Originally published as Ann Surg Oncol Early Release 10.1245/ASO.2004.09.012 on March 15, 2004
Annals of Surgical Oncology 11:420-425 (2004)
© 2004 Society of Surgical Oncology
Variability of Size and Shape of Necrosis Induced by Radiofrequency Ablation in Human Livers: A Volumetric Evaluation
Dirk L. Stippel, MD,
Hans G. Brochhagen, MD,
Mahesh Arenja,
Jens Hunkemöller,
Arnulf H. Hölscher, MD, FACS, FRCS and
K. Tobias E. Beckurts, MD
From the Department of Visceral and Vascular Surgery (DLS, MA, JH, AHH, KTEB) and Institute of Radiology (HGB), University Of Cologne, Cologne, Germany.
Correspondence: Address correspondence and reprint requests to: Dirk L. Stippel, MD, Department of Visceral and Vascular Surgery, University of Cologne, Joseph-Stelzmann-Straße 9, 50931 Cologne, Germany; Fax: 49-221-4786258; E-mail: Dirk.Stippel{at}uni-koeln.de
 |
ABSTRACT
|
|---|
Background: Definite size and shape of radiofrequency-induced ablations (RFAs) cannot be evaluated intraoperatively. Instead, surgeons choose a radiofrequency device that is supposed to cause a necrosis of a determined size greater than the malignant lesion. The aim of this study was to measure the variability of the induced necroses postoperatively and to define a reproducible ablation volume in human liver.
Methods: In 24 patients, 34 RFA procedures were performed with single applications of the device. The deployment was 3 cm (n = 16), 4 cm (n = 5), or 5 cm (n = 13). The induced necroses were analyzed by volumetric reconstructions of computed tomography (CT) scans. Measured volumes were compared with the expected volumes. Furthermore, the shape of the necrosis was classified according to an index of the diameters.
Results: The measured volumes of postoperative necroses were 14 ± 8 cm3 (deployment, 3 cm), 24 ± 12 cm3 (4 cm), and 45 ± 42 cm3 (5 cm). The diameter of a sphere fitted into the necroses reached 2.9 ± .5 cm (3 cm), 3.5 ± .7 cm (4 cm), and 4.1 ± 1.1 cm (5 cm), at P < .02, significantly smaller than the deployment. The classification of shapes yielded a spherical shape (n = 14), a teardrop shape (n = 13), or an irregular shape (n = 7). The energy consumption was 2.1 ± 1.5 kJ/cm3 (3 cm), 2.6 ± .5 kJ/cm3 (4 cm), or 3.5 ± 2.0 kJ/cm3 (5 cm).
Conclusions: The diameter of RFA-induced liver necrosis is significantly smaller than expected from needle deployment, especially with full-needle deployment. The shape of the lesion differs in more than half of the cases from the anticipated spherical pattern. The upper limit for reproducible necrosis induction is a tumor diameter of 3.4 cm.
Key Words: Computed tomography • Liver • Radiofrequency ablation • Shape • Volume
|
INTRODUCTION
|
|---|
Size and shape of radiofrequency-induced ablations (RFAs) of liver tissue cannot be objectively evaluated intraoperatively.1,2 Instead, surgeons plan the coagulation of a liver lesion according to its size by choosing a radiofrequency device that is supposed to yield a lesion of a certain size according to its deployment. Dual-phase, contrast-enhanced, helical computed tomography (CT) scans allow a reproducible and precise postoperative assessment of tissue coagulated by RFA. This has been shown by correlation of CT and histology in a porcine model3 and in humans.4 Volumetric reformation of the CT data allows a precise measurement of volume and a visualization of shape.5 In contrast, most of the published data about volumes after RFA have been calculated according to a formula that is based on the assumption that an RFA lesion has the shape of a rotational ellipsoid or a sphere. Thus, the volumes were calculated after measuring one to three diameters.
Major deviations of the ablated tissue volume from the assumed spherical shape have been demonstrated repeatedly.6–9 When RFA was done without a Pringle maneuver before the year 2000, a cloverleaf-like shape was often seen. In these cases, the necrosis was achieved only in direct vicinity to the prongs of the RFA device. The limited power that could be supplied by the generators was seen as the reason for this failure to achieve a homogeneous necrosis.6 The "heat sink" effect of large vessels is another reason why the resulting shape of the ablation can differ from a sphere.10 More recent evaluations of volumes of RFA necrosis in a porcine model described homogeneous necrosis with use of a state-of-the-art generator and a modern clustered device.11,12 These data suggest an influence of vascular occlusion on the size and shape of the induced lesions. Without a Pringle maneuver, the lesions tended to be of cylindrical shape and were two to three times smaller. All volumes were smaller than expected from the deployment of the RFA device. Volumetric data evaluating these new RFA devices in clinical practice are still lacking. These data are needed to provide surgeons with evidence for the reliable planning of ablation procedures utilizing a modern radiofrequency device.
The intention of this study was to investigate whether RFA lesions in human livers, with use of a state-of-the-art RFA device and generator, are homogeneous and which volume and shape of necrosis can be expected from a specific device deployment.
|
PATIENTS AND METHODS
|
|---|
Seventy-six liver lesions were treated by intraoperative RFA in 27 patients between June 2001 and August 2002. The diagnosis was colorectal liver metastasis in 24 cases. In all of these 24 patients a combined procedure (resection and ablation) was performed. Thirty-four lesions met the inclusion and exclusion criteria for this study. The inclusion criteria were defined as colorectal liver metastasis and single application of the RFA device at a specific lesion. Exclusion from the analysis was due to any of the following: multiple application of the RFA device at one site, location of ablation site in liver segment I, or use of a Pringle maneuver. Ablation was performed with the Starburst XL device (RITA Medical Systems, Mountain View, CA), deployed to an ablation diameter of 3 to 5 cm as needed. This is a multiarray system with nine prongs. The generator (model 1500) is capable of a maximum power output of 150 W. Two grounding pads were used in each case.
The generator settings and the ablation timing were set according to the recommendations of the manufacturer. If the cool-down procedure suggested an insufficient coagulation (temperature below 60°C), the device was rotated and a second ablation was performed. These cases were excluded from analysis. All patients were preoperatively examined by CT. Intraoperative device placement was done under ultrasound guidance. The Sonoline Sienna system, with a 7.5-MHz intraoperative probe (Siemens Medical Systems, Issaquah, WA), was used. The size of the lesion intended to treat was determined by measuring the maximum diameter in two planes with intraoperative ultrasound. The settings of the generator, the power output, and the temperatures measured by the thermistors of the device were recorded with use of an interface of the generator connected to a personal computer on which a dedicated database (Data Collection Software version 1.05, RITA Medical Systems, Mountain View, CA) had been implemented. Every second, the values for impedance, power output (W), and temperature were recorded for quality control. These data were used to calculate the energy consumption according to the equation J = W * s, where J is energy consumption in joules, W is power output in watts, and s is time in seconds.
The success of all ablations was evaluated by dual-phase, contrast-enhanced, helical CT performed between days 7 and 10 after the procedure (Volume Zoom or Somatom Plus 4, Siemens AG, Erlangen, Germany). This point of time was chosen to avoid enhancement around the lesion, as seen within the first 3 days.4,13 The timing between contrast medium administration and scan time was done by bolus tracking. The region of interest (ROI) for bolus tracking was defined as the descending aorta at the height of the diaphragm. The arterial phase was started when the contrast at the ROI reached 100 HU. The portal venous phase was started 60 seconds later. The technical configurations for the scanners were as follows: volume zoom, detector configuration 4 x 2.5 mm/pitch 15/increment 4 mm; Somatom Plus 4, slice thickness 5 mm/pitch 1.5/increment 4 mm. The effective slice thickness of the resulting images in the transverse plane was 4 mm for both scanners. The volumetric evaluation was carried out with use of the original CT information, which was transferred to a Leonardo work station and analyzed with the software package VA 40C (Siemens AG).
Segmentation of the ablation-induced lesions was performed stepwise. During the first step, a region of interest was defined in each slice by manually tracking the approximate borders of the lesion. In the second step, the precise border of the lesion was determined by a filter algorithm provided by the software package. The filter was based on density differences between the ablated tissue and the liver tissue. Thereafter, multiplanar reformation (MPR) in three dimensions and volumetric evaluation were performed automatically by the software. The lesions were evaluated by measurement of the total volume and the maximum diameter in each dimension. Additionally, an evaluation of the shape was done by visualization of the resulting shapes in the MPR mode.
Figure 1 shows a typical reformation in three perpendicular planes. Cursor movement in one of the dimensions leads to corresponding movements within the other two dimensions, thus enabling good visualization of the shape. For visualization purposes, the segmented lesions were filled with a monochrome color. The visual control was done also to identify possible "heat sink" effects that might result in an incomplete ablation. The expected ablation volumes were calculated as follows: volume = 
/ 6 * deployment3. The diameter of a sphere that could be fitted into a measured volume was calculated accordingly. In order to classify the deviation from the spherical shape, two indices were calculated: the fraction of the largest and smallest diameter (SIl) and the fraction of the middle and smallest diameter (SIm). A lesion was called spherical when SIl and SIm were 
1.5. A lesion was called teardrop when SIl was >1.5 and SIm was 1.5. A lesion was called irregular when SIl and SIm were >1.5. All calculations were performed with SPSS for Windows, version 10.0.7 (SPSS, Chicago). Values are given as mean ± standard deviation, unless mentioned otherwise. Normal distribution of values was assessed by the Kolmogorov-Smirnov test. Measured and expected values were compared with the two-sided t-test for one sample. A P value of <.05 was considered to be significant. The lower limit of the 95% confidence interval of the mean was calculated when appropriate to yield information on the minimum size to be expected.
View larger version (60K):
[in this window]
[in a new window]
|
FIG. 1. Segmentation of a radiofrequency ablation (RFA)-induced necrosis and visualization in three perpendicular planes. Example of a necrosis that was classified as a spherical shape (deployment, 5 cm; volume, 40 cm3 [62% of expected volume]). Upper row: native computed tomography (CT) scans; lower row: the same CT scans with a superimposed sphere 5 cm in diameter (transparent green). The position of the corresponding planes is indicated by vertical and horizontal lines (yellow) and a masking filter for the zone of necrosis (yellow).
|
|
|
RESULTS
|
|---|
Clinical Results
Liver resection and RFA were performed without any occurrences of death. One patient had biliary leakage due to RFA that was successfully treated by temporary endoscopic placement of a bile duct endoprosthesis. Clinical follow-up (mean, 19 ± 9 months) and radiological imaging revealed local recurrence in five lesions (15%; three patients). The diameter of the lesions that showed a local recurrence was significantly larger than that of lesions without recurrence (2.9 ± .6 cm versus 1.8 ± .8 cm; P < .005). Three of the five local recurrences had a diameter of more than 3 cm.
Quality Control of RFA Procedure
The mean electrical impedance was independent of localization within the liver and decreased with the increasing length of the RFA device: 67 ± 10 ohm (3 cm), 65 ± 21 ohm (4 cm), and 58 ± 11 ohm (5 cm). The variation in energy consumption was only minor after standardization for ablated tissue volume: 2.1 ± 1.5 kJ/cm3 (3 cm), 2.6 ± .6 kJ/cm3 (4 cm), and 3.4 ± 2.7 kJ/cm3 (5 cm). All cool-down cycles yielded temperatures above 60°C on all prongs.
Volume
The 34 ablation-induced lesions displayed a sharp border on the contrast-enhanced CT scans. The density within the lesion was homogeneous (44.5 ± 6.2; range, 34 to 59 Hounsfield units). This homogeneous density facilitated a reproducible recognition of the lesion by the filter algorithm. The expected volumes of ablated tissue were 14 cm3 (3-cm deployment), 33 cm3 (4-cm deployment), and 65 cm3 (5-cm deployment). The postoperatively measured volumes were 13.5 ± 8.0 cm3, 23.8 ± 12.3 cm3, and 45.2 ± 42.2 cm3, respectively. The distribution of the values is shown in Fig. 2. The calculated diameters of spheres that could be fitted into these volumes were 2.9 ± .5 cm (3-cm deployment), 3.5 ± .7 cm (4-cm deployment), and 4.1 ± 1.1 cm (5-cm deployment). With a 5-cm deployment, the diameter of a sphere fitted into the necrosis was significantly smaller than expected from the deployment (P < .02). The lower limits of the 95% confidence interval for the diameter of reproducible necrosis induction were 2.6 cm (3 cm), 2.6 cm (4 cm), and 3.4 cm (5 cm). In one case each the calculated diameter of a sphere fitted into the volume was equal or smaller than the diameter of the tumor; both cases were among the five cases with local recurrence. The localization within the liver segments did not influence the volume of tissue necrosis that could be achieved. An analysis comparing the measured volumes grouped by the liver segment with the expected, calculated volumes according to the deployment did not shown any differences between ablation-induced necrosis in liver segments with major vessels (segments IV, V, VIII) or without major vessels. Table 1 summarizes the parameters of an ablation-induced necrosis that can be expected as a minimum result.
View larger version (24K):
[in this window]
[in a new window]
|
FIG. 2. Volume of RFA-induced necrosis grouped by deployment of the device. Values are median, range, and interquartile range.
|
|
Shape
The mean values of the largest and smallest diameters were 5.1 ± 1.4 cm versus 3.0 ± .8 cm (3-cm deployment), 7.1 ± 2.6 cm versus 3.9 ± .8 cm (4-cm deployment), and 6.6 ± 2.2 cm versus 4.5 ± 1.8 cm (5-cm deployment). The upper range of the maximum diameter exceeded all expected values. The reason became clearly visible when the reconstructed dimensions of the CT image were viewed. Fig. 3 gives an example of an ablation where the shape is a stretched teardrop along the axis of the puncture. This shape is the usual cause in cases in which there is a great difference between largest and smallest diameter. The typical cause for an irregular shape is the vicinity of the liver surface; an example is shown in Fig. 3. The localization of the lesion within a specific liver segment did not correlate with a specific shape. An example of a spherical lesion is shown in Fig. 1. The classification of the lesion according to the indices SIl and SIm is shown in Table 2. The volume of lesions classified as spherical was 74 ± 54% of the expected volume; that of teardrop lesions, 88 ± 61% of the expected volume; and that of irregular lesions, 87 ± 64% of the expected volume.
View larger version (65K):
[in this window]
[in a new window]
|
FIG. 3. RFA-induced necrosis and visualization in three perpendicular planes (superimposed sphere, 5 cm in diameter [transparent green]; area of necrosis [yellow]). Upper row: example of a necrosis that was classified as a teardrop shape (deployment, 5 cm; volume, 25 cm3 [38% of expected volume]). Lower row: example of a necrosis that was classified as an irregular shape (deployment, 5 cm; volume, 17 cm3 [27% of expected volume]). The track ablation caused a significant part of the total necrosis.
|
|
There was only one lesion in which a heat sink effect caused incomplete necrosis of the targeted tissue volume. A main branch of the right portal vein running through the tissue volume shielded a small section of liver tissue from coagulation. This was the only case in which a heat sink effect could be verified. In many lesions, especially close to the main hepatic veins, the vessels stayed patent, as visualized with the intravenous contrast. However, there was no uptake of intravenous contrast in the surrounding tissue, which suggests that all tissue surrounding the vessel was coagulated.
Four of the five local recurrences were seen in RFA ablation with a spherical shape and one in an ablation with a teardrop shape. However, there was a bias, because the diameter of lesions treated was larger in the spherical ablations (2.5 cm ± .8 cm) than in the teardrop ablations (1.4 ± .7 cm) or irregular ablations (1.7 ± .5 cm).
|
DISCUSSION
|
|---|
Analysis of local recurrence after RFA of hepatic tumors revealed tumor size as a major factor14 in our initial clinical experience. Only three recurrences among 45 treated lesions were observed in tumors less than 3 cm in diameter, whereas the recurrence rate reached 56% among tumors larger than 3 cm in diameter (median follow-up, 18 months). This observation of a major influence of tumor size on local recurrence has been reported by others with regard to larger series as well.15–17 The cause of this clinical observation can be incorrect positioning of the device, inhomogeneous necrosis, or insufficient size of the necrosis. To analyze the latter two factors, multiplanar reformatted CT is well suited. Utilizing this technique, we found all but one RFA necrosis to be homogeneous. This inhomogeneous lesion was located close to the main branch of the right portal vein, functioning as a heat sink.
This observation of homogeneous necrosis validates an important advance in comparison with earlier studies.7,8 It is most likely due to the increased power output of the generator used these days. Irregular shapes of ablated tissue were seen only close to the liver surface. In some cases, the prongs perforated the surface of the liver and hence the volume of necrosis was smaller. In other cases, additional liver volume became necrotic, likely because of infarction caused by vessels occluded by the RFA procedure. Thus the resulting volumes were larger than planned. This occurred most often in segments VII and VIII. However, the volume of tissue ablation that could be achieved reproducibly was significantly smaller than expected from the deployment of the device. In particular, with the device fully deployed (5 cm), the resulting necrosis was one-third smaller than anticipated. The mean diameter of a sphere fitted into the necrosis was 4.1 cm. The lower limit of the 95% confidence interval was 3.4 cm. Of the five local recurrences in this study, three were in tumors between 3 cm and 4 cm in diameter. The observed volume was still larger than the mean volume coagulated in a porcine model with the same RFA device.12
Half of the zones of necrosis were more or less spherical, while most of the others manifested as a stretched teardrop; this might be caused partially by incomplete spread of the prongs during deployment. When the tumor is more fibrous than the surrounding liver tissue, a deviation of the prongs can easily occur. However, this same cylindrical shape was noted by Chang et al.12 in the porcine model without vascular occlusion and therefore might be secondary to heat dissipation by regional blood flow. Consequently, a puncture in the long axis of a tumor might increase the chance of a complete coagulation, if it is technically possible. The intraoperatively monitored temperature and energy output did not predict the shape of the ablation necrosis. The variability in ablation size and shape must also be taken into consideration when planning an ablation of a tumor by multiple insertions of the needle. Except for dual expansion of the system during a drawback procedure, the certainty about the achieved volume of tissue necrosis decreases exponentially with each expansion of the device. The effects of the Pringle maneuver in porcine models are well described,11,12 but the efficacy of this technical variation in humans still must be evaluated.
|
CONCLUSION
|
|---|
The evaluated RFA system produces homogeneous volumes of tissue necrosis. The general shape of the necrosis is close to a sphere or a teardrop. The necrosis volumes achieved are smaller than expected from the size of the device. The devices currently available allow only for a single puncture ablation of a tumor up to a maximum diameter of 3 cm without a Pringle maneuver. In larger lesions, there is a significant chance of an insufficient ablation effect, possibly resulting in local recurrence. These findings should be taken into consideration when tumor coagulation by RFA is planned. In surgical combined procedures with a curative intent, only lesions up to 3 cm should be treated by RFA with the 5-cm prongs currently available.
|
FOOTNOTES
|
|---|
The confidence interval (CI) for size and shape of necrosis due to radiofrequency-induced ablation (RFA) was analyzed with three-dimensional-reformatted computed tomography. The 95% CI of the diameter was 3.4 cm for 5-cm deployment. RFA with a curative intent is limited by this technical restriction.
Received for publication September 12, 2003.
Accepted for publication December 21, 2003.
|
REFERENCES
|
|---|
- Liu JB, Goldberg BB, Merton DA, Rawool NM, Needleman L, Forsberg F. The role of contrast-enhanced sonography for radiofrequency ablation of liver tumors. J Ultrasound Med 2001; 20: 517–23.[Abstract]
- Leyendecker JR, Dodd GD, Halff GA, et al. Sonographically observed echogenic response during intraoperative radiofrequency ablation of cirrhotic livers: pathologic correlation. AJR Am J Roentgenol 2002; 178: 1147–51.[Abstract/Free Full Text]
- Raman SS, Lu DS, Vodopich DJ, Sayre J, Lassman C. Creation of radiofrequency lesions in a porcine model: correlation with sonography, CT, and histopathology. AJR Am J Roentgenol 2000; 175: 1253–8.[Abstract/Free Full Text]
- Goldberg SN, Gazelle GS, Compton CC, Mueller PR, Tanabe KK. Treatment of intrahepatic malignancy with radiofrequency ablation: radiologic-pathologic correlation. Cancer 2000; 88: 2452–63.[CrossRef][Medline]
- Antoch G, Kuehl H, Vogt FM, Debatin JF, Stattaus J. Value of CT volume imaging for optimal placement of radiofrequency ablation probes in liver lesions. J Vasc Interv Radiol 2002; 13: 1155–61.[Medline]
- Chinn SB, Lee FT Jr, Kennedy GD, et al. Effect of vascular occlusion on radiofrequency ablation of the liver: results in a porcine model. AJR Am J Roentgenol 2001; 176: 789–95.[Abstract/Free Full Text]
- Patterson EJ, Scudamore CH, Owen DA, Nagy AG, Buczkowski AK. Radiofrequency ablation of porcine liver in vivo: effects of blood flow and treatment time on lesion size. Ann Surg 1998; 227: 559–65.[CrossRef][Medline]
- Rossi S, Garbagnati F, De Francesco I, et al. Relationship between the shape and size of radiofrequency induced thermal lesions and hepatic vascularization. Tumori 1999; 85: 128–32.[Medline]
- de Baere T, Denys A, Johns Wood B, et al. Radiofrequency liver ablation: experimental comparative study of water-cooled versus expandable systems. AJR Am J Roentgenol 2001; 176: 187–92.[Abstract/Free Full Text]
- Hansen PD, Rogers S, Corless CL, Swanstrom LL, Siperstien AE. Radiofrequency ablation lesions in a pig liver model. J Surg Res 1999; 87: 114–21.[CrossRef][Medline]
- Wiersinga WJ, Jansen MC, Straatsburg IH, et al. Lesion progression with time and the effect of vascular occlusion following radiofrequency ablation of the liver. Br J Surg 2003; 90: 306–12.[CrossRef][Medline]
- Chang CK, Hendy MP, Smith JM, Recht MH, Welling RE. Radiofrequency ablation of the porcine liver with complete hepatic vascular occlusion. Ann Surg Oncol 2002; 9: 594–8.[Abstract/Free Full Text]
- Morimoto M, Sugimori K, Shirato K, et al. Treatment of hepatocellular carcinoma with radiofrequency ablation: radiologic-histologic correlation during follow-up periods. Hepatology 2002; 35: 1467–75.[CrossRef][Medline]
- Stippel DL, Bohm S, Beckurts KT, Brochhagen HG, Holscher AH. Intraoperative radiofrequency ablation using a 3D navigation tool for treatment of colorectal liver metastases. Onkologie 2002; 25: 346–50.[Medline]
- Bleicher RJ, Allegra DP, Nora DT, Wood TF, Foshag LJ, Bilchik AJ. Radiofrequency ablation in 447 complex unresectable liver tumors: lessons learned. Ann Surg Oncol 2003; 10: 52–8.[Abstract/Free Full Text]
- Ruers T, Bleichrodt RP. Treatment of liver metastases: an update on the possibilities and results. Eur J Cancer 2002; 38: 1023–33.
-
This article has been cited by other articles:
|
|
|
|
 
F. F. Amersi, A. McElrath-Garza, A. Ahmad, T. Zogakis, D. P. Allegra, R. Krasne, and A. J. Bilchik
Long-term Survival After Radiofrequency Ablation of Complex Unresectable Liver Tumors
Arch Surg,
June 1, 2006;
141(6):
581 - 588.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. A. Ahmad
Editorial: Limitations of Radiofrequency Ablation in Treating Liver Metastases: A Lesson in Geometry
Ann. Surg. Oncol.,
April 1, 2004;
11(4):
358 - 359.
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
