Results: We found that intraoperative ultrasound angiography was easy to acquire and interpret, and that image quality was sufficient for neuronavigation. In 26 of 62 cases, ultrasound angiography was found to be helpful by visualizing hidden vessels adjacent to and inside the tumor, facilitating tailored approaches and safe biopsy sampling. Conclusions: Intraoperative 3D ultrasound angiography is straightforward to use, image quality is sufficient for image guidance, and it adds valuable information about hidden vessels, increasing safety and facilitating tailored approaches. Furthermore, with updated 3D ultrasound angiography imaging, accuracy of neuronavigation may be maintained in cases of brain shift. D 2006 Elsevier Inc. All rights reserved. Keywords: Neuronavigation; Image-guided surgery; Ultrasound; Doppler ultrasonography 1. Introduction During tumor surgery, avoiding damage to blood vessels is one of many concerns for the neurosurgeon. Neuronavigation may help the neurosurgeon to locate the position of important structures such as vessels; thus, tumor resection may be Abbreviations: 3D, 3-dimensional; AVM, arteriovenous malformation; BFI, blood flow imaging; CT, computed tomography; DSA, digital subtraction angiography; FMRI, functional magnetic resonance imaging; MRI, magnetic resonance imaging; PICA, posterior inferior cerebellar artery. 4 Corresponding author. St Olav University Hospital, Trondheim N7005 Trondheim Norway. Tel.: +47 73867596; fax: +47 73867977. E-mail address: ola.rygh@ntnu.no (O.M. Rygh). 0090-3019/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.surneu.2006.05.060 performed more safely and more radically. However, brain shift as well as registration errors may limit the overall accuracy of a neuronavigation system [3,7,9,16]. To compensate for brain shift, intraoperative imaging has been introduced in neuronavigation [1,4], thus maintaining accuracy [26,29]. In addition, residual tumor may be detected. In our clinic, we have used neuronavigation with integrated intraoperative ultrasound imaging for guidance of tumor resection for several years [4-6,25-28]. The equipment also has capability of Doppler imaging with power Doppler for visualization of vessels. Thus, 3D ultrasound angiography image data can be acquired intraoperatively for neuronavigation. The aim of the present 582 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 study was to evaluate neuronavigation with intraoperative 3D ultrasound angiography (power Doppler) for imageguided resection of intracranial tumors by visualizing the vascular anatomy adjacent to and inside the tumor. 2. Patients and methods Between January 2003 and May 2005, we have used intraoperative ultrasound angiography in 62 tumor cases. Cases selected for use of neuronavigation with 3D ultrasound angiography were those wherein neuronavigation (with 3D ultrasound) was already requested and wherein the surgeon assumed that ultrasound angiography might be of value. We used an intraoperative ultrasound-based neuronavigation system, SonoWand (MISON AS, Trondheim, Norway), with a 4- to 8-MHz flat phased array probe with optimal focusing properties at 3 to 6 cm and with the capability of acquiring power Doppler (ultrasound angiography) images. On the ultrasound angiography images, vessels were displayed as an overlay in shades of red over the tissue image according to the power of the Doppler signal. For neuronavigation with 3D ultrasound, only power Doppler image data were applied in the present study. The ultrasound platform (a GE Vingmed Vivid FiVe unit) in SonoWand also has triplex imaging, but this cannot be imported and used for neuronavigation with 3D ultrasound. For tracking of the ultrasound probe, a tracking frame was attached. Neuronavigation was performed with a tracked pointer device or the CUSA (Valleylab, Boulder, Colo, USA) with an attached tracking frame. Tracking of the CUSA enabled continuous monitoring of the position of the tip and the trajectory of this instrument, bonline resection.Q A biopsy forceps with tracking was also available for image guidance of biopsies. For acquisition of 3D ultrasound image data for navigation, the ultrasound probe was placed on the dura and tilted or translated over the area of interest by a freehand movement. The 2D ultrasound images acquired were reconstructed to a 3D ultrasound data set ready for navigation. Image acquisition and reconstruction typically took about 1 minute. The ultrasound probe was not in the operating field after the acquisition of a 3D ultrasound volume, unless additional 3D ultrasound data sets were acquired or real-time 2D imaging was needed. The ultrasound angiography images were displayed on the screen of the navigation system, usually in addition to the preoperative MRI, which was displayed simultaneously on the same screen. With a simple pointer device, the safest surgical route to the tumor was determined with the smallest risk of vascular damage. Fig. 1. Workflow in 3D ultrasound angiography. A: 3D ultrasound angiography image data are acquired by tilting or translating the probe over the area of interest. B: Then, the calibrated CUSA is used for online resection of the tumor. The navigation system displays the position of the CUSA in relation to the tumor margin and the vessels. C: After some resection, updated 3D ultrasound angiography data may be necessary because of brain shift (arrows) and is again acquired by tilting or translating the probe over the area of interest. 583 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 Table 1 List of cases Case Diagnosis Location of tumor Helpfulness of ultrasound angiography 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Chondrosarcoma Anaplastic astrocytoma Glioblastoma Low-grade astrocytoma Anaplastic astrocytoma Chordoma Meningioma Metastasis (malignant melanoma) Metastasis (lung cancer) Glioblastoma Chordoma Metastasis (malignant melanoma) Low-grade astrocytoma Metastasis (malignant melanoma) Trigeminal schwannoma Glioblastoma Low-grade astrocytoma Meningioma (sphenoid wing) Meningioma (olfactory) Meningioma (left convexity) Vestibular schwannoma Biopsy (result: normal tissue) Pineoblastoma Meningioma (sphenoid wing) Meningioma (olfactory) Anaplastic astrocytoma Glioblastoma Chordoma Ependymoma Meningioma (sphenoid wing) Low-grade astrocytoma Meningioma (foramen magnum) Glioblastoma (biopsy only) Unknown lesion Meningioma (posterior fossa) Anaplastic astrocytoma Meningioma (falx) Glioblastoma Meningioma (olfactory) Pilocytic astrocytoma Optic glioma Meningioma (sphenoid wing) Meningioma (parasagittal) Low-grade glioma Low-grade glioma Glioblastoma Vestibular schwannoma Low-grade astrocytoma Ganglioglioma Glioblastoma Meningioma Ganglioglioma Meningioma (falx) Meningioma (convexity) Meningioma (convexity) Glioblastoma Metastasis (breast cancer) Meningioma (convexity) Glioblastoma Meningioma (sphenoid wing) Vestibular schwannoma Meningioma (convexity) Cavernous sinus region (left side) Left frontal lobe Left frontal lobe Right parietotemporal region Left parietooccipital region Clivus Pineal region Right frontal lobe Left occipital lobe Right temporal lobe Cavernous sinus region (left side) Cavernous sinus region (right side) Left medial temporal lobe (hippocampal region) Occipital lobe Cavernous sinus region (right side) Left parietal lobe Right parietal lobe, close to central sulcus Sphenoid wing, skull base Anterior cranial fossa, midline Anterior cranial fossa (left side) Cerebellopontine angle (right) Left temporal lobe, hippocampal region Vermis of cerebellum Sphenoid wing (right side) Anterior cranial fossa, midline Right medial temporal lobe Left temporal lobe Posterior fossa Posterior fossa, fourth ventricle Sphenoid wing (left side) Temporal lobe (left) Foramen magnum Parietal lobe (right) Left temporal lobe Posterior fossa Parietal lobe, midline Falx, parietal region Right temporal lobe Anterior cranial fossa, skull base Posterior cranial fossa Extending from chiasma Sphenoid wing (right side) Occipital convexity (left side, parasagittal) Right temporal lobe Right frontal lobe Right occipital lobe Cerebellopontine angle (right side) Left medial temporal lobe Right temporal lobe Arising from left thalamus, intraventricular Sella region Intraventricular (left side ventricle) Occipital, arising from falx (left side) Occipital convexity (right) Anterior cranial fossa, right convexity Occipital lobe (right side) Right temporal lobe, sylvian fissure Parietal convexity meningioma, right side Left medial frontal lobe and corpus callosum Sphenoid wing, skull base Cerebellopontine angle, right side Parietal region, right side Yes No Yes Possibly No No Yes Yes No No Yes Yes Yes No Yes No Yes Possibly Yes No No No No Possibly Yes Yes No No Yes Yes Yes Yes Yes Yes No No Possibly No Possibly Yes Possibly Yes No No Yes Possibly No Yes Possibly Possibly No No Yes No Yes No Yes No Yes Possibly No No List of patients with type of lesion, location of lesion, and usefulness of ultrasound angiography as judged by the surgeon. 584 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 Fig. 2. Visualization of vessels. Snapshots from the neuronavigation system. A: Low-grade glioma in medial part of temporal lobe (case 26). Magnetic resonance imaging (T1 with gadolinium) (top) and ultrasound angiography (bottom). Vessels in the circle of Willis are visualized by ultrasound angiography; the tumor is partially resected. B: Glioma in right temporal lobe (case 38). Magnetic resonance imaging (T1 with gadolinium) (top) and ultrasound angiography (bottom). Vessels surrounding the tumor, probably both arterial and venous, are displayed on ultrasound angiography. The surgeon suspected that branches of the middle cerebral artery were among these vessels. From power Doppler alone, it is not possible to tell which vessels are arterial and which are venous. Being superficial, the vessels give a strong power Doppler signal. C: Meningioma originating from posterior part of falx (case 53). Magnetic resonance imaging (T1 with gadolinium) (top) and ultrasound angiography (bottom). Multiple veins inside the meningioma drain into the straight sinus. D: Thalamic glioma (case 50). Magnetic resonance imaging (T1 with gadolinium) (top) and ultrasound angiography. The thalamostriate vein is revealed by ultrasound angiography, crossing the tumor surface. E: Meningioma, foramen magnum (case 32). Magnetic resonance imaging (T1, top) and ultrasound angiography (bottom). Notice branches from vertebral arteries adjacent to the tumor. F: Pilocytic astrocytoma in posterior fossa (case 40). Ultrasound tissue (top) and ultrasound angiography (bottom). The vein of Galen is visualized (arrow) by ultrasound angiography. Additional ultrasound and ultrasound angiography image data were acquired after some resection, if required, because of suspected brain shift or for resection control. The workflow with the equipment is summarized in Fig. 1. Additional resection was usually carried out if residual tumor was detected, except in cases where this was decided against because of unacceptable risk of neurological damage to the patient. 585 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 Table 2 Cases where navigated 3D ultrasound angiography was found helpful Case Diagnosis Location of tumor Vessel visualized Role of navigation 1 Chondrosarcoma Internal carotid artery Visualization of hidden vessel 3 7 Glioblastoma Meningioma Cavernous sinus region (left side) Left frontal lobe Pineal region Middle cerebral artery Great cerebral vein 8 Metastasis (malignant melanoma) Chordoma Right frontal lobe Pericallosal arteries Cavernous sinus region (left side) Cavernous sinus region (right side) Internal carotid artery Visualization of hidden Visualization of hidden online CUSA resection Visualization of hidden online CUSA resection Visualization of hidden Left medial temporal lobe (hippocampal region) Cavernous sinus region (right side) Right parietal lobe, close to central sulcus Anterior cranial fossa, midline Anterior cranial fossa, midline Right medial temporal lobe Middle cerebral artery Visualization of hidden vessel, facilitation of approach, online CUSA resection Visualization of hidden vessel Internal carotid artery Visualization of hidden vessel Branches from middle cerebral artery Anterior cerebral artery Visualization of hidden vessel 11 12 Metastasis (malignant melanoma) 13 Low-grade astrocytoma 15 Trigeminal schwannoma 17 Low-grade astrocytoma 19 26 Meningioma (olfactory) Meningioma (olfactory) Anaplastic astrocytoma 29 Ependymoma 30 Meningioma (sphenoid wing) Low-grade astrocytoma 25 31 32 Posterior fossa, fourth ventricle Sphenoid wing (left side) Left temporal lobe 33 Meningioma (foramen magnum) Glioblastoma (biopsy only) 34 Unknown lesion Left medial temporal lobe 40 42 45 Pilocytic astrocytoma Meningioma (sphenoid wing) Low-grade glioma Posterior cranial fossa Sphenoid wing (right side) Right frontal lobe 48 Low-grade astrocytoma 53 Meningioma (falx) 55 Meningioma (convexity) Metastasis (breast cancer) Glioblastoma Left medial temporal lobe Occipital, arising from falx (left side) Right anterior cranial fossa Right temporal lobe in sylvian fissure Left medial frontal lobe and corpus callosum 57 59 Internal carotid artery Anterior cerebral artery Middle cerebral artery and internal cerebral vein Branches from PICA Anterior and middle cerebral arteries Middle cerebral artery Foramen magnum Vertebral arteries Right parietal lobe Smaller vessels Small cortical arteries on medial surface of temporal lobe Vein of Galen Middle cerebral artery and internal carotid Pericallosal artery The procedures were performed by 3 surgeons, and most of the procedures (51 cases) were performed by the senior author (GU). The value and role of navigated ultrasound angiography as judged by the surgeon were recorded postoperatively in most cases. In addition, we have retrospectively reviewed all cases. The subjective judgment Middle cerebral artery Posterior cerebral artery and straight sinus Pericallosal arteries Middle cerebral artery Pericallosal arteries Visualization of hidden online CUSA resection Visualization of hidden online CUSA resection Visualization of hidden tailored approach Visualization of hidden vessel vessel, vessel, vessel vessel, vessel, vessel, vessel Visualization of hidden vessel, online CUSA resection Tailored approach, visualization of hidden vessel, online CUSA resection Visualization of hidden vessel, online CUSA resection Safe biopsy sampling, visualization of hidden vessel Tailored approach, visualization of hidden vessel Visualization of hidden vessel Visualization of hidden vessel, online CUSA resection Visualization of hidden vessel, online CUSA resection Facilitation of approach Visualization of hidden vessel, online CUSA resection Visualization of hidden vessel, online CUSA resection Visualization of hidden vessel Visualization of hidden vessel, facilitation of approach, online CUSA resection by the surgeon of the value of navigated ultrasound angiography was graded into 3 categories: helpful, possibly helpful, and not helpful. The degree of helpfulness was graded as helpful when navigated ultrasound angiography was used actively during the whole procedure and was considered to contribute significantly to the safe completion 586 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 of the procedure. The degree of helpfulness was considered possibly helpful when navigated ultrasound angiography was considered useful for orientation but not considered essential for safe navigation. When ultrasound angiography did not give additional valuable information for image guidance, it was deemed not helpful. The cases are listed in Table 1. 3. Results Between 2003 and May 2005, in 62 of 108 patients operated with image guidance, 3D ultrasound angiography was applied and evaluated. Among these patients, there were 20 meningiomas, 24 gliomas, 5 metastases, and 11 tumors of other types. There were also 2 frameless Fig. 3. Chordoma in cavernous sinus region (case 11). A: Sagittal MRI. Green arrows outline the tumor, red arrow directed at internal carotid artery. B: Coronal MRI. Green arrows outline the tumor, red arrows directed at internal carotid artery. C: Photograph from the microscope. During surgery, the internal carotid artery was revealed. White arrows directed at the internal carotid artery. D: MRI image slice from a snapshot from the navigation system during surgery. Green arrows directed at the tumor borders. E: Ultrasound angiography from the same snapshot as panel (D) visualizing the internal carotid artery (in shades of red), which in this case was adjacent to the medial parts of the tumor. The tumor in this case was hypoechogenic. Red arrow directed at artery. A flash artifact is visible upward to the right of the artery (white arrow). F: Photograph from the microscope. The pointer is directed at the internal carotid artery (indicated by white and black arrows). The snapshot from which panels (D) and (E) originate was taken about the same time this photograph was taken, demonstrating good accuracy of the ultrasound angiography navigation, as the position of the pointer directed at the artery seem to correspond with the position as indicated by the navigation system. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article). O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 stereotactic biopsies performed with the use of navigated 3D ultrasound angiography. 3D ultrasound angiography image data were successfully acquired and visualized in all cases. Both arteries and veins (Fig. 2) were visualized via ultrasound angiography. Vessels in range of the probe gave a strong Doppler signal (Fig. 2). In some cases, the simultaneous display of preoperative MRI helped to decide the identity of vessels, by giving anatomical overview, as the 3D ultrasound angiography only displayed the part of the anatomy that was covered in the 3D ultrasound image acquisition. Image guidance with 3D ultrasound angiography was performed with a pointer device or tracked surgical instruments, such as the tracked CUSA or a tracked biopsy forceps. With tracking of the CUSA, the surgeon was able to see the position of the tip of the CUSA in relation to the tumor capsule or margin, as well as the vessels of interest (Fig. 1). We found this to be a useful safety measure during resection when there were vessels adjacent to or inside the tumor. In some of the cases (cases 26, 31, 34, 39, 44, 49, and 50), the role of navigation with 3D ultrasound angiographic 587 imaging was to tailor an approach with minimal invasiveness. With angiographic image guidance, a safe route to the tumor could be chosen, with less need for dissecting and visually inspecting vessels. With tracking of the biopsy forceps, frameless stereotactic biopsies were acquired with image guidance. Angiographic imaging with ultrasound was found to be helpful in deciding sites for biopsy sampling with minimal risk of bleeding. We found that ultrasound angiography sufficiently displayed small vessels (about z 0.5 mm) in the brain tissue for safe image-guided biopsy sampling. One patient in our series (case 57) had an infarction of the middle cerebral artery, which passed close to a metastasis. Except this case, we have not recorded any vascular complications in our series. The cases where navigated ultrasound angiography was found to be helpful and the role of navigated ultrasound angiography in those cases are presented in Table 2. In 26 (42%) of the cases, ultrasound angiography was found to be helpful, in 10 (16%) possibly helpful, and in 26 (42%) not helpful. Fig. 4. Large olfactory meningioma (case 25). Green arrows outline tumor. A: CT image showing the tumor. B: DSA showing branches from the anterior cerebral artery being displaced by the tumor (white arrows). C: Ultrasound angiography showing the same branches as indicated in panel (B) from the anterior cerebral artery (white arrows) adjacent to the tumor. D: Ultrasound angiography before the resection. E: Ultrasound angiography during resection. F: Ultrasound angiography after tumor removal. The branches from the anterior cerebral artery are clearly visualized during these phases of the operation. Updated 3D ultrasound angiography image data ensure that accuracy is good even if brain shift has occurred. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article). 588 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 4. Case descriptions 4.1. Case 11 A 58-year-old woman complained of hoarseness and swallowing difficulties. A left recurrent laryngeal nerve paresis was found on clinical examination, and MRI investigation showed a tumor in the left middle cranial fossa and jugular fossa with close anatomical relation to the internal carotid artery (Fig. 3A and B). She was operated using a pterional approach. The tumor was found to be a chordoma by frozen sections, confirmed later by histopathological analysis. Before reaching the tumor, 3D ultrasound and 3D ultrasound angiography image data were acquired. During surgery, navigating with the pointer and the calibrated CUSA showed the position of the internal carotid artery in relation to the tip of the instrument. This was found valuable for safe resection because the artery was not visible in this phase of the operation. Further dissection revealed the carotid artery (Fig. 3C), and we could verify that the navigation system showed the correct position of the artery when directing the pointer at the artery (Fig. 3E and F). Using neuronavigation with ultrasound angiography images, surgery was performed without any complications, and the patient had an uneventful recovery without any new neurological deficits. 4.2. Case 25 A 52-year-old man presented with progressive dementia. A CT scan revealed a large interhemispheric tumor in the anterior cranial fossa (Fig. 4A). Preoperative DSA showed branches of the anterior cerebral artery passing close to the tumor (Fig. 4B). An image-guided extirpation of the tumor via the bifrontal approach was performed. During surgery, intraoperative 3D ultrasound angiography showed branches from the anterior cerebral arteries and their close relation to the tumor (Fig. 4C). Debulking and resection of the tumor were performed with tracked CUSA, which enabled monitoring of the CUSA tip in relation to the vessels, clearly revealed by 3D ultrasound angiography. Several 3D ultrasound angiography image acquisitions were performed during different stages of the operation (Fig. 4D-F), ensuring that inaccuracy due to brain shift was minimal. Neuronavigation with intraoperative ultrasound angiography thus served to increase the safety of the procedure. The postoperative recovery was uneventful. At follow-up, the patient had a significant improvement in cognitive function. 5. Discussion In image-guided surgery, angiographic imaging may be valuable in addition to tissue imaging. In image guidance of aneurysm and AVM surgery, preoperative angiographic image data, MR Angiography and CT Angiography, have successfully been imported and used in neuronavigation [2,13,20,21,27] and have been found to be valuable for localization of aneurysms and AVM feeders. In addition, in tumor surgery, preoperative angiographic image data (CT angiography and MR angiography) may be used in neuronavigation. Advanced neuronavigation with CT angiography has been found to be valuable for tailoring surgical approaches and identification of hidden vessels (ie, vessels not exposed in the surgical field) during resection [17]. However, in cases with brain shift, preoperative images may become inaccurate during surgery [3,7,16]. Intraoperative imaging, tissue and angiographic, may therefore be necessary to maintain the accuracy in neuronavigation. Both intraoperative MRI and ultrasound are modalities for acquiring intraoperative angiographic image data for use in neuronavigation. Intraoperative ultrasound integrated with neuronavigation is now a convenient alternative, as modern ultrasound Doppler technology is capable of depicting intracranial vasculature with sufficient image quality [5,6,19,24,30,31]. Sure et al [23] has reported land-marking of vessels in a conventional neuronavigation system with the help of a tracked 2D ultrasound probe using color flow Doppler and found that marking the position of vessels adjacent to a tumor facilitated image-guided tumor resection. The solution we have used enables navigation directly in the 3D ultrasound angiography images without need for landmarking on a preoperative MRI data set. In addition, updated 3D ultrasound angiographic image data may quickly be acquired and used in neuronavigation. Not surprisingly, we found navigated 3D ultrasound angiography to be of value during surgery on tumor cases with large and important vessels adjacent to or embedded in the tumor (Table 2). Five of these cases (patients 19, 25, 30, 32, and 40) were patients with large meningiomas arising from the skull base, displacing important arteries. With online resection using the CUSA, monitoring the position of the tip of the CUSA relative to the position of vessels helped the surgeon to do a more complete resection of the meningiomas before starting the dissection of the capsule, probably reducing the traumatization of the cerebral cortex during this phase. Updated 3D ultrasound angiography image data gave angiographic images of sufficient quality for navigation during the different stages of the resection. In addition, in 9 cases (patients 10, 13, 22, 26, 31, 34, 44, 48, and 49) with tumors in the medial part of the temporal lobe, we found that visualization of vessels in the basal cistern (eg, the great cerebral vein and the middle cerebral artery) was helpful for safer resection and biopsy sampling. In other cases, ultrasound angiography was not found helpful, such as in superficial gliomas not close to larger vessels. Nevertheless, even in such cases, unexpected vessels may give the surgeon unpleasant surprises, and confirmation that important vessels were not present close to the tumor was found reassuring by the surgeon. In skull base surgery, the vessels are usually attached to the skull base and are therefore not subject to brain shift [3,22]. Nevertheless, in MRI-based neuronavigation, inaccuracy in image registration may still lead to inaccurate O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 targeting of skull base vessels. As no image registration is needed in neuronavigation based on 3D ultrasound angiography [28], this technique is still valuable because image registration errors are avoided. One limitation of the 3D ultrasound–based neuronavigation is that covering the whole area of interest during 3D ultrasound image data acquisition can be difficult, for example, in cases with large tumors. Furthermore, in skull base surgery, the skull base itself may hinder ultrasound imaging of the entire tumor, for example, in cases where a large skull base tumor is invading both the middle and the posterior cranial fossa. Still, it is our experience that careful planning of the surgical approach and keeping in mind optimal positioning of the ultrasound probe may reduce such problems to a minimum. Moreover, simultaneous display of MRI and 3D ultrasound in neuronavigation may be helpful, giving overview and anatomical orientation in cases with large tumors that are difficult to cover entirely with ultrasound. 589 6. Technical considerations For angiographic imaging with the Doppler technique, both color flow and power Doppler technique may be used. Power Doppler displays the vessels because of the power of the Doppler signal in shades of red as an overlay on the ultrasound tissue image. In contrast to color flow Doppler imaging, power Doppler does not have flow velocity or direction information but is less angle-dependent, there is no aliasing and the modality is more flow sensitive [15,18]. Because of less angle dependence, vessel continuity is also better with power Doppler than with color flow Doppler, and vessels will be visualized even if the ultrasound beam hits the vessel at angles close to 908. In an application with free-hand 3D ultrasound image acquisition, the ultrasound beam will almost always hit a vessel with several different angles; therefore, there is minimal risk of missing a vessel when acquiring 3D ultrasound angiography image data. Fig. 5. Multimodal visualization. A: Glioma (same patient as in Fig 2B, case 38). Ultrasound image slice placed in a 3D scene with a segmented representation of the MR-imaged tumor (yellow). Ultrasound angiography (green) and MR angiography (red) show the vessels in relation to the tumor. The ultrasound angiography and MR angiography do not completely overlap, and this may be due to brain shift or registration error. The ultrasound angiography displays more vessels than the MR angiography, and the vessels appear thicker. Still, the vascular anatomy is recognized on both ultrasound angiography and MR angiography. The ultrasound angiography only shows parts of the anatomy. B: Meningioma (same patient as in Fig 2C, case 53). Ultrasound angiography (segmented) here shown in red. Image slices from T1 MRI and ultrasound give orientation and show the tumor. C: Foramen magnum meningioma (same patient as in Fig 2E, case 32). A volume rendered representation of the tumor is shown in grayscale, whereas MR angiography is shown in red and ultrasound angiography in green. The visualization in a 3D scene demonstrates the spatial relationships between the tumor and vessels. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article). 590 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 In contrast to MR angiography, ultrasound angiography shows both arteries and veins at the same time. We consider this an advantage, as one usually is interested in preserving both arteries and veins during surgery. It is however not possible to discriminate between arteries and veins using the power Doppler imaging technique alone. Still, adjusting the blow velocity rejectQ setting on the ultrasound scanner may be helpful for showing only vessels with high flow velocities, in which case, most veins will be bfiltered out.Q Using triplex display may also be of help to discriminate between arteries and veins by evaluating the Doppler spectrum. Using the ultrasound probe as a pointer will enable targeting of a vessel with neuronavigation while using triplex display, and in this way, it is possible to evaluate the targeted vessel with triplex. We did however use triplex in the present study. With larger vessels, it was often possible to decide whether the targeted vessel was an artery or a vein by using anatomical knowledge and comparing with arterial MR angiography in cases where this was available. In ultrasound angiography imaging, the ultrasound data are used for simultaneous imaging of both tissue and angiographic imaging. We have not experienced wherein the quality of the ultrasound tissue part of the image is significantly decreased on ultrasound angiography images. There are, however, some drawbacks of 3D ultrasound angiography with power Doppler. The high sensitivity of power Doppler may lead to the visualization of too many vessels, as small and surgically less important vessels are depicted. Sometimes, this may result in confusing images that are difficult to interpret. In our experience, reducing the gain setting on the ultrasound scanner may reduce this problem by filtering out the smaller vessels. For better orientation, simultaneous display of preoperative MRI and MR angiography is also useful in such cases. Blooming may also be a problem with power Doppler, as the power Doppler signal tends to expand beyond the vessel walls; consequently, the vessels may seem larger and thicker than they really are. In addition, the 2D ultrasound image plane itself has a certain thickness that varies with the depth; thus, the resolution in the elevation direction (normal to the scan plane) of the ultrasound beam affects the image quality in 3D ultrasound. Therefore, the vessels in reconstructed (ultrasound angiography) images tend to appear bulkier when viewed on slices orthogonal to the scan plane. Finally, power Doppler is relatively sensitive to flash artifacts, which occur as red fields or stripes (Fig. 2). Flash artifacts may, for example, occur when the probe is in a cavity filled with saline, when motion in the saline is detected by the power Doppler (Fig. 3C). This is another consequence of the high motion sensitivity of power Doppler, which can occur relatively often during free-hand 3D ultrasound angiography acquisition. Flash artifacts can be reduced with gentle movement of the probe during image acquisition and by adjustments on the ultrasound scanner, such as adjusting the lower limit for detectable flow. The overall clinical accuracy of the SonoWand system may be as low as 2 mm in a clinical setting when using intraoperative imaging to compensate for brain shift [11]. We did not systematically measure accuracy in this study. Nonetheless, in each case a vessel was exposed in the surgical field, we directed the pointer at the artery for accuracy testing, as in Fig. 3F. In each case, we observed that the position of the pointer corresponded with the position indicated by the navigation system, as shown in Fig. 3E. Although not a precise measurement of accuracy, we found these accuracy tests to indicate that the clinical accuracy of the navigation system was satisfactory. 7. Limitations of this study This technical evaluation study of navigated 3D ultrasound angiography demonstrates the straightforward application of neuronavigation with this imaging modality in 62 cases. To prove any benefit for the patient, a study with a different design is needed. The cases selected for evaluation of ultrasound angiography were cases wherein ultrasound-based neuronavigation was chosen in advance. Small, extra-axial tumors that were easy to localize were not included in this material because they were not operated with image guidance. In other tumor cases, which were operated with image guidance, ultrasound angiography was not used because the surgeon felt that it would not be of value. Most of the operations, and thus also the evaluation of the helpfulness of ultrasound angiography, were performed by one surgeon. Still, the 3 surgeons who participated seemed to judge the usefulness of 3D ultrasound angiography similarly. Nevertheless, different surgeons may judge the helpfulness of navigated ultrasound angiography differently, for example, some surgeons would not trust the navigation system enough to leave out dissection of important vessels. All this has to be kept in mind when interpreting the results of this study. 8. Future prospects Ultrasound angiography still has potential for significant improvements, and developments along several venues are to be expected: New signal processing methods such as BFI [12] may better visualize flow inside vessels and hopefully reduce the problem of blooming. Ultrasound contrast-enhancing agents have been reported to be helpful in tumor neurosurgery for assessing the vasculature close to and inside the tumor [8,14] and may improve the quality of ultrasound angiographic imaging. New multirow probes will have a more optimal beam shape with better resolution in the elevation direction (eg, normal to the scan plane). This will further improve the image quality of ultrasound angiography because elevation resolution is a limiting factor for image quality with the present technology. Real-time 3D probes will be able to O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 acquire 3D ultrasound data sets directly without free hand movement, and this will minimize flash artifacts. New multimodal visualization techniques where preoperative MR and intraoperative ultrasound are integrated in the same 3D scene are already available [10]. This may probably further enhance the surgeon’s perception of anatomic and spatial relationships between tumor and adjacent vessels (Fig. 5). Diffusion tensor imaging of tracts and FMRI of eloquent cortex may also be visualized. Furthermore, robust volume-to-volume registration techniques for registration of preoperative MR angiography data to intraoperative ultrasound angiography data may make it possible to adjust preoperative MR image data in cases of brain shift. 9. Conclusions On the basis of our experience with navigated intraoperative ultrasound angiography, we believe that this imaging modality is sufficient for intraoperative imaging of arteries and veins. Image guidance with 3D ultrasound angiography may increase the safety in surgery of tumor cases where an important vessel is adjacent to or inside a tumor. It may facilitate tailored approaches, and frameless stereotactic biopsies may be performed with increased safety. Furthermore, with updated 3D ultrasound angiography imaging, accuracy of angio-based neuronavigation may be maintained in cases of brain shift. Acknowledgments This work was supported by the Norwegian Research Council. References [1] Albayrak B, Samdani AF, Black PM. Intra-operative magnetic resonance imaging in neurosurgery. Acta Neurochir (Wien) 2004;146:543 - 56 [discussion 57]. [2] Coenen VA, Dammert S, Reinges MH, et al. Image-guided microneurosurgical management of small cerebral arteriovenous malformations: the value of navigated computed tomographic angiography. Neuroradiology 2005;47:66 - 72. [3] Dorward NL, Alberti O, Velani B, et al. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg 1998;88:656 - 62. [4] Gronningsaeter A, Kleven A, Ommedal S, et al. SonoWand, an ultrasound-based neuronavigation system. Neurosurgery 2000; 47:1373 - 9 [discussion 79-80]. [5] Gronningsaeter A, Unsgard G, Ommedal S, et al. Ultrasound-guided neurosurgery: a feasibility study in the 3-30 MHz frequency range. Br J Neurosurg 1996;10:161 - 8. [6] Hernes TA, Ommedal S, Lie T, et al. Stereoscopic navigationcontrolled display of preoperative MRI and intraoperative 3D ultrasound in planning and guidance of neurosurgery: new technology for minimally invasive image-guided surgery approaches. Minim Invasive Neurosurg 2003;46:129 - 37. [7] Hill DL, Maurer Jr CR, Maciunas RJ, et al. Measurement of intraoperative brain surface deformation under a craniotomy. Neurosurgery 1998;43:514 - 26 [discussion 27-28]. 591 [8] Kanno H, Ozawa Y, Sakata K, et al. Intraoperative power Doppler ultrasonography with a contrast-enhancing agent for intracranial tumors. J Neurosurg 2005;102:295 - 301. [9] Letteboer MM, Willems PW, Viergever MA, et al. Brain shift estimation in image-guided neurosurgery using 3-D ultrasound. IEEE Trans Biomed Eng 2005;52:268 - 76. [10] Lindseth F, Kaspersen JH, Ommedal S, et al. Multimodal image fusion in ultrasound-based neuronavigation: improving overview and interpretation by integrating preoperative MRI with intraoperative 3D ultrasound. Comput Aided Surg 2003;8:49 - 69. [11] Lindseth F, Lango T, Bang J, et al. Accuracy evaluation of a 3D ultrasound-based neuronavigation system. Comput Aided Surg 2002;7:197 - 222. [12] Loevstakken L, Bjaerum S,Martens D, et al. Real-time blood motion imaging a 2D blood flow visualization technique. IEEE International Ultrasonics symposium, Montreal, Canada, Aug. 24-27, 2004. 1: 602. [13] Muacevic A, Steiger HJ. Computer-assisted resection of cerebral arteriovenous malformations. Neurosurgery 1999;45:1164 - 70 [discussion 70-1]. [14] Otsuki H, Nakatani S, Yamasaki M, et al. Intraoperative ultrasound arteriography with the bCoded Harmonic AngioQ technique. Report of three cases. J Neurosurg 2001;94:992 - 5. [15] Riccabona M, Resch B, Eder HG, et al. Clinical value of amplitudecoded colour Doppler sonography in paediatric neurosonography. Childs Nerv Syst 2002;18:663 - 9. [16] Roberts DW, Hartov A, Kennedy FE, et al. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 1998;43:749 - 58 [discussion 58-60]. [17] Rohde V, Spangenberg P, Mayfrank L, et al. Advanced neuronavigation in skull base tumors and vascular lesions. Minim Invasive Neurosurg 2005;48:13 - 8. [18] Rubin JM. Power Doppler. Eur Radiol 1999;9(Suppl 3):S318-22. [19] Rubin JM, Hatfield MK, Chandler WF, et al. Intracerebral arteriovenous malformations: intraoperative color Doppler flow imaging. Radiology 1989;170:219 - 22. [20] Russell SM, Woo HH, Joseffer SS, et al. Role of frameless stereotaxy in the surgical treatment of cerebral arteriovenous malformations: technique and outcomes in a controlled study of 44 consecutive patients. Neurosurgery 2002;51:1108 - 16 [discussion 16-18]. [21] Schmid-Elsaesser R, Muacevic A, Holtmannspotter M, et al. Neuronavigation based on CT angiography for surgery of intracranial aneurysms: primary experience with unruptured aneurysms. Minim Invasive Neurosurg 2003;46:269 - 77. [22] Sure U, Alberti O, Petermeyer M, et al. Advanced image-guided skull base surgery. Surg Neurol 2000;53:563 - 72 [discussion 72]. [23] Sure U, Benes L, Bozinov O, et al. Intraoperative landmarking of vascular anatomy by integration of duplex and Doppler ultrasonography in image-guided surgery. Technical note. Surg Neurol 2005; 63:133 - 41. [24] Tekula F, Pritz MB, Kopecky K, et al. Usefulness of color Doppler ultrasound in the management of a spinal arteriovenous fistula. Surg Neurol 2001;56:304 - 7. [25] Unsgaard G, Gronningsaeter A, Ommedal S, et al. Brain operations guided by real-time two-dimensional ultrasound: new possibilities as a result of improved image quality. Neurosurgery 2002;51:402 - 11 [discussion 11-12]. [26] Unsgaard G, Ommedal S, Muller T, et al. Neuronavigation by intraoperative three-dimensional ultrasound: initial experience during brain tumor resection. Neurosurgery 2002;50:804 - 12 [discussion 12]. [27] Unsgaard G, Ommedal S, Rygh O, et al. Operation of arteriovenous malformations assisted by stereoscopic navigation-controlled display of preoperative magnetic resonance angiography and intraoperative ultrasound angiography. Neurosurgery 2005;56:281 - 90. [28] Unsgaard G, Rygh OM, Selbekk T, et al. Intra-operative 3D ultrasound in neurosurgery. Acta Neurochir (Wien) 2006;148: 235 - 53. 592 O.M. Rygh et al. / Surgical Neurology 66 (2006) 581–592 [29] Wirtz CR, Knauth M, Staubert A, et al. Clinical evaluation and follow-up results for intraoperative magnetic resonance imaging in neurosurgery. Neurosurgery 2000;46:1112 - 20 [discussion 20-2]. [30] Woydt M, Greiner K, Perez J, et al. Intraoperative color duplex sonography of basal arteries during aneurysm surgery. J Neuroimaging 1997;7:203 - 7. [31] Woydt M, Perez J, Meixensberger J, et al. Intra-operative colourduplex-sonography in the surgical management of cerebral AVmalformations. Acta Neurochir (Wien) 1998;140:689 - 98. Commentary In the past several years, intraoperative imaging has gained increasing importance. With the newest medicotechnical achievements such as high-field (1.5–3 T) MRI and multislice CT, a certain renaissance of intraoperative imaging has started. However, the setting up of such diagnostic tool requires a major financial background. Consequently, Unsgaard et al started to promote the intraoperative navigated ultrasound technology by introducing a 3D ultrasound navigation system. In previous publications, this group described a very nice system that is capable of producing a 3D imaging ultrasound data set that is used for navigation and can be supported by the display of preoperatively acquired corresponding MRI images. In their article, the authors demonstrate the integration of power Doppler (Duplex or color mode) signals in their 3D navigation data set. The figures, as shown by the authors, are of very high quality; let us hope that the drawbacks of the intraoperative ultrasound imaging of vessels, such as the exaggeration of the caliper (blooming), can be reduced in future. Frequently, for us neurosurgeons, the vascular anatomy is the most interesting part to recognize in a navigational data set. Therefore, we believe that particularly the 3D visualization of the vascular anatomy, as presented here, is of major benefit during surgery of a broad variety of intracranial pathologies. In our experience, such technology is of particular benefit when vessels are encased or distorted by tumors or when evaluation of hemodynamics within a complex AVM is intended [1]. The usefulness of a similar technique for treatment of vascular lesions has also been previously illustrated by the same group [2]. One of the most interesting questions to answer in the future will be whether the ultrasound technology will allow to evaluate the neck of a clipped aneurysm to exclude a remnant. However, alongside the economic aspect, acquisition time and quality of intraoperative visualization of the vascular anatomy are further advantages of the ultrasound technology when compared with the MRI and/or CT intraoperative imaging solutions. Dorothea Miller, MD Oliver Bozinov, MD Ulrich Sure, MD Clinic for Neurosurgery, University Hospital Philipps-University Marburg Marburg, 35033 Germany References [1] Tirakotai W, Miller D, Heinze S, Benes L, Bertalanffy H, Sure U. A novel platform for image-guided ultrasound. Neurosurgery 2006;58: 710 - 8. [2] Unsgaard G, Ommedal S, Rygh OM, Lindseth F. Operation of arteriovenous malformations assisted by stereoscopic navigationcontrolled display of preoperative magnetic resonance angiography and intraoperative ultrasound angiography. Neurosurgery 2005; 56(2 Suppl):281 - 90.