CT-Guided Liver Biopsy With Electromagnetic Tracking: Results From a Single-Center Prospective Randomized Controlled Trial

Transcription

1 Vascular and Interventional Radiology Original Research Kim et al. of CT-Guided Liver Biopsy Vascular and Interventional Radiology Original Research Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved Edward Kim Thomas J. Ward Rahul S. Patel Aaron M. Fischman Scott Nowakowski Robert A. Lookstein Kim E, Ward TJ, Patel RS, Fischman AM, Nowakowski S, Lookstein RA Keywords: CT guidance, electromagnetic tracking, liver biopsy DOI:0./AJR..06 Received October, 0; accepted after revision March 6, 0. This trial was supported by an unrestricted educational grant from Philips Healthcare. Site visits to monitor protocol compliance were provided. E. Kim is a consultant for Philips Healthcare and has given industrysponsored lectures for Philips Healthcare. R. S. Patel is a consultant for Arstasis and Sirtex. A. M. Fischman is a consultant for Arstasis and Terumo. R. A. Lookstein is a consultant for Cordis and Medrad, has given industrysponsored lectures for Boston Scientific, and serves on the scientific advisory board for Boston Scientific. All authors: Department of Interventional Radiology, Mount Sinai Medical Center, One Gustave L. Levy Pl, Box, New York, NY 009. Address correspondence to E. Kim (Edward.Kim@mountsinai.org). Supplemental Data Available online at WEB This is a web exclusive article. AJR 0; 0:W W 06 0X//06 W American Roentgen Ray Society CT-Guided Liver Biopsy With : Results From a Single-Center Prospective Randomized led Trial OBJECTIVE. The purpose of this study is to evaluate the effectiveness of electromagnetic tracking in assisting CT-guided liver biopsies. MATERIALS AND METHODS. This was a single-center prospective randomized controlled trial comparing nonfluoroscopic CT guided liver biopsy using an advance-andscan technique with and without electromagnetic tracking. Fifty patients with a liver lesion referred for biopsy (women, %; mean age, 9. years; mean lesion size,.6 cm) were enrolled in the study and were randomly assigned to either arm. The primary and secondary objectives were to assess and quantify differences in the number of intraprocedural scans, cumulative effective radiation dose, number of needle manipulations, and procedure time from skin-stick to the target lesion with and without assistance. RESULTS. Electromagnetic tracking significantly decreased the number of scans, effective radiation dose, number of manipulations per procedure, and time from skin-stick to the target lesion. The ratio of the number of scans (electromagnetic tracking to control) was 0. (9% CI, 0. 0.; p < 0.000). The mean difference in effective radiation dose (electromagnetic tracking control) was. msv (9% CI,.0 to. msv; p = 0.000), and the median difference was. msv (9% CI,.0 to.6 msv; p < 0.000). The ratio of the number of manipulations (electromagnetic tracking to control) was 0.6 (9% CI, 0. 0.; p < 0.000). The mean difference for the time from skin-stick to the target lesion was.6 seconds (9% CI, 9. to 00. seconds; p = 0.00) and the median difference was.0 seconds (9% CI,.00 to.00 seconds; p = 0.000). CONCLUSION. Electromagnetic tracking assistance has the potential to decrease the number of intraprocedural CT scans and needle manipulations and to reduce patient radiation dose during CT-guided liver biopsy. P ercutaneous procedures are routinely performed by diagnostic and interventional radiologists for a variety of indications. Percutaneous ultrasound-, CT-, or MRI-guided biopsy has replaced open surgical biopsy in most settings, when feasible. Each guiding modality has its advantages and disadvantages. Ultrasound may be cost effective and provide real-time feedback without radiation, although lesions can be obscured by bone or bowel gas and may be occult on ultrasound. CT offers improved spatial resolution and is not limited by adjacent bones or bowel gas. CT scanning or fluoroscopy causes radiation exposure to the patient. Many lesions may also be inconspicuous without contrast agent, making confident targeting difficult. MRI has excellent contrast resolution without radiation but is expensive and requires special equipment with limited availability at many centers. An electromagnetic tracking system is similar to a car s global positioning system navigation system; it shows the real-time location of a tracked instrument on previously obtained images and its relationship to the goal destination. An electromagnetic tracking system consists of an electromagnetic field generator and multiple electromagnetic field sensors. Sensors are located in fiducial markers placed on the patient s skin, as well as procedural devices, such as a coaxial introducer needle. The field generator induces currents in the sensors, with the known position of the fiducial markers used to locate the position of the procedural device within the patient s body. The orientation of the tracked instrument within the magnetic field affects the tracking ability. Tracked instruments, AJR:0, December 0 W

2 Kim et al. Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved A C Fig. Setup of electromagnetic tracking system. A, Line diagram of setup of CT suite using electromagnetic tracking system. (Illustration by Bluvol N) B, Photograph of electromagnetic tracking system. Red arrow denotes tracked coaxial introducer needle, and black arrow denotes field generator. During CT-guided liver biopsy, electromagnetic tracking screen (inset) identifies needle tip location by position of crosshairs placed on historical images. Distance to lesion and size of circle decrease as needle is advanced from subcutaneous tissues to target lesion. Yellow crosshairs indicate position of needle tip. C and D, Screenshots of case with level- difficulty, with needle at skin surface (C) and with needle at lesion (D), are shown. B D W6 AJR:0, December 0

3 of CT-Guided Liver Biopsy Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved A Fig. Screenshots for case with level- difficulty. A and B, Images are shown with needle at skin surface (A) and with needle at lesion (B). Yellow crosshairs indicate position of needle tip. when oriented in a perpendicular orientation toward the field generator, will not be tracked as effectively because of inherent limitations of electromagnetic tracking. This may result in intermittent tracking signal in the worst case. The electromagnetic tracking system used in this study (PercuNav. Tx, Philips Healthcare) is a U.S. Food and Drug Administration 0(k) cleared computer-assisted image-guided diagnostic and intervention system. The software displays the location and orientation of a tracked instrument in real-time onto a historical dataset from cross-sectional imaging. Phantom and animal models have shown the accuracy and utility of such tracking systems [ ]. Clinical reports have shown the safety and effectiveness, although no prospective randomized trial has been conducted to evaluate and quantify any benefit of the technology in routine clinical practice compared with conventional CT-guided interventions [ ]. The purpose of this study is to evaluate the effect of electromagnetic tracking assistance on nonfluoroscopic CT guided liver biopsies with respect to the number of intraprocedural scans, effective radiation dose, number of needle manipulations, and procedure time from skin-stick to the target lesion. Materials and Methods This was a single-center prospective randomized controlled trial, with study design approved by the performing center s institutional review board. All patients provided informed consent to be included in this study. This study was performed with HIPAA compliance. Study Enrollment and Patient Selection Study design Between November 0 and March 0, 0 consecutive patients (mean age, 9. years; SD,. years; % women) referred for nonfluoroscopic CT biopsy at a single center were prospectively enrolled in this randomized controlled trial to assess and quantify differences in the number of intraprocedural scans, effective radiation dose, number of needle manipulations, and procedure time from skin-stick to the target lesion during nonfluoroscopic CT guided liver biopsy using an advance-and-scan technique performed with and without electromagnetic tracking. A liver lesion that was sonographically occult or difficult to biopsy was not part of the inclusion criteria. The time from skin-stick to the target lesion was defined as the time between skin puncture with the introducer needle to arriving at the target lesion. Total procedural time was not recorded. The randomization schedule was generated by study statisticians and was stored in a folder with access restricted to the statisticians. Treatment allocation was concealed in a series of sequentially numbered opaque sealed envelopes given to the investigators. All biopsies were performed by a single operator, a board-certified radiologist with a certificate of added qualification in interventional radiology with 6 years of experience performing conventional CT-guided biopsies. The operator practiced using electromagnetic tracking on an acrylic phantom approximately 0 times over the course of a week before using the device in clinical practice. The operator had 6 months of experience with the electromagnetic tracking system in routine clinical practice before the start of the trial, approximately biopsies. In the study group, the operator was shown the electromagnetic tracking screen, allowing him to locate and correct the coaxial biopsy introducer trajectory in real time. In the control group, the operator was blinded to the electromagnetic tracking screen, although the screen was recorded for later review. Inclusion criteria Patients were included in the study if they had been referred for biopsy of a liver lesion and had undergone a preprocedural CT scan of the abdomen, were older than years, and had the ability to understand and the willingness to sign a written informed consent form, as well as comply with the study protocol. Exclusion criteria Patients were precluded from a biopsy on the basis of standard exclusions B AJR:0, December 0 W

4 Kim et al. Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved No. of Scans or Manipulations Intraprocedural Scans (e.g., irreversible coagulopathy or lack of percutaneous window), if they had an adhesive allergy (because of the application of active fiducial markers with adhesive backing), if they had a pacemaker or automatic implantable cardiac defibrillator, if they had a body weight above 6 lb (0. kg) for women or 0 lb (. kg) for men, if they were pregnant, and if they were unable to follow procedural instructions, including, but not limited to, holding their breath and remaining reasonably motionless during the procedure. The subject exclusion criteria were tailored to local demographic data for a 9th percentile body habitus. The 9th percentile weight variable was calculated according to data from the National Health and Nutrition Examination Survey III, which contains anthropometric data [9]. The data were adjusted to the ethnic distribution for the state of New York. Clinical Evaluation There were two planned study visits for each patient. During visit, all eligible patients had a preenrollment meeting and consultation with an investigator, where the patient s and family s questions were answered. The patient s history was further reviewed to confirm eligibility. The study and equipment were explained in detail to the patient and their family during the consent process. Informed consent was obtained and patients were given the option of ultrasound, as well as CT guidance, with or without participation in the research study. Prior imaging was also reviewed. Each patient read and signed the informed consent form at this visit. During visit, patients who met all the inclusion criteria and none of the exclusion criteria were randomized to either nonfluoroscopic CT guided biopsy with or without electromagnetic tracking. Needle Manipulations Fig. Box plot of number of intraprocedural scans and needle manipulations in electromagnetic tracking and control groups. Lines denote first quartile, median, and third quartile. Whiskers denote minimum and maximum. Circles denote mean. Effective Dose (msv) This study was performed with : randomization, with treatment allocation concealed in sequentially numbered opaque sealed envelopes. Randomization was performed before the trial began. The level of difficulty for the biopsy was scored using both subjective and objective grading schemes. The subjective level of difficulty of the biopsy was graded by the operator on a scale of to 0, with 0 being most difficult. The objective level of difficulty was scored using a previously described scheme modified for the liver [0], where indicates a lesion less than cm deep or larger than cm in diameter, indicates a lesion smaller than cm in diameter but more than cm deep, indicates a lesion smaller than cm in diameter but more than 0 cm deep, and indicates a lesion that is difficult to biopsy because of small size (< cm). Procedural Details Before initial CT scanning of the liver, registration fiducial markers and a radiopaque grid were Fig. Box plot of time from skin-stick to target lesion in electromagnetic tracking and control groups. Lines denote first quartile, median, and third quartile. Whiskers denote minimum and maximum. Circles denote mean. Fig. Box plot of effective radiation dose in electromagnetic tracking and control groups. Lines denote first quartile, median, and third quartile. Whiskers denote minimum and maximum. Circles denote mean. Time to Target (s) applied to the patient s skin around the upper abdomen in the area of interest. A CT scan of the liver was then performed with a CT scanner (Somatom Sensation, Siemens Healthcare) using a standardized protocol for all patients (0 kvp; slice thickness, mm; matrix, ; tube current, 6 ma). If prior imaging indicated the lesion to be occult without contrast agent, the initial scan was performed with the administration of 6% iopamidol injection ( Isovue 00, Bracco Diagnostics). For arterially enhancing lesions, an injection rate of.0 ml/s and a 0- to -second delay was chosen. For nonarterially enhancing lesions, an injection rate of..0 ml/s and an 0-second delay was used. Contrast agent was needed for five patients in the electromagnetic tracking group and nine patients in the control group. The biopsy was performed immediately after this initial reference CT. Images were then transferred to the electromagnetic tracking system. No other electromagnetic tracking systems were available at the per- W AJR:0, December 0

5 of CT-Guided Liver Biopsy Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved No. of Subjects forming institution to allow comparison; the system used for this study did not allow fusion with ultrasound. The CT scan was reviewed on the electromagnetic tracking system, and the locations of the fiducial markers were verified. The software maps the position of a tracking device in this case, a 6-gauge coaxial biopsy introducer on the graphic overlay allowing real-time trajectory modification. The setup and procedure differed between the electromagnetic tracking and control groups in one way; the electromagnetic tracking screen was shown to the operator for the electromagnetic tracking group patients but not the control patients. Setting up the electromagnetic tracking system took approximately minutes for each case. The operator performed all liver biopsies with the goal of safely achieving a diagnosticquality sample from all patients with as few needle manipulations and intraprocedural scans as possible. After the initial skin-stick and coaxial introducer needle advancement, an intraprocedural scan was obtained to plan the next needle manipulation or to confirm a satisfactory location for biopsy. This was repeated until the needle was at the target lesion, as described earlier. Examples are provided of the procedural display (Fig. ), sample cases (Fig. ), and movies of the procedure (Videos and, which can be viewed by clicking Supplemental at the top of this article and then clicking the video number on the Supplemental page). The numbers of intraprocedural CT scans and needle manipulations and the time from first skinstick with the coaxial introducer needle until the needle was at the target lesion (in-line trajectory and within mm) were recorded. The numbers of needle manipulations and intraprocedural scans 6 0 Total No. of Intraprocedural Scans Fig. 6 Histogram of number of intraprocedural scans in electromagnetic tracking and control groups. No. of Subjects varied according to clinical judgment depending on the size of the lesion, length of the throw, and adjacent critical structures. Intraprocedural CT scans were defined as the number of helical CT scans performed during a given biopsy, with the FOV set two or three axial slices above and below the target slice number. CT fluoroscopy is not available at the performing institution. All needle manipulations and intraprocedural scans were performed with the patient at end expiration. When the needle was at the lesion, three biopsy samples were obtained with an -gauge core biopsy needle (lengths, 0 0 cm; Adjustable Coaxial Temno biopsy needle, CareFusion) advanced through the coaxial introducer needle. As a result, only a single capsule puncture was required for all cases. If an intraprocedural scan was required, the z-axis scan length was reduced to two or three slices above and below the level of the needle. A electromagnetic tracking screen capture was acquired before all intraprocedural scans, as well as before resuming the procedure. At the end of the procedure, a verification CT scan was obtained to evaluate for complications. All patients were called at home within hours to make sure no delayed complications occurred. Study Endpoints and Objectives The primary endpoint was the number of intraprocedural CT scans performed during the course of the nonfluoroscopic CT guided liver biopsy. The primary objective was to assess and quantify differences in the number of intraprocedural scans in biopsies performed with and without electromagnetic tracking assistance. The secondary objectives were to assess and quantify differences in radiation dose, number of needle manipulations, and time from skin-stick 6 9 No. of Manipulations Fig. Histogram of number of needle manipulations in electromagnetic tracking and control groups. to the target lesion with and without assistance. Safety was evaluated by prospectively recording adverse events in both groups. Technical success was defined as concordance between pathologic findings, imaging, and clinical history. Postprocedure Protocol After the biopsy was concluded, the patient was monitored in an independent interventional radiology postanesthetic care unit. Vital signs, including blood pressure, pulse, and oxygen saturation, were monitored for hours. At this time, the patient was advised to ambulate as tolerated and was discharged to home when at baseline. The patient was called within hours of the biopsy to evaluate for possible complications. Statistical Analysis Assuming an underlying Poisson distribution for each treatment modality, we performed a simulation to aid in the study design. When we assumed a type error rate of %, the simulation revealed that subjects per study group were sufficient to detect a difference 9% of the time. Overall, a total of approximately 0 participants ( participants per group) were to be randomly assigned. The primary and secondary analyses were performed on a modified intent-to-treat basis, which included all randomized subjects who completed the CT-guided biopsy procedure. Poisson regression model with a log link was used on scan count data to assess the ratio of scans between the electromagnetic tracking and control groups. Least squares means (a theoretic better estimate of the true population mean than the arithmetic mean), standard errors of the mean (SEMs), and 9% CIs for each group and for the treatment ratio (electromagnetic tracking to control) were evaluated. AJR:0, December 0 W9

6 Kim et al. Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved Parametric and nonparametric ANOVA were used to assess all the primary and secondary endpoints for each group. Comparisons between the groups were performed using an F-test. Least squares means, SEMs, and 9% CIs for each group and for the difference in scans by treatment group (electromagnetic tracking control) were presented for ANOVA. Median, standard error of the median, Hodges-Lehmann estimates, 9% CIs for median shift, and Kruskal-Wallis test were presented for nonparametric ANOVA. Bootstrap CIs using 0,000 samples were constructed to support robustness of the analysis for mean and median difference using parametric and nonparametric ANOVA. All analyses were conducted using SAS/STAT software (version 9., SAS Institute). TABLE : Patient Demographics Characteristic Electromagnetic Tracking Group (n = ) Results Fifty patients were randomly assigned, to each study arm. Two patients were excluded from analysis (one from each group) because of technical difficulties with the electromagnetic tracking system at the time of the scheduled procedure. Forty-eight patients are included in our final analysis; patient demographics and procedural details are provided in Tables and, respectively. All biopsies provided diagnostic samples for pathologic examination, with all biopsy samples concordant with the imaging and clinical history. A descriptive summary of pathology results and the objective level of biopsy difficulty are provided in Table. Results of the study show the number of intraprocedural scans, effective radiation dose, number of needle manipulations, and procedure time from skin-stick to the target lesion to be significantly less in the electromagnetic tracking group. A box plot of number of scans, effective dose, number of manipulations, and time from skin-stick to the target lesion for each group is provided in Figures, respectively. A histogram of number of scans and manipulations for each group are provided in Figures 6 and, respectively. The objective level of difficulty was not significantly different between the electromagnetic tracking group and the control group (level of difficulty score:.0 and.9, respectively, on a scale of, with the most difficult; p = 0.). The subjective level of difficulty, as described by the operator at time of biopsy on a scale of 0 (with 0 the most difficult), was not significantly different between the electromagnetic tracking group and the control group (.6 vs.; p = 0.). Descriptive summary of primary and secondary endpoints are provided in Table. Efficacy conclusions were drawn on the basis of the guidelines of the superiority tests constructed. For differences (electromagnetic tracking control) analysis, if both the lower and upper limits for the two-sided 9% CI were less than 0, the electromagnetic tracking group was declared to be superior to Group (n = ) Total (n = ) p Age (y).0 (.) 6.9 (0.) 9. (.) 0. Sex, no. (%) of patients Female (6.) 0 (.) (.) 0.9 Male 9 (.) (.) (.9) Height (cm) 6. (9.) 66. (.) 6. (9.) 0.6 Weight (kg) 6.9 (.). (.) 0. (6.9) 0.9 Body mass index (kg/m ). (.6).9 (.0).6 (.6) 0.9 Lesion size (cm).0 (.0). (.00).6 (.66) 0. Note Except for patient sex and p values, all data are mean (SD). TABLE : Procedural Details Characteristic Electromagnetic Tracking Group (n = ) Group (n = ) Total (n = ) Sedation Conscious sedation (9.) 9 (9.) (.) General anesthetic (.) (6.) (0.) None (.) (.) (.) Patient position Left anterior oblique (.) 0 (0.0) (.) Supine (9.) (00) (9.9) Type of contrast agent IV (0.) 9 (.) (9.) None 9 (9.) (6.) (0.) Contrast agent amount (ml) 0 (.) 6 (.0) 9 (.) 00 (.) (.) (0.) Overall fit of registration a Mean (SD). (0.90). (.0). (0.9) 9% CI Median... Minimum, maximum 0.,.0 0.,. 0.,. Case outcome successful (00) (00) (00) Note Except where noted otherwise, data are number (%) of patients. a The overall fit of the registration is meant to represent the root-mean square error for all matched registered landmarks, both external and internal to the patient images. It is a method of communicating the quality of the registration between the patient and imaging space. the control group. For ratio (electromagnetic tracking to control) analysis, if both the lower and upper limits for the two-sided 9% CI were less than, the electromagnetic tracking group was declared to be superior to the control group. For the total number of intraprocedural scans, the ratio of scans (electromagnetic tracking to control) was 0. (9% CI, 0. W0 AJR:0, December 0

7 of CT-Guided Liver Biopsy Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved 0.; p < 0.000). The least squares mean difference was.6 scans (9% CI,.6 to.9 scans; p < 0.000), and the median difference was.0 scans (9% CI,.00 to.00 scans; p < 0.000). The results show that electromagnetic tracking decreases the number of scans by at least % and requires at least.9 mean fewer scans and at least median fewer scans than the control group. For the effective radiation dose, the least squares mean difference was. msv (9% CI,.0 to. msv; p = 0.000), and the median difference was. msv (9% CI,.0 to.6 msv; p < 0.000). The results show that electromagnetic tracking decreases the mean effective dose by at least. msv and the median effective dose by at least.6 msv compared with the control group. For the number of manipulations, the ratio of manipulations (electromagnetic tracking to control) was 0.6 (9% CI, 0. 0.; p < 0.000). The least-squares mean difference was. manipulations (9% CI,. to. manipulations; p < 0.000), and the median difference was.0 manipulations (9% CI,.00 to.00 manipulations; p < 0.000). The results show that electromagnetic tracking decreases the number of manipulations by at least 6% and requires at least. fewer mean manipulations and at least fewer median manipulation manipulations. For the time from skin-stick to the target lesion, the least-squares mean difference was.6 seconds (9% CI, 9. to 00. seconds; p = 0.00) and the median difference was.0 seconds (9% CI,.00 to.00 seconds; p = 0.000). The results show electromagnetic tracking to require at least 00. seconds less mean time to the target lesion and at least seconds less median time to the target lesion. Bootstrap analysis supported all these results. No adverse events were observed in either group. The mean subjective level of difficulty of the biopsy on a scale of to 0, as graded by the operator, was.6 (SEM, 0.9; 9% CI,. 6.) and. (SEM, 0.; 9% CI,. 6.9) for the electromagnetic tracking and control groups, respectively. This difference was not significantly significant (p = 0.). The mean objective level of difficulty, based on the previously described grading scheme, was not significantly different either (electromagnetic tracking,.0; control,.9; p = 0.9). These results confidently identify the level of biopsy difficulty to be equivalent in the two groups, as would be suspected given the prospective randomized nature of the examination. TABLE : Descriptive Summary of Pathologic Findings and Biopsy Level of Difficulty Assisted CT-Guided Biopsy Pathologic Finding Objective Level of Difficulty Score Discussion The effectiveness and potential applications for electromagnetic tracking systems have been investigated in phantom and animal models, as well as limited clinical practice. The existing research has shown electromagnetic tracking to be accurate and to provide useful procedural information [,,, ]. Decreased procedural time, number of needle manipulations, and radiation dose have also been shown in phantom and animal models [, ]. To our knowledge, no prospective randomized controlled trial has been conducted to investigate the use of electromagnetic tracking in clinical practice. Unassisted CT-Guided Biopsy Pathologic Finding Objective Level of Difficulty Score Dysplastic nodule Metastatic colon cancer Metastatic pancreatic cancer Chronic hepatitis and fibrosis Lipofuscinosis Cholangiocarcinoma Sclerosing hemangioma Sclerosing hemangioma Cholangiocarcinoma Metastatic uterine cancer Sclerosing hemangioma Sarcoid Cholangiocarcinoma Cholangiocarcinoma Inflammatory pseudotumor Focal nodular hyperplasia Metastatic pancreatic cancer Sclerosing hemangioma Hepatocellular carcinoma Cholangiocarcinoma Neuroendocrine tumor Focal nodular hyperplasia Neuroendocrine tumor Dysplastic nodule Neuroendocrine tumor Hepatocellular carcinoma Hepatocellular carcinoma Cholangiocarcinoma Cholangiocarcinoma Hepatocellular carcinoma Metastatic squamous cell Dysplastic nodule carcinoma Hepatocellular carcinoma Hepatocellular carcinoma Metastatic squamous cell Poorly differentiated carcinoma carcinoma Lymphoma Metastatic melanoma Inflammatory pseudotumor Dysplastic nodule Metastatic gallbladder Metastatic pancreatic cancer adenocarcinoma Hepatocholangiocarcinoma Atypical lymphoid infiltrate Organizing abscess Hepatocellular carcinoma Metastatic gastric cancer Hepatocellular carcinoma Mean level of difficulty.0 of Mean level of difficulty.9 of In this single-operator experience, electromagnetic tracking during nonfluoroscopic CT guided liver biopsy reduced the number of intraprocedural scans, radiation dose to the patient, and the number of needle manipulations. CT-guided liver biopsy was safe with or without electromagnetic tracking assistance; no adverse events were observed in either group. The primary endpoint analysis showed % fewer intraprocedural CT scans with electromagnetic tracking compared with the control group (9% CI, %). This translated into a statistically significant reduction in a mean effective radiation dose to the patient of. msv. The observed radiation savings are equivalent to the effective dose of AJR:0, December 0 W

8 Kim et al. Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved approximately three unenhanced CT scans of the head or one unenhanced CT scan of the body []. The amount is also greater than the annual background radiation, from natural sources, to people living in the United States ( msv/year). The impact of small radiation doses of this magnitude on the risk of malignancy has been debated. Some suggest that a small excess cancer risk at very low doses likely exists, whereas others say that these estimates are without rigorous scientific basis [ 6]. Regardless, decreasing radiation exposure when clinically possible is a widely accepted goal. Although biopsies may be performed with ultrasound guidance, which results in no radiation to the patient, not all lesions are visible on ultrasound. To this end, electromagnetic tracking provides a mean to achieve such goals in the appropriate setting. The reduction in the number of manipulations by 6% also has significant potential implications for patient care. When targeting lesions that are in close proximity to vital structures, such as the heart or major blood vessels, fewer instrument manipulations could translate into a lower rate of complications. For example, an increased number of pleural punctures during CT-guided lung biopsy has been shown to result in a higher risk of pneumothoraces []. Tumor seeding is a theoretic risk with percutaneous biopsy. The rate at which this occurs after biopsy of a hepatocellular carcinoma has been reported to be around % []. Needle diameter, amount of normal parenchyma traversed by the biopsy, and number of biopsy passes has been said to affect the rate at which tumor seeding occurs [9]. As a result, a decreased risk of seeding related to a lower number of biopsy manipulations is a theoretic advantage of CT-guided biopsy with electromagnetic tracking assistance technologies. Fewer device manipulations may also translate into benefits with regard to percutaneous ablations. If multiple ablations were to be performed in a single setting or if multiple electrodes needed to be placed with precise orientation relative to one another (irreversible electroporation), the number of procedural manipulations saved could be greater. The amount of time from skin-stick to the target lesion was significantly lower in the electromagnetic tracking group. Although the difference in means (.6 seconds; 9% CI, 9. to 00. seconds) was statistically TABLE : Descriptive Summary of Primary and Secondary Endpoints Endpoint, Treatment Group Mean (SD) Median Minimum, Maximum 9% CI No. of scans Electromagnetic tracking. (.9),.. (.)., 0 6. Effective dose (msv) Electromagnetic tracking. (.0)..,.6 9 (.)..,. 0.. No. of manipulations Electromagnetic tracking. (0.),.6. (.9), 9.9. Time (s) Electromagnetic tracking. (0.) 6, 0... (9.)., 0. significant, it is unlikely that this difference would routinely translate into a clinically significant benefit. Again, if multiple ablations were to be performed or multiple devices needed to be placed at a single time, any time savings would be greater, making electromagnetic tracking more appealing. Randomization of patients to either electromagnetic tracking assisted biopsy or nonfluoroscopic CT guided biopsy minimized allocation bias and allowed the validity assumptions required for statistical analysis. The two groups were not significantly different with regard to lesion size or level of difficulty, the two factors that would most likely affect the endpoints of interest. Given these considerations, we think the results observed in this study would mimic those seen in clinical practice. These results are in line with recent results by Hakime et al. [0], who found that electromagnetic tracking shortened needle placement time and number of manipulations, regardless of operator experience. Some strengths of our study compared with that study include objective level of difficulty grading, sealed-envelope randomization, and an analysis of radiation dose with and without electromagnetic tracking. There are limitations to this study. One major limitation is the inability of the operator to be blinded to which group the patient was randomly assigned. The nature of the treatment makes this difficult or impossible. The inability to blind the operator to the patient s randomization arm has the potential to introduce significant bias. This was controlled for as much as possible by objectively and subjectively grading the biopsy level of difficulty. Furthermore, the operator performed all biopsies with the goal of safely achieving a diagnostic sample in all patients with as few needle manipulations and intraprocedural scans as possible. All biopsies being performed by a single operator is also a major limitation of this study. With any novel technology, there is a learning curve involved while acclimating to a new device. In this study, the single operator became facile with the device within a few hours of use on an acrylic phantom. The operator in this study is a fellowship-trained interventional radiologist with extensive experience in percutaneous image-guided interventions. It is stipulated that gaining comfort with this technique may take longer for operators with less experience or those who perform image-guided percutaneous interventions less frequently. Because this was a single-operator study, a comparison of the learning curve between operators is not possible. This may be investigated in future studies if warranted. Another limitation is that patients with defibrillators and patients who cannot follow directions are not candidates for imageguided intervention with the electromagnetic tracking system, because cooperation and the ability to remain still are critical for good electromagnetic tracking. All these factors limit the generalizability of the observed results to all patients and all operators. The final limitations involve the study inclusion criteria. A liver lesion that was sonographically occult was not a requirement for inclusion. It is true that one or more includ- W AJR:0, December 0

9 of CT-Guided Liver Biopsy Downloaded from by... on 0/0/ from IP address... Copyright ARRS. For personal use only; all rights reserved ed lesions could have been biopsied under ultrasound or MRI guidance. We stipulate that ultrasound guidance offers real-time visualization of the needle during manipulation with zero radiation dose, something that nonfluoroscopic CT guidance with electromagnetic tracking assistance does not provide. MRI guidance may also offer an alternate approach, although MRI guidance is available at relatively few institutions. We do not think, however, that these points make the observed single-operator decrease in scans, manipulations, and dose less valid. We think that institutional capabilities and operator preference should dictate the guiding modality. Nonfluoroscopic CT guided biopsy was necessitated because CT fluoroscopy was not available at the performing institution; this is also a limitation. Another limitation of the study is that cost was not included as a study endpoint. Patients who weighed greater than 6 lb (0. kg) were not included in the study. Depth and body size are challenges that affect any interventional procedure and would not necessarily reflect a limitation of the navigation system. It is known, however, that electromagnetic tracking systems have a constrained volume within which instruments are tracked. Instruments that are outside this tracking volume are not tracked by the navigation system. Thus, setup of the tracked volume to ensure coverage of the target area is important. In conclusion, in selected patients, electromagnetic tracking assistance has the potential to decrease the number of intraprocedural CT scans and needle manipulations and to reduce effective radiation dose to the patient during CT-guided liver biopsy. References. Bruners P, Penzkofer T, Nagel M, et al. Electromagnetic tracking for CT-guided spine interventions: phantom, ex-vivo and in-vivo results. Eur Radiol 009; 9: Banovac F, Wilson E, Zhang H, Cleary K. Needle biopsy of anatomically unfavorable liver lesions with an electromagnetic navigation assist device in a computed tomography environment. J Vasc Interv Radiol 006; :6 6. Hung AJ, Ma Y, Zehnder P, Nakamoto M, Gill IS, Ukimura O. Percutaneous radiofrequency ablation of virtual tumours in canine kidney using global positioning system-like technology. BJU Int 0; 09:9 0. Wood BJ, Zhang H, Durrani A, et al. Navigation with electromagnetic tracking for interventional radiology procedures: a feasibility study. J Vasc Interv Radiol 00; 6:9 0. Krücker J, Xu S, Glossop N, et al. Electromagnetic tracking for thermal ablation and biopsy guidance: clinical evaluation of spatial accuracy. J Vasc Interv Radiol 00; : 0 6. Krücker J, Xu S, Venkatesan A, et al. Clinical utility of real-time fusion guidance for biopsy and ablation. J Vasc Interv Radiol 0; :. Liu FY, Yu XL, Liang P, et al. Microwave ablation assisted by a real-time virtual navigation system for hepatocellular carcinoma undetectable by conventional ultrasonography. Eur J Radiol 0; : 9. Nakai M, Sato M, Sahara S, et al. Radiofrequency ablation assisted by real-time virtual sonography and CT for hepatocellular carcinoma undetectable by conventional sonography. Cardiovasc Intervent Radiol 009; : Centers for Disease and Prevention. The third national health and nutrition examination survey (NHANES III 9 99) reference manuals and reports. Bethesda, MD: National Center for Health Statistics; Singh AK, Leeman J, Shankar S, Ferrucci JT. Core biopsy with curved needle technique. AJR 00; 9: 0. Appelbaum L, Solbiati L, Sosna J, Nissenbaum Y, Greenbaum N, Goldberg SN. Evaluation of an electromagnetic image-fusion navigation system for biopsy of small lesions: assessment of accuracy in an in vivo swine model. Acad Radiol 0; 0:09. Jung EM, Friedrich C, Hoffstetter P, et al. Volume navigation with contrast enhanced ultrasound and image fusion for percutaneous interventions: first results. PLoS ONE 0; :e96. Huda W, Vance A. Patient radiation doses from adult and pediatric CT. AJR 00; :0 6. Cardis E, Vrijheid M, Blettner M, et al. Risk of cancer after low doses of ionising radiation: retrospective cohort study in countries. BMJ 00; :. Mathews JD, Forsythe AV, Brady Z, et al. Cancer risk in people exposed to computed tomography scans in childhood or adolescence: data linkage study of million Australians. BMJ 0; 6:f60 6. Charles M. UNSCEAR report 000: sources and effects of ionizing radiation United Nations Scientific Committee on the Effects of Atomic Radiation. J Radiol Prot 00; : 6. Moore EH, Shepard JA, McLoud TC, Templeton PA, Kosiuk JP. Positional precautions in needle aspiration lung biopsy. Radiology 990; :. Silva MA, Hegab B, Hyde C, Guo B, Buckels JA, Mirza DF. Needle track seeding following biopsy of liver lesions in the diagnosis of hepatocellular cancer: a systematic review and meta-analysis. Gut 00; : Chhieng DC. Fine needle aspiration biopsy of liver: an update. World J Surg Oncol 00; : 0. Hakime A, Barah A, Deschamps F, et al. Prospective comparison of freehand and electromagnetic needle tracking for US-guided percutaneous liver biopsy. J Vasc Interv Radiol 0; :6 69 AJR:0, December 0 W