AJR 2004; 183:1785-1789
© American Roentgen Ray Society
Assessment of Cerebral Microembolism During Percutaneous Radiofrequency Ablation of Lung Tumors Using Diffusion-Weighted Imaging
Akira Yamamoto1,
Toshiyuki Matsuoka1,
Masami Toyoshima1,
Tomohisa Okuma1,
Yoshimasa Oyama1,
Masao Hamuro1,
Keiko Nakayama1,
Kiyotoshi Inoue2,
Kenji Nakamura1 and
Yuichi Inoue1
1 Department of Radiology, Osaka City University Graduate School of Medicine,
1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan.
2 Second Department of Surgery, Osaka City University Graduate School of
Medicine, Abeno-ku, Osaka 545-8585, Japan.
Received January 25, 2004;
accepted after revision May 6, 2004.
Address correspondence to A. Yamamoto.
Abstract
OBJECTIVE. It is well known that radiofrequency ablation generates
microbubbles in the liver. We hypothesized that microbubbles generated during
percutaneous radiofrequency ablation of lung tumors flow into the pulmonary
veins and are distributed to the systemic arteries, as with radiofrequency
ablation of liver tumors. To assess the risk of cerebral infarction during
radiofrequency ablation of lung tumors, we performed diffusion-weighted
imaging and, if possible, monitored microemboli in the carotid artery during
radiofrequency ablation.
SUBJECTS AND METHODS. We prospectively studied 20 patients (19 men
and one woman) who underwent radiofrequency ablation of lung tumors. Pre- and
postoperative MRI examinations were performed in all 20 patients, and during
17 radiofrequency ablation sessions, sonography was used to monitor whether
microemboli were generated.
RESULTS. Radiofrequency ablation was technically feasible for the
treatment of selected pulmonary tumors. Microemboli, which were believed to
represent microbubbles, were seen on sonography during three of the 17
radiofrequency ablation sessions. They were rarely observed when a lung tumor
was small, the treatment session was brief, and the radiofrequency emission
power was low. No new area of abnormal intensity was seen on postoperative MRI
in all 20 patients. Although the microemboli were observed, MRI could not
confirm infarction.
CONCLUSION. We concluded that cerebral infarction as a result of
microbubbles generated during radiofrequency ablation of lung tumors has a low
possibility of becoming a clinical problem.
Introduction
Imaging-guided percutaneous radiofrequency ablation has been successful in
the local control of solid cancers such as hepatoma
[1]. The application of
radiofrequency ablation to the treatment of lung cancers has been reported in
animals
[2–4]
and humans
[5–8].
Sonography shows that radiofrequency ablation of liver tumors generates highly
echoic lesions around the radiofrequency probe. These highly echoic lesions
are believed to be microbubbles and have been observed to flow into the
hepatic vein [9]. If
microbubbles are generated during radiofrequency ablation of lung tumors, as
with radiofrequency ablation of liver tumors, it is assumed that they flow
into the pulmonary veins and are distributed to the systemic arterial system.
We hypothesized that the risk of causing systemic infarctions including
cerebral infarction was high if microbubbles were generated during
radiofrequency ablation of lung tumors. To test this hypothesis, we
investigated the risk of cerebral infarction using diffusion-weighted imaging
after radiofrequency ablation of lung tumors, and, if possible, we monitored
generation of microemboli in the carotid artery through imaging. To our
knowledge, this study is the first to assess microemboli generated during
radiofrequency ablation of lung tumors using MRI.
Subjects and Methods
Patients
Between June 2002 and July 2003, after obtaining written informed consent
from the patients and approval from the institution's review board, we
prospectively enrolled 23 patients in our study. Because of patient refusal or
clinical status, we actually included 20 patients (19 men and one woman;
median age, 69.8 years; age range, 59–87 years) who underwent
radiofrequency ablation for lung tumors in the present study.
Radiofrequency Ablation Technique
Radiofrequency ablation was indicated in these patients because of medical
comorbidities, prohibitive pulmonary reserve, or refusal of surgery.
Twenty-five lung tumors in 20 patients were treated with radiofrequency
ablation. The treated tumors included lung cancer recurrence (n = 5)
(non–small cell lung carcinoma) and metastatic lung tumor from
extrathoracic malignancy (n = 20; renal cell carcinoma, esophageal
carcinoma, leiomyosarcoma, colon carcinoma, thyroid carcinoma, bladder
carcinoma, and malignant myxoid tumor)
(Table 1). Radiofrequency
ablation was performed in a CT room with the patient under local anesthesia.
Scans were obtained through the lung at a slice thickness of 2 mm using a CT
unit (Xvigor, Toshiba). Radiofrequency was generated by a RF2000 unit (Boston
Scientific).
After a 23-gauge localizing needle was used to define the proper route for
the subsequent tandem insertion, a needle electrode (LeVeen Needle Electrode,
Boston Scientific) was inserted. We used a 15-gauge LeVeen Needle Electrode
with either eight retractable distal hooks that can deploy to the full length
of 2 cm or 10 retractable distal hooks that can deploy to the full length of 3
cm. After a patient was placed in the prone or supine position on the CT
table, grounding pads were placed on both of the patient's thighs.
Radiofrequency ablation was performed with the emission power initially set at
20–30 W and increased by 5 W every 2 min. The radiofrequency energy was
applied until it automatically stopped (roll-off) because of increased
resistance caused by tissue dehydration. The maximum emission power was 80
W.
Sonography
The carotid artery was monitored during 17 of the 20 radiofrequency
ablation sessions using a sonography unit (LOGIQ500, GE Yokogawa Medical; a
probe [LA39, GE Yokogawa Medical]; B-mode, 8.7 MHz; Doppler, 5.0-MHz linear
transducer). If possible (depending on the patient's position), the right or
left common carotid artery was monitored before and throughout radiofrequency
ablation and recorded on videotape for later analysis. Of the 17 sessions,
four sessions were monitored using B-mode sonography only and 13 using B-mode
and Doppler sonography (Table
1). On B-mode sonography, emboli were observed visually. On
pulsed-wave Doppler sonography, emboli with high signal intensities above and
below the background Doppler signal (vertical spike) and a characteristic
harmonic quality were confirmed both visually and auditorily. The relationship
between the existence of microemboli and the basic data for the radiofrequency
ablation session—that is, patient age, tumor size, maximum emission
power, treatment time, and distance from the major pulmonary vein (> 5 mm
on CT)—was compared and assessed.
MRI
Diffusion-weighted imaging and FLAIR imaging were performed in all 20
patients both preoperatively and postoperatively (
24 hr after operation
[median time, 19.2 hr; range, 16–22 hr]). Diffusion-weighted imaging was
performed at 1.5 T (Signa, GE Healthcare) using a single-shot echo-planar
spin-echo technique. A maximum diffusion sensitivity of a b value of 1,000
sec/mm2 was applied to three orthogonal planes (x, y, z).
Axial images were obtained at 20 levels (TR/TE, 5,000/102; number of
excitations, 1; section thickness, 5 mm; intersection gap, 1.5 mm; field of
view, 20 cm; matrix, 128 x 128; bandwidth, 78 kHz). FLAIR imaging was
also performed at 1.5 T (Signa, GE Yokogawa Medical), and axial images were
obtained at 20 levels (9,002/150; number of excitations, 1; inversion time,
2,000 msec; section thickness, 5 mm; intersection gap, 1.5 mm; field of view,
20 cm; matrix, 256 x 192; bandwidth, 16 kHz).
Two radiologists who were blinded to the clinical status of the patients
independently interpreted the images. The criterion for a new
abnormal-intensity area on postoperative MRI was a high-signal area of more
than 10 mm2 excluding the existing area shown on preoperative MRI
or a high-signal area at two contiguous levels.
Statistical Analysis
Interobserver variability was determined by kappa statistical tests
[10]. The relation between the
existence of microemboli and the basic radiofrequency ablation session data
was investigated. Between the patients with microemboli and those without
microemboli, a pretest for comparing variance of the two distributions
(homoscedasticity) was performed. The Student's t test was chosen
when the variances of the two distributions were equal, and the Welch
t test was chosen when not equal. The nominal level was 0.05.
Confidence interval limits were derived using standard statistical
methods.
Results
Radiofrequency ablation was technically feasible for the treatment of
selected pulmonary tumors. The median maximum emission power was 48.7 W
(range, 20–80 W), and the median treatment time (ablation time) was 45.0
min (range, 7–105 min). The maximum emission power and treatment time
varied depending on the size and position of the tumor
(Table 1). None of the 20
patients had a transient ischemic attack or stroke after the procedure.
None of the patients was shown to have carotid artery microemboli on
preoperative sonography. In three of the 17 patients (patients 7, 13, and 20),
highly echoic emboli were displayed and a visible vertical spike was observed
above the background Doppler signal. A microembolus was seen before roll-off
in one patient (patient 13), and microemboli were seen approximately 5 min
after the start of ablation in two patients (patients 7 and 20). One
microembolus was observed in patient 13, and multiple microemboli were
observed throughout the procedure in patients 7 and 20 (e.g., 
100 per
minute in patient 7) (Fig. 1A).
When ablation was stopped, the microemboli were not observed. Preoperative
results of diffusion-weighted imaging and FLAIR imaging were normal. On
postoperative diffusion-weighted imaging and FLAIR imaging, no new area of
abnormal intensity was identified in any of the patients by either of the two
observers (
= 1.00).
View larger version (76K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A. —65-year-old man with 25-mm lung metastasis from
leiomyosarcoma (patient 7 in Table
1). Sonogram in carotid artery during pulmonary radiofrequency
ablation shows microemboli and materials (white arrow) that are
considered to be microbubbles. Visible vertical spike (black arrows)
indicating passage of emboli is observed.
|
|
Sonography and MRI studies were performed in 13 patients. No new area of
abnormal high intensity was observed on MRI in any of the patients including
patient 13 in whom one microembolus was seen and patients 7 and 20 in whom
multiple microemboli were seen on sonography (Figs.
1B and
1C). We found a relationship
between the existence of microemboli and the basic data
(Table 2) about the
radiofrequency ablation sessions—that is, tumor size
(Fig. 2A), maximum emission
power (Fig. 2B), and treatment
time (Fig. 2C).
View larger version (172K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B. —65-year-old man with 25-mm lung metastasis from
leiomyosarcoma (patient 7 in Table
1). Diffusion-weighted image (B) and FLAIR image (C)
of head obtained 20 hr after radiofrequency ablation show no new area of
abnormal intensity and no evidence of acute stroke. High-intensity area in
bilateral posterior periventricular white matter was present on preoperative
MR image (not shown).
|
|
View larger version (175K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C. —65-year-old man with 25-mm lung metastasis from
leiomyosarcoma (patient 7 in Table
1). Diffusion-weighted image (B) and FLAIR image (C)
of head obtained 20 hr after radiofrequency ablation show no new area of
abnormal intensity and no evidence of acute stroke. High-intensity area in
bilateral posterior periventricular white matter was present on preoperative
MR image (not shown).
|
|
View larger version (7K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —Graphs illustrate findings. Graphs show relation between
existence of microemboli (white bars) and tumor size (A),
maximum power (B), and treatment time (C). Black bars indicate
number of patients with no microemboli. Analyses show difference between
patients with microemboli and those with no microemboli is statistically
significant for average tumor size (p = 0.0100), average maximum
power (p = 0.0149), and average treatment time (p =
0.0183).
|
|
View larger version (9K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B. —Graphs illustrate findings. Graphs show relation between
existence of microemboli (white bars) and tumor size (A),
maximum power (B), and treatment time (C). Black bars indicate
number of patients with no microemboli. Analyses show difference between
patients with microemboli and those with no microemboli is statistically
significant for average tumor size (p = 0.0100), average maximum
power (p = 0.0149), and average treatment time (p =
0.0183).
|
|
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C. —Graphs illustrate findings. Graphs show relation between
existence of microemboli (white bars) and tumor size (A),
maximum power (B), and treatment time (C). Black bars indicate
number of patients with no microemboli. Analyses show difference between
patients with microemboli and those with no microemboli is statistically
significant for average tumor size (p = 0.0100), average maximum
power (p = 0.0149), and average treatment time (p =
0.0183).
|
|
Discussion
It is well known that radiofrequency ablation generates microbubbles in the
liver. Sonography monitoring of microbubbles guides the ablated area.
Microbubbles flowing into the pulmonary vein are distributed to the systemic
arterial system; therefore, the risk of cerebral infarction is assumed to be
high. Bubbles flowing into the pulmonary vein can cause stroke, which is a
common complication after surgery with cardiopulmonary bypass
[11] or lung surgery
[12]. Laser surgery
(laser-induced hyperthermia) for hepatoma has caused death due to an air
embolism [13]. Rose et al.
[14] reported that sonography
showed microbubbles during radiofrequency ablation of lung tumors in all three
patients in their study group. Cerebral infarction is also a possible
complication.
In three (17.6%) of our 17 patients, microemboli were seen during ablation.
In our study, microemboli with high signal intensities above and below the
background Doppler signal (vertical spike) and a characteristic harmonic
quality were confirmed both visually and auditorily
(Fig. 1A). These findings are
similar to those reported by Rose et al.
[14] and are considered to be
microbubbles.
Rose et al. [14] reported
that microbubbles were observed during radiofrequency ablation in all three
patients in their study, which does not correspond to our findings. This
discrepancy might be because the generating frequency in our study was low
compared with that reported by Rose et al. They set the initial emission power
at 25–40 W and increased it by 5–10 W per minute, whereas we set
the initial emission power at 20–3 W and increased it by 5 W every 2
min.
We set the maximum power output to be less than 80 W in order not to expose
the tumor to high temperature for a long time, and we raised the output more
gradually than Rose et al.
[14]. We think that the
differences in the emission power settings seen in the study conducted by Rose
et al. and our study contributed to the different frequencies of microemboli
generation. The tumor size was larger (p < 0.0100), maximum
emission power was higher (p < 0.0149), and treatment time was
longer (p < 0.0183) in patients with microemboli than in those
without microemboli. Moreover, the distance from the major pulmonary vein
(> 5 mm on CT) was considered to be unrelated to microemboli generation.
These findings suggest that the possibility of microbubbles being generated is
increased when the maximum emission power is high. Rose et al. reported that
unenhanced CT of the brain performed after lung ablation shows no abnormal
findings in patients with microemboli.
We performed diffusion-weighted imaging because a more sensitive diagnosis
of acute cerebral infarction could be made using diffusion-weighted imaging
than using CT or conventional MRI
[15]. Diffusion-weighted
imaging can detect parenchymal changes in humans within 40 min after onset
[16] and can continue to
detect them for at least 1 week
[17]. We performed
diffusion-weighted imaging within 24 hr after the lung radiofrequency ablation
and think that this timing is appropriate for determining whether any areas of
unusual signal intensity are present. Preoperative MRI was performed to
exclude the presence of a silent embolism
[18].
In our study, postoperative MRI showed no new abnormal high-signal area in
any patients including those in whom microemboli were observed on sonography.
We believe that there are four possible explanations for this discrepancy.
First, there is a possibility that a transient arterial occlusion was improved
at an early stage. Second, the early dissolution of microbubbles in the plasma
was observed in the carotid artery. The composition of microbubbles was
unknown. They are reported to occur from vapor formation of blood or water,
nitrogen or nitrous oxide gas, and carbon dioxide. Intrahepatic vascular gas
has been observed on real-time hepatic sonograms, but not on preoperative
cardial Doppler sonograms
[19]. Third, emboli that
reached the brain were very small, and they did not cause arterial infarction.
In a mouse model, microbubbles were reported to be 3–8 µ m
[9]. Microbubbles are used as
sonographic contrast agents, and their average size is 35.5 µ m
[20]. When the size of
microbubbles is approximately 35.5 µ m, the possibility that they
might cause cerebral infarction is low. Fourth, embolic events in patients
with early reperfusion would not be detected on MRI performed within 24 hr of
radiofrequency ablation. There are many reports about acute stroke syndrome
with fixed neurologic deficit and false-negative findings of
diffusion-weighted imaging in human and animal models
[21–23].
Negative findings on diffusion-weighted imaging of the brain cannot always
exclude the possibility that a patient has acute cerebral ischemia.
An upper bound for the incidence of an event with zero occurrences in a
study sample is estimated to be less than 10% in patients with an
abnormal-intensity area detected on diffusion-weighted imaging using a
Bayesian estimate with a noninformative beta (p < 0.05)
[24]. In our institute,
radiofrequency ablation of lung tumors had been performed 60 times. Further
study using a larger number of subjects is required, but our results indicate
that cerebral infarction due to microbubbles during radiofrequency ablation of
lung tumors has a low possibility of becoming a clinical problem.
References
- De Sanctis JT, Goldberg SN, Mueller PR. Percutaneous treatment of
hepatic neoplasms: a review of current techniques. Cardiovasc
Intervent Radiol 1988;21:273
–296
- Goldberg SN, Gazelle GS, Compton CC, McLoud TC. Radiofrequency
tissue ablation in the rabbit lung: efficacy and complications.
Acad Radiol1995; 2:776
–784[Medline]
- Goldberg SN, Gazelle GS, Compton CC, Mueller PR, McLoud TC.
Radio-frequency tissue ablation of VX2 tumor nodules in the rabbit lung.
Acad Radiol1996; 3:929
–935[Medline]
- Miao Y, Ni Y, Bosmans H, et al. Radiofrequency ablation for
eradication of pulmonary tumor in rabbits. J Surg Res2001; 99:265
–271[Medline]
- Asai T, Tanigawa N, Tanabe MJ, et al. Radiofrequency thermal
coagulation therapy for lung tumors: an experimental study [in Japanese].
Ryukyu Med1997; 17:203
–209
- Dupuy DE, Zagoria RJ, Akerley W, Mayo-Smith WW, Kavanagh PV, Safran
H. Percutaneous radiofrequency ablation of malignancies in the lung.
AJR 2000;174:57
–59[Free Full Text]
- Toyoshima M, Matsuoka T, Tanaka S, et al. Percutaneous
radiofrequency ablation for metastatic lung tumors: a case report [in
Japanese]. Gan To Kagaku Ryoho2001; 28:1604
–1606[Medline]
- Toyoshima M, Matsuoka T, Okuma T, et al. Radiofrequency ablation of
pulmonary malignancies [in Japanese]. Nippon Igaku Hoshasen Gakkai
Zasshi 2003;62:836
–838
- Kruskal JB, Oliver B, Huertas JC, Goldberg SN. Dynamic intrahepatic
flow and cellular alterations during radiofrequency ablation of liver tissue
in mice. J Vasc Interv Radiol2001; 12:1193
–1201[Medline]
- Posner KL, Sampson PD, Caplan RA, Ward RJ, Cheney FW. Measuring
interrater reliability among multiple raters: an example of methods for
nominal data. Stat Med1990; 9:1103
–1115[Medline]
- Blauth CI. Macroemboli and microemboli during cardiopulmonary
bypass. Ann Thorac Surg1995; 59:1300
–1303[Abstract/Free Full Text]
- Licker M, de Perrot M, Hohn L, et al. Perioperative mortality and
major cardio-pulmonary complications after lung surgery for non-small cell
carcinoma. Eur J Cardiothorac Surg1999; 15:314
–319[Abstract/Free Full Text]
- Hahl J, Haapiainen R, Ovaska J, Puolakkainen P, Schroder T.
Laser-induced hyperthermia in the treatment of liver tumors. Lasers
Surg Med 1990;10:319
–321[Medline]
- Rose SC, Fotoohi M, Levin DL, Harrell JH. Cerebral
microembolization during radiofrequency ablation of lung malignancies.
J Vasc Interv Radiol2002; 13:1051
–1054[Medline]
- Kidwell CS, Alger JR, Di Salle F, et al. Diffusion MRI in patients
with transient ischemic attacks. Stroke1999; 30:1174
–1180[Abstract/Free Full Text]
- Yoneda Y, Tokui K, Hanihara T, Kitagaki H, Tabuchi M, Mori E.
Diffusion-weighted magnetic resonance imaging: detection of ischemic injury 39
minutes after onset in a stroke patient. Ann Neurol1999; 45:794
–797[Medline]
- Gonzalez RG, Schaefer PW, Buonanno FS, et al. Diffusion-weighted MR
imaging: diagnostic accuracy in patients imaged within 6 hours of stroke
symptom onset. Radiology1999; 210:155
–162[Abstract/Free Full Text]
- Forbes KP, Shill HA, Britt PM, Zabramski JM, Spetzler RF, Heiserman
JE. Assessment of silent embolism from carotid endarterectomy by use of
diffusion-weighted imaging: work in progress. AJNR2001; 22:650
–653[Abstract/Free Full Text]
- Malone DE, Lesiuk L, Brady AP, Wyman DR, Wilson BC. Hepatic
interstitial laser photocoagulation: demonstration and possible clinical
importance of intravascular gas. Radiology1994; 193:233
–237[Abstract/Free Full Text]
- Soetanto K, Chan M. Fundamental studies on contrast images from
different-sized microbubbles: analytical and experimental studies.
Ultrasound Med Biol2000; 26:81
–91[Medline]
- Wang W, Goldstein S, Scheuer ML, Branstetter BF. Acute stroke
syndrome with fixed neurological deficit and false-negative diffusion-weighted
imaging. J Neuroimaging2003; 13:158
–161[Medline]
- Uchino A, Sawada YA, Imaizumi T, Mineta T, Kudo S. Report of
fogging effect on fast FLAIR magnetic resonance images of cerebral
infarctions. Neuroradiology2004; 46:40
–43[Medline]
- Lin SP, Schmidt RE, McKinstry RC, Ackerman JJ, Neil JJ.
Investigation of mechanisms underlying transient T2 normalization in
longitudinal studies of ischemic stroke. J Magn Reson
Imaging 2002;15:130
–136[Medline]
- Basu AP, Gaylor DW, Chen JJ. Estimating the probability of
occurrence of tumor for a rare cancer with zero occurrence in a sample.
Regul Toxicol Pharmacol1996; 23:139
–144[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
