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

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kuszyk, B. S. Articles by Fishman, E. K. Search for Related Content PubMed PubMed Citation Articles by Kuszyk, B. S. Articles by Fishman, E. K.

Local Tumor Recurrence Following Hepatic Cryoablation: Radiologic-histopathologic Correlation in a Rabbit Model¹

¹ From the Russell H. Morgan Department of Radiology and Radiological Science (B.S.K., D.A.B., S.S., C.A.M., K.M.H., J.E., E.K.F.), and the Departments of Pathology (J.K.B.) and Surgery (M.A.C.), the Johns Hopkins Medical Institutions, 601 N Caroline St, Baltimore, MD 21287. From the 1998 RSNA scientific assembly. Received March 10, 1999; revision requested April 23; revision received February 25, 2000; accepted March 30. Supported in part by the National Cancer Institute, grant no. CA 55641-01A1. Address correspondence to E.K.F. (e-mail: efishman@jhmi.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To use radiologic-histopathologic correlation in an animal model to distinguish normal postoperative findings from evidence of residual tumor after cryoablation of malignant hepatic tumors.

MATERIALS AND METHODS: Hepatic cryoablation was performed in 12 rabbits with VX2 tumors and in two healthy rabbits. Nonenhanced and dynamic contrast material–enhanced computed tomography (CT) and magnetic resonance (MR) imaging and power and color Doppler flow ultrasonography (US) were performed 7–8 days after cryoablation. Histopathologic findings were correlated with imaging findings.

RESULTS: Twenty tumors of 5–20 mm (mean, 10 mm) and seven areas of normal liver were treated with cryolesions of 11–21 mm (mean, 15 mm). All cryolesions exhibited arterial phase rim enhancement at CT and MR imaging, and 13 (57%) of 23 lesions demonstrated peripheral flow at US because of granulation tissue. There was macroscopic recurrence in 15 (75%) of 20 treated tumors; 14 (93%) appeared as peripheral nodularity with low-grade enhancement. Necrotic tissue did not enhance. Intact vessels extended up to 6 mm inside cryolesion margins and caused focal internal enhancement and Doppler flow. Areas of high signal intensity on T2-weighted MR images correlated with liquefaction necrosis, granulation tissue, and tumor.

CONCLUSION: In this animal model, recurrent tumor typically appeared as focal nodules at the cryolesion periphery. Rim and central foci of enhancement, Doppler flow, and increased signal intensity on T2-weighted MR images can be normal findings after hepatic cryoablation.

Index terms: Animals • Cryotherapy, 761.12168 • Liver, interventional procedures, 761.12168 • Liver, CT, 761.12111, 761.12113, 761.12115 • Liver, MR, 761.121411, 761.121412, 761.121415, 761.12143 • Liver, US, 761.12981, 761.12983, 761.12984 • Liver neoplasms, 761.12168

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Initial clinical results have shown that cryoablation of malignant hepatic tumors is associated with reduced morbidity and similar 5-year survival compared with those associated with hepatic resection (1–7). Computed tomography (CT) is typically performed in the 1st week after cryoablation to rule out postprocedural complications, to grossly assess the adequacy of treatment, and to serve as a baseline for subsequent follow-up imaging (8–10). The normal CT appearance of hepatic cryolesions at that time is that of a low-attenuating lesion that extends to the liver capsule and may exhibit rim enhancement (8,9). This appearance can be similar to that of hepatic tumors, which makes early detection of residual or recurrent tumor at the cryoablation site difficult. It is impossible to precisely correlate imaging findings with local tumor recurrence in patients because complete histopathologic correlation is not available. As a result, imaging features of recurrent or residual tumor after cryoablation are poorly understood.

The problem of accurately assessing the adequacy of tumor treatment extends to a number of emerging techniques for liver tumor ablation. Ablation techniques are attractive because they are amenable to minimally invasive approaches that can reduce morbidity, as compared with resection. Percutaneous ethanol ablation (11–13), radio-frequency ablation (14–16), microwave coagulation therapy (17–20), transcatheter chemoembolization (21,22), laser therapy (23), and interstitial radiation therapy (24) have all been evaluated, with varying degrees of success. Unlike surgical resection, ablation does not allow direct visualization of the tumor or provide a histopathologic specimen for evaluation of treatment margins. Ablative therapies rely on imaging for procedural guidance and for assessment of treatment adequacy. An understanding of normal and abnormal imaging characteristics is needed to assess treatment adequacy. The purpose of this study was to correlate findings at CT, magnetic resonance (MR) imaging, and ultrasonography (US) with histopathologic findings to identify findings that suggest local tumor recurrence after the cryoablation of malignant hepatic tumors.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Animal and Tumor Model
The animal care and use committee at our institution approved this study protocol. The VX2 tumor model used in this study was initially a virus-induced papilloma first seen in the domestic rabbit in 1937 (25). With sequential transplantations, the tumor line became increasingly anaplastic. VX2 carcinoma (EG&G Mason Research Institute, Worchester, Mass) has been carried for approximately the past 15 years at our institution by repeatedly transferring tumor every 14–21 days by way of intramuscular or subcutaneous implantation into the thighs of male New Zealand White rabbits. This widely available tumor model was chosen in part because it is aggressive and grows rapidly, making macroscopic recurrences at the treatment site more likely.

Sixteen male New Zealand White rabbits (Robinson Services, Clemens, NC) weighing 3.38–4.24 kg were used. Prior to all procedures, including tumor implantation, cryoablation, and imaging, rabbits received an intramuscular injection of 1.0 mg/kg of body weight of acepromazine maleate (Fermenta Animal Health, Kansas City, Mo) and 35 mg/kg of ketamine hydrochloride (Ketaject; Phoenix Scientific, St Joseph, Mo). Intravenous access was then acquired via a marginal ear vein. Animals were intubated with a noncuffed endotracheal tube and connected to a small animal ventilator (Harvard Apparatus, South Natick, Mass). Anesthesia was maintained by using 35 mg/kg of intravenously administered sodium pentothal (Thiopental; Abbott Laboratories, North Chicago, Ill).

Twenty-five separate tumors were implanted into the livers of 14 rabbits. The liver was exposed by performing midline laparotomy. VX2 tumor fragments of approximately 1 mm³ were injected intraparenchymally into the right lobe of the liver, in two rabbits; into the left lobe, in one rabbit; or into both lobes, in 11 rabbits by using a 16-gauge angiocatheter with the needle removed and with a .045-inch guide wire as a pusher.

Cryoablation
Cryoablation was performed 13–25 days after tumor implantation (mean, 18 days) in the 12 tumor-bearing animals and in the two healthy animals. The animals were placed under general anesthesia as described and were positioned on a heating pad. A hand-held cryotherapy unit (Cryostar; Cryogenic Technologies, Golden Valley, Minn) designed for animal use and not available commercially was used. After a 4-cm midline incision was made, the 2-mm cryoprobe of the cryotherapy unit was inserted into the tumor or the normal liver by using direct visualization and palpation. Intraoperative US was not performed to guide or monitor the freezing, because of the small size of the rabbit liver. Liquid nitrogen (-190°C) was circulated through the probe for 2–3 minutes per freezing. A short freezing time was intentionally used because the goal of this study was to characterize the imaging appearance of local tumor recurrence and not to document the efficacy of hepatic cryoablation. Freezing was monitored by using direct visualization and palpation to ensure that the resulting ice ball encompassed the tumor and extended at least 3 mm beyond the tumor margin, where possible. Tumors were subjected to two freeze-thaw cycles, with an interfreezing thaw of approximately 2 minutes. Hemostasis was achieved after the final thaw by using compression and absorbable gelatin sponge (Gelfoam; Upjohn, Kalamazoo, Mich) as needed. Two cryolesions were created in each animal: one in the right lobe and one in the left lobe. The three animals with only one tumor underwent cryoablation of the normal liver parenchyma in the contralateral lobe. Two animals that were implanted with tumor did not undergo cryoablation after implantation because of the absence of detectable tumor at preoperative imaging (n = 1) or anesthesia-related death (n = 1). Only one of two tumor implantation attempts was successful in one animal. A total of 20 tumors in 12 animals were ablated.

Imaging
Liver imaging was performed with CT, MR imaging, and US 1 day prior to cryoablation and 7 or 8 days after cryoablation. The two animals without tumor also underwent imaging at 1 and 7 months after cryoablation. All three examinations were performed within 24 hours of each other in all cases. Animals were anesthetized and intubated as described earlier and were paralyzed by using intravenous pancuronium (Elkins-Sinn, Cherry Hill, NJ) immediately prior to all imaging.

CT was performed by using a Somatom Plus 4 scanner (Siemens Medical Systems, Iselin, NJ). An initial nonenhanced helical study of the entire liver was performed by using 3-mm collimation, 3 mm/sec table speed in a craniocaudal direction, 0.75 seconds per gantry rotation, a field of view of 100–164 mm, 292 mA, and 120 kVp. Ventilation was suspended for the entire helical study. Transverse reconstructions were performed every 4 mm by using 180° linear interpolation. The level of the lesion(s) was determined from the nonenhanced helical images. A dynamic contrast-enhanced examination was then performed at that level after the intravenous power injection of 8–9 mL of nonionic iodinated contrast material (300 milligrams of iodine per mL iohexol [Omnipaque; Nycomed, Princeton, NJ], 2 mL/kg, 2.4–2.7 g iodine) at 2 mL/sec via an ear vein. Dynamic scanning began at the completion of contrast material injection, with 2-mm collimation, 143 mA, and 140 kVp. Scanning was performed every 5 seconds for the 1st 3 minutes 20 seconds and was followed by additional scanning at 4 and 5 minutes. Ventilation was suspended for 20-second intervals alternated with 10-second periods of hyperventilation.

MR imaging was performed by using a 1.5-T system (Signa Horizon 5.6; GE Medical Systems, Milwaukee, Wis) with an extremity coil. Transverse conventional spin-echo T1-weighted and fast spin-echo T2-weighted images were obtained by using a 4-mm section thickness, 1-mm intersection spacing, four signals acquired, and a 16 x 16-cm field of view. For T1-weighted spin-echo images, the repetition time was 450–500 msec, the echo time was 14–20 msec (450–500/14–20), and the matrix was 256 x 160. For T2-weighted fast spin-echo images (4,000/105), the matrix was 256 x 224, and the echo train length was 16. Nonenhanced and dynamic contrast-enhanced T1-weighted fast multiplanar spoiled gradient-echo images with fat saturation (100/4.2, 60° flip angle) were obtained with suspended ventilation and a 5-mm section thickness, 1.5-mm intersection spacing, a 256 x 128 matrix, and a 16 x 12–16-cm field of view. Five sections were obtained in 12 seconds by using this sequence. Acquisitions were made 0, 20, 40, 60, 90, 120, 180, and 240 seconds after a bolus intravenous administration of 0.1 mmol/kg gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) at 2 mL/sec followed by a 5-mL normal saline flush.

US of the liver was performed by using a model 3000 imager (ATL, Bothell, Wash), with a linear L10-5-MHz 38-mm transducer. Imaging was performed by a radiologist (S.S.), who was aided by a technologist. Lesion vascularity was imaged by using color Doppler flow and power Doppler imaging. Color settings were adjusted for optimal slow flow detection. Color gain was adjusted at the beginning of each examination by selecting the highest gain value at which artifacts did not degrade the image. Ventilation was suspended for all Doppler US and as needed for gray-scale US.

Pathologic Analysis
All animals were sacrificed by means of intravenous pentobarbital (Veterinary Laboratories, Lenexa, Kan) overdose immediately after the completion of postcryoablation imaging (7 or 8 days after cryoablation in tumor-bearing animals, 7 months after cryoablation in non–tumor-bearing animals). The livers were removed and fixed in formalin. Each liver was then sectioned in the transverse plane at 1–3-mm increments, and the resulting slices were photographed for gross correlation with imaging findings. Samples from each cryoablation site were then obtained, embedded in paraffin, and stained with hematoxylin-eosin. A liver pathologist (J.K.B.) and a radiologist (B.S.K.) evaluated the gross and microscopic findings at each cryoablation site, and a consensus was reached.

Image Analysis
Pre- and postcryoablation images in all 14 animals were reviewed by three radiologists (D.A.B., S.S., E.K.F.), who read side by side and reached a consensus in each case. Histopathologic findings were then correlated with imaging findings in a lesion-by-lesion fashion (B.S.K.). Each lesion was evaluated for size; shape; extension to the liver periphery; attenuation, echogenicity, or signal intensity; enhancement; and air, hemorrhage, or vascular changes. The imaging appearance of each area of recurrent tumor seen at histopathologic analysis was tabulated.

Statistical Analysis
Comparisons of imaging finding percentages were performed with the two-way Fisher exact test (Stata Version 6.0, Stata, College Station, Tex). A P value of .05 or less was considered to indicate a significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Precryoablation Imaging
CT depicted all 20 untreated VX2 tumors as low-attenuating liver masses with low-level central enhancement of approximately 30 HU. Moderate peripheral rim enhancement on arterial phase images was present in 17 (85%) of the 20 lesions.

MR imaging also depicted all tumors on precryoablation images. All untreated tumors had predominantly low signal intensity on T1-weighted images and intermediate signal intensity on T2-weighted images. Central areas of high signal intensity on T2-weighted images and low signal intensity on T1-weighted images that were compatible with central necrosis were seen in four (20%) of the 20 tumors, all of which were larger than 1 cm. Gadolinium-enhanced images showed low-level central enhancement and peripheral arterial phase rim enhancement in all 20 (100%) tumors.

US depicted 15 (75%) of the 20 tumors: Lesions appeared as predominantly hypoechoic in nine (60%), as hyperechoic in five (33%), and as mixed in one (7%). Peripheral increased flow at color Doppler flow and power Doppler US was seen in nine (60%) of the 15 tumors.

Surgical and Histopathologic Findings
Laparotomy at the time of cryoablation revealed 20 liver tumors 5–20 mm in diameter (mean diameter, 10 mm). The tumors were treated with cryolesions of 11–21 mm in diameter (mean diameter, 15 mm). Seven cryolesions measuring 15 mm in diameter were created in normal liver. All cryolesions appeared to encompass the treated tumor at visual inspection and palpation performed at the time of cryoablation.

Histopathologic analysis of the cryolesions showed mainly coagulation necrosis (Figs 1–5). Central areas of liquefaction necrosis and cavitation were present in larger lesions (Fig 4). A rim of granulation tissue and inflammation that was 1–5-mm thick was present at the cryolesion margin in all cases. Scar tissue with minimal active granulation tissue was seen at the interface between macroscopic tumor recurrences and adjacent liver. Several cryolesions had intact portal venous tracks surrounded by necrosis, with patent hepatic arterial and/or portal venous branches extending as much as 6 mm into the cryolesion. In two lesions, microscopic foci of viable tumor were seen immediately adjacent to intact portal venous tracks within otherwise necrotic cryolesions (Fig 5).

View larger version (109K): [in this window] [in a new window] [Download PPT slide]   Figure 1a. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (103K): [in this window] [in a new window] [Download PPT slide]   Figure 1b. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (137K): [in this window] [in a new window] [Download PPT slide]   Figure 1c. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (122K): [in this window] [in a new window] [Download PPT slide]   Figure 1d. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (129K): [in this window] [in a new window] [Download PPT slide]   Figure 1e. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (111K): [in this window] [in a new window] [Download PPT slide]   Figure 1f. Images obtained 1 week after procedure of cryolesion in normal liver. (a) Contrast-enhanced transverse arterial phase CT scan shows no enhancement in the majority of the central cryolesion (arrow), with attenuating rim enhancement (arrowhead) and focal central enhancement. (b) Contrast-enhanced transverse portal venous phase CT scan shows resolution of the rim enhancement and no enhancement in most of the low-attenuating lesion (arrow). (c) Gadolinium-enhanced transverse T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) shows rim enhancement (arrowhead) and no enhancement in most of the central lesion (arrow). (d) Transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105). The majority of the lesion demonstrates low signal intensity (arrow), but there is a rim of increased signal intensity (arrowhead) that corresponds to the area of enhancement seen in a. (e) Transverse color Doppler US image shows a complex hypoechoic cryolesion (arrow), with foci of flow at the periphery (arrowhead) and a small focus of flow within the lesion. (f) Gross specimen sectioned in the transverse plane shows that the majority of the central cryolesion is composed of coagulation necrosis (arrow). The pearly rim is granulation tissue (arrowhead) at the cryolesion periphery, which corresponds to the areas of enhancement and high signal intensity seen at T2-weighted MR imaging.

View larger version (134K): [in this window] [in a new window] [Download PPT slide]   Figure 2a. Hepatic cryoablation, with large local recurrence. (a) Contrast-enhanced transverse portal venous phase CT scan shows a low-attenuating tumor (arrow) in the left lobe of the liver. Transverse (b) contrast-enhanced arterial-phase and (c) portal venous phase CT scans obtained 1 week after cryoablation show a large peripheral nodule that suggests recurrent tumor (T), with low-grade central enhancement and outward displacement of the arterial phase rim enhancement that are compatible with a large local recurrence. No enhancement is present in the necrotic portion of the cryolesion (straight arrow). Subtle arterial-phase rim enhancement (arrowhead in b) is seen. The cryolesion created in normal liver in the right lobe contains foci of intense enhancement that are compatible with vascular malformations (curved arrows). Wedge-shaped areas of enhancement in the liver parenchyma are adjacent to the right lobe lesion. The gallbladder (G) has a thick, enhancing wall and central areas of high attenuation, which are compatible with hemorrhage due to accidental cryoinjury. (d) Gross specimen sectioned in the transverse plane confirms a large local recurrence (T) with central necrosis on the left. The necrotic portion of the cryolesion contains coagulation necrosis peripherally (straight arrow) and early liquefaction necrosis centrally. A rim of granulation tissue (arrowhead) that corresponds to enhancement is present. Vascular malformation (curved arrow) is present within the cryolesion on the right, and hemorrhagic cholecystitis (G) is seen.

View larger version (148K): [in this window] [in a new window] [Download PPT slide]   Figure 2b. Hepatic cryoablation, with large local recurrence. (a) Contrast-enhanced transverse portal venous phase CT scan shows a low-attenuating tumor (arrow) in the left lobe of the liver. Transverse (b) contrast-enhanced arterial-phase and (c) portal venous phase CT scans obtained 1 week after cryoablation show a large peripheral nodule that suggests recurrent tumor (T), with low-grade central enhancement and outward displacement of the arterial phase rim enhancement that are compatible with a large local recurrence. No enhancement is present in the necrotic portion of the cryolesion (straight arrow). Subtle arterial-phase rim enhancement (arrowhead in b) is seen. The cryolesion created in normal liver in the right lobe contains foci of intense enhancement that are compatible with vascular malformations (curved arrows). Wedge-shaped areas of enhancement in the liver parenchyma are adjacent to the right lobe lesion. The gallbladder (G) has a thick, enhancing wall and central areas of high attenuation, which are compatible with hemorrhage due to accidental cryoinjury. (d) Gross specimen sectioned in the transverse plane confirms a large local recurrence (T) with central necrosis on the left. The necrotic portion of the cryolesion contains coagulation necrosis peripherally (straight arrow) and early liquefaction necrosis centrally. A rim of granulation tissue (arrowhead) that corresponds to enhancement is present. Vascular malformation (curved arrow) is present within the cryolesion on the right, and hemorrhagic cholecystitis (G) is seen.

View larger version (142K): [in this window] [in a new window] [Download PPT slide]   Figure 2c. Hepatic cryoablation, with large local recurrence. (a) Contrast-enhanced transverse portal venous phase CT scan shows a low-attenuating tumor (arrow) in the left lobe of the liver. Transverse (b) contrast-enhanced arterial-phase and (c) portal venous phase CT scans obtained 1 week after cryoablation show a large peripheral nodule that suggests recurrent tumor (T), with low-grade central enhancement and outward displacement of the arterial phase rim enhancement that are compatible with a large local recurrence. No enhancement is present in the necrotic portion of the cryolesion (straight arrow). Subtle arterial-phase rim enhancement (arrowhead in b) is seen. The cryolesion created in normal liver in the right lobe contains foci of intense enhancement that are compatible with vascular malformations (curved arrows). Wedge-shaped areas of enhancement in the liver parenchyma are adjacent to the right lobe lesion. The gallbladder (G) has a thick, enhancing wall and central areas of high attenuation, which are compatible with hemorrhage due to accidental cryoinjury. (d) Gross specimen sectioned in the transverse plane confirms a large local recurrence (T) with central necrosis on the left. The necrotic portion of the cryolesion contains coagulation necrosis peripherally (straight arrow) and early liquefaction necrosis centrally. A rim of granulation tissue (arrowhead) that corresponds to enhancement is present. Vascular malformation (curved arrow) is present within the cryolesion on the right, and hemorrhagic cholecystitis (G) is seen.

View larger version (96K): [in this window] [in a new window] [Download PPT slide]   Figure 2d. Hepatic cryoablation, with large local recurrence. (a) Contrast-enhanced transverse portal venous phase CT scan shows a low-attenuating tumor (arrow) in the left lobe of the liver. Transverse (b) contrast-enhanced arterial-phase and (c) portal venous phase CT scans obtained 1 week after cryoablation show a large peripheral nodule that suggests recurrent tumor (T), with low-grade central enhancement and outward displacement of the arterial phase rim enhancement that are compatible with a large local recurrence. No enhancement is present in the necrotic portion of the cryolesion (straight arrow). Subtle arterial-phase rim enhancement (arrowhead in b) is seen. The cryolesion created in normal liver in the right lobe contains foci of intense enhancement that are compatible with vascular malformations (curved arrows). Wedge-shaped areas of enhancement in the liver parenchyma are adjacent to the right lobe lesion. The gallbladder (G) has a thick, enhancing wall and central areas of high attenuation, which are compatible with hemorrhage due to accidental cryoinjury. (d) Gross specimen sectioned in the transverse plane confirms a large local recurrence (T) with central necrosis on the left. The necrotic portion of the cryolesion contains coagulation necrosis peripherally (straight arrow) and early liquefaction necrosis centrally. A rim of granulation tissue (arrowhead) that corresponds to enhancement is present. Vascular malformation (curved arrow) is present within the cryolesion on the right, and hemorrhagic cholecystitis (G) is seen.

View larger version (145K): [in this window] [in a new window] [Download PPT slide]   Figure 3a. Hepatic cryolesions with microscopic local recurrence in the right lobe and arteriovenous malformation but no evidence of recurrence in the left lobe. (a) Contrast-enhanced transverse arterial phase CT scan obtained 1 week after cryoablation shows low-attenuating peripheral wedge-shaped cryolesions in the right and left lobes, with rim enhancement (arrowhead). The left lobe lesion shows early central enhancement (long arrow) that is compatible with arteriovenous malformation. Wedge-shaped enhancement (short arrow) is present in the liver parenchyma adjacent to the left lobe lesion; thrombus was found within small portal venous branches in this region. (b) Transverse Doppler US image obtained in the left lobe lesion 1 week after cryoablation confirms central flow, with a low resistive index in the left lobe lesion that is compatible with arteriovenous shunting. (c) Transverse histopathologic specimen of right lobe lesion shows a microscopic focus of tumor recurrence (T) that is centered within the granulation tissue (arrowheads) at the cryolesion periphery and does not distort the peripheral contour; therefore, it is not visible in images. (Hematoxylin-eosin stain; original magnification, x48.) (d) Transverse histopathologic specimen of left lobe lesion shows a vascular malformation that is similar to peliosis (long arrow) and extends into the center of the cryolesion. No recurrent or residual tumor was identified. (Hematoxylin-eosin stain; original magnification, x30.)

View larger version (124K): [in this window] [in a new window] [Download PPT slide]   Figure 3b. Hepatic cryolesions with microscopic local recurrence in the right lobe and arteriovenous malformation but no evidence of recurrence in the left lobe. (a) Contrast-enhanced transverse arterial phase CT scan obtained 1 week after cryoablation shows low-attenuating peripheral wedge-shaped cryolesions in the right and left lobes, with rim enhancement (arrowhead). The left lobe lesion shows early central enhancement (long arrow) that is compatible with arteriovenous malformation. Wedge-shaped enhancement (short arrow) is present in the liver parenchyma adjacent to the left lobe lesion; thrombus was found within small portal venous branches in this region. (b) Transverse Doppler US image obtained in the left lobe lesion 1 week after cryoablation confirms central flow, with a low resistive index in the left lobe lesion that is compatible with arteriovenous shunting. (c) Transverse histopathologic specimen of right lobe lesion shows a microscopic focus of tumor recurrence (T) that is centered within the granulation tissue (arrowheads) at the cryolesion periphery and does not distort the peripheral contour; therefore, it is not visible in images. (Hematoxylin-eosin stain; original magnification, x48.) (d) Transverse histopathologic specimen of left lobe lesion shows a vascular malformation that is similar to peliosis (long arrow) and extends into the center of the cryolesion. No recurrent or residual tumor was identified. (Hematoxylin-eosin stain; original magnification, x30.)

View larger version (116K): [in this window] [in a new window] [Download PPT slide]   Figure 3c. Hepatic cryolesions with microscopic local recurrence in the right lobe and arteriovenous malformation but no evidence of recurrence in the left lobe. (a) Contrast-enhanced transverse arterial phase CT scan obtained 1 week after cryoablation shows low-attenuating peripheral wedge-shaped cryolesions in the right and left lobes, with rim enhancement (arrowhead). The left lobe lesion shows early central enhancement (long arrow) that is compatible with arteriovenous malformation. Wedge-shaped enhancement (short arrow) is present in the liver parenchyma adjacent to the left lobe lesion; thrombus was found within small portal venous branches in this region. (b) Transverse Doppler US image obtained in the left lobe lesion 1 week after cryoablation confirms central flow, with a low resistive index in the left lobe lesion that is compatible with arteriovenous shunting. (c) Transverse histopathologic specimen of right lobe lesion shows a microscopic focus of tumor recurrence (T) that is centered within the granulation tissue (arrowheads) at the cryolesion periphery and does not distort the peripheral contour; therefore, it is not visible in images. (Hematoxylin-eosin stain; original magnification, x48.) (d) Transverse histopathologic specimen of left lobe lesion shows a vascular malformation that is similar to peliosis (long arrow) and extends into the center of the cryolesion. No recurrent or residual tumor was identified. (Hematoxylin-eosin stain; original magnification, x30.)

View larger version (123K): [in this window] [in a new window] [Download PPT slide]   Figure 3d. Hepatic cryolesions with microscopic local recurrence in the right lobe and arteriovenous malformation but no evidence of recurrence in the left lobe. (a) Contrast-enhanced transverse arterial phase CT scan obtained 1 week after cryoablation shows low-attenuating peripheral wedge-shaped cryolesions in the right and left lobes, with rim enhancement (arrowhead). The left lobe lesion shows early central enhancement (long arrow) that is compatible with arteriovenous malformation. Wedge-shaped enhancement (short arrow) is present in the liver parenchyma adjacent to the left lobe lesion; thrombus was found within small portal venous branches in this region. (b) Transverse Doppler US image obtained in the left lobe lesion 1 week after cryoablation confirms central flow, with a low resistive index in the left lobe lesion that is compatible with arteriovenous shunting. (c) Transverse histopathologic specimen of right lobe lesion shows a microscopic focus of tumor recurrence (T) that is centered within the granulation tissue (arrowheads) at the cryolesion periphery and does not distort the peripheral contour; therefore, it is not visible in images. (Hematoxylin-eosin stain; original magnification, x48.) (d) Transverse histopathologic specimen of left lobe lesion shows a vascular malformation that is similar to peliosis (long arrow) and extends into the center of the cryolesion. No recurrent or residual tumor was identified. (Hematoxylin-eosin stain; original magnification, x30.)

View larger version (141K): [in this window] [in a new window] [Download PPT slide]   Figure 4a. Cryolesion with central liquefaction necrosis, a thick rind of granulation tissue, and macroscopic tumor recurrence. (a) Transverse postcryoablation gadolinium-enhanced T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) and (b) transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105) show a cryolesion (L) with a peripheral nodule that is compatible with local tumor recurrence (T) and distorts the otherwise smooth lesion contour and displaces the rim enhancement and increased signal intensity (arrowheads) outward. The majority of the cryolesion has signal intensity characteristics of fluid, with no evidence of enhancement. (c) Transverse color Doppler US image shows a peripheral soft-tissue nodule (T) at the margin of a complex hypoechoic cryolesion (L). (d) Transverse gross specimen confirms an 11-mm recurrence (T) at the medial margin of the lesion. The center of the lesion was fluid due to liquefaction necrosis (L). A thick rind of granulation tissue (arrowhead) surrounds the remainder of the lesion and corresponds to a rim of contrast enhancement and increased signal intensity seen at T2-weighted MR imaging.

View larger version (152K): [in this window] [in a new window] [Download PPT slide]   Figure 4b. Cryolesion with central liquefaction necrosis, a thick rind of granulation tissue, and macroscopic tumor recurrence. (a) Transverse postcryoablation gadolinium-enhanced T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) and (b) transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105) show a cryolesion (L) with a peripheral nodule that is compatible with local tumor recurrence (T) and distorts the otherwise smooth lesion contour and displaces the rim enhancement and increased signal intensity (arrowheads) outward. The majority of the cryolesion has signal intensity characteristics of fluid, with no evidence of enhancement. (c) Transverse color Doppler US image shows a peripheral soft-tissue nodule (T) at the margin of a complex hypoechoic cryolesion (L). (d) Transverse gross specimen confirms an 11-mm recurrence (T) at the medial margin of the lesion. The center of the lesion was fluid due to liquefaction necrosis (L). A thick rind of granulation tissue (arrowhead) surrounds the remainder of the lesion and corresponds to a rim of contrast enhancement and increased signal intensity seen at T2-weighted MR imaging.

View larger version (147K): [in this window] [in a new window] [Download PPT slide]   Figure 4c. Cryolesion with central liquefaction necrosis, a thick rind of granulation tissue, and macroscopic tumor recurrence. (a) Transverse postcryoablation gadolinium-enhanced T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) and (b) transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105) show a cryolesion (L) with a peripheral nodule that is compatible with local tumor recurrence (T) and distorts the otherwise smooth lesion contour and displaces the rim enhancement and increased signal intensity (arrowheads) outward. The majority of the cryolesion has signal intensity characteristics of fluid, with no evidence of enhancement. (c) Transverse color Doppler US image shows a peripheral soft-tissue nodule (T) at the margin of a complex hypoechoic cryolesion (L). (d) Transverse gross specimen confirms an 11-mm recurrence (T) at the medial margin of the lesion. The center of the lesion was fluid due to liquefaction necrosis (L). A thick rind of granulation tissue (arrowhead) surrounds the remainder of the lesion and corresponds to a rim of contrast enhancement and increased signal intensity seen at T2-weighted MR imaging.

View larger version (119K): [in this window] [in a new window] [Download PPT slide]   Figure 4d. Cryolesion with central liquefaction necrosis, a thick rind of granulation tissue, and macroscopic tumor recurrence. (a) Transverse postcryoablation gadolinium-enhanced T1-weighted fast multiplanar spoiled gradient-echo MR image with fat saturation (100/4.2, 60° flip angle) and (b) transverse T2-weighted fast spin-echo MR image with fat saturation (4,000/105) show a cryolesion (L) with a peripheral nodule that is compatible with local tumor recurrence (T) and distorts the otherwise smooth lesion contour and displaces the rim enhancement and increased signal intensity (arrowheads) outward. The majority of the cryolesion has signal intensity characteristics of fluid, with no evidence of enhancement. (c) Transverse color Doppler US image shows a peripheral soft-tissue nodule (T) at the margin of a complex hypoechoic cryolesion (L). (d) Transverse gross specimen confirms an 11-mm recurrence (T) at the medial margin of the lesion. The center of the lesion was fluid due to liquefaction necrosis (L). A thick rind of granulation tissue (arrowhead) surrounds the remainder of the lesion and corresponds to a rim of contrast enhancement and increased signal intensity seen at T2-weighted MR imaging.

View larger version (126K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Viable tumor adjacent to intact vessel is within the gross cryoablation margins. Histopathologic specimen shows a viable tumor (open arrow) in the center of the cryolesion, 6 mm inside its gross margins. This tumor focus is in the same region as the original tumor and is centered on intact portal venous track blood vessels that extend through adjacent necrotic tissue and into the center of the lesion (curved arrows). L = normal liver, arrowheads = granulation tissue. (Hematoxylin-eosin stain; original magnification, x30.)

Organized bland thrombus was present within small portal venous branches in the untreated liver parenchyma that surrounded several cryolesions. There was no evidence of thrombosis of large arterial or venous branches outside the ablated area. Blood lakes without endothelial lining that were similar to the lesions seen in patients with peliosis were seen within cryolesions at 7–8 days in treated tumor and in treated areas of normal liver (Fig 3).

In the two non–tumor-bearing animals that were sacrificed 7 months after cryoablation, no histologic evidence of residual necrosis, inflammation, or granulation tissue was found. No gross or histologic abnormality was identified at the cryoablation sites in one animal. In the other animal, capsular retraction was present at both treatment sites, with adjacent blood lakes similar to those seen within the cryolesions at 1 week.

Postcryoablation Imaging
Table 1 summarizes the features at CT, MR, and US that corresponded with the primary histopathologic components of the hepatic cryolesions 1 week after cryoablation.

All 14 (100%) focal macroscopic recurrences appeared as low-attenuating peripheral nodules (Figs 2, 4, 5). This finding was not seen in any of the seven cryolesions in normal liver or in treated tumor without macroscopic evidence of recurrence. This difference was significant (Table 2). Twelve (100%) of 12 focal lesions evaluated by using contrast material showed low-grade enhancement and focal outward displacement of the rim. Central water attenuation areas were seen in larger recurrences that corresponded to areas of necrosis. The single nonfocal macroscopic recurrence appeared at CT as a rind of intense arterial phase enhancement, without focal displacement of the cryolesion contour.

Serial imaging in the two non–tumor-bearing animals at 1 month showed small areas of decreased attenuation, with adjacent retraction of the liver capsule. The lesions were smaller at 1 month than at initial imaging at 1 week and showed no evidence of rim enhancement. CT findings at 7 months were normal in one animal and showed only focal retraction of the liver capsule at the cryolesion sites in the other animal.

MR imaging.—All cryolesions were wedge shaped and extended to the liver capsule. Lesions varied from having low signal intensity to having slightly increased signal intensity on T1-weighted MR images (Fig 1). Areas of increased signal intensity on T1-weighted images that were compatible with hemorrhage (methemoglobin) were seen in 20 (74%) of 27 cryolesions overall, which included 15 (75%) of 20 treated tumors and five (71%) of seven cryolesions in normal liver. On T2-weighted images, areas of mixed (high, intermediate, and low) signal intensity were observed in cryolesions with and without recurrent tumor (Fig 3). Central areas of high signal intensity on T2-weighted images correlated with areas of liquefaction necrosis, which were more common in larger lesions. A peripheral rim of high signal intensity with varying thickness on T2-weighted images correlated with granulation tissue, which exhibited intense enhancement on contrast-enhanced images that was similar to its appearance on CT scans (Fig 1). Areas of low to intermediate signal intensity within the lesion on T2-weighted images corresponded to coagulation necrosis. Macroscopic foci of recurrent tumor appeared as areas of peripheral nodularity, with outward displacement of the cryolesion contour, intermediate to high signal intensity on T2-weighted images, low-grade central enhancement, and focal outward displacement of the cryolesion rim enhancement. Recurrent tumor had lower signal intensity on T2-weighted images than areas of liquefaction necrosis but had signal characteristics similar to those of granulation tissue. The macroscopic recurrence that appeared as a rind of tumor had signal intensity characteristics similar to those of the focal recurrences, without distortion of the lesion contour.

Wedge-shaped arterial phase enhancement of the adjacent liver was seen on gadolinium-enhanced images in 12 (48%) of 25 lesions, similar to the enhancement seen at CT. Subtle increased signal intensity on T2-weighted images was noted in a similar distribution in 10 (83%) of 12 areas of enhancement. The blood lakes identified within the cryolesions at histopathologic examination demonstrated intense arterial phase enhancement and high signal intensity on T2-weighted images.

Serial MR imaging in the two non–tumor-bearing animals at 1 month showed small areas of low signal intensity on T1-weighted images and high signal intensity on T2-weighted images at the cryoablation site, with adjacent retraction of the liver capsule. The lesions were smaller at 1 month than at 1 week and had no rim enhancement. MR images obtained at 7 months were normal in one animal and showed focal retraction of the liver capsule in the other animal, with no abnormal signal intensity or enhancement.

US.—US depicted 23 (92%) of 25 cryolesions. Two lesions in one animal were not evaluated with US because of anesthesia-related death. The remaining two lesions, which did not have histopathologic evidence of macroscopic recurrence, could not be identified, even with prior knowledge of their location. Cryolesions appeared predominantly hyperechoic, as compared with liver, in 13 (57%) of 23 lesions and appeared predominantly hypoechoic in 10 (43%). Central hypoechoic areas corresponding to liquefaction necrosis were seen in 18 (78%) lesions (Fig 4). Increased flow at the margin of the cryolesion was identified at power and color Doppler flow US (Fig 1) in 13 (57%) lesions, which included five (50%) of 10 lesions without macroscopic recurrence and eight (62%) of 13 lesions with macroscopic recurrence. Nodularity corresponding with recurrence of tumor was seen at the cryolesion margin in eight (62%) of 13 lesions with proved macroscopic recurrences. Nodularity was not seen at the margin of any lesions without proved macroscopic recurrences. Blood lakes identified at histopathologic analysis appeared as areas of low-resistance flow within the cryolesions, which was compatible with arteriovenous shunting (Fig 3). None of the four lesions in the two non–tumor-bearing animals were seen at US at 1 or 7 months after cryosurgery. No abnormal vascularity was seen in the region of the cryolesions at delayed imaging.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Cryolesion Imaging
The results of this study show that recurrent or residual tumors at hepatic cryoablation sites typically occur as a focal nodule at the cryolesion margin that is centered within the rim of granulation tissue; this tissue is found at the interface between the cryolesion and the adjacent untreated liver at 1 week. Local recurrences were best seen at contrast-enhanced CT and MR imaging as areas of outward displacement of the normally smooth cryolesion border. The CT appearance of VX2 carcinoma is similar to that of hypovascular tumors such as hepatic colorectal metastases in humans. In contrast with areas of necrosis, foci of recurrent tumor exhibited low-grade enhancement in this animal model. High-attenuation peripheral arterial phase enhancement at CT and MR imaging and peripheral areas of increased flow at US were normal findings at 1 week that were caused by granulation tissue. Central areas of enhancement and Doppler US flow that were caused by peliosis-like lesions and intact portal venous tracks were seen extending up to 6 mm inside the gross margins of the cryolesions in the absence of recurrent tumor. While the cryolesions contained large areas of low signal intensity on T2-weighted images because of coagulation necrosis, areas of high signal intensity on T2-weighted images were found within the cryolesion in the absence of recurrent tumor and corresponded with peripheral granulation tissue and central liquefaction necrosis.

The wedge-shaped area of enhancement seen surrounding cryolesions at CT and MR imaging (Figs 2, 3) has also been described after ethanol ablation (26,27). Yoshikawa et al (26) found similar areas of enhancement and areas of focal atrophy that correlated with decreased enhancement at CT during arterial portography and suggested that these changes were a result of damage to the portal venous system. We found organized thrombus in small portal venous branches that surrounded hepatic cryolesions, which was similar to changes reported adjacent to ethanol ablation sites (28). While large blood vessels are resistant to freezing injury, smaller vessels and bile ducts are more sensitive (29).

These results suggest that recurrent tumor cannot be differentiated from normal postcryoablation findings by using signal intensity characteristics or enhancement patterns alone at 1 week after ablation. The single most important imaging finding associated with recurrent disease was morphologic: a peripheral nodule that distorted the otherwise smooth margin of the cryolesion. In addition, follow-up images obtained months after the procedure should show a progressive decrease in cryolesion size and resolution of arterial phase rim enhancement due to granulation tissue.

Implications for Detection and Management of Local Recurrences
Our results show that although intraoperative US is commonly used to guide hepatic cryoablation, contrast-enhanced CT and MR imaging are superior to transabdominal US for detecting recurrent tumor at follow-up imaging. We rely primarily on helical CT for follow-up imaging in most of our patients who undergo cryoablation and employ MR imaging as a problem-solving tool in difficult cases. CT or MR imaging performed immediately after cryoablation and before granulation tissue has formed may be advantageous. We have not yet evaluated imaging performed during this time in animals or humans. Normal granulation tissue in this animal model has signal intensity and enhancement characteristics, similar to those of recurrent tumor, that make the detection of subtle recurrences more difficult. Our results suggest that granulation tissue is less of a problem on images obtained later: Rim enhancement was not seen in any of the four lesions imaged at 1 and 7 months.

The focal nature of the majority (14 [93%] of 15) of the local recurrences has important implications. These focal lesions have a characteristic appearance on images, provide a target for percutaneous biopsy to definitively diagnose recurrence, and are amenable to repeat treatment with imaging-guided percutaneous tumor ablation techniques. Local recurrence appeared as a rind of tumor at the cryolesion margin in one (7%) of 15 cases. This form of recurrence is difficult or impossible to distinguish from normal granulation tissue on images and is more difficult to treat percutaneously.

Mechanisms of Local Recurrence and Potential Strategies for Reducing Local Recurrence
Cryoablation is believed to kill cells by using several mechanisms: extracellular or intracellular ice formation, solute-solvent shifts causing cell dehydration and rupture, and small-vessel obliteration and resulting hypoxia (30–32). Cell kill may be increased with more rapid cooling rates, lower temperatures, slow thawing, and more than one freezing-thawing cycle (33–35). Tumor cells and normal hepatocytes are destroyed with approximately equal efficacy by freezing. Although partial hepatectomy has been associated with accelerated growth of residual tumor in animal models (36–38), this has not been observed after hepatic cryoablation (38).

The purpose of this study was to evaluate the radiologic and histopathologic features of local tumor recurrence after hepatic cryoablation and not to evaluate the efficacy of cryoablation. Investigators in prior studies in both animals and humans (1–7,34) have shown that cryoablation is effective in treating hepatic tumors. The 95% local recurrence rate is largely a result of the study design and cannot be extrapolated to clinical practice for the following reasons: (a) The small size of the rabbit liver and the goal of imaging recurrent tumor necessitated the use of margins smaller (3 mm) than the 1-cm margins used clinically and shorter freezing times, (b) intraoperative US was not performed to monitor the ice ball margins relative to the tumor because of the small size of the animal model (the lesions in this study were easily palpated and directly visualized), and (c) VX2 tumor is aggressive, and its response to cryoablation may not be representative of hepatic tumors in humans.

Despite important differences between this animal model and clinical practice that likely affect the rate of local recurrence, mechanisms of local recurrence may be similar. Although investigators in initial clinical studies (1–6,39) have generally shown low local recurrence rates after hepatic cryoablation, at least one group of surgeons has suggested that local recurrence can be an important problem. Adam et al (7) reported a 44% local recurrence rate for hepatic metastases treated with cryoablation, as compared with a 0% local recurrence rate for hepatocellular carcinoma. Our results suggest several mechanisms for local tumor recurrence and strategies for reducing the local recurrence rate associated with cryosurgery and other ablative therapies.

First, all macroscopic local recurrences in this study occurred at the cryolesion periphery. The tumor margin may be left untreated if an error is made in determining the position of the treatment margins relative to the tumor, and the adjacent normal liver provides a vascular supply to any surviving tumor cells. Intraoperative US is limited in its use in delineating the adequacy of treatment margins because of the acoustic shadowing produced by the ice ball (40). The deep margin, therefore, cannot be visualized. Improved methods of imaging guidance such as CT and MR imaging and assessment of treatment margins are key areas of research (41,42).

Second, intact portal venous tracks were identified as much as 6 mm within gross cryolesion margins, with microscopic foci of residual tumor adjacent to intact portal venous tracks in two lesions. These findings highlight a fundamental difference between resection and ablative therapies: While resection is essentially 100% effective in treating all cancer cells up to the resection margin, cryoablation results in cell kill that varies within the gross treatment margins as a function of cooling rate and absolute temperature (35). The "heat sink" effect caused by patent blood vessels that extend into the freezing zone can cause sublethal conditions well inside the gross margin (Fig 5). The cryolesion margin is subjected to slower freezing rates and higher minimum temperatures, both of which reduce cell kill (35). Given that we intentionally undertreated the tumors in this study in terms of treatment margins and freezing duration, as compared with accepted clinical protocols (43), the frequency of patent vessels and viable tumor inside the gross margin is probably overestimated. Nevertheless, we propose that larger treatment margins may be needed for ablative therapies to ensure adequate treatment, since cell kill and blood supply disruption may be incomplete near the margin.

Third, an intense reaction of granulation tissue was seen at the periphery of all cryolesions in this study. Arterial phase rim enhancement at CT and MR imaging, increased flow at US, and a rim of increased signal intensity on T2-weighted MR images are the imaging correlates of granulation tissue and reflect capillary proliferation (angiogenesis) and increased capillary permeability. Angiogenesis is central to wound healing and tumor stroma formation (44). This is born out by the similarity of imaging findings of granulation tissue and angiogenesis at tumor margins (Figs 2, 4). Surviving tumor cells at the cryolesion periphery are left in an environment of intense angiogenesis that is created by host’s normal healing response (Fig 3c) and may facilitate tumor growth. While inflammation is also an important component of granulation tissue, it is generally nonspecific and therefore may be limited in its antitumor activity. The hypervascularity of the granulation tissue at the treatment margin could be exploited by performing transcatheter chemoembolization. Chemoembolic material should be taken up by hypervascular granulation tissue at the cryolesion margin, regardless of the vascularity of the target tumor. Tanaka et al (45) found that the addition of transcatheter embolization to ethanol ablation significantly improved survival in patients with hepatocellular carcinoma, as compared with patients who underwent only ethanol ablation.

Practical application: Recurrent tumor in this animal model appeared most commonly as a focal nodule at the cryolesion periphery and was best seen at contrast-enhanced CT and MR imaging. Rim and central foci of enhancement, Doppler flow, and increased signal intensity on T2-weighted MR images can be normal findings after hepatic cryoablation and do not necessarily indicate recurrent tumor. The findings seen in normal cryolesions and local tumor recurrences suggest possible mechanisms for local recurrence and potential strategies for reducing the local recurrence rate after hepatic cryoablation.

	FOOTNOTES

 
D.A.B. is a consultant to GE Medical Systems and has received research support from Berlex Laboratories.

Author contributions: Guarantors of integrity of entire study, B.S.K., E.K.F.; study concepts, B.S.K., J.K.B., M.A.C., D.A.B., E.K.F.; study design, B.S.K., M.A.C., D.A.B., E.K.F.; definition of intellectual content, B.S.K., J.K.B., M.A.C., D.A.B., E.K.F.; literature research, B.S.K.; experimental studies, B.S.K., J.K.B., M.A.C., D.A.B., S.S., C.A.M., E.K.F.; data acquisition, B.S.K., J.K.B., D.A.B., S.S., C.A.M., K.M.H., E.K.F.; data analysis, B.S.K., J.K.B., K.M.H., J.E.; statistical analysis, J.E., B.S.K.; manuscript preparation, B.S.K.; manuscript editing and review, all authors.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Ravikumar TS, Kane R, Cady B, Jenkins R, Clouse M, Steele G. A 5-year study of cryosurgery in the treatment of liver tumors. Arch Surg 1991; 126:1520-1524.[Abstract]
2. Zhou XD, Tang ZY, Yu YQ, Ma ZC. Clinical evaluation of cryosurgery in the treatment of primary liver cancer: report of 60 cases. Cancer 1988; 61:1889-1892.[Medline]
3. Ravikumar TS, Steele G, Kane R, King V. Experimental and clinical observations on hepatic cryosurgery for colorectal metastases. Cancer Res 1991; 51:6323-6327.[Medline]
4. Onik G, Rubinsky B, Zemel R, et al. Ultrasound-guided hepatic cryosurgery in the treatment of metastatic colon carcinoma: preliminary results. Cancer 1991; 67:901-907.[Medline]
5. Weaver ML, Atkinson D, Zemel R. Hepatic cryosurgery in treating colorectal metastases. Cancer 1995; 76:210-214.[Medline]
6. Weaver ML, Atkinson D, Zemel R. Hepatic cryosurgery in the treatment of unresectable metastases. Surg Oncol 1995; 4:231-236.[Medline]
7. Adam R, Akpinar E, Johann M, Kunstlinger F, Majno P, Bismuth H. Place of cryosurgery in the treatment of malignant liver tumors. Ann Surg 1998; 225:39-50.
8. Kuszyk BS, Choti MA, Urban BA, et al. Hepatic tumors treated by cryosurgery: normal CT appearance. AJR Am J Roentgenol 1996; 166:363-367.[Abstract/Free Full Text]
9. McLoughlin RF, Saliken JF, McKinnon G, Wiseman D, Temple W. CT of the liver after cryotherapy of hepatic metastases: imaging findings. AJR Am J Roentgenol 1995; 165:329-332.[Abstract/Free&