HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2242011269v1 224/2/452    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nour, S. G. Articles by Lewin, J. S. Search for Related Content PubMed PubMed Citation Articles by Nour, S. G. Articles by Lewin, J. S.

MR Imaging–guided Radio-frequency Thermal Ablation of the Lumbar Vertebrae in Porcine Models¹

¹ From the Depts of Radiology (S.G.N., A.J.A., I.C.S.M., J.L.D., J.S.L.), Pathology (S.N.E.), and Biomedical Engineering (J.L.D.), Univ Hospitals of Cleveland/Case Western Reserve Univ School of Medicine, 11100 Euclid Ave, Cleveland, OH 44106-5056. Received Jul 24, 2001; revision requested Sep 12; revision received Dec 21; accepted Feb 25, 2002. Univ Hospitals of Cleveland/Case Western Reserve Univ Interventional MR Program supported in part through research collaborations with Siemens Medical Systems and Radionics. Also supported by grants from Whitaker Foundation, American Cancer Society, and NIH 1R01 CA81431-01A1 and 1R01-CA84433-01. Address correspondence to J.S.L. (e-mail: lewin@uhrad.com).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To test the hypotheses that (a) magnetic resonance (MR) imaging–guided radio-frequency (RF) thermal ablation of the vertebrae is feasible in porcine models, (b) procedure safety depends on the location of ablation within the vertebra, and (c) MR imaging allows accurate monitoring of induced thermal lesion size and shape.

MATERIALS AND METHODS: Ten percutaneous MR imaging–guided RF thermal ablations were randomized over various lumbar vertebral levels and locations in seven pigs. Animals were followed up for 2, 7, or 14 days before sacrifice. Thermal lesion size and shape as measured on MR images obtained immediately after ablation and at follow-up were compared with gross pathologic findings. Mean absolute differences between lesion diameters at pathologic examination and MR imaging were evaluated by using a paired t test, as were differences between lesion-to-vertebra contrast-to-noise ratios obtained for each sequence. Clinical and imaging data were correlated with histologic findings.

RESULTS: Successful RF electrode placement in the targeted part of the vertebra was achieved in all procedures. Ablations performed away from neural elements were safe to perform. Pedicular ablations resulted in radiculopathy, whereas ablations performed directly over the posterior cortex resulted in paraplegia. Lesion sizes measured on T2-weighted images were closest to those measured at gross pathologic examination (mean absolute difference, 0.72 mm ± 0.83 [SD]), followed by those measured on contrast material–enhanced T1-weighted (1.27 mm ± 0.83) and short inversion time inversion-recovery (STIR) (1.5 mm ± 1.84) images. Size measurements obtained on T2-weighted images were significantly closer to gross pathologic measurements than were those obtained on contrast-enhanced T1-weighted images (P = .013) but were not different from those obtained on STIR (P = .27) images. The contrast-to-noise ratio was significantly higher for contrast-enhanced T1-weighted images than for T2-weighted (P < .001) or STIR (P < .001) images.

CONCLUSION: MR imaging–guided RF thermal ablation of the vertebrae is feasible in porcine models, but the safety of the procedure depends on the location of ablation within the vertebra. MR imaging allows accurate monitoring of thermal lesion size and shape.

© RSNA, 2002

Index terms: Interventional procedures, experimental studies, 33.1269 • Magnetic resonance (MR), guidance, 33.121412 • Radiofrequency (RF) ablation, 33.1269 • Spine, interventional procedures, 33.1269

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Radio-frequency (RF) thermal ablation has been applied effectively and safely to treat a wide range of benign and malignant conditions (1). Target organs have included the liver (2), brain (3), and prostate (4), among others. In RF thermal ablation, the RF current is deployed into the target tissue through an electrode connected to a RF generator. As the current flows from the source to the return electrode, the ions in the tissues surrounding the monopolar source electrode start to agitate, resulting in frictional (resistive) heating with consequent formation of an ovoid necrotic lesion surrounding the RF electrode tip.

Such thermal therapy has been practiced with conventional fluoroscopic (5), ultrasonographic (6), and computed tomographic (CT) (7) guidance. Reports from magnetic resonance (MR) imaging–guided thermal ablation procedures (8–10) highlight the unique capability of MR imaging to depict tissue necrosis instantaneously during the ablation session, thereby allowing the interventionalist to tailor the induced thermal lesion so that it covers the entire targeted tumor.

This feature of MR imaging, coupled with sporadic reports addressing initial clinical success in the percutaneous treatment of spinal lesions such as osteoid osteomas (11,12), hemangiomas (13,14), and, recently, vertebral metastases (15), provided the impetus to conduct the current study to evaluate whether MR imaging can be used to guide and monitor safe vertebral RF ablation as a potential future application of minimally invasive therapy of spinal abnormalities. Specifically, our goal was to test the hypotheses that (a) MR imaging–guided RF thermal ablation of the vertebrae is feasible in porcine models, (b) procedure safety depends on the location of the ablation within the vertebra, and (c) MR imaging allows accurate monitoring of induced vertebral thermal lesion size and shape.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Animal Models and Anesthesia
Experiments were performed on 10 lumbar vertebrae in seven female American farm pigs weighing 20–25 kg. One ablation was performed in each of the first three pigs, and two ablations were performed in each pig thereafter, except for pig number 5, in which only one ablation was performed because of an intraprocedural complication. Following a protocol approved by the animal use and care committee of our institution, all procedures were performed with the use of general anesthesia. Induction of anesthesia was achieved by means of intramuscular injection with a combination of acepromazine maleate (25 mg per kilogram of body weight; Fermenta Animal Health, Kansas City, Mo) and ketamine hydrochloride (7.5 mg/kg, Ketaject; Phoenix Scientific, St Joseph, Mo). This was followed by intravenous administration of thiopental sodium (15 mg/kg, Pentothal; Abbott Laboratories, North Chicago, Ill) to facilitate tracheal intubation. Inhalation anesthesia was maintained during all procedures by using halothane 1% (Halocarbon Laboratories, River Edge, NJ). No premedications were administered prior to the procedures.

MR Imaging System
All procedures were performed in an interventional MR imaging suite by using a 0.2-T open C-arm MR imaging system (Magnetom Open; Siemens Medical Systems, Erlangen, Germany) with a 21-cm-diameter belt-shaped solenoid surface receiver coil (Siemens Medical Systems). Near real-time needle and/or RF probe tracking was made possible by (a) the ability to acquire rapid gradient-echo images—a fast imaging with steady-state precession (FISP) (17.8/8.1 [repetition time msec/echo time msec]; flip angle, 90°; number of signals acquired, three) technique was used to track the needles in six procedures with a temporal resolution of three frames per 22 seconds, while true FISP (12.5/5.9; flip angle, 90°; number of signals acquired, five) was used in the remaining four procedures, giving a temporal resolution of three frames per 24 seconds; and (b) the ability to both operate the imager and view images at the imager side with an in-room high-spatial-resolution 1,024 x 1,280-pixel RF-shielded liquid crystal monitor controlled by an MR-compatible mouse and foot pedal.

RF Ablation
The animals were placed in the left lateral decubitus position with the shoulder regions shaved and covered with two 8 x 12-cm wire mesh grounding pads (Radionics, Burlington, Mass), one on each side, to serve as a return path for the RF ablation current. Ablation sites were randomized to include most of the lumbar vertebral levels (one site at L2, four at L3, one at L4, three at L5, and one at L6), as well as various locations within the vertebra (four in the center of the vertebral body, one in the body close to the anterior cortex, one in the body close to the posterior cortex, one in the body close to the inferior cortex, one in the body close to the posterior and inferior cortices, and two in the right pedicle). Since the porcine lumbar spine consists of six or seven vertebrae (16,17), we always considered the vertebra above the sacrum to be number 6. After marking the skin entry site and shaving and scrubbing the area, an 11-gauge (3 x 100-mm) MR-compatible bone biopsy needle (Somatex, Berlin, Germany) was introduced into the planned part of the vertebra, always from the right side, with a transpedicular approach in eight procedures and with direct puncture of the lateral cortex in two procedures. No drills were used. Needle insertion was guided by means of direct MR fluoroscopy with FISP and true FISP sequences, as detailed above. The ability to detect vertebral outlines and spinal canal contents with the two sequences was compared subjectively. The stylet of the biopsy needle was then replaced with a custom-made 2-cm exposed-tip 17-gauge MR-compatible monopolar titanium RF electrode (Radionics). The RF electrode position was confirmed by means of fast spin-echo (SE) T1-weighted (680/24; echo train length, five; number of signals acquired, two) imaging (Fig 1). Electrode placement was considered successful when the exposed (2-cm-long) electrode tip was documented to be centered in the targeted part of the vertebra on the confirmatory T1-weighted images. Vertebral ablation was then performed by using a 100-W maximum-power RF generator operating at 500 KHz (RFG-3C; Radionics). For all ablation procedures, the power was manually controlled to keep the electrode tip temperature constant at 90° C ± 2, as measured continuously with a thermocouple located within the electrode tip, for a 10-minute period. The applied current and tissue impedance were recorded (by I.C.S.M.) at 1-minute intervals throughout the entire ablation session in all 10 procedures. The time spent for preprocedural imaging, RF electrode placement, and the actual ablation was also recorded (by I.C.S.M.) in all experiments.

View larger version (79K): [in this window] [in a new window] [Download PPT slide]   Figure 1. A, Transverse and B, sagittal fast SE T1-weighted (680/24; echo train length, five; number of signals acquired, two) images obtained to confirm RF electrode position prior to ablation. The electrode is seen as a linear signal void (arrows) traversing the right pedicle into the vertebral body. The sagittal image plane allows the electrode tip (arrowhead in B) to be directed inferiorly to reach the middle or lower part of the vertebral body, thereby providing rapid confirmation of the electrode position in all three dimensions. The electrode diameter appears larger than its real diameter, since we applied frequency encoding in a perpendicular direction to the electrode shaft to increase the electrode conspicuity.

The animals were then imaged again (by using the same pulse sequences used to obtain the initial postablation images) and sacrificed, and the vertebrae were harvested. The formalin-fixed nondecalcified vertebrae were then sliced transversely with an electric bone saw to have the same section thickness used for MR imaging (5 mm), while attempting to cut exactly through the plane of the RF electrode track. Sections were then photographed, and the maximum transverse diameter of each thermal lesion perpendicular to the RF electrode track (short axis) was measured. The maximum transverse diameter was also measured (by S.G.N.) for each sequence of the initial postablation and follow-up images by using electronic calipers on a satellite viewer (Siemens Medical Systems) of the main imaging system. In peripherally located vertebral lesions that also had soft-tissue components, we restricted our measurement on MR images to the intravertebral portion to validate the comparison with the gross pathologic vertebral specimens. In the gross specimens, we considered every marrow change inside the thin, dark outer margin to be part of the thermal lesion, whereas on MR images, we measured the inside of the enhancing rim on contrast-enhanced images and the inside of the hyperintense rims on fast SE T2-weighted and STIR images. The mean of the absolute difference between the lesion diameters as measured in gross pathologic specimens and on each of the follow-up images was calculated, and statistical significance was evaluated by using the paired Student t test. The diameter of each lesion was also compared on both initial postablation and follow-up images. The lesion-to-vertebra contrast-to-noise ratio (CNR) was calculated for each pulse sequence on both the initial postablation and follow-up images by using the formula CNR = (SI_lesion - SI_vertebra)/SD_noise, where SI_lesion is the signal intensity of the lesion, SI_vertebra is the signal intensity of the nonablated part of the vertebra, and SD_noise is the SD of the signal intensity of the background noise. Background noise was measured along the phase-encoding direction on the same images on which the lesion and vertebra signal intensity measurements were obtained. The diameters of the regions of interest used to measure these signal intensities were chosen to encompass the largest possible artifact-free parts of the areas being evaluated. Again, the statistical significance of the differences between the measured CNRs on various pulse sequences was evaluated by using the paired Student t test. For histologic correlation, the 5-mm-thick vertebrae were then decalcified, sectioned, and stained with hematoxylin-eosin. Histologic sections of the induced vertebral lesions were examined for viability and density of hematopoietic cells, viability of the cancellous bone, and the presence of inflammation, granulation tissue, or hemorrhage.

Criteria Used to Evaluate Ablation Safety
For the ablation performed close to the anterior vertebral cortex, we evaluated the pig (a) for evidence of a developing retroperitoneal hematoma by using MR fluoroscopy during the procedure and (b) for the relation of the anterior boundary of the created thermal lesion to the aorta and inferior vena cava on the initial postablation and follow-up MR images. For ablations performed close to the posterior vertebral cortex, we evaluated the initial postablation and follow-up MR images for evidence of focal myelopathy in the adjacent spinal cord segment, seen as abnormal hyperintensity on T2-weighted and STIR images. We also evaluated those pigs during the follow-up period for any paresis or paralysis of the hind limbs. After sacrifice, we evaluated the gross pathologic specimens for morphologic changes of the spinal cord segment adjacent to the ablation site and for histologic evidence of nerve cell or fiber degeneration or necrosis (eg, focal demyelination, axonal edema, cellular swelling, loss of nuclei, and cytoplasmic precipitates). For ablations performed within the vertebral pedicles, the animals were evaluated for evidence of temporary or permanent limping in the hind limb on the side of the ablated pedicle during the follow-up period. The ipsilateral nerve roots were examined for evidence of swelling, spongiotic degeneration, or adhesion at gross pathologic examination and for axonal edema, dural adhesion, inflammatory infiltration, or gliosis at histologic examination. For ablations performed close to the vertebral endplates, the initial postablation and follow-up MR images were evaluated for signal intensity changes in the adjacent intervertebral disks. In addition, the follow-up images of the entire animal population were evaluated for evidence of osseous or soft-tissue infection.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Feasibility and Safety
Successful introduction of the bone biopsy needle, followed by RF electrode placement in the targeted part of the vertebra, was achieved with direct MR imaging guidance in all procedures. Subjectively, the true FISP sequence offered better visualization of both vertebral outlines and spinal canal contents than the FISP sequence for guiding device placement. In one animal, a retroperitoneal hematoma complicated attempts to place the needle through the lateral cortex as close to the anterior vertebral cortex as possible with FISP guidance. Partial resorption of the hematoma was observed on the follow-up images. Apart from slight postprocedural pain that was relieved with pain medications as needed (0.12 mg of buprenorphine twice daily; Buprenorphine hydrochloride; Reckitt & Colman Products, Hull, England), ablation procedures performed away from the neural elements (Figs 2, 3) were generally well tolerated by all of the animals. In one of the two ablation procedures performed within the vertebral pedicles (Fig 4, D–F), the pig limped on the ipsilateral leg for a few days before its condition gradually improved. The pig in which we performed two ablation procedures flush with the posterior cortex of the vertebral body (Fig 5) developed initial paraplegia that necessitated early sacrifice 2 days after the procedure. In this case, areas of focal myelopathy at the ablated levels were evident on both the initial postablation and follow-up MR images as areas of hyperintensity on both T2-weighted and STIR images (Fig 5, A, C). A CT scan of the specimen was obtained in this case only (Mx8000 Quad; Marconi Medical Systems, Cleveland, Ohio). The intervening cortex appeared intact on both MR and CT images (Fig 5, A–D). Ablation performed close to the vertebral endplates did not result in intervertebral disk compromise. No signs of osseous or soft-tissue infection were noticed on the follow-up images. The Table summarizes the various complications of vertebral ablations according to the ablated part of the vertebra.

View larger version (106K): [in this window] [in a new window] [Download PPT slide]   Figure 2. RF ablation within the central part of L3 vertebral body. Transverse A, contrast-enhanced SE T1-weighted (528/26; flip angle, 90°; number of signals acquired, three), B, fast SE T2-weighted (2,600/96; echo train length, seven; number of signals acquired, seven), and C, fast SE STIR (2,700/48; echo train length, seven; number of signals acquired, five) MR images acquired in vivo on day 14 after ablation demonstrate a well-circumscribed oval thermal lesion located totally within the confinement of the vertebral body. The thermal lesion displays signal hypointensity in all pulse sequences (*, A-C), with a hyperintense rim on T2-weighted and STIR images (arrowheads, B and C) that enhances on the contrast-enhanced image (arrowheads, A). D, Transverse multiplanar reformatted CT image of the same lesion acquired after sacrifice shows reactive bone marrow sclerosis (arrowhead) marginating the lesion and surrounding the position of the electrode tip (arrow). E, Gross pathologic specimen obtained at the same level as the images shows the dark red electrode track (arrow), surrounded by a pale ovoid area of tissue necrosis with a thin, dark erythematous rim (arrowheads) outlining the periphery of the necrotic region. F, Histologic section (hematoxylin-eosin stain; original magnification, x250) obtained from the thermal lesion shows the hemorrhagic needle tract (*), surrounded by coagulative necrosis of hematopoietic cells (white arrowhead) and the osteoblasts of bony trabeculae (black arrowhead). Within the necrotic zone that occupies most of the field, the nuclei of viable infiltrating monocytes and/or macrophages are visible, scattered within necrotic debris. Viable trabecular bone and marrow hematopoietic cells (white arrow) and adipocytes (black arrow) are evident outside the necrotic zone along the left margin of the field.

View larger version (60K): [in this window] [in a new window] [Download PPT slide]   Figure 3. A, Transverse STIR (2,700/48; echo train length, seven; number of signals acquired, five) MR image and B, corresponding gross pathologic section show a successful ablation performed within the posterior part of the L4 vertebral body. Although the nearest point of the RF electrode track (curved arrows) is located 2 mm from the spinal canal, this pig did not develop any neurologic deficit after the procedure, and neurohistologic examination revealed C, a completely normal spinal cord with intact nerve cells in the gray matter, seen preserving their normal nuclei (arrows, C) (hematoxylin-eosin stain; original magnification, x125). Arrowheads in A indicate the margins of the induced thermal lesion within the vertebral body. The difference between this ablation and that shown in Figure 5 is that the RF electrode is not placed entirely over the posterior vertebral cortex, but it approximates it at a more localized area. Also note the epidural vessel (B, straight arrow) located just at the point where the electrode is closest to the spinal canal. The blood flow within this vessel, coupled with the cerebrospinal fluid pulsations, might have contributed to some heat dissipation effect that protected the spinal cord.

View larger version (124K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Two ablation procedures performed within the right pedicles of the L5 (top) and L6 (bottom) vertebrae. A, D, Transverse fast SE T2-weighted (2,600/96; echo train length, seven; number of signals acquired, seven) MR images acquired in vivo on the the 7th day after ablation show the hypointense thermal lesions in the right vertebral pedicles, surrounded by hyperintense rims (arrowheads, A and D) and traversed by linear areas of hyperintensity representing the RF electrode tracks (arrows, A and D). Note that the right exiting nerve root on A lies within the extent of the hyperintense margin of the lesion, while on D it lies within the periphery of the hypointense lesion itself. B, E, Gross pathologic sections from both lesions show the pale necrotic lesions involving the right pedicles (*, B) and bounded by the hypointense erythematous margins (arrowheads, B and E). The RF electrode track is shown in E (curved arrow). Note the crumpling and adhesion of the exiting nerve root (straight arrow) on the side of the lesion in E. C, Histologic section of the right nerve root of the lesion depicted on A and B (hematoxylin-eosin stain; original magnification, x250) demonstrates mild axonal edema (arrowheads), and F, histologic section (hematoxylin-eosin stain; original magnification, x125) of the right nerve root of the lesion depicted on D and E shows adhesion (between arrows) of the nerve root (N) to the perineurium (dura mater, DM), along with an inflammatory monocyte and glial infiltrate (arrowheads) within the nerve root adjacent to the site of adhesion. Examination results of the adjacent spinal cord segments were normal for both cases.

View larger version (106K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Ablation performed in the most posterior part of the L3 vertebral body with the RF electrode placed directly over the posterior cortex of the vertebral body. This animal developed immediate paraplegia following ablation and was sacrificed after 48 hours. Transverse A, T2-weighted and B, STIR MR images obtained just prior to sacrifice show a less-defined hypointense area (arrowheads, A and B) ventral to the RF electrode track (straight arrows, A and B), with hyperintensity (curved arrows, A and B) involving the spinal cord denoting associated myelopathy (better seen on the STIR image). C, Sagittal STIR MR image of the same ablation demonstrates that the area of myelopathy (arrowheads) extends well above and below the level of the thermal lesion. The thermal lesion is seen as a round area of hypointensity (arrow) that is better defined than on transverse images, with the electrode track seen on end as a bright dot in the center of the lesion. (A-C, Imaging parameters are the same as those in Figure 2). D, Transverse multiplanar reformatted CT image acquired after sacrifice documents the complete integrity of the posterior vertebral body cortex (arrowhead) adjacent to the RF electrode track (arrow). E, Gross pathologic section from the same pig shows a pale area of bone necrosis (*) at the posterior vertebral body. The margins of the lesion are not very well defined in this 2-day-old lesion when compared with the previous 7- and 14-day-old lesions. The arrow indicates the site of the RF electrode track. F, Histologic section (hematoxylin-eosin stain; original magnification, x125) of the spinal cord at the assumed area of myelopathy on MR images demonstrates diffuse gray matter necrosis with loss of the nuclei of nerve cells (arrows).

Applied Current and Tissue Impedance
The mean RF current applied to the vertebrae during the 10-minute ablation procedure was 310 mA ± 20. The mean tissue impedance during the applied current was 86.5 ± 0.7.

Figure 6 illustrates the mean RF current and tissue impedance profiles plotted as functions of time during the 10-minute ablation procedures. We excluded the current and impedance recordings from one of the 10 ablation procedures, since the RF generator was momentarily switched off during RF application.

View larger version (37K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Plot of the mean RF current and tissue impedance as functions of time during the 10-minute ablations. Deposited energy, represented by the RF current, reached a maximum during the 1st minute of ablation, then continued to decline slowly as the ablation progressed. Tissue impedance decreased during the 1st minute to maintain a constant value throughout the entire ablation procedure.

View larger version (120K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Transverse contrast-enhanced SE T1-weighted (528/26; flip angle, 90°; number of signals acquired, three) MR image clearly demonstrates the nonprotective property of the cortical bones during in vivo RF thermal ablation. Despite placement of the RF electrode within the pedicle of this vertebra, the induced thermal lesion continued to grow outside the vertebra to develop an even larger intramuscular component (white arrows), with the intact cortical bone (black arrow) conspicuous as a linear area of hypointensity within the less hypointense coagulated area. Arrowheads represent the intravertebral component of the thermal lesion.

The lesion-to-vertebra CNR was significantly higher for the contrast-enhanced T1-weighted images than for the T2-weighted (P < .001) or STIR (P < .001) images (on both initial postablation and follow-up images). On the initial postablation images, the mean CNR was 620 ± 131 for contrast-enhanced T1-weighted images, 358 ± 73 for T2-weighted images, and 186 ± 35 for STIR images. On the follow-up images, the mean CNR was 652 ± 81 for contrast-enhanced T1-weighted images, 424 ± 89 for T2-weighted images, and 225 ± 41 for STIR images.

Thermal Lesion Size Evolution in the 7-day Follow-up Group
Follow-up imaging of six of the 10 thermal lesions was performed on the 7th day. For this group, the mean maximum diameter of the intravertebral parts of the lesions perpendicular to the probe track (short axis) as measured on the initial postablation MR images was 8.2 mm ± 1.5 for fast SE T2-weighted images, 8.2 mm ± 1.5 for fast SE STIR images, and 9.5 mm ± 1.4 for contrast-enhanced SE T1-weighted images. As measured on images obtained on the 7th day, the mean maximum diameter was 11.2 mm ± 1.9 for fast SE T2-weighted images, 10.5 mm ± 1.9 for STIR images, and 11.3 mm ± 2.4 for contrast-enhanced SE T1-weighted images. The overall mean lesion diameters in the initial postablation and follow-up groups are provided below.

Histopathologic Correlation
At gross pathologic examination, an induced thermal lesion typically showed a dark red linear puncture wound penetrating the cortical bone and extending obliquely into the vertebral body; the specific anatomic site of the puncture of the cortical bone and the specific distribution within the vertebral body, as well as involvement of the pedicle, varied with the individual specimen. There was an elliptical zone of dull pallor extending radially for 3–5 mm around the needle tract. This area of pallor showed an irregular, densely erythematous border. In addition, there was an inconsistent narrow zone of hemorrhage or erythema at the interface of the needle tract and the necrotic cancellous bone.

Lesion sizes measured on T2-weighted images were closest to those measured in gross pathologic specimens (mean absolute difference, 0.72 mm ± 0.83), followed by those measured on contrast-enhanced SE T1-weighted (1.27 mm ± 0.83) and STIR (1.50 mm ± 1.84) images. The 0.55-mm difference between T2- and T1-weighted images was found to be statistically significant (P = .013). The absolute error for the T2-weighted technique was lower than that for the T1-weighted technique in five of nine cases, and it was identical in four of nine cases. For T2-weighted versus STIR imaging, the difference was not statistically significant by using a paired t test (P = .27). Absolute deviation from gross pathologic findings was the same in seven of nine cases, slightly lower in one case, and 6 mm lower in the remaining case. The latter case accounts for the fact that the mean differences between T2-weighted and STIR errors are larger than those between T2- and T1-weighted techniques. Given the unusual pattern of results (seven with no difference, one with a small difference, and one with an extreme outlier) and the small sample, these results are probably best described qualitatively rather than statistically. For the above calculations, we excluded one lesion from our measurements, since we could not obtain the corresponding pathologic section perfectly in the plane of the RF electrode track.

The overall mean lesion diameter on the initial postablation MR images was 9.1 mm ± 2.0 for fast SE T2-weighted images, 8.5 mm ± 1.4 for fast SE STIR images, and 9.9 mm ± 1.6 for contrast-enhanced SE T1-weighted images. On the follow-up images, the mean lesion diameter was 10.7 mm ± 1.6 for fast SE T2-weighted images, 10.1 mm ± 1.5 for fast SE STIR images, and 10.6 mm ± 2.0 for contrast-enhanced SE T1-weighted images. The mean lesion diameter at gross pathologic examination was 11.0 mm ± 1.8.

Examination of histologic sections from the lesions demonstrated the needle tracts to be filled with recent, or, in some cases, organizing blood clots. The areas surrounding the tracts showed fat or coagulation necrosis in the marrow, and no nuclei were detected either in the hematopoietic cells in the marrow or in the lacunae of cancellous bone. Monocytes and phagocytes were seen infiltrating the necrotic bone tissue (Fig 2, F). Focal hemorrhage was identified in some fields at the margins of the puncture wound with the necrotic bone. The margins of the lesions, while somewhat irregular, were sharp, with viable bone and marrow that generally appeared normal; there was focal hemorrhage at the margins of necrosis in some fields, but little granulation tissue was identified. Reactive bone formation was seen at the margins in a few fields.

Sections from the spinal cord of the pig in which we performed our ablation procedures flush with the posterior vertebral cortex showed focal demyelination and severe diffuse axonal edema in the ventral vertical tracts of the white matter of the spinal cord, with less severe but similar changes in the dorsal spinal tracts and in the axons within the white matter of the nerve root. Anterior horn neurons exhibited diffuse necrosis and appeared swollen and rounded, with loss of nuclei (Fig 5, F) and coarse granular precipitates within the cytoplasm. Posterior horn neurons were severely degenerated or necrotic. There was no appreciable gliosis. These changes, while more advanced on the right side where the RF electrode was introduced, were also clearly evident on the left side.

In the two ablation procedures performed within the vertebral pedicles, the active tip of the RF electrode was kept at a distance from the lateral boundary of the spinal canal so that the ipsilateral exiting nerve root was barely within the circumference of the hypointense thermal lesion "core" area in one ablation (Fig 3, D) and within the hyperintense rim of the thermal lesion in the other (Fig 3, A). The pig in the former ablation limped on the ipsilateral limb for few days, whereas the pig in the latter ablation developed mild weakness of both hind limbs that improved rapidly and seemed to be due to postprocedural pain rather than nerve root involvement. Neurohistologic examination of these pigs revealed mild axonal edema in the ipsilateral nerve roots (Fig 4, C), while the axons in all tracts within the white matter of the cord appeared normal. Gliosis of the lateral portion of the nerve root with focal adhesion to the dura mater was also noted in the pig that limped on the ipsilateral limb (Fig 4, E, F).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Minimally invasive therapy has been an increasingly important topic in the radiology literature during the past several years. The field of interventional radiology has broadened to include almost every body organ. Most image-guided percutaneous surgeries of the spine are focused on laser disk decompression (18,19), pain management procedures such as periradicular infiltration and facet joint block (20), and percutaneous vertebroplasty (21,22).

In addition to being the most common site for skeletal metastases (23), the spine is also the seat for a number of primary benign and malignant tumors, such as hemangioma, osteoid osteoma, osteoblastoma, malignant fibrous histiocytoma, and osteo- and chondrosarcoma (24).

The mission of spinal tumor treatment is to prevent complications, such as vertebral collapse, spinal instability, and progressive neurologic compromise (25). In advanced cases, palliation of pain to enhance the patient’s quality of life is the primary goal of treatment. Rosenthal and colleagues first described the use of RF thermal ablation to treat osteoid osteomas of the extremities (26,27) and reported an outcome essentially equivalent to surgical excision of extraspinal osteoid osteomas (28). Dupuy and colleagues (15) recently reported the successful treatment of one patient with vertebral body metastasis and another patient with a vertebral osteoid osteoma by using RF ablation with CT guidance. Initial clinical success in CT-guided alcohol ablation of symptomatic vertebral hemangiomas was also reported by Goyal et al (13) and then Doppman et al (14).

The main advantage of using MR imaging in RF thermal ablation is being able to depict the immediate effect of therapy, thus defining the treatment end point without moving the patient from the interventional suite. This is an attribute of MR imaging that cannot be duplicated reliably with any other current imaging modality, giving MR imaging an edge over other imaging modalities for monitoring vertebral thermal ablation procedures. In the vertebra, this is of particular importance to ensure that a developing thermal lesion does not approach the adjacent neural structures.

During the guidance phase of percutaneous vertebral intervention, MR imaging offers additional advantages when compared with other imaging modalities. These include (a) the ability to guide needle insertion through the pedicle with continuous, near real-time imaging so that the instrument can be redirected to avoid critical structures in a time-efficient manner, and (b) the multiplanar imaging capabilities that ensure precise needle centralization along the transverse and craniocaudal dimensions of the target. In addition, imaging in any arbitrary plane allows the needle trajectory to be modified according to the individual case. While accurate electrode placement is certainly possible with other imaging modalities—CT in particular—the ability to monitor the electrode position continuously in any arbitrary image plane makes insertion a simple and time-efficient procedure with MR imaging guidance.

In our series, we used MR imaging both to guide the procedures and to monitor the effect of the ablation. We randomized our ablation sites to involve various lumbar vertebral levels and numerous locations within the vertebra so that the full range of technical challenges, procedure complications, and animal morbidity could be evaluated. Although the porcine paraspinal muscles are bulky enough to facilitate a safe lateral vertebral approach, we chose to pursue the transpedicular route in most of our procedures to simulate future ablation procedures in humans.

By using fast gradient-echo sequences, particularly the true FISP technique, with frequent shifts between the transverse and sagittal imaging planes, the needle could be readily and precisely positioned within the desired ablation site. It is worth noting that when using the true FISP sequence, the needle plane must be as close to the center of the coil as possible to avoid masking the plane of interest with the band artifacts inherent to the true FISP sequence.

Induced thermal lesions in the vertebrae displayed the same MR signal intensity characteristics described in previous studies involving the pancreas (29), liver (9,10), and brain (3,30). The gross pathologic and histologic appearances of the induced vertebral thermal lesions corresponded with published data on thermal ablation procedures in other organs (9,29,31). Corroborating previous reports (29,32), our lesion-to-vertebra CNR measurements showed that thermal lesions were significantly better depicted on contrast-enhanced SE T1-weighted images.

Fast SE T2-weighted images offered the most precise assessment of thermal lesion sizes in comparison to gross pathologic findings. This finding supports previously published data on hepatic thermal lesion assessment (9) but varies from the lesion size measurements in previous pancreatic RF ablation procedures (29). As described by previous investigators (29,33), thermal lesions continued to grow during the follow-up period, and we recorded 37% (11.2 mm ± 1.9 from 8.2 mm ± 1.5), 19% (11.3 mm ± 2.4 from 9.5 mm ± 1.4), and 29% (10.5 mm ± 1.9 from 8.2 mm ± 1.5) increases in mean lesion diameter on T2-weighted, contrast-enhanced T1-weighted, and STIR images, respectively, in the 1-week follow-up group.

Much of the basic investigation of the different parameters that controlled RF ablation procedures was performed by Goldberg et al (34–37), who recommended the measurement of RF current (in milliamperes) rather than power (in watts) as the more important factor for predicting thermal lesion size during ablation procedures (34). The RF current and tissue impedance profiles created from recordings during our ablation procedures showed that the mean tissue impedance decreased from approximately 97 before ablation to a rather constant value ranging between approximately 85 and 87 during RF application. After reaching its maximum level during the 1st minute, the RF current (representing the deployed energy) continued to decline slowly as the ablation progressed. These findings corroborate the earlier work of Chung et al (38), with the only difference in our current study being a mild rise of the mean RF current during the last minute of ablation.

The average RF current required to maintain the electrode tip temperature at 90°C ± 2 for the 10-minute duration of our vertebral ablation procedures ranged between approximately 285 and 345 mA, which is less than the values reported for hepatic (10,39) and pancreatic (29) ablation procedures.

In their ex vivo experiments, Dupuy et al (15) observed decreased heat transmission in cancellous bone and an insulative effect of cortical bone. They also concluded, on the basis of temperature recordings within the epidural space during in vivo ablation procedures, that a margin of safety is provided when preserved cancellous or cortical bone is present.

On the basis of our observations from the initial postablation and follow-up MR images and gross pathologic and histologic findings, we tend to be more conservative regarding the protective effect of intact bone positioned between the RF electrode and the vital tissues. In our series, induced thermal lesions created in the vicinity of the anterior and/or lateral vertebral cortices consistently evolved into their full oval configuration, regardless of the vertebral boundaries. In the two ablation procedures we performed within the vertebral pedicles, histologic examination showed mild axonal edema in the ipsilateral nerve roots in both pigs, in addition to nerve root adhesion to the dura mater in one. By placing the RF electrode within the vertebral body with an anteroposterior orientation, we could safely create an intravertebral thermal lesion while keeping the electrode tangential to the spinal canal with its closest point located 2 mm from the anterolateral aspect of the spinal canal and its extreme tip located 6 mm from the center of the canal. The spinal cord and nerve root in this pig appeared completely normal at histologic examination. The safety of the ablation in this case is probably attributed to the heat dissipation effect created by the flowing blood within the epidural vessels, as well as by the cerebrospinal fluid pulsations. However, when we placed the entire active tip of the RF electrode directly over the intact posterior vertebral body cortex, the animal developed immediate postablation paraplegia, and histologic examination of the spinal cord of this pig confirmed the permanent nature of the cord damage.

As a result of this experience, we propose that RF thermal ablation is most suitable for treating circumscribed tumors confined to the vertebra in which the thermal lesion can be planned to encompass the entire tumor without entering the spinal canal and without placing the RF electrode directly adjacent to the posterior cortex. Treatment of tumors that erode into the spinal canal could present even greater risk of spinal cord injury from direct RF heating or from developing edema within the treated tumor, and, on the basis of our observations, should be avoided. Lesions violating the anterior and/or lateral cortices will probably pose less of a technical challenge.

In conclusion, we accept the hypothesis that MR imaging–guided RF thermal ablation of the vertebrae is feasible in porcine models, that the safety of the procedure depends on the location of the ablation within the vertebra, and that MR imaging allows accurate monitoring of induced vertebral thermal lesion size and shape. Ablation procedures performed away from neural elements were generally safe. Pedicular ablation procedures performed adjacent to the nerve root resulted in radiculopathy, whereas ablations performed directly over the posterior vertebral cortex resulted in severe spinal cord damage.

Practical applications: When considering the feasibility and safety of RF thermal ablation of the vertebrae in animal models, together with the reported clinical success in controlling localized malignancies in various body organs by using RF ablation, the technique appears to be a promising method for the future treatment of spinal tumors.

Since metastases are much more common than primary benign or malignant tumors of the spine, patients with metastatic disease are expected to constitute the primary candidates for future clinical studies on investigation of vertebral RF ablation.

Radiation therapy has been well established as the first line of noninvasive treatment for painful spinal metastases. Often, patients suffer recurrent vertebral metastases after they have already received their maximum radiation dose. We propose that it is this group of patients who might benefit from the new mode of treatment (Fig 8).

View larger version (29K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Flowchart highlights the proposed role of RF thermal ablation among the major treatment options currently available for patients with spinal metastases.

The rationale for RF ablation treatment of circumscribed benign spinal lesions, such as osteoid osteoma, hemangioma, and osteoblastoma, will vary, whereas RF thermal ablation has a great potential to become the immediate, minimally invasive alternative to surgical resection, aiming for complete cure.

Finally, an issue of concern when applying RF ablation to the vertebrae in the clinical setting is that the created area of vertebral necrosis can cause vertebral collapse when a patient resumes an upright position. Therefore, the treatment may need to be combined with external support and/or percutaneous vertebroplasty.

	ACKNOWLEDGMENTS

 
The authors thank Nanette Kleinman, Tami McCourt, Jean Janesz, and Markeya Owens for their assistance with animal observation and anesthesia; Claudia Hillenbrand, PhD, for helpful conversations and help with manuscript preparation; Bonnie Hami, MA, and Elena DuPont for assistance in manuscript preparation; and all the members of the Interventional MRI research group and the imaging laboratory at CWRU for their outstanding support.

	FOOTNOTES

 
² Current address: Dept of Diagnostic Radiology, Cairo Univ Hospitals, Egypt.

³ Current address: Dept of Diagnostic Radiology, Univ of Ulm, Germany.

Abbreviations: CNR = contrast-to-noise ratio, FISP = fast imaging with steady-state precession, RF = radio frequency, SE = spin echo, STIR = short inversion time inversion-recovery

Author contributions: Guarantors of integrity of entire study, S.G.N., J.S.L.; study concepts, J.S.L., S.G.N., J.L.D.; study design, S.G.N., J.S.L., A.J.A.; literature research, I.C.S.M., S.G.N.; experimental studies, S.G.N., I.C.S.M., A.J.A.; data acquisition, I.C.S.M., S.G.N., S.N.E.; data analysis/interpretation, I.C.S.M., S.G.N., J.S.L., S.N.E.; statistical analysis, S.G.N.; manuscript preparation, S.G.N., I.C.S.M., S.N.E.; manuscript definition of intellectual content, J.S.L., J.L.D.; manuscript editing, J.S.L., S.G.N.; manuscript revision/review, J.S.L., A.J.A.; manuscript final version approval, all authors.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Zlotta AR, Wildschutz T, Raviv G, et al. Radiofrequency interstitial tumor ablation (RITA) is a possible new modality for treatment of renal cancer: ex vivo and in vivo experience. J Endourol 1997; 11:251-258.[Medline]
2. Rossi S, Di Stasi M, Buscarini E, et al. Percutaneous RF interstitial thermal ablation in the treatment of hepatic cancer. AJR Am J Roentgenol 1996; 167:759-768.[Abstract/Free Full Text]
3. Anzai Y, Lufkin R, DeSalles A, Hamilton DR, Farahani K, Black KL. Preliminary experience with MR-guided thermal ablation of brain tumors. AJNR Am J Neuroradiol 1995; 16:39-48; discussion 49–52.[Abstract]
4. Zlotta AR, Peny MO, Matos C, Schulman CC. Transurethral needle ablation of the prostate: clinical experience in patients in urinary acute retention. Br J Urol 1996; 77:391-397.[Medline]
5. Tillotson CL, Rosenberg AE, Rosenthal DI. Controlled thermal injury of bone: report of a percutaneous technique using radiofrequency electrode and generator. Invest Radiol 1989; 24:888-892.[CrossRef][Medline]
6. Francica G, Marone G. Ultrasound-guided percutaneous treatment of hepatocellular carcinoma by radiofrequency hyperthermia with a ’cooled-tip needle’: a preliminary clinical experience. Eur J Ultrasound 1999; 9:145-153.[CrossRef][Medline]
7. Trubenbach J, Konig CW, Duda SH, et al. Percutaneous radiofrequency ablation of hepatic neoplasms using a "cluster" electrode—first clinical results. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2000; 172:905-910[German].[Medline]
8. Huppert PE, Trubenbach J, Schick F, Pereira P, Konig C, Claussen CD. MRI-guided percutaneous radiofrequency ablation of hepatic neoplasms: first technical and clinical experiences. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2000; 172:692-700[German].[Medline]
9. Boaz TL, Lewin JS, Chung YC, Duerk JL, Clampitt ME, Haaga JR. MR monitoring of MR-guided radiofrequency thermal ablation of normal liver in an animal model. J Magn Reson Imaging 1998; 8:64-69.[Medline]
10. Lewin JS, Connell CF, Duerk JL, et al. Interactive MRI-guided radiofrequency interstitial thermal ablation of abdominal tumors: clinical trial for evaluation of safety and feasibility. J Magn Reson Imaging 1998; 8:40-47.[Medline]
11. Cove JA, Taminiau AH, Obermann WR, Vanderschueren GM. Osteoid osteoma of the spine treated with percutaneous computed tomography-guided thermocoagulation. Spine 2000; 25:1283-1286.[CrossRef][Medline]
12. Gangi A, Dietemann JL, Guth S, et al. Percutaneous laser photocoagulation of spinal osteoid osteomas under CT guidance. AJNR Am J Neuroradiol 1998; 19:1955-1958.[Abstract]
13. Goyal M, Mishra NK, Sharma A, Gaikwad SB, Mohanty BK, Sharma S. Alcohol ablation of symptomatic vertebral hemangiomas. AJNR Am J Neuroradiol 1999; 20:1091-1096.[Abstract/Free Full Text]
14. Doppman JL, Oldfield EH, Heiss JD. Symptomatic vertebral hemangiomas: treatment by means of direct intralesional injection of ethanol. Radiology 2000; 214:341-348.[Abstract/Free Full Text]
15. Dupuy DE, Hong R, Oliver B, Goldberg SN. Radiofrequency ablation of spinal tumors: temperature distribution in the spinal canal. AJR Am J Roentgenol 2000; 175:1263-1266.[Free Full Text]
16. Dondelinger RF, Ghysels MP, Brisbois D, et al. Relevant radiological anatomy of the pig as a training model in interventional radiology. Eur Radiol 1998; 8:1254-1273.[CrossRef][Medline]
17. Sisson S. Porcine osteology. In: Getty R, eds. Sisson and Grossman’s the anatomy of the domestic animals. 4th ed. Philadelphia, Pa: Saunders, 1975; 1216-1252.
18. Gevargez A, Groenemeyer DW, Czerwinski F. CT-guided percutaneous laser disc decompression with Ceralas D, a diode laser with 980-nm wavelength and 200-microm fiber optics. Eur Radiol 2000; 10:1239-1241.[CrossRef][Medline]
19. Botsford JA. Radiological considerations: patient selection for percutaneous laser disc decompression. J Clin Laser Med Surg 1994; 12:255-259.[Medline]
20. Gangi A, Dietemann JL, Mortazavi R, Pfleger D, Kauff C, Roy C. CT-guided interventional procedures for pain management in the lumbosacral spine. RadioGraphics 1998; 18:621-633.[Abstract]
21. Mathis JM, Barr JD, Belkoff SM, Barr MS, Jensen ME, Deramond H. Percutaneous vertebroplasty: a developing standard of care for vertebral compression fractures. AJNR Am J Neuroradiol 2001; 22:373-381.[Free Full Text]
22. Levine SA, Perin LA, Hayes D, Hayes WS. An evidence-based evaluation of percutaneous vertebroplasty. Manag Care 2000; 9:56-60, 63.
23. Hatrick NC, Lucas JD, Timothy AR, Smith MA. The surgical treatment of metastatic disease of the spine. Radiother Oncol 2000; 56:335-339.[CrossRef][Medline]
24. Bridge JA, Schwartz HS, Neff JR. Specific malignancies, sarcomas of bone. In: Abeloff MD, eds. Clinical oncology. 2nd ed. New York, NY: Churchill Livingstone, 2000; 2160-2272.
25. Merk H, Koch H, Liebau C, Baltzer A, Dragendorf L, Grasshoff H. Implantation of a Harms titanium mesh cylinder for vertebral body replacement in spinal metastases. Z Orthop Ihre Grenzgeb 2000; 138:169-173[German].[CrossRef][Medline]
26. Rosenthal DI, Alexander A, Rosenberg AE, Springfield D. Ablation of osteoid osteomas with a percutaneously placed electrode: a new procedure. Radiology 1992; 183:29-33.[Abstract/Free Full Text]
27. Rosenthal DI, Springfield DS, Gebhardt MC, Rosenberg AE, Mankin HJ. Osteoid osteoma: percutaneous radio-frequency ablation. Radiology 1995; 197:451-454.[Abstract/Free Full Text]
28. Rosenthal DI, Hornicek FJ, Wolfe MW, Jennings LC, Gebhardt MC, Mankin HJ. Percutaneous radiofrequency coagulation of osteoid osteoma compared with operative treatment. J Bone Joint Surg Am 1998; 80:815-821.[Abstract/Free Full Text]
29. Merkle EM, Haaga JR, Duerk JL, Jacobs GH, Brambs HJ, Lewin JS. MR imaging-guided radio-frequency thermal ablation in the pancreas in a porcine model with a modified clinical C-arm system. Radiology 1999; 213:461-467.[Abstract/Free Full Text]
30. De Salles AA, Brekhus SD, De Souza EC, et al. Early postoperative appearance of radiofrequency lesions on magnetic resonance imaging. Neurosurgery 1995; 36:932-936; discussion 936–937.[Medline]
31. McGahan JP, Brock JM, Tesluk H, Gu WZ, Schneider P, Browning PD. Hepatic ablation with use of radio-frequency electrocautery in the animal model. J Vasc Intervent Radiol 1992; 3:291-297.[Medline]
32. Aschoff AJ, Rafie N, Jesberger JA, Duerk JL, Lewin JS. Thermal lesion conspicuity following interstitial radiofrequency thermal tumor ablation in humans: a comparison of STIR, fast spin-echo T2-weighted, and contrast-enhanced T1-weighted MR images at 0.2 T. J Magn Reson Imaging 2000; 12:584-589.[CrossRef][Medline]
33. Tacke J, Adam G, Speetzen R, et al. MR-guided interstitial cryotherapy of the liver with a novel, nitrogen-cooled cryoprobe. Magn Reson Med 1998; 39:354-360.[Medline]
34. Goldberg SN, Gazelle GS, Solbiati L, Rittman WJ, Mueller PR. Radiofrequency tissue ablation: increased lesion diameter with a perfusion electrode. Acad Radiol 1996; 3:636-644.[CrossRef][Medline]
35. Goldberg SN, Gazelle GS, Halpern EF, Rittman WJ, Mueller PR, Rosenthal DI. Radiofrequency tissue ablation: importance of local temperature along the electrode tip exposure in determining lesion shape and size. Acad Radiol 1996; 3:212-218.[CrossRef][Medline]
36. Goldberg SN, Gazelle GS, Dawson SL, Rittman WJ, Mueller PR, Rosenthal DI. Tissue ablation with radiofrequency: effect of probe size, gauge, duration, and temperature on lesion volume. Acad Radiol 1995; 2:399-404.[CrossRef][Medline]
37. Goldberg SN, Hahn PF, Tanabe KK, et al. Percutaneous radiofrequency tissue ablation: does perfusion-mediated tissue cooling limit coagulation necrosis? J Vasc Interv Radiol 1998; 9((1 pt 1)):101-111.[Medline]
38. Chung YC, Duerk JL, Lewin JS. Generation and observation of radio frequency thermal lesion ablation for interventional magnetic resonance imaging. Invest Radiol 1997; 32:466-474.[CrossRef][Medline]
39. Solbiati L, Goldberg SN, Ierace T, et al. Hepatic metastases: percutaneous radio-frequency ablation with cooled-tip electrodes. Radiology 1997; 205:367-373.[Abstract/Free Full Text]

This article has been cited by other articles:

				A. Adachi, T. Kaminou, T. Ogawa, T. Kawai, Y. Takaki, K. Sugiura, Y. Ohuchi, and M. Hashimoto Heat Distribution in the Spinal Canal during Radiofrequency Ablation for Vertebral Lesions: Study in Swine Radiology, May 1, 2008; 247(2): 374 - 380. [Abstract] [Full Text] [PDF]

				C. P. Cantwell, J. Kerr, J. O'Byrne, and S. Eustace MRI Features After Radiofrequency Ablation of Osteoid Osteoma with Cooled Probes and Impedance-Control Energy Delivery. Am. J. Roentgenol., May 1, 2006; 186(5): 1220 - 1227. [Abstract] [Full Text] [PDF]

				C. P. Cantwell, J. O'Byrne, and S. Eustace Radiofrequency ablation of osteoid osteoma with cooled probes and impedance-control energy delivery. Am. J. Roentgenol., May 1, 2006; 186(5 Suppl): S244 - S248. [Abstract] [Full Text] [PDF]

				S. G. Nour, J. S. Lewin, M. Gutman, C. Hillenbrand, F. K. Wacker, J. W. Wong, I. C. Mitchell, C. B. Armstrong, M. M. Hashim, J. L. Duerk, et al. Percutaneous MR Imaging- guided Radiofrequency Interstitial Thermal Ablation of Tongue Base in Porcine Models: Implications for Obstructive Sleep Apnea Syndrome Radiology, February 1, 2004; 230(2): 359 - 368. [Abstract] [Full Text] [PDF]

				O. Schaefer, C. Lohrmann, M. Markmiller, P. Uhrmeister, and M. Langer Combined Treatment of a Spinal Metastasis with Radiofrequency Heat Ablation and Vertebroplasty Am. J. Roentgenol., April 1, 2003; 180(4): 1075 - 1077. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2242011269v1 224/2/452    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend