Electromagnetic Navigation System for CT-Guided Biopsy of Small Lesions

Transcription

1 Vascular and Interventional Radiology Original Research Vascular and Interventional Radiology Original Research Liat Appelbaum 1 Jacob Sosna Yizhak Nissenbaum Alexander Benshtein S. Nahum Goldberg Appelbaum L, Sosna J, Nissenbaum Y, Benshtein A, Goldberg SN Keywords: CT-guided biopsy, electromagnetic navigation system, phantom study DOI: /AJR Received June 13, 2010; accepted after revision August 26, Support for this study was provided by Veran Medical Technologies (St. Louis, MO). The authors are not employees of the company and had complete and independent control of the study, data gathering, data analysis, and manuscript preparation. 1 All authors: Department of Radiology, Hadassah- Hebrew University Medical Center, POB 91120, Jerusalem, Israel Address correspondence to L. Appelbaum (liata@hadassah.org.il). AJR 2011; 196: X/11/ American Roentgen Ray Society Electromagnetic Navigation System for CT-Guided Biopsy of Small Lesions OBJECTIVE. The purpose of this study was to evaluate an electromagnetic navigation system for CT-guided biopsy of small lesions. MATERIALS AND METHODS. Standardized CT anthropomorphic phantoms were biopsied by two attending radiologists. CT scans of the phantom and surface electromagnetic fiducial markers were imported into the memory of the 3D electromagnetic navigation system. Each radiologist assessed the accuracy of biopsy using electromagnetic navigation alone by targeting sets of nine lesions (size range, 8 14 mm; skin to target distance, cm) under eight different conditions of detector field strength and orientation (n = 117). As a control, each radiologist also biopsied two sets of five targets using conventional CT-guided technique. Biopsy accuracy, number of needle passes, procedure time, and radiation dose were compared. RESULTS. Under optimal conditions (phantom perpendicular to the electromagnetic receiver at highest possible field strength), phantom accuracy to the center of the lesion was 2.6 ± 1.1 mm. This translated into hitting 84.4% (38/45) of targets in a single pass (1.1 ± 0.4 CT confirmations), which was significantly fewer than the 3.6 ± 1.3 CT checks required for conventional technique (p < 0.001). The mean targeting time was 38.8 ± 18.2 seconds per lesion. Including procedural planning (~5.5 minutes) and final CT confirmation of placement (~3.5 minutes), the full electromagnetic tracking procedure required significantly less time (551.6 ± 87.4 seconds [~9 minutes]) than conventional CT (833.3 ± seconds [~14 minutes]) for successful targeting (p < 0.001). Less favorable conditions, including nonperpendicular relation between the axis of the machine and weaker field strength, resulted in statistically significant lower accuracy (3.7 ± 1 mm, p < 0.001). Nevertheless, first-pass biopsy accuracy was 58.3% (21/36) and second-pass (35/36) accuracy was 97.2%. Lesions farther from the skin than cm were out of range for successful electromagnetic tracking. CONCLUSION. Virtual electro mag netic tracking appears to have high accuracy in needle placement, potentially reducing time and radiation exposure compared with those of conventional CT techniques in the biopsy of small lesions. P recise device placement and positioning are the basis of every successful imaging-guided biopsy, tumor ablation, and most other interventional radiology procedures [1]. This process usually is undertaken with real-time imaging guidance (CT, ultrasound, or fluoroscopy) and more often than not is heavily dependent on operator experience. Precise placement and positioning, however, often are not possible owing to radiation burden and limitations in visualization. Furthermore, information not presented on the realtime guiding screen, such as the relation between a nonenhancing lesion (or poorly visible lesion without contrast administration) and anatomic landmarks, or in the case of ablation regions of unablated tumor, often is needed for optimal lesion depiction. Thus mental registration is needed between the anatomic information acquired with offline modalities (such as contrast-enhanced CT, MRI, and PET/CT) and the guiding modality findings. As procedures become more complex, synthesis of the requisite information becomes more challenging. Various devices have been developed to assist in this spatial navigation. Electromagnetic tracking is one potential method of acquiring spatial navigation information in real time that is based on reconstruction of and interaction with datasets of previously acquired images [2 12]. This method can enable accurate positioning of therapeu AJR:196, May 2011

2 tic and monitoring devices with an electromagnetic navigation system alone, without continual imaging of the patient. In addition, preoperative plans and images from other modalities, such as PET, contrast-enhanced CT, and MRI, can function as the source data and therefore in essence be combined with realtime imaging modalities such as fluoroscopy and unenhanced CT. Because of the perceived benefits, several electromagnetic and optical tracking and navigation devices are in research and development, and some are available for clinical use. In spite of the importance of establishing key endpoints of accuracy and precision for these devices, to the best of our knowledge, little has been published regarding this key issue. In this study, we evaluated the feasibility and efficacy of an electromagnetic navigation system for CT-guided procedures. We performed the study using a well-defined experimental system that can be readily translated to evaluation of all competitive systems. A C Materials and Methods Experimental System Biopsy navigation system We used a U.S. Food and Drug Administration approved electromagnetic biopsy navigation system (Veran IG4, Veran Medical Technologies) (Fig. 1) that has full 4D visualization capability (i.e., 3D resolution in real time). The centerpiece and technologic platform is an electromagnetic field generator system that interacts with both CT and MRI datasets and positional sensors on the patient and at the tip of a needle or other applicator. To create the virtual reference work space within a reference 3D imaging dataset, the system identifies six fiducial markers (i.e., sensors) in the vicinity of the target region of interest. For CT datasets, these markers are tiny metallic (0.5 8 mm) rods embedded in three pads placed on the surface of the patient: four markers in one pad and one each in two accessory pads that come into alignment at specified intervals of respiration. The fiducial markers are identified and automatically segmented from a thin-slice helical CT dataset obtained before the procedure. The electromagnetic field generator Fig. 1 Electromagnetic biopsy navigation system. A, Photograph shows electromagnetic detector (solid arrow), screen and computer (white dashed arrow), and phantom on CT table (black dashed arrow). B, Photograph shows metallic rods embedded in three pads placed on surface of phantom to serve as fiducial markers for system. C, Photograph shows needle and sensor in target (arrow) within phantom. D, Photograph shows needle and sensor components in detail. B D AJR:196, May

3 Fig. 2 Depiction of electromagnetic signal strength with color target circles. Black circle (inner 15 cm of electromagnetic field) represents highest field strength; concentric outer yellow (30 cm) and red zones (50 cm) represent progressively lower field strengths. Distance slider in lower part of screen works in conjunction with bull s-eye target for positioning of electromagnetic generator at correct distance from target. A, Diagram shows cross-hairs directly centered, indicating maximum electromagnetic strength. B, Diagram shows cross-hairs on yellow band, denoting reduction in electromagnetic strength from 80 to 9 A/m. provides a 50-cm cubic volume for navigation, enabling definition of a 25-cm side-to-side region of space centered and under the four main fiducial markers in which a moving sensor device placed coaxially at the tip of the needle can be tracked and displayed along its insertion path within the reconstructed images. To facilitate visualization, the electromagnetic system reconstructs and displays the CT dataset according to operator preferences in axial, sagittal, coronal, and oblique true needle path planes. To ensure optimal tracking, the electromagnetic generator informs the operator about the overall field strength of electromagnetic detection in two distinct ways on the screen. The first is color target circles. An inner black circle presents the highest field strength, and concentric outer yellow and red zones represent progressively lower field strengths of the 30- and 50-cm rings of the electromagnetic generator field. The other on-screen notification is a distance slider that works in conjunction with the bull seye target to position the electromagnetic generator at the correct distance from the target (Fig. 2). The optimal distance between the fiducial marker and the face of the electromagnetic generator is cm. The electromagnetic sensor operates with the voltage induced by the electromagnetic flux field, which depends on the distance between the sensor and the electromagnetic generator in an inverse cubic relation. Therefore, moving the main fiducial pad 10 cm from the face of the electromagnetic generator decreases the electromagnetic field strength 30 times (6000 to 200 A/m). In addition, an out-of-field or out-of-volume message flashes on the screen when the tracker needle is more than 25 cm from the fiducial markers. Phantom For these experiments, we used standardized CT abdominal biopsy phantoms (model 071, CIRS) consisting of 12 randomly positioned small radiodense target lesions (diameter, 8 14 mm; mean, 9.5 ± 2.2 [SD] mm) embedded in solid polymer gel (Fig. 3). The phantom also includes A ribs and a simulated spine. The skin to target distance was cm (mean, 8.8 ± 2.3 cm). An opaque screen was placed over the side of the phantom to ensure that visual identification of the needle or target could not be performed while the biopsy specimen was obtained. B C Fig. 3 Biopsy accuracy measurements in phantom. A, Screen of electromagnetic navigation device shows position of needle path in both axial and oblique projections (i.e., simulating ultrasound appearance). Needle tip is depicted as green cross-hair and is in center of lesion. Gas bubbles represent previous needle insertions. B, CT scan shows tip of Chiba needle in lesion (arrow). C, Photograph obtained at visual inspection of phantom shows needle tips were embedded in targeted lesions (arrows). B A 1196 AJR:196, May 2011

4 Data acquisition and biopsy The phantom was positioned and stabilized on the CT table. Fiducial pads were placed on its surface. The main pad was placed at the center of the top of the phantom, and the auxiliary pads were placed cm laterally, except where otherwise noted (Fig. 1). Baseline MDCT scans (Brilliance version scanner, Philips Healthcare) of the biopsy phantom with electromagnetic fiducial markers were acquired and imported into the 3D electromagnetic navigation system memory. CT parameters were collimation, mm; slice thickness, 1.0 mm; slice increment, 1.0 mm; pitch, 0.72; 140 kv; and 200 mas/slice. Targets were biopsied by two attending radiologists (6 and 12 years of experience) using 15-cm 20-gauge Chiba needles. Each needle contained a stylet that held the electromagnetic sensor in its tip (Fig. 1D). Each radiologist assessed the accuracy of biopsy using the electromagnetic navigation system alone by targeting lesions under different conditions of detector field strength and orientation of the detector to the targets. For all experiments, signal strength was in the green band of the distance slider on the screen. In addition, the phantom and electromagnetic detection device were placed at least 1 m from the CT gantry to eliminate potential interference of the metallic gantry with the magnetic field. Parameters Studied The following eight parameters were assessed (Table 1). 1. CT acquisition with optimally aligned phantom Phantoms were placed generally perpendicular (i.e., ± 10 ) to the electromagnetic receiver at the highest possible field strength (cross-hairs at the center of the black center of the bull s-eye [Fig. 1]), usually within cm of the fiducial marker and phantom skin surface). This test case was considered the most restrictive for assessing the accuracy of the optimal position CT The phantom was scanned at a 45 angle to a electromagnetic receiver that was perpendicular (highest field strength) to the angled phantom. This step was used to determine the effect of nonperpendicular CT acquisition on fiducial identification and reconstruction of the virtual workspace generator The electromagnetic receiver was placed at a 30 angle to the phantom, and CT images of the phantom were acquired in standard perpendicular alignment. This step enabled determination of the effect of nonperpendicular interaction between the electromagnetic generator and the fiducial markers. 4. Lower field strength The phantom was placed perpendicular, but the electromagnetic receiver was placed cm from the phantom to induce lower field strength. Cross-hairs were placed at the yellow TABLE 1: Biopsy Accuracy Under Differing Conditions Parameter No. Parameter No. of Sets Average (mm) SD 1 Exactly perpendicular CT generator Lower field strength with lower field strength Double phantom Triple phantom 1 8 Separate pads a Note Dash ( ) indicates electromagnetic navagation was not possible. a Lesions that could be targeted. band interface (Fig. 2B), which reduced electromagnetic field strength approximately 9 times (from 80 to 9 A/m) for the main fiducial pad. This step was used to determine the effect of lower field strength on accuracy and outcome with lower field strength A combination of CT of the phantom at 45 and tracking and targeting at the lower, yellow-band field strength was used to determine whether potential degradation of accuracy (from experimental parameters 2 and 4) is additive. 6. Double phantom Two phantom studies were conducted to determine accuracy in the z- plane. Phantoms were positioned next to each other, and CT of both phantoms, and electromagnetic navigation was performed according to parameter 1. Fiducial markers were placed on one phantom, TABLE 2: Number of Passes and First-Pass Accuracy Parameter No. and targets were biopsied off the z-axis in the second phantom. 7. Triple phantom The z-axis was extended beyond cm by generation of a CT dataset from three phantoms affixed to one another; otherwise parameters 1 and 6 were used. 8. Separate pads The surface pads were separated from 1.0 cm apart to cm apart for evaluation of the effect of fiducial positioning on system accuracy; otherwise parameter 1 was used. Study Procedure Each radiologist performed biopsies on sets of nine lesions. Radiologist A performed seven sets (n = 63), and radiologist B performed six sets (n = 54) for a total of 117 simulated biopsies (Table 1). Each needle insertion was timed, and at the end of TABLE 3: Comparison of Conventional CT and Electromagnetic Guidance Technique Parameter Preparation Time (min) No. of Targets Targeting Time (s) No. of Passes Total Average First-Pass Percentage 1 Exactly perpendicular CT generator Lower field strength with lower field strength Double phantom Triple phantom 9 8 Separate pads a Note Dash ( ) indicates electromagnetic navagation was not possible. a Lesions that could be targeted. No. of Passes Average SD Average SD Total Radiation Dose (mgy) Electromagnetic guidance Conventional CT AJR:196, May

5 each set of biopsies, actual targeting was assessed with CT and visual inspection of the phantom by a second blinded observer to determine whether the needle tip was positioned in the target. At the conclusion of each biopsy set, the phantom was scanned with conventional CT with the needles in place to measure the distance between the needle tip and the target center of each target. If the needle was determined not to be within the lesion target, second and subsequent passes were attempted until correct positioning was attained. Comparison With Conventional Biopsy Technique In an additional experiment, each radiologist also biopsied a set of five lesion targets using identical 15-cm 20-gauge Chiba needles and conventional CT-guided technique without the navigation system. The phantom was scanned, the surface was marked, and the biopsy needle was introduced through the surface in the angle planned by the radiologist according to the prebiopsy scan. Additional conventional images were obtained at the radiologist s discretion to assist in accurate positioning. This step necessitated that the radiologist and technician repeatedly leave the CT suite while the phantom was sequentially scanned. After reviewing the images, the radiologist advanced the needle according to the new scan findings and corrected the angle as necessary. This process was repeated until CT showed that correct targeting was achieved. Study Endpoints The following endpoints were measured for each of the eight electromagnetic navigation and CT parameters. By definition, only the last three endpoints were measured for the conventional CT biopsy technique. Biopsy accuracy Biopsy accuracy was defined as the distance between needle tip and target center measured at CT after the electromagnetic navigation device showed the needle tip was at the lesion center (Fig. 3). Number of passes Number of passes was defined as the number of passes required to achieve successful positioning of the needle for sufficient core biopsy. Procedure time The time from the beginning of the biopsy procedure (placement of the needle on the phantom surface before puncture of the phantom) to final needle positioning (decided by the radiologist according to the data on the screen) was recorded for each lesion. In addition, the total time, including preparatory and confirmatory CT, was measured for each targeting set for estimation of total time on a lesion by lesion basis. Radiation exposure Total radiation exposure included the dose of the initial planning scan and all subsequent sets of localized CT scans that showed the needle course and tip position, including the final confirmation study. Exposure was calculated per lesion and expressed as the doselength product in milligrays. Statistical Analysis Several forms of statistical analysis were used. Mean value and SD were calculated for all four endpoints. Comparisons of endpoints in different parameter sets were performed by multivariate analysis of variance with multiple comparisons or paired Student t test as appropriate. SAS version 9.2 (SAS) and Excel (Microsoft) statistical software packages were used. Results Electromagnetic tracking for biopsy of small phantom targets under the eight parameters was successfully performed. The results are summarized in Table 1. Biopsy Accuracy Under optimal conditions (parameter 1, all targets and sensors within 20 cm of a perpendicular detector set to maximum calibration), phantom accuracy for the center of the lesion was 2.6 ± 1.1 mm (Table 1). Similar results were achieved in performance of biopsy up to cm along the z-axis (parameter 6) and in moving the surface pads apart (parameter 8) if the targets remained within the 20-cm target zone. Scanning the phantom with CT at a 45 angle (parameter 2) and placing the electromagnetic receiver at a 30 angle in relation to the phantom (parameter 3) had accuracies of 3.5 ± 0.8 mm and 4.1 ± 1.2 mm (p < 0.01 for each comparison with the optimal group). Decreasing field strength by placing the electromagnetic receiver at a distance of cm with the phantom either generally perpendicular (parameter 4) or at 45 (parameter 5) had accuracies of 3.7 ± 1.4 mm and 4.0 ± 1.0 mm. An error message (out of field or out of volume) appeared for further extension of the z-axis (i.e., with use of a triple phantom, parameter 7) and during attempts to biopsy lesions more than 25 cm from all pads when these were placed far apart (parameter 8), rendering electromagnetic navigation impossible for these conditions. Number of Passes and First-Pass Accuracy The average number of passes required was minimum under the optimal conditions of parameter 1 and in the double phantom model (parameter 6), both averaging passes (Table 2). Under both of these conditions, the percentage of biopsies successful in a single pass was %. Separating the pads led to an average of 1.2 passes with 78% success on the first pass. Firstpass success decreased in all the other conditions described with an average of passes and a 58.3% (21/36) rate of first-pass success. Nevertheless, navigation was sufficiently accurate to achieve a 97.2% (35/36) second-pass success rate. All lesions were successfully targeted in three passes. Procedure Time The total anticipated targeting time from the beginning of the initial CT acquisition to completion of the confirmatory CT scan was estimated at less then 10 minutes. This time comprised 3 minutes for the initial planning CT scan, 30 seconds for fiducial pad placement, 1 minute of CT data transfer to the navigation device, 1 minute for calibration of the navigation device, 38.8 ± 19.4 seconds of actual measured biopsy time, and 3 minutes 36 seconds of posttargeting imaging (3 minutes of CT time 1.2 passes). The mean targeting time per lesion was not significantly different between the optimal and nonoptimal conditions (Table 3). Comparison With Conventional CT Biopsy Technique All lesions were successfully targeted with conventional CT (Table 3). The procedure, however, took an average of 3.6 passes, which was statistically significantly greater than the 1.2 passes for the electromagnetic navigation system (p < 0.001). Furthermore, the electromagnetic guidance procedure was much shorter than the conventional CT procedure. With electromagnetic guidance, preparation time was 7.5 minutes and targeting time 38.8 ± 18.2 seconds. On the other hand, for the conventional CT technique, although CT preparation time was only 3 minutes, targeting time (including the additional scanning) was dramatically greater at seconds (13.9 minutes) (p < 0.001). Therefore, time was reduced from to 9.52 minutes (51.2%). The total dose-length product for electromagnetic navigation biopsy was 219 mgy and for the conventional technique was mgy, translating into an approximately 42.3% reduction in radiation dose. Discussion Minimally invasive procedures guided by cross-sectional imaging have become firmly established in daily practice because they are increasingly offered with either diagnostic or 1198 AJR:196, May 2011

6 therapeutic intent [13, 14]. Ultrasound and CT are the most common tools for imaging guidance, each having inherent advantages and disadvantages. Combining conventional ultrasound or CT technology with electromagnetic tracking may be beneficial to the performance of imaging-guided procedures. Specifically, electromagnetic tracking can be used to generate a real-time display of the position and orientation of an otherwise poorly visualized inserted instrument or a real-time combination of previously acquired images to gain additional anatomic and functional information. For example, navigation based on CT data acquired during distinct phases of contrast enhancement may help in the targeting of lesions visible only during certain phases of perfusion. Electromagnetic tracking also can be used with all relevant imaging modalities so that imaging modalities can be combined to advantage while the influence of their individual limitations is minimized. Theoretically, with electromagnetic guidance, the steep angulation approach of ultrasound can be combined with the avoidance of bone and air interference afforded by CT guidance. Furthermore, the use of a tracking device can improve accuracy, save time, and for CT, potentially decrease radiation exposure. Despite the potential benefits, it is important to emphasize that results in optimal settings, such as the engineering bench [2], are unlikely to be duplicated in some patients because of the inherent limitations of daily clinical practice. For electromagnetic imaging guidance, beyond the often considered respiratory motion [15], optimal field strength may require positioning of the detector or fiducial markers too close to the target to enable straightforward performance of the procedure. For example, a close detector can restrict the space needed for optimal flexibility when a probe or biopsy needle is moved, jeopardizing the sterility of the procedure. These devices must be evaluated under conditions that realistically simulate clinical practice. Our results show that less than 3-mm accuracy can be achieved under optimal conditions with at least one electromagnetic guidance device when high-quality CT datasets are acquired from a standardized phantom. We found that these conditions include placing the phantom generally perpendicular to the electromagnetic receiver with the electromagnetic signal positioned to maximum field strength (i.e., having the electromagnetic receiver positioned within cm of the fiducial marker and phantom skin surface). Optimal accuracy can be achieved when the pads are placed outside the direct path, that is, along the z-axis on the skin up to cm from the target. This extended flexibility facilitates freedom of movement when deeper lesions require steep angulation of the biopsy needle. Because procedures on live patients do not always approximate the rigid conditions achievable in biopsy of a phantom, accuracy was evaluated under less than optimal conditions. These clinically achievable conditions in which, for example, patient habitus may preclude achieving ideal electromagnetic fiducial spatial geometry include the following: nonperpendicular CT acquisition (testing the system up to 45 nonperpendicular interaction between the electromagnetic generator and the fiducial markers), increasing the distance between the electromagnetic receiver and the phantom (which resulted in lower electromagnetic field strength), and a combination of these factors. Although we observed some degradation of accuracy, we found acceptable first-pass accuracy (< 4 mm) and 97% successful targeting of the lesion in two passes (an initial pass and one repositioning). Because the window for access can be quite limited under real-life conditions, we acknowledged that the fiducial pads of the electromagnetic system can get in the way and even compromise the sterility of a clinical procedure. Therefore, we examined the possibility of altering fiducial positioning by increasing the distance between the surface pads from 1 cm to cm, and we found two possible outcomes. Accuracy was preserved when lesions remained within 25 cm of all pads, and out-of-field error messages were generated for lesions that would have been in field (i.e., within 25 cm of all pads) when pads were spaced closer together. In other words, separating the three pads reduced the effective region of tracking because more distant lesions became out of field. Therefore, for the platform assessed, clinicians should balance the need to have the pads out of the sterile field with the geometry and relations of the fiducial markers and the target location and path to ensure that all are contained within a defined 25-cm distance. Our results are compatible in terms of accuracy with those obtained with at least one other electromagnetic tracking system for which registration error and tracking error were less than 5 mm in phantom and a pig models [2]. In a later phase, a basic tracking error of 5.8 ± 2.6 mm which improved to 3.5 ± 1.9 mm with nonrigid registrations in which previous internal needle positions were used as additional fiducial markers was seen in 19 patients [2]. The many nonoptimal scenarios of our study that produced 3.6-mm accuracy therefore were similar to the results of Wood et al. [2]. Results of additional early clinical studies of electromagnetic guidance are being reported, including results on a heterogeneous sample of body areas and target sizes [16 18]. Additional results are needed, and they must be compared with results obtained with optical navigation systems, the advantages of which may include MRI guidance (which electromagnetic systems cannot enable). Because not all electromagnetic tracking systems are based on the same technology, comparison of this device with other electromagnetic and optical tracking navigation systems is needed to determine which features are most useful in different clinical scenarios. In addition to finding the electromagnetic guidance system accurate, we found that the use of this navigation system has potential to save time compared with conventional CT biopsy technique. Much of this time saving can be explained by the reduced number of steps. Instead of being advanced incrementally and checked with interval scanning, the needle or probe can be introduced directly and a confirmation scan obtained only after the needle is in position. Another important aspect of our study is the potential to reduce radiation dose. The reduction in the number of CT confirmations of tip position from 3.6 to 1 rendered a radiation dose reduction from to 219 mgy (42.3%), all of which might be considered additional CT radiation exposure. Our results show that because use of the navigation system can avoid recurrent scanning, the radiation dose to which a patient is exposed is reduced. Our study had several limitations. The issue of motion, which is particularly relevant to imaging-guided procedures, was not addressed. Motion in terms of both respiration and patient movement is likely to further degrade the results. It is important to note, however, the value of defining baseline best-case scenarios likely to be representative of procedures on nonmoving organs before addressing this issue in anticipated in vivo studies. Further limitations of the study included the use of only one type of needle and one procedure type. Future studies therefore will be expanded to include other needle types and AJR:196, May

7 diameters and procedures other than biopsy, including ablation. However, because the underlying robustness of the imaging-guided navigation technology represents the basis of the accuracy of all these procedures, we anticipate similar results for a wide range of devices. Another aspect that will have to be studied is the ideal and acceptable parameters of the initial CT (or other imaging modality) data sets. In this study, we selected a baseline dataset of high diagnostic quality because a major aim was to evaluate the targeting aspects of the system on the basis of optimal imaging datasets. Nevertheless, further study may establish equal accuracy with a reduction in baseline CT resolution and radiation dose. Conclusion Our results show that at least one virtual electromagnetic tracking device can be used with a high degree of accuracy in needle placement and a reduction in time and radiation exposure for biopsy of small lesions compared with conventional CT modalities. Future research to assess the clinical scenarios in which these navigation devices have greatest utility, including small lesions, lesions that are difficult to see, and lesions in difficult positions is warranted. References 1. Goldberg SN, Grassi CJ, Cardella JF, et al. Imageguided tumor ablation: standardization of terminology and reporting criteria. Society of Interventional Radiology Technology Assessment Committee, International Working Group on Image-Guided Tumor Ablation. Radiology 2005; 235: Wood BJ, Zhang H, Durrani A, et al. Navigation with electromagnetic tracking for interventional radiology procedures: a feasibility study. J Vasc Interv Radiol 2005; 16: Seiler PG, Blattmann H, Kirsch S, Muench RK, Schilling C. A novel tracking technique for the continuous precise measurement of tumour positions in conformal radiotherapy. Phys Med Biol 2000; 45:N103 N Frantz DD, Wiles AD, Leis SE, Kirsch SR. Accuracy assessment protocols for electromagnetic tracking systems. Phys Med Biol 2003; 48: Solomon SB, White P, Wiener CM, Orens JB, Wang KP. Three-dimensional CT-guided bronchoscopy with a real time electromagnetic position sensor. Chest 2000; 118: Solomon SB, Dickfield T, Calkins H. Real-time cardiac catheter navigation on three-dimensional CT images. J Interv Card Electrophysiol 2003; 8: Solomon SB, Magee CA, Acker DE, Venbrux AC. Experimental nonfluoroscopic placement of inferior vena cava filters: use of an electromagnetic navigation system with previous CT data. J Vasc Interv Radiol 1999; 10: Solomon SB, Magee C, Acker DE, Venbrux AC. TIPS placement in swine, guided by electromagnetic real-time needle tip localization displayed on previously acquired 3-D CT. Cardiovasc Intervent Radiol 1999; 22: Banovac F, Glossop N, Lindisch D, Tanaka D, Levy E, Cleary K. Liver tumor biopsy in a respiring phantom with the assistance of a novel electromagnetic navigation device. Part 1. In: Dohi T, Kikinis R, eds. Medical image computing and computer-assisted intervention 5th international conference. New York, NY: Springer, 2002: Wood B, Lindisch D, Ranjan S, Glossop N, Cleary K. Electromagnetically tracked guidewires for interventional procedures. In: International congress series: program and abstracts of Computer Aided Radiology and Surgery (CARS 2004). New York, NY: Excerpta Medica, 2004: Glossop N, Cleary K, Burgess J, Corral G. Magnetically tracked bone screws for image guided intervention. In: International congress series: program and abstracts of Computer Aided Radiology and Surgery (CARS 2004). New York, NY: Excerpta Medica, 2004: Sacolick L, Patel N, Tang J, Levy E, Cleary K. Electromagnetically tracked placement of a peripherally inserted central catheter. In: Galloway RL Jr, ed. Medical imaging 2004: visualization, image-guided procedures, and display proceedings of SPIE, vol Bellingham, WA: SPIE, 2004: Mueller PR, vansonnenberg E. Interventional radiology in the chest and abdomen. N Engl J Med 1990; 322: Vannier MW, Marsh JL. Three-dimensional imaging, surgical planning and image-guided therapy. Radiol Clin North Am 1996; 34: Santos RS, Gupta A, Ebright MI, et al. Electromagnetic navigation to aid radiofrequency ablation and biopsy of lung tumors. Ann Thorac Surg 2010; 89: Wallace MJ, Gupta S, Hicks ME. Out-of-plane computed-tomography-guided biopsy using a magnetic field based navigation system. Cardiovasc Intervent Radiol 2006; 29: Krücker J, Sheng X, Glossop N, et al. Electromagnetic tracking for thermal ablation and biopsy guidance: clinical evaluation of spatial accuracy. J Vasc Interv Radiol 2007; 18: Bruners P, Penzkofer T, Nagel M, et al. Electromagnetic tracking for CT-guided spine interventions: phantom, ex-vivo and in-vivo results. Eur Radiol 2009; 19: AJR:196, May 2011