Abstract
Objective: CT-guided biopsy still plays a decisive role in the management of liver tumors, especially if the lesions are not visible or accessible by ultrasound. As CT-guided stereotaxy appears to be a very accurate targeting technique, the aim of this study was to evaluate the targeting accuracy, diagnostic yield, and complications of CT-guided stereotactic liver biopsy of primary and secondary liver tumors.
Methods and Materials: Prior to stereotactic liver biopsy, patients under general anesthesia were immobilized using a vacuum cushion. Respiratory motion was controlled by temporary disconnections of the endotracheal tube. An optical-based navigation system was used for 3D trajectory planning and placement of a 15-G coaxial needle via a stereotactic aiming device. The histological samples were obtained using a 16-G Tru-Cut™ biopsy needle system. For evaluation of targeting accuracy the control CT image with the needles in place was fused with the planning CT image. The lateral error at the tip and skin entry point and the angular error were calculated. In addition, the skin-to-liver-surface (SL) distance, the needle-to-liver-surface (NL) angle, and the presence of liver cirrhosis were evaluated. The diagnostic yield was evaluated by histological reports from the institutional pathologists.
Results: The median lateral error was 2.5 mm (range: 0–6.5 mm) at the needle entry point and 3.2 mm (range: 0.01–9.4 mm) at the needle tip. The median angular error was 1.06° (range: 0–6.64°). Liver cirrhosis, NL angle and SL distance showed no significant impact on the targeting accuracy. Forty-five of the 46 liver biopsies (97.8%) were diagnostic according to the histological reports. No puncture-related complications such as bleeding or perforation of intestinal organs or lung tissue were recorded.
Introduction
Percutaneous image guided liver biopsy still plays a decisive role in the management of liver tumors. In the literature, ultrasound (US) guidance is most commonly used for liver biopsy and is considered safe for both in- and outpatients, with complication rates ranging from 6 to 11% Citation[1]. Complications, such as pleural and gastrointestinal perforations, laceration of vessels with bleeding, bile duct injuries with subsequent bilioma, and insufficient histological assessment, are mainly the result of imprecise needle placement.
Surgical navigation systems have recently been adapted to improve trajectory planning and targeting Citation[2–5]. Navigation systems allow real-time tracking of surgical instruments projected in the patient's corresponding CT or MR images. This technology may improve the technical options for and accuracy of percutaneous CT-guided liver targeting through 3D imaging, computerized trajectory planning in arbitrarily oriented tracks, and navigated or stereotactic aiming device-guided needle placement Citation[2–3]. Previous studies have confirmed high targeting accuracy with such systems in phantoms and animal experiments Citation[6]; however, stereotactic image guided biopsy has so far only been reported for diagnosis of brain or breast tumors Citation[7], Citation[8]. The aim of the present work was to clinically assess the targeting accuracy, diagnostic yield, and complications of CT-guided stereotactic liver biopsy.
Materials and Methods
Patients
Forty-five consecutive patients undergoing stereotactic liver biopsy of primary or secondary liver tumors between January 2008 and August 2009 were included in the study. The patients comprised 33 males and 12 females, and their median age was 63 years (range: 41–83 years). The study was approved by the Institutional Review Board and informed consent was obtained from all patients.
Procedure
The entire procedure was performed with the patient under general anesthesia with muscle relaxation. Patient immobilization was guaranteed by using a BlueBAG vacuum cushion (Medical Intelligence GesmbH, Schwabmünchen, Germany) Citation[9], Citation[10]. Compensation for respiratory motion during the planning CT scan, stereotactic targeting and subsequent control CT scan was made feasible by temporary disconnection of the endotracheal tube (ETT), which was performed and controlled by the anesthetist in the intervention room Citation[11].
For image-to-patient registration, 15 widely spaced registration markers (Beekley SPOTS®, Beekley Corporation, Bristol, CT) were attached to the skin in the region of interest on the thorax/upper abdomen. A contrast-enhanced (90–120 ml Iodixanol, 320 mg I/ml, 3 ml/sec) helical planning CT scan, with a slice thickness of 3 mm, was then performed using a Siemens SOMATOM Sensation Open CT scanner with a sliding gantry 82 cm in diameter (Siemens AG, Erlangen, Germany).
The CT data was subsequently sent via the hospital's picture archiving and communication systems (PACS) to an optical-based navigation system (StealthStation® TREON™ plus; Medtronic, Inc., Louisville, KY).
Three-dimensional visualization of the patient's image data by the navigation software enabled advanced trajectory planning and adjustment using the different viewing modes.
After calibration, 7 widely spaced registration markers were manually defined in the image data and subsequently identified on the patient with the navigation probe for paired-point image-to-patient registration during end-expiration. The navigation system immediately calculated the root mean square error (RMSE) of the fiducial registration error (FRE). The FRE/RMSE is the error of corresponding registered points and an estimate of the registration accuracy. Only values lower than 1 mm were accepted. However, the FRE depends on the geometric distribution of the fiducials (i.e., the registration markers) and may underestimate target registration errors. Finally, a short security check was performed by touching the patient's markers with the probe and comparing their positions to those calculated.
After disinfection, the entire interventional field was draped with a translucent, partially self-adhesive plastic foil. The sterilized ATLAS aiming device (Medical Intelligence GesmbH, Schwabmünchen, Germany) was mounted to the draped patient-covering bridge. This device consists of an adjustable mechanical arm and two pivot joints, enabling trajectory entry and angulations to be adjusted separately. Guided by the navigation system, the aiming device was fixed when the calculated trajectory alignment was ≤0.5 mm and 0°. The distance to target was automatically calculated and marked on a 15-G/17.2 cm coaxial needle (Bard, Inc., Covington, GA). Rigidly guided by the aiming device, coaxial needle insertion was performed during temporary ETT disconnection without imaging control.
Following completion, a native control CT scan was performed during ETT disconnection to confirm the position of the coaxial needle and rule out puncture-related complications. Using the coaxial technique, a 16-G Tru-Cut™ biopsy needle with 2.2 cm ejection was used to obtain a minimum of three complete cores.
All procedures were performed by radiologists with extensive experience in CT-guided and frameless stereotactic interventions. In the present study, all lesions were detectable by the contrast-enhanced planning CT, and thus no fusion of different modalities was necessary.
Time expenditure: The period of time required for the registration (including placement of registration markers, acquisition of the contrast-enhanced planning CT scan, data transfer to the navigation system, marker definition on the navigation system, and marker indication with the navigation probe) is approximately 15 min. Trajectory planning takes approximately 2–5 min, depending on the anatomical situation. The trajectory alignment using the stereotactic aiming device, including the placement of the coaxial needle during temporary ETT disconnection, takes approximately 2–3 min, and the biopsy itself approximately 2–4 min.
Evaluation
Targeting accuracy: The native control CT scan was fused with the planning CT data using the navigation system's 3D image registration algorithm (ImMerge™, Medtronic, Inc.) (see ). The lateral error, i.e., the normal distance at the tip and skin entry point, and the angular error, i.e., the angle of deviation, between the planned and inserted needle were calculated as described by Widmann et al. Citation[12].
Figure 1. Screenshot of the native control CT data with the needles in place fused with the contrast-enhanced planning CT data showing the planned trajectory as well as the inserted coaxial needle. The upper and lower left images correspond to longitudinal cuts along the planned trajectory (“trajectory views”). The upper right image shows an orthogonal cut at the tip of the coaxial needle (the “probe's eye view”), and the lower right image is a 3D rendering of the trajectory and the patient.
As possible impact factors, the NL angle and SL distance were evaluated. The NL angle represents the angle between the inserted needle and the corresponding liver surface and was measured in two coplanarities perpendicular to one another and taken as the mean difference from 90°. The SL distance is the distance between the skin and liver surface.
Diagnostic yield: All histological reports from the institutional pathologists were evaluated and defined as diagnostic if sufficient histological material could be obtained to make a definite diagnosis.
Complications: Only puncture-related complications were recorded by evaluation of the native control CT scan after needle placement.
Statistical analysis
Data analysis and descriptive statistics were performed using PASW Statistics 15.0 (SPSS, Inc., Chicago, IL). Quantitative data were described with mean values, standard deviation and range. Student's t-test was used to determine significant statistical differences for paired groups. Factor analysis was performed by simple linear regression. A p-value less than 0.05 was defined as statistically significant.
Results
In the 45 patients a total of 46 stereotactic liver biopsies were performed of 26 hepatocellular carcinomas (HCCs) (56.5% of the total), 5 cholangiocellular carcinomas (CCCs) (10.9%), and 15 liver metastases (32.6%; comprising 8 colorectal carcinomas, 2 bronchial carcinomas, 2 mamma carcinomas, 1 neuroendocrine tumor, 1 urothelial carcinoma and 1 melanoma). The median tumor size was 3.5 cm (range: 1.5–8.0 cm). Of the 46 tumors, 18 (39.1%) were subcapsular, 12 (26.1%) were in proximity to larger vessels, 2 (4.3%) were in proximity to organs (such as the stomach, bowel, gall bladder, etc.) and 14 (30.4%) were located in the central lobe. Twenty-three of the patients (51.1%) had liver cirrhosis.
All planned liver biopsies were completed successfully. No puncture-related complications such as bleeding, perforation of intestinal organs or lung tissue were recorded, and 45/46 (97.8%) of the biopsies were diagnostic according to histological reports.
Targeting errors: The median lateral error was 2.5 mm (range: 0–6.5 mm) at the needle entry point and 3.2 mm (range: 0.01–9.4 mm) at the needle tip. The median angular error was 1.06° (range: 0–6.64°) (see and ).
Figure 2. Box-plot of calculated lateral errors at the needle entry (LE e) and tip (LE t) and of the calculated angular errors.
Table I. Overall statistics for 46 stereotactic needle placements in 45 patients.
Possible impact factors: Liver cirrhosis, NL angle and SL distance showed no significant impact on the targeting accuracy (see and ).
Table II. Group comparison.
Table III. Factor analysis.
Discussion
Since the first liver biopsy was performed by Paul Ehrlich in 1883 Citation[13], indications, peri-interventional management and, especially, technical procedures have developed considerably. Nowadays, liver biopsies are performed either with ultrasound, CT, or MR guidance, or “blind”, i.e., without imaging control. As second-line procedures (e.g., in patients with coagulation disorders), transvenous or stereoscopic biopsies may be performed Citation[14], Citation[15].
Weigand and Weigand Citation[1] reported diagnostic accuracy rates of approximately 99.4% with conventional ultrasound (though lesion size was not specified). In terms of small liver lesions (<3 cm), success rates are 91 to 95.8% with US Citation[16], Citation[17], 86 to 95% with conventional CT Citation[16], Citation[18] and 96% with MRI Citation[19]. However, Ma et al. Citation[16] showed significantly lower success rates for cirrhotic livers and caudate lobe locations using conventional US and CT, while Rhim and Dodd Citation[20] indicated that many liver lesions may not be accessible by US guidance due to an inconspicuous tumor or inadequate path.
Importantly, when using conventional CT-guided techniques, target lesions must be visible in native scans. Furthermore, sensitive structures such as small vessels are only clearly visible on the contrast-enhanced planning CT images, leading to an increased risk of damage during the intervention. Another important drawback of conventional targeting is the necessity of advancing the needle stepwise to perform control scans, resulting in a decisive increment of irradiation and puncture duration.
Improving CT-guided interventions
Laser guidance devices, consisting of a laser unit mounted to the CT gantry, allow for easy, quick and accurate adjustment of the planned trajectory by projecting a laser beam on the entry point with the proper angulations. Pereles et al. Citation[21] showed significantly improved accuracy with laser guidance compared to freehand punctures (5 ± 0.4 mm versus 11 ± 0.8 mm). Puncture guides Citation[22], composed of a hemispheric fluid-filled device with a tubular needle guide, lead to an improvement in terms of stabilization of the needle advancement. However, these techniques still have in-plane and, to some extent, trajectory angle limitations.
To further improve puncture techniques, stereotactic navigation systems have been adjusted to meet the needs of radiologists Citation[2]. Three-dimensional imaging, computerized trajectory planning at arbitrary angulations, and navigated needle placement yield the best access to lesions limited to standard in-plane approaches; even lesions close to the pleura, hollow viscera, vessels or bile duct become accessible. Another advantage of computer-assisted stereotaxy is the possibility of combining imaging modalities by the use of multi-modality image fusion Citation[2]. Hence, lesions that are undetectable by CT but detectable by, for example, MRI or PET, become accessible.
The tracking method used by current navigation systems is either optical or electromagnetic Citation[2], Citation[23]. One of the biggest advantages of an electromagnetic system, apart from the avoidance of radiation exposure, is that a visual connection between the navigation probe and sensor is not necessary. However, the electromagnetic tracking field is much smaller compared to that of optical systems, and also more delicate Citation[24], Citation[25]. Maier-Hein et al. Citation[26] reported total puncture errors of 3.7 ± 2.3 mm (max. 11.6 mm) for in vivo targeting of artificial liver lesions in a ping model based on invasive registration needles. Das et al. Citation[27] showed total targeting errors of 3.5 mm (range: 2.7–4.2 mm) with an optical navigation system based on CT-guided, frame-attached marker paired-point registration in a phantom, and achieved 100% success in initial attempts to target simulated liver lesions ranging in size from 5 to 15 mm. Levy et al. Citation[28] showed total errors of 4 mm in an abdominal phantom and Zhang et al. Citation[25] reported errors of 8.3 ± 3.7 mm in a pig model, both with electromagnetic tracking systems.
In contrast to tracked needle placement, which is still prone to manual error during freehand navigation, stereotactic aiming devices may stabilize the probe advancement and allow for a quick and accurate trajectory alignment. In general, such devices consist of freely adjustable mechanical arms connected to the intervention table. The drawback of a cumbersome trajectory alignment was overcome by modifications in subsequent generations.
Real-time tracking of the biopsy needle Atlas aiming device (Medical Intelligence, Schwabmünchen, Germany) is not necessary. Measurements from the aiming device to the planned target are automatically calculated, and can be easily marked on the biopsy needle. In a phantom study, optically navigated stereotactic CT-guided (3-mm slice thickness), surface marker paired-point registration-based and aiming device-supported punctures showed total targeting errors of 2.2 ± 1.1 mm (range: 0.6–5.5 mm) and lateral errors of 1.8 ± 1.2 mm (range: 0.1–5.0 mm) Citation[6].
Use of conventional techniques requires multiple control CT scans, whether by “real-time” or “quick-check” methods. In contrast, stereotactic CT-guided systems do not require step-wise imaging control, resulting in a clear reduction in radiation exposure. Contrast medium is only needed for the planning CT scan. Image fusion of the native control CT scan (with the needle in place) with the pre-interventional planning CT image data allows for a quick and precise accuracy evaluation Citation[2]. However, Widmann et al. Citation[11] showed fusion errors of 0.39 ± 0.31 mm (range: 0–1.4 mm) for external targets and 0.71 ± 0.50 mm (range: 0.1–3.8 mm) for internal targets, which must be kept in mind when position corrections are necessary.
Robotic systems may lead to further improvements, especially in the future, facilitating an intervention that is not dependent on human targeting skills, such as perceiving navigational data or needle handling. Stoffner et al. Citation[6] reported a similar targeting accuracy (using 1- and 3-mm slice thicknesses) for a robot-assisted system to that obtained with stereotaxy using an aiming device. However, most current robotic systems are cumbersome and not yet ready for clinical implementation.
In the clinical setting of our study, a vacuum cushion and temporary ETT disconnections were used for patient immobilization and respiratory motion control, both of which have proven to be safe with an accuracy of 2 ± 0.93 mm for external targets and 1.41 ± 0.75 mm for internal targets Citation[11]. CT-guided stereotactic, optically navigated, surface marker paired-point registration-based targeting in our clinical series of 46 interventions showed median lateral errors of 3.2 mm (range: 0.1–9.4 mm) at the needle tip and angular errors of 1.1° (range: 0.0–6.6°). The possible impact factors evaluated, such as the skin-to-liver-surface distance, presence of liver cirrhosis or NL angle, were shown to have no significant influence on the evaluated errors. These results imply a decisive robustness of deviation errors in terms of different approaches.
The mortality rate of blind (i.e., without imaging control) liver biopsies is 0.01–0.1% Citation[29–31]. In US-guided percutaneous liver biopsies, Weigand and Weigand Citation[1] reported total complication rates ranging from 6.3% to 11.3% (no death) with a success rate of 99.4%. In terms of approach, Perrault et al. Citation[31] reported a significantly higher complication rate of 4.1% for subcostal routes, compared to 2.7% for transthoracic routes.
In our clinical setting, all planned biopsies were completed successfully and no peri-interventional complications were recorded. Forty-five of the 46 liver biopsies (97.8%) were diagnostic according to histological reports. Taking three complete cores with 2.2 cm ejection seems to be safe and efficient.
Limitations
The number of patients or biopsies evaluated in this study may be too small. To our knowledge, however, no comparable clinical studies have been published.
In the error calculations, the image fusion error and the influence of the draping with translucent plastic foil were included Citation[12].
Finally, 3D navigation/stereotaxy has substantially greater requirements in terms of infrastructure and staff than conventional targeting techniques. Patient immobilization and respiratory motion control are essential Citation[2], Citation[3], Citation[11]. Moreover, careful system set-up, trained staff, probe calibration and a rigid aiming device are mandatory. A control CT scan must verify proper needle positioning prior to an intervention.
Conclusion
The results presented indicate the potential of CT-guided stereotaxy for precise and secure liver biopsy on all possible arbitrary trajectories. This technique is especially helpful for biopsies of small lesions that are not clearly visible with ultrasound or on native CT scans and lesions that are in “difficult” locations, demanding double angulated trajectories.
Acknowledgments
The authors would like to express their sincere gratitude to the the radiation technicians Martin Fasser, RT, Florian Schanda, RT, and Julia Mahlknecht, RT, from the Department of Microinvasive Therapy, Department of Radiology.
Declaration of interest: Prof. Bale is a (co)inventor of the ATLAS aiming device (Medical Intelligence GesmbH, Schwabmünchen, Germany) and a co-shareholder in its financial returns.
References
- Weigand K, Weigand K. Percutaneous liver biopsy: Retrospective study over 15 years comparing 287 inpatients with 428 outpatients. J Gastroenterol Hepatol 2009; 24(5)792–799
- Bale R, Widmann G. Navigated CT-guided interventions. Minim Invasive Ther Allied Technol 2007; 16(4)196–204
- Bale R, Widmann G, Stoffner DI. Stereotaxy: Breaking the limits of current radiofrequency ablation techniques. Eur J Radiol 2010; 75(1)32–36
- Bale RJ, Hoser C, Rosenberger R, Rieger M, Benedetto KP, Fink C. Osteochondral lesions of the talus: Computer-assisted retrograde drilling - feasibility and accuracy in initial experiences. Radiology 2001; 218(1)278–282
- Bale RJ, Laimer I, Martin A, Schlager A, Mayr C, Rieger M, Czermak BV, Kovacs P, Widmann G. Frameless stereotactic cannulation of the foramen ovale for ablative treatment of trigeminal neuralgia. Neurosurgery 2006; 59(4 Suppl 2)ONS394–ONS401, discussion ONS2
- Stoffner R, Augschöll C, Widmann G, Böhler D, Bale R. Accuracy and feasibility of frameless stereotactic and robot-assisted CT-based puncture in interventional radiology: A comparative phantom study. Rofo 2009; 181(9)851–858
- Lomoschitz FM, Helbich TH, Rudas M, Pfarl G, Linnau KF, Stadler A, Jackman RJ. Stereotactic 11-gauge vacuum-assisted breast biopsy: Influence of number of specimens on diagnostic accuracy. Radiology 2004; 232(3)897–903
- Widmann G, Eisner W, Kovacs P, Fiegele T, Ortler M, Lang TB, Stoffner R, Bale R. Accuracy and clinical use of a novel aiming device for frameless stereotactic brain biopsy. Minim Invasive Neurosurg 2008; 51(6)361–369
- Bale RJ, Vogele M, Rieger M, Buchberger W, Lukas P, Jaschke W. A new vacuum device for extremity immobilization. AJR 1999; 172(4)1093–1094
- Bale RJ, Lottersberger C, Vogele M, Prassl A, Czermak B, Dessl A, Sweeney RA, Waldenberger P, Jaschke W. A novel vacuum device for extremity immobilisation during digital angiography: Preliminary clinical experiences. Eur Radiol 2002; 12(12)2890–2894
- Widmann G, Schullian P, Haidu M, Wiedermann F, Bale R. Respiratory motion control for stereotactic computer-assisted and robotic liver interventions. Int J Med Robot 2010; 6(3)343–349
- Widmann G, Stoffner R, Sieb M, Bale R. Target registration and target positioning errors in computer-assisted neurosurgery: Proposal for a standardized reporting of error assessment. Int J Med Robot 2009; 5(4)355–365
- Von Frerichs F. Über den Diabetes. Hirschwald, Berlin 1884
- Weickert U, Siegel E, Schilling D, Eickhoff A, Jakobs R, Bohrer MH, Riemann JF. [The diagnosis of liver cirrhosis: A comparative evaluation of standard laparoscopy, mini-laparoscopy and histology] [in German]. Z Gastroenterol 2005; 43(1)17–21
- Sawyerr AM, McCormick PA, Tennyson GS, Chin J, Dick R, Scheuer PJ, Burroughs AK, McIntyre N. A comparison of transjugular and plugged-percutaneous liver biopsy in patients with impaired coagulation. J Hepatol 1993; 17(1)81–85
- Ma X, Arellano RS, Gervais DA, Hahn PF, Mueller PR, Sahani DV. Success of image-guided biopsy for small (</=3 cm) focal liver lesions in cirrhotic and noncirrhotic individuals. J Vasc Interv Radiol 2010; 21(10)1539–1547, quiz 47
- Reading CC, Charboneau JW, James EM, Hurt MR. Sonographically guided percutaneous biopsy of small (3 cm or less) masses. AJR 1988; 151(1)189–192
- Stattaus J, Kuehl H, Ladd S, Schroeder T, Antoch G, Baba HA, Barkhausen J, Forsting M. CT-guided biopsy of small liver lesions: Visibility, artifacts, and corresponding diagnostic accuracy. Cardiovasc Intervent Radiol 2007; 30(5)928–935
- Stattaus J, Maderwald S, Baba HA, Gerken G, Barkhausen J, Forsting M, Ladd ME. MR-guided liver biopsy within a short, wide-bore 1.5 Tesla MR system. Eur Radiol 2008; 18(12)2865–2873
- Rhim H, Dodd GD. 3rd. Radiofrequency thermal ablation of liver tumors. J Clin Ultrasound 1999; 27(5)221–229
- Pereles FS, Baker M, Baldwin R, Krupinski E, Unger EC. Accuracy of CT biopsy: Laser guidance versus conventional freehand techniques. Acad Radiol 1998; 5(11)766–770
- Meyer JM, Schmitz-Rode T, Krombach G, Wildberger JE, Pfeffer J, Günther RW. Navi-ball: A new guidance device for CT-directed punctures. Invest Radiol 2001; 36(5)299–302
- Wood BJ, Locklin JK, Viswanathan A, Kruecker J, Haemmerich D, Cebral J, Sofer A, Cheng R, McCreedy E, Cleary K, et al. Technologies for guidance of radiofrequency ablation in the multimodality interventional suite of the future. J Vasc Interv Radiol 2007; 18(1 Pt 1)9–24
- Khadem R, Yeh CC, Sadeghi-Tehrani M, Bax MR, Johnson JA, Welch JN, Wilkinson EP, Shahidi R. Comparative tracking error analysis of five different optical tracking systems. Comput Aided Surg 2000; 5(2)98–107
- Zhang H, Banovac F, Lin R, Glossop N, Wood BJ, Lindisch D, Levy E, Cleary K. Electromagnetic tracking for abdominal interventions in computer aided surgery. Comput Aided Surg 2006; 11(3)127–136
- Maier-Hein L, Tekbas A, Seitel A, Pianka F, Müller SA, Satzl S, Schawo S, Radeleff B, Tetzlaff R, Franz AM, et al. In vivo accuracy assessment of a needle-based navigation system for CT-guided radiofrequency ablation of the liver. Med Phys 2008; 35(12)5385–5396
- Das M, Sauer F, Schoepf UJ, Khamene A, Vogt SK, Schaller S, Kikinis R, van Sonnenberg E, Silverman SG. Augmented reality visualization for CT-guided interventions: System description, feasibility, and initial evaluation in an abdominal phantom. Radiology 20061; 240(1)230–235
- Levy EB, Tang J, Lindisch D, Glossop N, Banovac F, Cleary K. Implementation of an electromagnetic tracking system for accurate intrahepatic puncture needle guidance: Accuracy results in an in vitro model. Acad Radiol 2007; 14(3)344–354
- McGill DB, Rakela J, Zinsmeister AR, Ott BJ. A 21-year experience with major hemorrhage after percutaneous liver biopsy. Gastroenterology 1990; 99(5)1396–1400
- Sherlock S, Dick R, Van Leeuwen DJ. Liver biopsy today. The Royal Free Hospital experience. J Hepatol 1985; 1(1)75–85
- Perrault J, McGill DB, Ott BJ, Taylor WF. Liver biopsy: Complications in 1000 inpatients and outpatients. Gastroenterology 1978; 74(1)103–106