Radiofrequency Ablation: Post-ablation Assessment Using CT Perfusion with Pharmacological Modulation in a Rat Subcutaneous Tumor Model

Abstract

Inflammatory reaction surrounding the ablated area is a major confounding factor in the early detection of viable tumor after radiofrequency (RF) ablation. A difference in the responsiveness of normal and tumor blood vessels to vasoactive agents may be used to distinguish these regions in post-ablation follow-up. The goal of this study was to examine longitudinal perfusion changes in untreated viable tumor and the peripheral hyperemic rim of RF-ablated tumor in response to a vasoconstrictor (phenylephrine) or vasodilator (hydralazine) in a subcutaneous rat tumor model.
Bilateral subcutaneous shoulder tumors were inoculated in 24 BDIX rats and evenly divided into two groups (phenylephrine and hydralazine groups). One tumor in each animal was completely treated with RF ablation (at 90 +/- 2 degrees C for 3 minutes), and the other remained untreated. Computed tomographic perfusion scans before and after phenylephrine (10 microg/kg) or hydralazine (5 mg/kg) administration were performed 2, 7, and 14 days after ablation. Four rats per group were euthanized on each scan day, and pathologic evaluation was performed. The changes of blood flow in the peripheral rim of ablated tumor and untreated viable tumor in response to phenylephrine or hydralazine at each time point were compared. The diagnostic accuracy of viable tumor using the percentage change of blood flow in response to phenylephrine and hydralazine was compared using receiver-operating characteristic analysis.
The peripheral rim of ablated tumor presented with a hyperemic reaction with dilated vessels and congestion on day 2 after ablation, numerous inflammatory vessels on day 7, and granulation tissue formation on day 14. Phenylephrine significantly decreased the blood flow in the peripheral hyperemic rim of ablated tumor on days 2, 7, and 14 by 16.3 +/- 9.7% (P = .001), 24.0 +/- 22.6% (P = .007), and 31.1 +/- 25.4% (P = .045), respectively. In untreated viable tumor, the change in blood flow after phenylephrine was irregular and insignificant. Hydralazine decreased the blood flow in the peripheral rim of both ablated tumor and untreated viable tumor. Receiver-operating characteristic analysis showed that reliable tumor diagnosis using the percentage change of blood flow in response to phenylephrine was noted on days 2 and 7, for which the areas under the curve were 0.82 (95% confidence interval, 0.64-1.00) and 0.81 (95% confidence interval, 0.56-1.00), respectively. However, tumor diagnosis using the blood flow change in response to hydralazine was unreliable.
Phenylephrine markedly decreased blood flow in the peripheral hyperemic rim of ablated tumor but had little effect on the untreated viable tumor. Computed tomographic perfusion with phenylephrine may be useful in the long-term treatment assessment of RF ablation.

Full text

Radiofrequency Ablation:Post-ablation Assessment Using CT
Perfusion with Pharmacological Modulation in a Rat
Subcutaneous Tumor Model
Hanping Wu, MD, PhD
Department of Radiology, University Hospitals, Case Medical Center, 11100 Euclid Avenue,
Cleveland, OH 44106
Agata A. Exner, PhD
Department of Radiology, University Hospitals, Case Medical Center, 11100 Euclid Avenue,
Cleveland, OH 44106
Tianyi M. Krupka, BS
Department of Biomedical Engineering, Case Western Reserve University, 11100 Euclid Avenue,
Cleveland, OH 44106
Brent D. Weinberg, PhD
Department of Biomedical Engineering, Case Western Reserve University, 11100 Euclid Avenue,
Cleveland, OH 44106
Ravi Patel, BS
Department of Biomedical Engineering, Case Western Reserve University, 11100 Euclid Avenue,
Cleveland, OH 44106
John R. Haaga, MD
Department of Radiology, University Hospitals, Case Medical Center, 11100 Euclid Avenue,
Cleveland, OH 44106
Abstract
Rationale and Objectives—Inflammatory reaction surrounding the ablated area is a major
confounding factor in the early detection of viable tumor after radiofrequency (RF) ablation. A
difference in the responsiveness of normal and tumor blood vessels to vasoactive agents may be
used to distinguish these regions in post-ablation follow-up. The goal of this study was to examine
longitudinal perfusion changes in untreated viable tumor and the peripheral hyperemic rim of RF-
ablated tumor in response to a vasoconstrictor (phenylephrine) or vasodilator (hydralazine) in a
subcutaneous rat tumor model.
Materials and Methods—Bilateral subcutaneous shoulder tumors were inoculated in 24 BDIX
rats and evenly divided into two groups (phenylephrine and hydralazine groups). One tumor in
each animal was completely treated with RF ablation (at 90 ± 2°C for 3 minutes), and the other
remained untreated. Computed tomographic perfusion scans before and after phenylephrine (10
μg/kg) or hydralazine (5 mg/kg) administration were performed 2, 7, and 14 days after ablation.
Four rats per group were euthanized on each scan day, and pathologic evaluation was performed.
The changes of blood flow in the peripheral rim of ablated tumor and untreated viable tumor in
response to phenylephrine or hydralazine at each time point were compared. The diagnostic
© AUR, 2009
Address correspondence to: J.R.H. john.haaga@uhhospitals.org.
NIH Public Access
Author Manuscript
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
Published in final edited form as:
Acad Radiol
. 2009 March ; 16(3): 321–331. doi:10.1016/j.acra.2008.09.008.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

accuracy of viable tumor using the percentage change of blood flow in response to phenylephrine
and hydralazine was compared using receiver-operating characteristic analysis.
Results—The peripheral rim of ablated tumor presented with a hyperemic reaction with dilated
vessels and congestion on day 2 after ablation, numerous inflammatory vessels on day 7, and
granulation tissue formation on day 14. Phenylephrine significantly decreased the blood flow in
the peripheral hyperemic rim of ablated tumor on days 2, 7, and 14 by 16.3±9.7% (P = .001),
24.0±22.6% (P = .007), and 31.1±25.4% (P = .045), respectively. In untreated viable tumor, the
change in blood flow after phenylephrine was irregular and insignificant. Hydralazine decreased
the blood flow in the peripheral rim of both ablated tumor and untreated viable tumor. Receiver-
operating characteristic analysis showed that reliable tumor diagnosis using the percentage change
of blood flow in response to phenylephrine was noted on days 2 and 7, for which the areas under
the curve were 0.82 (95% confidence interval, 0.64–1.00) and 0.81 (95% confidence interval,
0.56–1.00), respectively. However, tumor diagnosis using the blood flow change in response to
hydralazine was unreliable.
Conclusion—Phenylephrine markedly decreased blood flow in the peripheral hyperemic rim of
ablated tumor but had little effect on the untreated viable tumor. Computed tomographic perfusion
with phenylephrine may be useful in the long-term treatment assessment of RF ablation.
Keywords
Radiofrequency ablation; CT perfusion; vasoactive agents; tumor perfusion; postablation
assessment
Interventional oncologic procedures such as chemical ablation, laser ablation,
radiofrequency (RF) ablation, and microwave ablation have been widely used in
unresectable solid tumor treatment. These techniques can provide efficacious, cost-effective
management of localized cancer sites (1,2). However, local recurrence due to factors such as
large tumor size (>3 cm), irregular contours, and affluent blood supply continues to be a
main cause of treatment failure (3,4). The early detection of residual or locally recurrent
tumor after interventional treatment is critical and can facilitate successful retreatment at an
early stage.
Contrast-enhanced computed tomography and magnetic resonance (MR) are currently used
to evaluate RF ablation efficacy. However, an inflammatory peripheral rim induced by
ablation, which presents as enhancement surrounding the coagulated zone, can obscure
viable residual tumor. The histologic examination of ablated tumors has shown that the
sensitivity of computed tomography or MR for the detection of viable tumor after ablation
ranges from 36% to 86% (5–7). Recent reports that positron emission tomography/computed
tomography can increase the sensitivity and accuracy for the detection of residual tumor are
promising (8,9), but this modality may not be sufficiently cost effective to become a routine
follow-up method. New imaging methods are needed for the successful and accurate follow-
up of local tumor ablation therapy.
Previous research has shown that tumor vessels are immature, lack normal smooth muscle
and pericyte structure, and do not react to vasoactive drugs (10,11). This characteristic has
been used for more than two decades to differentiate malignant from benign lesions by
angiography, a procedure called pharmacoangiography (12). However, angiography is
invasive and of low spatial resolution and is seldom used for viable tumor detection after RF
ablation. Computed tomographic (CT) perfusion has been recently used to quantify the
blood perfusion of a lesion or organ (13–15). This technique requires only an intravenous
administration of contrast agent and has excellent spatial and temporal resolution.
Wu et al. Page 2
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

We hypothesized that because of the difference between tumor and normal vessels, the
perfusion in inflammatory tissue and viable residual tumor surrounding an ablated zone
would change differently in response to vasoactive drugs. That is, vasoconstrictors should
produce constriction of normal vessels and accordingly decrease blood flow in inflammatory
tissue. Vasodilators should dilate normal vessels and increase blood flow in inflammatory
tissue. In contrast, tumor blood flow should show no change or behave paradoxically in
response to the same vasoactive drugs. Here, the constriction of surrounding normal vessels
should “push” blood into the tumor and increase tumor blood flow, while the dilation of
surrounding normal vessels should “divert” blood away from tumor and decrease tumor
blood flow. This difference could be useful in differentiating viable tumor from the
peripheral inflammatory hyperemic rim after RF ablation. Accordingly, the purpose of the
present study was to evaluate perfusion changes in viable tumor and the peripheral
hyperemic rim of ablated tumor using CT perfusion imaging. A vasoconstrictor,
phenylephrine, and a vasodilator, hydralazine, were used to modulate blood flow during a 2-
week post-ablation follow-up period.
MATERIALS AND METHODS
Overall Experimental Design
Twenty-four rats with 48 bilateral shoulder tumors were used in this study. The rats were
randomly divided into phenylephrine (n = 12) and hydralazine (n = 12) groups. All right-
side tumors were treated with RF ablation. The left-side tumors were untreated and served as
internal controls. On days 2, 7, and 14 after RF ablation, animals were examined with CT
perfusion scans before and after phenylephrine or after hydralazine. Four animals in each
group were euthanized immediately after each scan day. Corresponding tumor slices in
orientation to the CT image data were obtained and evaluated with histology. The changes in
blood flow in untreated tumors and the peripheral rim of ablated tumors in response to
phenylephrine or hydralazine at each time point were compared.
Animal Model
A well-established rat tumor model, DHD/K12/TRb adenocarcinoma in the subcutaneous
tissue of BDIX rats, was used in this study (16). The experimental protocols were approved
by the Institutional Animal Care and Use Committee at our institution. For all animal
experiments, procedures were carried out under general gas anesthesia with 1% isoflurane
and an oxygen flow rate of 1 L/min (EZ150 Isoflurane Vaporizer; Euthanex Corporation,
Palmer, PA). The DHD/K12/TRb colorectal carcinoma cell line originates from a 1,2-
dimethylhydrazine-induced colon adenocarcinoma in BDIX rats (the original source was the
European Collection of Cell Cultures). Cells were cultured in RPMI-1640 with 10% fetal
bovine serum and passaged weekly. Six-week-old male rats (Charles River Laboratories,
Inc, Wilmington, MA) were used for inoculation. Tumors were inoculated subcutaneously
with a bilateral injection of 0.05 mL DHD/K12 cell suspension (2 × 106 cells/mL) on the
shoulders using a 26-gauge hypodermic needle. Tumor size was measured weekly using
calipers. Tumors were grown for 6 weeks, and the maximum diameter was 9.0 to 13.5 mm.
RF Ablation
RF ablation was performed using a 480-kHz RF generator (Radionics, Inc, Burlington, MA)
and a custom-designed 21-gauge monopolar needle electrode. Here, the RF antenna and a
thermocouple were located in the central core of a disposable, uninsulated hypodermic
needle to heat the tissue and monitor temperature during ablation. A grounding pad was
placed on the abdomen, and ultrasonic gel was used to ensure proper contact. The skin was
cut open about 5 to 7 mm in length with a scalpel to avoid skin burning, and the electrode
was inserted into the tumor center to 0.8 cm in depth. During ablation, the temperature was
Wu et al. Page 3
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

manually adjusted to 90 ± 2°C for 3 minutes. The power was approximately 2 to 3 W. This
ablation protocol resulted in complete tumor ablation in our preliminary experiments.
CT Perfusion
CT perfusion scans were performed with a 24-slice CT scanner (Sensation Open; Siemens
Medical Systems, Erlangen, Germany) 2, 7, and 14 days after RF ablation. The rat position
on the CT table was adjusted so that the centers of both tumors were parallel to the scan
plane. An initial planar scan was carried out to determine the optimal position for the
perfusion scan. Then, the perfusion scan was performed with a 0.5-mL bolus of contrast
(Optiray 240; Mallinckrodt, Inc, St Louis, MO) administered through a 24-gauge catheter
(Safelet Cath; Exel Internatoinal Medical Products, Los Angeles, CA) in the tail vein. The
perfusion protocol, which imaged 12 consecutive slices every 0.5 seconds for 40 seconds,
was executed using the following parameters: axial scan, 80 kV; 150 mAs; rotation time, 0.5
seconds; detector width, 1.2 mm; reconstruction width, 2.4 mm; field of view, 60 mm; and
matrix, 512 × 512. Thus, each perfusion scan obtained 960 images. Phenylephrine (10 μg/
kg; Sigma-Aldrich, Inc, St Louis, MO) or hydralazine (5 mg/kg; MP Bio-Medicals, Inc,
Solon, OH) was injected through the tail vein 1 or 5 minutes before the post-perfusion scan.
The delay times for phenylephrine and hydralazine were based on the effect-time curves
reported in the literature (17,18). The interval between baseline and post-drug perfusion
scans was 10 minutes to allow for contrast agent clearance.
Image Analysis
Image analysis was performed by one observer (H.W.) with 5 years of experience in CT
diagnosis and 3 years of experience in CT image postprocessing. All preperfusion and post-
perfusion studies were analyzed on the Wizard workstation using commercial
compartmental analysis-based perfusion software (Body Perfusion CT, Syngo CT 2006A-W
Wizard VB20B-W; Siemens Medical Solutions) and a tumor perfusion algorithm. Blood
flow was measured in milliliters per 100 milliliters per minute on the basis of the maximum
slope method. A single 2.4-mm slice that best depicted both of the bilateral tumors was
chosen from the 12 perfusion slices. The analyzed slices of pre- and postper-fusion of each
animal were consistent.
CT images of the chosen slice were loaded into the software. The image stack was
examined, and images with severe motion artifacts were removed before perfusion
calculation. A rectangular reference region of interest (ROI) that contained the tumors was
drawn, and two-dimensional motion correction was performed automatically. The arterial
input curve of contrast medium concentration was obtained from an ROI (50–70 pixels) in
the aorta or common carotid artery. A processing threshold of −50 to 200 Hounsfield units
was selected. The time-attenuation curves of artery and mean tissue were generated by the
software and displayed in an arterial input function and optimization dialog box. The times
of arterial shift and Patlak start for Patlak modeling and the start time, rise time, peak time
for maximum slope modeling were adjusted as necessary on the basis of the time-attenuation
curves of artery and mean tissue. A grayscale maximum-intensity projection image and
pseudocolor perfusion map were generated corresponding to each voxel. ROIs of untreated
viable tumor, peripheral rim of ablated tumor, and adjacent normal muscle were drawn on
the maximum-intensity projection image. Because only negligible blood flow was found in
the center of the untreated tumor and the coagulation zone of the ablated tumor, the
enhanced rim of untreated tumor was determined as the untreated viable tumor, and the
enhanced rim of the ablated zone was determined as the peripheral rim of the ablated tumor.
Care was taken to exclude the surrounding vessels and fat when drawing the tumor
boundary. The mean values of blood flow for every ROI were recorded.
Wu et al. Page 4
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Histopathologic Examination
Rats were euthanized with carbon dioxide inhalation immediately after the perfusion scans.
The tumors were excised, and the samples corresponding to the CT perfusion scans were
obtained and fixed with 10% formalin. The samples were then embedded in paraffin, and 5-
μm slices were obtained and stained with hematoxylin and eosin. All tissue processing was
carried out by the Tissue Procurement and Histology Core Facility of the Case
Comprehensive Cancer Center (P30 CA43703).
Statistical Analysis
Unless otherwise noted, data are presented as mean ± standard error. A paired t test was
used for the comparison of blood flow before and after drug administration. The differences
of predrug blood flow in the peripheral rim of ablated tumor and untreated viable tumor at
different time points were compared using unpaired t tests. One-way analysis of variance
was used to compare blood flow among three time points. All statistical tests were evaluated
by Minitab version 15.0 (Minitab, Inc, State College, PA). A P value < .05 indicated a
statistically significant difference.
The diagnostic performance for the detection of viable tumor using the percentage change of
blood flow in response to phenylephrine or hydralazine was tested with receiver-operating
characteristic (ROC) analysis. The detection accuracy of each observation time point and
pooled data were measured according to the area under the ROC curve (AUC). ROC curves,
AUC values, and 95% confidence intervals (CIs) for each observation time point were
analyzed with Analyse-it (Analyse-it Software, Ltd, Leeds, United Kingdom). An AUC of
0.5 indicated that a diagnosis was obtained by guessing.
RESULTS
CT Perfusion Findings and Corresponding Histologic Correlation
The perfusion images and corresponding histology of a typical untreated tumor are shown in
Figure 1. The central zone of the tumor exhibited necrosis, and hyperdensity in this zone in
the CT image indicates possible calcification (Fig 1a). No perfusion was noted in this area
(Fig 1b). In contrast, the outer part of the tumor consisted of viable tumor cells enclosed by a
fibrous capsule (Fig 1c). Microvascular morphology was noted in this region and was
located primarily in the capsule (Fig 1d). Correspondingly, contrast enhancement and higher
perfusion were noted in this area.
After ablation, the skin in the immediate vicinity of the ablated tumor showed swelling and
edema on day 2, but by day 7, the swelling and edema had mostly resolved. Although some
scar formation was noted on day 14, the ablated tissue appeared fully healed under gross
examination. Histology showed extensive cellular damage in the coagulation zone. The
peripheral rim of ablated area presented as a hyperemic reaction zone with dilated vessels
and congestion on day 2 (Figs 2a and 2b). On day 7, increased vascularization was seen in
the peripheral rim, but the lumen of these vessels had returned to normal (Figs 2c and 2d).
On day 14, granulation tissue formation and lymphocyte infiltration were seen at the
boundary of ablated zone, and the quantity of blood vessels had decreased (Figs 2e and 2f).
The changes in blood flow in the viable tumors of the untreated control and in the peripheral
rim of ablated tumor are shown in Figure 3. The blood flow in the untreated tumors was 43.3
± 11.9, 50.9 ± 11.0, and 45.3 ± 9.0 mL/100 mL/min on days 2, 7, and 14, respectively, with
no significance differences between the days (analysis of variance, P = .178). The blood
flow in the peripheral rim of the ablated tumor was greater than that of untreated tumor on
days 2, 7, and 14 (49.4 ± 19.5, 70.3 ± 21.4, and 69.0 ± 17.2 mL/100 mL/min, respectively).
Wu et al. Page 5
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

The differences in blood flow in the peripheral rim of the ablated tumor and the untreated
viable tumor were significant on days 7 and 14 (P = .008 and .004, respectively).
Blood Flow Changes in the Peripheral Rim of Ablated Tumor and Untreated Viable Tumor
in Response to Phenylephrine
Table 1 summarizes the blood flow changes in the peripheral rim of the ablated tumor,
untreated viable tumor, and muscle in response to phenylephrine during the follow-up
period. After the administration of phenylephrine, the blood flow change in the untreated
viable tumor was irregular and nonsignificant. The percentage changes were 11.0 ± 32.7%
(P = .181), −4.5 ± 9.0% (P = .174), and 1.8 ± 35.0% (P = .483) on days 2, 7, and 14. In 13
of 24 tested untreated tumors, the tumor blood flow decreased in response to phenylephrine
(Fig 4a). In contrast, phenylephrine significantly and consistently decreased the blood flow
in the peripheral rim of the ablated tumor. The change rates were −16.3 ± 9.7% (P = .001),
−24.0 ± 22.6% (P = .007), and −31.1 ± 25.4% (P = .045) on days 2, 7, and 14. In 23 of 24
tested ablated tumors, blood flow in the peripheral rim decreased in response to
phenylephrine (Fig 4b). Figure 5 shows a series of representative CT perfusion images
before and after phenylephrine on days 2, 7, and 14. The blood flow in muscle decreased at
every time point after phenylephrine administration, but the changes had no statistical
significance.
Perfusion Changes in the Peripheral Rim of Ablated Tumor and Untreated Viable Tumor in
Response to Hydralazine
The blood flow changes in response to hydralazine in untreated viable tumor and the
peripheral rim of ablated tumor are summarized in Table 2. Hydralazine decreased the blood
flow in the peripheral rim of both the ablated tumor and untreated viable tumor at every time
point, and all changes, except in the peripheral rim of the ablated tumor on day 14, were
statistically significant. The decrease was more pronounced in the untreated viable tumor.
The change rates were −27.5 ± 19.1%, −24.4 ± 32.1%, and −30.6 ± 5.1% in the untreated
viable tumor and −15.0 ± 25.9%, −17.7 ± 16.8%, and −10.3 ± 30.7% in the peripheral rim
of the ablated tumor on days 2, 7, and 14, respectively. In 22 of 24 tested untreated tumors
and 18 of 24 tested ablated tumors, blood flow in the viable tumor or the peripheral rim of
the ablated tumor decreased in response to hydralazine (Fig 6). Figure 7 shows
representative blood flow images before and after hydralazine administration on days 2, 7,
and 14. Hydralazine decreased normal muscle blood flow at every time point, and these
decreases were statistically significant on days 2 and 14.
ROC Analysis
ROC curves of the changes of blood flow in response to phenylephrine and hydralazine on
days 2, 7, and 14 are shown in Figure 8, and the AUC values and 95% CIs are shown in
Table 3. Reliable tumor diagnosis using the percentage change of blood flow in response to
phenylephrine was noted on days 2 and 7, for which the AUCs were 0.82 (95% CI, 0.64–
1.00) and 0.81 (95% CI, 0.56–1.00), respectively. However, tumor diagnosis using the blood
flow change in response to phenylephrine on day 14 and using the percentage change in
response to hydralazine was unreliable.
DISCUSSION
The early detection of residual and recurrent tumor in patients undergoing local thermal
ablation or chemical ablation is vital. Many imaging modalities, such as computed
tomography, MR, and ultrasound are currently used in treatment follow-up to assess tumor
response after ablation. In all of these modalities, contrast enhancement is crucial for
differentiating the ablated tumor and residual viable tumor on the basis of the fact that the
Wu et al. Page 6
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

ablated tumor is void of contrast enhancement, whereas the viable tumor is enhanced
because of existing tumor vessels. However, contrast enhancement fails to provide desirable
sensitivity (5,6).
The main reason for the low specificity of contrast-enhanced CT or MR examinations in
differentiating viable tumor is the physiologic response (nonspecific inflammation) to
thermal injury. It is well documented that this response will occur surrounding the ablation
zone after ablation therapy. Previous clinical imaging and pathologic correlation studies
have shown that this inflammatory rim presents benign periablational enhancement, usually
uniform in thickness, and this status will remain 2 to 6 months after ablation. In contrast,
viable residual or recurrent tumor has shown focal irregular peripheral enhancement (19,20).
However, it is difficult to predict tumor regrowth in the early stage solely on the basis of
these morphologic findings. Dromain et al (21) reported that CT and MR imaging may at the
earliest depict tumor recurrence at 4 months after RF ablation, because the peripheral rim
disappeared with time and was present in only 8% of the RF-ablated areas at this time point.
In our study, the CT perfusion and histologic changes of subcutaneous rat tumor after RF
ablation were explored. During the 2-week observation period, the peripheral rim of ablated
tumor presented, with a hyperemic reaction with dilated vessels and congestion on day 2,
numerous inflammatory vessels on day 7, and granulation tissue formation on day 14. The
maximum concentration of blood vessels was observed on day 7. The corresponding CT
perfusion study also demonstrated that the blood flow in the peripheral rim increased after
RF ablation, was highest on day 7, and had decreased only slightly by day 14. These
findings confirm that the inflammatory reaction persists >2 weeks after RF ablation in this
animal model, and CT perfusion can successfully monitor blood vessel formation after RF
ablation.
Many investigators have used vasoactive drugs to modulate tumor or normal tissue blood
flow on the basis of the fact that tumor vessels lack the anatomy necessary to elicit a
response following vasoactive drug administration (22,23). In this study, we used vasoactive
drugs—phenylephrine as a vasoconstrictor and hydralazine as a vasodilator—to modulate
tumor and inflammatory tissue blood flow. Phenylephrine is a powerful post-synaptic α1
receptor stimulant with little effect on the β receptors of the heart. It produces
vasoconstriction that lasts longer than that of epinephrine. It is now the most common over-
the-counter decongestant as an oral medicine or as a nasal spray. When used by injection,
phenylephrine is used to maintain adequate blood pressure (24). Hydralazine was one of the
first available oral antihypertensive drugs and is currently used primarily to treat pregnancy-
associated hypertension. It lowers blood pressure by exerting a peripheral vasodilatory effect
through a direct relaxation of vascular smooth muscle (25). Both of the agents are
commonly used in clinics. They have also been extensively investigated to modulate blood
flow in animal experiments (22,26).
We found that phenylephrine significantly and consistently decreased blood flow in
peripheral hyperemic rim of the ablated tumor at each time point. This differed considerably
from the untreated viable tumor response, which showed no significant changes. These
findings are consistent with our hypothesis that the vasoconstrictor phenylephrine can
constrict inflammatory vessels, decrease the blood flow in the peripheral rim of tumor after
RF ablation, and have little effect on blood flow in the untreated tumor rim. This
phenomenon can be helpful to differentiate viable tumor from inflammatory tissue after RF
ablation in the early stage.
We also found that hydralazine decreased the blood flow in the peripheral rim of both the
ablated tumor and untreated viable tumor. This appears to be opposite to our hypothesis that
Wu et al. Page 7
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

a vasodilator would increase inflammatory and normal tissue blood flow but would have
little effect on or decrease tumor blood flow. However, with further comparison of the
change rates, we found that the response of peripheral hyperemic rim of the ablated tumor
was less than that of untreated viable tumor at every time point (−15.0% vs −27.5% on day
2, −17.7% vs −24.4% on day 7, and −10.3% vs −30.6% on day 14). This indicates that the
inflammatory vessels surrounding the ablation zone dilated in response to hydralazine,
leading to a competing effect and limited blood flow reduction. Therefore, hydralazine
decreased the blood flow in the peripheral rim of the ablated tumor to a lesser degree than in
untreated viable tumor. The observed changes may also have been a result of hydralazine
dose. Significant decreases of normal muscle blood flow on days 2 and 14 indicate that the
dose of hydralazine we used was high. A high dose may decrease systemic blood pressure
and subsequently reduce blood flow in both ablated and untreated tumor.
ROC analysis confirmed that tumor diagnosis using the change in blood flow in response to
phenylephrine was reliable on days 2 and 7. However, tumor diagnosis using the perfusion
change in response to hydralazine was not reliable. This further supports the assertion that
CT perfusion with phenylephrine might potentially serve as a useful tool in differentiating
viable tumor from inflammation. We also found that the test accuracy of perfusion change in
response to phenylephrine had no statistical significance on day 14. This may have been due
to the limited sample size (n = 8). More animals are needed to enhance the statistical power.
A large variability of blood flow in the tumor was noted in this study. For example, the
blood flow in the untreated viable tumor ranged from 21.4 to 74.3 mL/100 mL/min. No
previous studies have reported the normal range of blood flow in our tumor model.
However, we expect that intrinsic factors, such as blood vessel density and blood pressure,
and extrinsic factors, such as environment temperature and anesthesia, could all influence
baseline flow. Importantly, because we observed the blood flow changes in the same rat and
in the same time period (within 10 minutes) in response to phenylephrine or hydralazine, the
external factors were minimized, and the observed blood flow changed in response to
vasoactive drugs should be reliable.
There were also several limitations to our study. First, we compared the changes in blood
flow in response to vasoactive drugs in untreated viable tumor to those in the peripheral rim
of completely ablated tumor. This may not strictly represent the clinical scenario, in which
the residual viable tumor embeds in the peripheral rim after RF ablation. To discern the
residual tumor from the inflammatory tissue in this rat tumor model using clinical computed
tomography is difficult because of the limited spatial resolution. The commercially available
micro-CT imaging cannot provide sufficient temporal resolution to perform perfusion scans
at this stage.
Second, the subcutaneous tumor model used in the present study has some limitations. Our
preliminary study showed that the ablated tumor would heal with scab formation in 8 to 14
days after treatment. The scab with ablated tissue would fall off after day 14. This allowed
us to follow-up the perfusion changes of the tumor for up to 2 weeks after RF ablation,
although a longer period of observation would be critical in a clinical scenario. On the other
hand, the pattern of blood supply can be unique in each type of tumor model and in
individual organs; hence, we cannot assume that the same outcome will be observed in other
tumor models in different organs. Future studies will be performed with liver or kidney
tumor models in a larger animal model, such as the rabbit, and follow-up with a longer
period to validate the results.
Wu et al. Page 8
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Third, we only tested one vasoconstrictor, phenylephrine, and one vasodilator, hydralazine,
and one dose of each agent. More experiments are needed to explore other vasodilators or
vasoconstrictors and their corresponding dose response.
Fourth, we evaluated the blood vessels with routine histology (hematoxylin and eosin stain).
Although immunohistochemical methods may provide specific information about
microvessel density, prior studies and our own unpublished data suggest that blood flow is
not highly correlated with microvessel density in similar studies (27).
Many parameters, imaging modalities (eg, MR, ultrasound), and calculation models have
been used in perfusion studies (28,29). Other parameters, such as permeability, blood
volume, time to peak, and mean transit time have been used to describe tumor perfusion.
The changes in blood volume and permeability in response to phenylephrine or hydralazine
were investigated as well in this study. However, the difference of the changes of these
parameters between viable tumor and peripheral rim of ablated tumor were not significant
(data not shown). In terms of perfusion models, deconvolution theory is also widely used.
Comparison of the pros and cons of these models and imaging modalities was outside the
scope of the present study. However, using different imaging modalities and perfusion
models to duplicate this study might aid in validating our hypothesis.
CONCLUSION
We found that perfusion in post-ablation peripheral rim and untreated viable tumors reacts
differently to vasoconstrictors, with phenylephrine leading to a marked decrease in blood
flow in the peripheral rim of the ablated tumor and little effect on that in the untreated viable
tumor. CT perfusion with phenylephrine might potentially serve as a useful tool in the
treatment assessment of RF ablation.
REFERENCES
1. Hong K, Georgiades CS, Geschwind JF. Technology insight: Image-guided therapies for
hepatocellular carcinoma—intra-arterial and ablative techniques. Nat Clin Pract Oncol. 2006;
3:315–324. [PubMed: 16757969]
2. Gazelle GS, McMahon PM, Beinfeld MT, Halpern EF, Weinstein MC. Metastatic colorectal
carcinoma: cost-effectiveness of percutaneous radiofrequency ablation versus that of hepatic
resection. Radiology. 2004; 233:729–739. [PubMed: 15564408]
3. Harrison LE, Koneru B, Baramipour P, et al. Locoregional recurrences are frequent after
radiofrequency ablation for hepatocellular carcinoma. J Am Coll Surg. 2003; 197:759–764.
[PubMed: 14585410]
4. Curley SA. Radiofrequency ablation versus resection for resectable colorectal liver metastases: time
for a randomized trial? Ann Surg Oncol. 2008; 15:11–13. [PubMed: 17955298]
5. Lim HK, Choi D, Lee WJ, et al. Hepatocellular carcinoma treated with percutaneous radio-
frequency ablation: evaluation with follow-up multi-phase helical CT. Radiology. 2001; 221:447–
454. [PubMed: 11687689]
6. Lu DS, Yu NC, Raman SS, et al. Radiofrequency ablation of hepatocellular carcinoma: treatment
success as defined by histologic examination of the explanted liver. Radiology. 2005; 234:954–960.
[PubMed: 15681691]
7. Lencioni R, Cioni D, Bartolozzi C. Percutaneous radiofrequency thermal ablation of liver
malignancies: techniques, indications, imaging findings, and clinical results. Abdom Imaging. 2001;
26:345–360. [PubMed: 11441546]
8. Veit P, Antoch G, Stergar H, Bockisch A, Forsting M, Kuehl H. Detection of residual tumor after
radiofrequency ablation of liver metastasis with dual-modality PET/CT: initial results. Eur Radiol.
2006; 16:80–87. [PubMed: 15868122]
Wu et al. Page 9
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

9. Okuma T, Okamura T, Matsuoka T, et al. Fluorine-18-fluorodeoxyglucose positron emission
tomography for assessment of patients with unresectable recurrent or metastatic lung cancers after
CT-guided radiofrequency ablation: preliminary results. Ann Nucl Med. 2006; 20:115–121.
[PubMed: 16615420]
10. Morikawa S, Baluk P, Kaidoh T, Haskell A, Jain RK, McDonald DM. Abnormalities in pericytes
on blood vessels and endothelial sprouts in tumors. Am J Pathol. 2002; 160:985–1000. [PubMed:
11891196]
11. Baluk P, Morikawa S, Haskell A, Mancuso M, McDonald DM. Abnormalities of basement
membrane on blood vessels and endothelial sprouts in tumors. Am J Pathol. 2003; 163:1801–1815.
[PubMed: 14578181]
12. Ekelund L. Pharmacoangiography of the kidney: an overview. Urol Radiol. 1980; 2:9–15.
[PubMed: 7233648]
13. Miles KA, Griffiths MR. Perfusion CT: a worthwhile enhancement? Br J Radiol. 2003; 76:220–
231. [PubMed: 12711641]
14. Zima A, Carlos R, Gandhi D, Case I, Teknos T, Mukherji SK. Can pretreatment CT perfusion
predict response of advanced squamous cell carcinoma of the upper aerodigestive tract treated
with induction chemotherapy? AJNR Am J Neuroradiol. 2007; 28:328–334. [PubMed: 17297007]
15. Cuenod CA, Leconte I, Siauve N, et al. Deconvolution technique for measuring tissue perfusion by
dynamic CT: application to normal and metastatic liver. Acad Radiol. 2002; 9(suppl):S205–S211.
[PubMed: 12019870]
16. Krupka TM, Weinberg BD, Wu H, Ziats NP, Exner AA. Effect of intratumoral injection of
carboplatin combined with pluronic P85 or L61 on experimental colorectal carcinoma in rats. Exp
Biol Med (Maywood). 2007; 232:950–957. [PubMed: 17609512]
17. Tozer GM, Maxwell RJ, Griffiths JR, Pham P. Modification of the 31P magnetic resonance spectra
of a rat tumour using vasodilators and its relationship to hypotension. Br J Cancer. 1990; 62:553–
560. [PubMed: 2223572]
18. Ohyama N, Yamaguchi S. Effects of phenylephrine and prazosin on axial movement of the rat
incisor and arterial blood pressure. Jpn J Pharmacol. 1999; 80:271–274. [PubMed: 10461774]
19. Goldberg SN, Gazelle GS, Compton CC, Mueller PR, Tanabe KK. Treatment of intrahepatic
malignancy with radiofrequency ablation: radiologic-pathologic correlation. Cancer. 2000;
88:2452–2463. [PubMed: 10861420]
20. Hsu CP, Razavi MK, So SK, Parachikov IH, Benaron DA. Liver tumor gross margin identification
and ablation monitoring during liver radiofrequency treatment. J Vasc Interv Radiol. 2005;
16:1473–1478. [PubMed: 16319153]
21. Dromain C, de Baere T, Elias D, et al. Hepatic tumors treated with percutaneous radio-frequency
ablation: CT and MR imaging follow-up. Radiology. 2002; 223:255–262. [PubMed: 11930075]
22. Chaplin DJ, Acker BD, Horsman MR. Reduction of tumour blood flow by vasoactive drugs: a role
in cancer therapy. Biomed Biochim Acta. 1989; 48:S264–S268. [PubMed: 2730621]
23. Chan RC, Babbs CF, Vetter RJ, Lamar CH. Abnormal response of tumor vasculature to vasoactive
drugs. J Natl Cancer Inst. 1984; 72:145–150. [PubMed: 6582294]
24. Bradley JG. Nonprescription drugs and hypertension. Which ones affect blood pressure? Postgrad
Med. 1991; 89:195–197. [PubMed: 2020648]
25. Arce C, Segura-Pacheco B, Perez-Cardenas E, Taja-Chayeb L, Candelaria M, Duennas-Gonzalez
A. Hydralazine target: from blood vessels to the epigenome. J Transl Med. 2006; 4:10. [PubMed:
16507100]
26. Nielsen FU, Topp S, Horsman MR, Overgaard J, Stodkilde-Jorgensen H, Maxwell RJ. Localized in
vivo 1H NMR spectroscopy of murine tumours: effect of blood flow reduction. NMR Biomed.
1999; 12:175–183. [PubMed: 10421908]
27. Phongkitkarun S, Kobayashi S, Kan Z, Lee TY, Charnsangavej C. Quantification of angiogenesis
by functional computed tomography in a Matrigel model in rats. Acad Radiol. 2004; 11:573–582.
[PubMed: 15147622]
28. Lassau N, Chami L, Benatsou B, Peronneau P, Roche A. Dynamic contrast-enhanced
ultrasonography (DCE-US) with quantification of tumor perfusion: a new diagnostic tool to
Wu et al. Page 10
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

evaluate the early effects of antiangiogenic treatment. Eur Radiol. 2007; 17(suppl):F89–F98.
[PubMed: 18376462]
29. Bisdas S, Medov L, Baghi M, et al. A comparison of tumour perfusion assessed by deconvolution-
based analysis of dynamic contrast-enhanced CT and MR imaging in patients with squamous cell
carcinoma of the upper aerodigestive tract. Eur Radiol. 2008; 18:843–850. [PubMed: 18175123]
Wu et al. Page 11
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1.
Correlation of perfusion images ([a] maximum-intensity projection, [b] blood flow) with
histology ([c] 4×, [d] 200×[hematoxylin and eosin]) in an untreated tumor. The central zone
(asterisks) of the tumor presents hyperdense on the maximum-intensity projection image (a),
shows no perfusion (b), and appears necrotic on the corresponding histologic image (c). The
tumor rim consists of densely packed viable tumor cells encased by a hypervascular fibrous
capsule (d).
Wu et al. Page 12
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 2.
Representative hematoxylin and eosin histology ([a,c,e] 4×, [b,d,f] 200×)of tumor
progression after radiofrequency ablation on days 2 (a,b), 7 (c,d), and 14 (e,f). (a) On day 2,
coagulative necrosis was seen in the ablation zone. In the peripheral rim (b) (white square),
a hyperemic reaction zone with dilated vessels and red blood cell infiltration is present. (c)
On day 7, numerous blood vessels were found in the peripheral rim of the ablation zone (d)
(white square). The lumen of these vessels shrank to normal. (e) On day 14, granulation
tissue formed in the margin of the ablation zone (f) (white square), and the quantity of blood
vessels decreased.
Wu et al. Page 13
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 3.
Blood flow in the peripheral rim of ablated tumor (solid line) and untreated viable tumor
(dashed line). Error bar represents the standard error of the mean. *P < .01 versus untreated
tumor (Student's unpaired t test).
Wu et al. Page 14
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 4.
Plots of percentage change of blood flow in untreated viable tumor (a) and peripheral rim of
ablated tumor (b) induced by phenylephrine to baseline blood flow of 24 tested tumors. SD,
standard deviation.
Wu et al. Page 15
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 5.
Maximum-intensity projection (MIP) and blood flow images before and after phenylephrine
injection on days 2, 7, and 14 show that phenylephrine had little effect on blood flow in
untreated tumor (left) and decreased flow in the peripheral rim of ablated tumor (right).
Dotted lines mark the region of untreated viable tumor (left) and the peripheral rim of
ablated tumor (right). L, left; RF, radiofrequency.
Wu et al. Page 16
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 6.
Plots of percentage change of blood flow in untreated tumor (a) and the peripheral rim of
ablated tumor (b) induced by hydralazine to baseline blood flow of 24 tested tumors. SD,
standard deviation.
Wu et al. Page 17
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 7.
Maximum-intensity projection (MIP) and blood flow images before and after hydralazine
injection on days 2, 7, and 14 show that hydralazine decreased the blood flow in both
untreated viable tumor (left) and the peripheral rim of ablated tumor (right). Dotted lines
mark the region of untreated viable tumor (left) and peripheral rim of ablated tumor (right).
L, left; RF, radiofrequency.
Wu et al. Page 18
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 8.
Receiver-operating characteristic curves of the percentage change of blood flow in response
to phenylephrine (solid line) and hydralazine (dashed line) on days 2 (a), 7 (b), and 14 (c).
Wu et al. Page 19
Acad Radiol. Author manuscript; available in PMC 2012 March 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Wu et al. Page 20
Table 1
Changes in Blood Flow (mL/100 mL/min) in the Peripheral Rim of Ablated Tumor, Untreated Viable Tumor, and Muscle in Response to Phenylephrine
Peripheral Rim of Ablated Tumor Untreated Viable Tumor Muscle
Day nPre Post % Change Pre Post % Change Pre Post % Change
2 12 52.8 ± 19.5 44.1 ± 18.2*−16.3 ± 9.7 46.9 ± 12.9 52.2 ± 24.0 11.0 ± 32.7 4.7 ± 3.3 2.9 ± 2.3 −36.8 ± 54.7
2 8 69.2 ± 26.3 52.2 ± 25.1*−24.0 ± 22.6 51.4 ± 8.2 49.0 ± 8.4 −4.5 ± 9.0 4.8 ± 4.7 3.7 ± 3.7 −1.6 ± 56.8
14 4 73.5 ± 14.5 49.5 ± 17.3†−31.1 ± 25.4 46.9 ± 10.0 46.5 ± 16.9 1.8 ± 35.0 5.1 ± 5.3 2.1 ± 1.9 −61.2 ± 17.8
*P < .05.
†P < .01 (paired t test).
Acad Radiol. Author manuscript; available in PMC 2012 March 21.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Wu et al. Page 21
Table 2
Changes in Blood Flow (mL/100 mL/min) in the Peripheral Rim of Ablated Tumor, Untreated Viable Tumor, and Muscle in Response to Hydralazine
Peripheral Rim of Ablated Tumor Untreated Viable Tumor Muscle
Day nPre Post % Change Pre Post % Change Pre Post % Change
2 12 47.2 ± 20.1 38.1 ± 12.8*−15.0 ± 25.9 40.9 ± 11.1 28.8 ± 9.1†−27.5 ± 19.1 5.6 ± 2.3 3.6 ± 2.0†−23.3 ± 52.7
7 8 71.6 ± 17.0 58.1 ± 14.5*−17.7 ± 16.8 50.6 ± 13.4 36.2 ± 10.5*−24.4 ± 32.1 3.4 ± 1.9 2.5 ± 1.7 −16.9 ± 58.5
14 4 64.4 ± 20.6 58.3 ± 31.2 −10.3 ± 30.7 43.7 ± 8.9 30.6 ± 8.7†−30.6 ± 5.1 6.6 ± 3.1 6.6 ± 3.1 2.5 ± 2.1†
*P < .05.
†P < .01 (paired t test).
Acad Radiol. Author manuscript; available in PMC 2012 March 21.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Wu et al. Page 22
Table 3
AUCs and 95% CIs of the Receiver-operating Characteristic Curves of Percentage Changes of Blood Flow in
Response to Phenylephrine and Hydralazine
Phenylephrine Hydralazine
AUC 95% CI AUC 95% CI
Day 2 0.82 0.64–1.00 0.58 0.33–0.82
Day 7 0.81 0.56–1.00 0.66 0.37–0.95
Day 14 0.81 0.46–1.00 0.75 0.26–1.00
Pooled data 0.83 0.71–0.95 0.63 0.46–0.80
AUC, area under the curve; CI, confidence interval.
Acad Radiol. Author manuscript; available in PMC 2012 March 21.