- Email: [email protected]
Hepatic Infarction after Radiofrequency Ablation of Hepatocellular Carcinoma with an Internally Cooled Electrode
Hepatic Infarction after Radiofrequency Ablation of Hepatocellular Carcinoma with an Internally Cooled Electrode Young-sun Kim, Hyunchul Rhim, Hyo K. Lim, Dongil Choi, Won Jae Lee, and Seung Hoon Kim
PURPOSE: To elucidate the attributes of hepatic infarction after radiofrequency ablation (RFA) of hepatocellular carcinoma (HCC) with an internally cooled electrode. MATERIALS AND METHODS: The authors retrospectively reviewed follow-up computed tomographic (CT) scans (follow-up period, 1– 60.1 months; mean, 15.4 months) in 872 consecutive patients (male:female ratio, 672:200; mean age, 59.5 years) who had undergone 1,120 sessions of RFA for 1,335 HCCs with an internally cooled electrode. Diagnosis of hepatic infarction was made on the basis of CT findings. The authors evaluated the frequency of hepatic infarction, clinical features, initial and follow-up CT findings, accompanied complications, and prognosis. Potential risk factors were evaluated with multiple logistic regression analysis. RESULTS: The frequency of hepatic infarction was 1.8% (20 of 1,120 sessions). Common presenting symptoms were abdominal pain (16 of 20 patients) and fever (11 of 20 patients). All infarctions were found at the first follow-up CT examination. Gas collections were noted in 65% of patients. All lesions showed progressive shrinkage. Accompanied complications were biloma (n ⴝ 2), abscess (n ⴝ 2), and portal vein thrombosis (n ⴝ 1). One patient with a lobar infarction died from hepatic failure. Older age (P ⴝ .048) and larger tumor size (P ⴝ .026) were statistically significant risk factors by multivariate analysis. CONCLUSION: RFA complicated by hepatic infarction is uncommon. Although hepatic infarction can be managed conservatively in most cases, possible extensive involvement should be considered seriously because it may cause mortality. J Vasc Interv Radiol 2007; 18:1126 –1133 Abbreviations:
HCC ⫽ hepatocellular carcinoma, RFA ⫽ radiofrequency ablation, TACE ⫽ transcatheter arterial chemoembolization
RADIOFREQUENCY ablation (RFA) for hepatocellular carcinoma (HCC) is accepted as a complement or even alternative to surgical resection because of its therapeutic efficacy and safety (1). Although RFA of the liver is known to be safe, there is a spectrum
From the Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 IIwon-dong, Kangnam-ku, Seoul 135-710, Korea. Received January 11, 2007; final revision received May 28, 2007; accepted June 1, 2007. Address correspondence to H.R.; E-mail: [email protected] None of the authors have identified a conflict of interest. © SIR, 2007 DOI: 10.1016/j.jvir.2007.06.005
1126
of major and minor complications associated with it that has an incidence of 2.4%–10.6% according to recent large series studies (2–5). Important complications include hemorrhage, hepatic abscess, biliary tree injury, bowel injury, and cancer seeding along the electrode tract. These complications have been studied by many investigators. The attributes of these problems are well established (2–9). Infarction of the hepatic parenchyma as a complication of RFA has been known to be a less common complication than those listed above; its incidence has been reported to be 0%– 0.07% (2–5). Previous studies, however, have only included cases with substantial clinical problems. Therefore, a considerable number of cases
with hepatic infarction that may cause an abnormality only seen at imaging or mild signs and/or symptoms might have been overlooked or misclassified as something else (eg, side effects or manifestations of postablation syndrome). The purpose of our study was to elucidate the frequency, clinical features, risk factors, findings at longterm follow-up computed tomography (CT), and prognosis of hepatic infarction caused by RFA of HCC with an internally cooled electrode.
MATERIALS AND METHODS Patients and Tumors Written informed consent for treatment and data review for research purposes was obtained from all pa-
Volume 18
Number 9
tients at therapy. Institutional review board approval was not required for this retrospective clinical study. From July 2000 to April 2006, we performed a total of 1,120 sessions of RFA in 872 patients with 1,335 HCCs by using internally cooled electrodes. The study population consisted of 672 men and 200 women aged 24 –91 years (mean age, 59.5 years). Clinical characteristics of the study patients and tumors are summarized in Table 1. The diagnosis of HCC was confirmed with percutaneous needle biopsy in 231 tumors from 178 patients. The remaining 1,104 tumors from 694 patients were considered to be HCC on the basis of clinical criteria from the Barcelona-2000 European Association for the Study of the Liver (EASL) conference (10). Of the 1,335 tumors, 741 (55.5%) in 478 patients were residual or recurrent lesions after surgical resection (186 tumors in 124 patients), transcatheter arterial chemoembolization (TACE) (510 tumors in 312 patients), or previous RFA (352 tumors in 224 patients). The remaining 594 tumors (44.5%) in 476 patients were newly detected. RFA Procedures All procedures were performed on an inpatient basis. Patients were discharged 1 day after the procedure if there was no subsequent problem. RFA was performed percutaneously when all of the following criteria were met: (a) the patient had a single HCC 5 cm or less in maximal diameter, (b) the patient had a multinodular HCC (ⱕ3) with each tumor measuring up to 3 cm in maximal diameter, (c) the patient had Child-Pugh class A or B, (d) tumors could be visualized at ultrasonography (US) and were accessible with a percutaneous route, (e) the prothrombin time ratio was greater than 50% (prothrombin time with international normalized ratio ⬍1.7), and (f) the platelet count was greater than 70,000 cells/mm3 (70 cells ⫻ 109/L). In cases in which the size and/or number of tumors were beyond these criteria or when the location of the tumor was inappropriate for a US-guided percutaneous approach, intraoperative procedures after open laparotomy were adopted. We performed RFA percutaneously under US guidance with use of intra-
Kim et al
•
1127
Table 1 Clinical Characteristics of Patients with HCC Treated with RFA by using an Internally Cooled Electrode No. of Patients (n ⫽ 872)
Clinical Features Type of liver disease Liver cirrhosis Child-Pugh classification Class A Class B Chronic hepatitis No disease Healthy carrier of hepatitis type B virus Cause of liver disease (n ⫽ 841) Viral hepatitis Type B Type C Others Alcohol abuse Miscellaneous
765 574 191 76 31 15 801 613 174 14 33 7
Note.—Each patient had 1– 6 tumors (mean, 1.27). The maximal tumor size at US was 0.5– 6.5 cm (range, 2.17 cm).
venous sedation and local anesthesia in 1,005 sessions for 1,163 tumors. We ablated the tumors after laparotomy under US guidance with general anesthesia in 113 sessions for 170 tumors. For two cases, we adopted laparoscopic guidance due to abutted colonic loops. We used an internally cooled electrode system (Cool-tip RF system; Valleylab, Boulder, Colo) for all tumors. Three different types of electrodes were used: a single 17-gauge straight electrode with a 2- and 3-cm active tip, and a cluster of three electrodes with a 2.5-cm active tip mounted on a common handle. We used a single 3-cm electrode for 980 tumors, a single 2-cm electrode for 159 tumors, a cluster electrode for 183 tumors, both a single 3-cm and a cluster electrode for 11 tumors, and both a single 3- and 2-cm electrode for two tumors. Our strategy for complete necrosis of the tumor was to ablate at least 0.5 cm of normal hepatic parenchyma surrounding the tumor for a safety margin. Therefore, for tumors larger than 2.5 cm in diameter (n ⫽ 184), we adopted a multiple overlapping technique (n ⫽ 71) and/or used a cluster electrode (n ⫽ 117). A multiple overlapping technique was also adopted for tumors less than or equal to 2.5 cm if the operator determined it to be appropriate. This technique was adopted for 136 tumors in 65 patients (2.21 times ⫾ 0.64; range, 2– 6 times). Total
ablation time varied (13.2 minutes ⫾ 5.6; range, 3–75 minutes). Follow-up after RFA For early evaluation of the therapeutic response or possible immediate complications after the percutaneous procedure, contrast medium– enhanced US (n ⫽ 140), CT (n ⫽ 693), or both (n ⫽ 172) were performed within 24 hours. An immediate examination was omitted in cases with an intraoperative procedure (n ⫽ 115). All patients were followed-up with contrast-enhanced three-phase CT 1 month later and every 3 months thereafter. However, the interval of follow-up was possibly changed to 2–6 months according to the opinion of the clinicians. Contrast-enhanced CT was performed with one of five helical scanners (HiSpeed CT/i, LightSpeed QX/i, LightSpeed Ultra, or LightSpeed 16, GE Healthcare, Milwaukee, Wis; Brilliance 40, Philips Medical Systems, Best, the Netherlands). A total of 120 mL of nonionic contrast medium (iopromide; Ultravist 300 [300 mg of iodine per milliliter], Schering, Berlin, Germany) was administered intravenously with an automatic injector (OP 100; Medrad, Indianola, Pa) at a rate of 3 mL/sec. Images were obtained 25– 35, 60 –70, and 180 seconds after the initiation of contrast medium injection, representing the hepatic arterial, portal venous, and equilibrium phases,
1128
•
Hepatic Infarction after RF Ablation of HCC
September 2007
JVIR
Figure 1. Images in a 62-year-old woman treated with percutaneous RFA by using an internally cooled electrode (3 cm, single type) for 10 minutes. (a) CT scan obtained in the portal phase before treatment shows a lowattenuation HCC (arrows) in segment VII of the liver. (b–d) Triple-phase contrast-enhanced CT scans obtained 2 hours after RFA in the (b) hepatic arterial, (c) portal venous, and (d) equilibrium phases show a newly developed, wedge-shaped nonenhanced area peripheral to the ablation zone (solid arrows) with its base on the hepatic capsule, a finding that is representative of infarcted hepatic parenchyma (open arrows), in segment VII of the liver. The infarcted area contains gas collections (arrowheads) of a curvilinear configuration. The patient complained of severe abdominal pain and a febrile sensation and was treated with intravenous pethidine hydrochloride and antibiotics for 10 days. (e) CT scan obtained in the equilibrium phase 11 months after treatment demonstrates prominent atrophy of the infarcted hepatic parenchyma (open arrows) compared with a relatively mild degree of shrinkage of the ablation zone (solid arrows).
respectively. By using a single-detector helical CT unit, we obtained images in the craniocaudal direction with 7-mm-thick sections and a 7-mm interval. For multi-detector CT, we used 2.5–5.0-mm-thick sections and 2.5–5.0-mm intervals. We conducted 46 sessions of additional RFA because an unablated residual tumor was found at immediate
follow-up. If it was not possible to perform additional RFA due to poor delineation or new concurrent multiple lesions, we preferred to use TACE for repeat procedures. Imaging and Clinical Data Analysis We retrospectively reviewed follow-up CT scans obtained immediately
(n ⫽ 693) and/or 1 month (n ⫽ 1,074) after all RFA sessions. We focused on the detection of lesions suggestive of hepatic parenchymal infarction, defined as a newly developed, wedge-shaped, round or oval, consistently nonenhanced lesion peripheral to the ablation zone with its base on the hepatic capsule at all three phases of contrast-enhanced CT (Fig 1) (11, 12).
Volume 18
Number 9
Kim et al
•
1129
Table 2 Comparison of Potential Risk Factors between Groups with and without Complicated Hepatic Infarction after RFA for HCC and Results of Multivariate Analysis Variable Age (y) Underlying liver disease No disease Chronic hepatitis Liver cirrhosis Child-Pugh classification Class A Class B History of TACE Mean size of HCC (cm) Type of radiofrequency electrode† Single, 2 cm Single, 3 cm Cluster Adoption of overlapping ablation technique Mean ablation time (min)
Patients without Infarction (n ⫽ 852)
Patients with Infarction (n ⫽ 20)
58.7 ⫾ 10.0
64.1 ⫾ 8.4
30 (3.5) 74 (8.7) 748 (87.8)
1 (5) 2 (10) 17 (85)
557 (65.4) 188 (22.1) 306 (35.9) 2.2 ⫾ 0.9
17 (85) 3 (15) 6 (30) 2.8 ⫾ 1.1
157 (11.9)‡ 970 (73.8)‡ 188 (14.3)‡ 136 (10.3)‡ 13.2 ⫾ 5.5
2 (10) 12 (60) 6 (30) 6 (30) 16.2 ⫾ 8.7
P Value .048* .616
.924 .635 .026* .994
.141 .657
Note.—Numbers in parentheses are percentages. * Statistically significant with the multiple logistic regression model. † In cases in which multiple types of electrode were used, the electrode that made the largest ablation area was used for data analysis. ‡ Data were determined from 1,315 HCCs.
In cases in which hepatic infarction was detected initially, we examined the following data associated with imaging features at CT: (a) absolute volume (in cubic centimeters) by using a tool to measure area and the summation-of-areas technique at the portal venous phase on a picture archiving and communications system (Centricity; GE Healthcare) with a 2,000 ⫻ 2,000 pixels monitor (13), (b) relative volume compared with that of the hepatic segment in which the lesion had developed (subsegmental, segmental, and suprasegmental) by using visual assessment, (c) thrombosis in the large portal vein up to a second-order branch, (d) shape and location of gas collection in the infarcted and/or ablated area, and (e) other associated complications. We also evaluated the following imaging features at follow-up CT: (a) the period during which hepatic infarction was present (disappearance of the lesion was diagnosed when the calculated volume was less than 1 cm3), (b) change in the size of hepatic infarction, (c) change in the pattern of contrast enhancement, and (d) evolution of capsular retraction. Data related to RFA included (a) type of electrode used, (b) duration of radiofrequency energy application, (c)
approach method (percutaneous vs intraoperative), and (d) whether overlapping ablation was adopted or not. Clinical features evaluated included (a) the cause and status of underlying liver disease, (b) the size and location of the index tumor, (c) Child-Pugh classification, (d) history of TACE, (e) symptoms and signs after the procedure and their durations based on the nursing medical record, (f) changes in laboratory findings (aspartate aminotransferase, alanine aminotransferase, and total bilirubin levels) within 3 days of the procedure, (g) management, (h) duration of hospitalization, and (i) prognosis. Statistical Analysis for Risk Factors Several data were evaluated as potential risk factors (Table 2). Multiple logistic regression models were used as a multivariate analysis to determine which variables were independently significant. Data based on patients and tumors were analyzed separately. A P value of less than .05 was considered statistically significant. Data analyses were performed by using software (SPSS for Windows, version 11.0; SPSS, Chicago, Ill).
RESULTS Frequency and Clinical Data of Hepatic Infarction Hepatic infarctions were found in 20 patients (male:female ratio, 15:5; mean age, 64.1 years; age range, 50 – 80 years; 1.8% per session [20/1120]; 1.5% per tumor [20/1335]). Seventeen of the 20 patients had one treated tumor and three patients had two treated tumors. The tumors in those 20 patients measured 1.2–5.5 cm (mean, 2.8 cm). Eight of the 20 patients had previously undergone treatment for HCC (TACE, n ⫽ 4; TACE and RFA, n ⫽ 2; surgical resection, n ⫽ 1; surgical resection and RFA, n ⫽ 1). The time between the last treatment and RFA was 41–371 days (mean, 141 days). None of the patients was found to have evidence of hepatic infarction at CT performed before RFA. The procedure was performed percutaneously in 17 patients and intraoperatively in three. One of the intraoperative procedures was performed with the Pringle maneuver. Other important data are described in Table 2. Follow-up CT Findings All hepatic infarctions were found at the first follow-up CT, which was
1130
•
Hepatic Infarction after RF Ablation of HCC
performed 2 hours–52 days (mean, 10.1 days) after RFA. Immediate follow-up CT scans obtained within 24 hours of the procedure were available in only 12 patients because CT was replaced with contrast-enhanced US (n ⫽ 5) or omitted due to recent surgery (n ⫽ 3). The calculated volume for hepatic infarction at their initial manifestations was 8.6 – 679.1 cm3 (mean ⫾ standard deviation, 87.1 cm3 ⫾ 153.5). The hepatic infarction was classified as subsegmental in 11 patients (55%), segmental in seven (35%), and suprasegmental in two (10%). At the initial CT examination, gas collections were noted in 13 of the 20 patients (65%) in the area of the ablation zone or infarction. Nine of the 13 gas collections (69%) were estimated to be in the branches of the hepatic artery or portal vein because of a curvilinear shape and concordant orientation with the vessels. Follow-up CT was performed one to 24 times (mean, 7.1 times) for 1– 60.1 months (mean, 15.4 months) after the diagnosis of infarction. All lesions showed progressive shrinkage (Fig 1). All infarctions in 13 patients who were followed-up for more than 6 months disappeared with only a residual scar. In 12 of those 13 patients (92%), this was accompanied by progressive retraction of the hepatic capsule. In three patients (15%), the contrast enhancement pattern was changed. Contrast enhancement at the equilibrium phase was seen at the second follow-up CT (1– 4 months after initial presentation), which is a finding suggestive of incomplete reperfusion (Fig 2). Four of the 20 patients with hepatic infarction (20%) had complications: one each had biloma at the ablation zone, abscess at the infarcted area (Fig 3), both biloma at the ablation zone and abscess at the infarcted area, and extensive portal vein thrombosis. Clinical Manifestations Clinically, 17 of the 20 patients with hepatic infarction (85%) complained of symptoms. Abdominal pain was most common and occurred in 16 of the 17 patients (94%). Analgesics (acetaminophen and/or pethidine hydrochloride) were used in seven of those 16 patients (44%). Fever was the second most common symptom, occurring in 11 of the 17 patients (65%). The peak body temperature in those patients ranged
from 37.6 to 39.2°C (mean, 38.4°C). Other symptoms were chills (n ⫽ 3, 18%), nausea (n ⫽ 2, 12%), abdominal distention (n ⫽ 1, 5.9%), right shoulder pain (n ⫽ 1, 5.9%), and generalized itching (n ⫽ 1, 5.9%). Laboratory data obtained within 3 days of the procedure were available in 18 patients. The mean serum aspartate aminotransferase/alanine aminotransferase level was 555.0 IU/L ⫾ 825.8/541.0 IU/L ⫾ 863.7 (preoperative values, 59.4 IU/L ⫾ 41.0/64.6 IU/L ⫾ 86.8). The mean serum total bilirubin level was 5.5 mg/dL ⫾ 11.8 (94 mol/L) (preoperative value, 0.8 mg/dL ⫾ 0.7). Eight patients (47.1%) who underwent the percutaneous procedure were discharged after 1 overnight stay without substantial symptoms. Two patients (66.7%) who underwent an intraoperative procedure were discharged 7 and 9 days after RFA, respectively. The duration of the hospital stay of the remaining 10 patients (50%) was increased. A 65-year-old man who underwent intraoperative RFA with a cluster-type electrode for 12 minutes for HCC that had been treated partially with TACE had a large infarction (volume, 679.1 cm3) involving the right hepatic lobe with portal venous thrombosis from the main to the segmental branches. He refused liver transplantation. Despite intensive care, he died 36 days later due to hepatic failure. Therefore, the mortality rate of patients with hepatic infarction as a complication of RFA was 5% (one of 20 patients). Risk Factor Analysis Among the eight potential risk factors evaluated, old age (P ⫽ .048) and large tumor size (P ⫽ .026) turned out to be independently significant risk factors for developing hepatic infarction after RFA for HCC according to multiple logistic regression analysis (Table 2).
DISCUSSION Previous studies have described the CT appearance of hepatic infarction (11,12,14). One of two recent reports (11) described hepatic infarction as a lowattenuation, peripheral, and wedgeshaped lesion at CT. The other report (12) described three patterns: (a) peripheral, wedge-shaped; (b) central or peripheral, round or oval; and (c)
September 2007
JVIR
irregular and parallel to bile duct. In both studies, proof of the hepatic infarction had been made with histopathologic, angiographic, or clinical methods. In our study, we used the imaging definition of hepatic infarction as a newly developed, wedge-shaped, round or oval, nonenhanced lesion peripheral to the ablation zone with its base on the hepatic capsule at all three phases of contrast-enhanced CT. Recently, several multi-institutional, large series studies (2–5) have been performed to evaluate the complications of RFA for hepatic tumors. Hepatic infarction was so uncommon that the incidence was reported to be no more than 0%– 0.07%. There were only two cases of hepatic infarction noted after RFA in 4,379 patients with 6,524 hepatic tumors. Overall, the incidence was only 0.045%. Conversely, 20 cases of hepatic infarction were found in 1,120 sessions (1.8%) of RFA for 872 patients (2.3%) with 1,335 HCCs (1.5%) in our study; this frequency was higher than that reported in previous studies. We believe that this discrepancy might be caused by the method of identifying cases. We reviewed all CT scans obtained after RFA to search for lesions suggestive of hepatic infarction irrespective of clinical symptoms. However, previous studies might have included only cases that caused a substantial clinical problem or a noticeable change on imaging studies. According to standardization of terminology and reporting criteria developed the International Working Group on Image Guided Tumor Ablation (15), only 11 patients in our study actually could be classified as having a major complication due to a lengthening of the hospital stay (n ⫽ 10) and/or an increase in the level of care (n ⫽ 7). Moreover, some patients classified as having a major complication might even be regarded as merely having postablation syndrome clinically (16), in which similar symptoms and laboratory changes can occur. Therefore, previous studies might have overlooked asymptomatic cases, cases with minor symptoms, or cases with similar features to postablation syndrome; therefore, the real incidence must be higher than reported previously. Because the liver has a dual blood supply from the hepatic artery and
Volume 18
Number 9
Kim et al
•
1131
Figure 2. Images in a 50-year-old man treated with percutaneous RFA by using an internally cooled electrode (2 cm, single type) for 8 minutes. (a– c) CT scans obtained in the (a) arterial, (b) portal venous, and (c) equilibrium phases 2 hours after RFA show a nonenhanced RFA zone (solid arrow) and complicated hepatic infarction (open arrows) in segment VI of the liver. The patient complained of tolerable abdominal pain and a mild febrile sensation and was discharged 2 days later after undergoing only conservative treatment. (d–f) Follow-up CT scans obtained 1 month after treatment demonstrate atrophic change of the low-attenuation infarcted area (open arrows) with mild capsular retraction at the hepatic arterial (d) and portal venous (e) phases. The equilibrium phase scan (f), however, shows reperfusion of the previously infarcted lesion. RFA zone (solid arrow) showed a relatively small change. This lesion disappeared completely at 4-month follow-up CT, with only minimal residual scar and further capsular retraction.
portal vein as well as an ability to develop extensive collateral pathways, hepatic infarction is generally thought
to be relatively rare (17–19). When hepatic infarction occurs, however, it is known to result from an insult either
to the hepatic arterial supply or to both the hepatic arterial and portal venous systems, but not to the portal
1132
•
Hepatic Infarction after RF Ablation of HCC
Figure 3. Images in a 56-year-old man treated with percutaneous RFA by using two types of internally cooled electrode (both the cluster and single 3-cm type) for 8 minutes. This patient complained of mild abdominal pain and high fever (39.3°C) 2 days after treatment. (a) CT scan obtained during the portal venous phase immediately after RFA shows the typical RFA zone (solid arrow) and complicated hepatic infarction (open arrows). (b) Follow-up CT scan obtained 4 days later shows that multifocal, small, low-enhanced lesions with peripheral rim enhancement suggestive of hepatic microabscesses (arrowheads) had developed at the infarcted area. Purulent aspirate from the lesion confirmed the diagnosis (arrow: RFA zone). Symptoms were relieved after a 9-day course of intravenous antibiotic therapy.
venous supply alone (20 –22). Therefore, the hepatic artery appears to have a key role in the development of hepatic infarction. This concept might help explain the risk factors that were proved to be significant in our current study. Older patients have age-related reduced status of their arterial system; therefore, the hepatic artery in older patients is more prone to insult by the same stimulation (ie, hyperthermia) than that of younger patients. In cases with larger tumors, the chance of vascular injury increases with the ablation of a larger area. Although other factors (eg. total ablation time and history of TACE) may be related to an increased risk of arterial injury, they were proved not to be independently significant with multivariate analysis. The clinical features associated with hepatic infarction, especially in mild cases, overlap those of postablation syndrome. According to a recent prospective study (16), fever and malaise were the most common symp-
toms of postablation syndrome; abdominal pain and fever were most the common symptoms in our patients. As for fever, patterns are very similar (postablation syndrome vs hepatic infarction: onset, 1–9 days [mean, 3.0 days] vs 1–3 days [mean, 2.0 days]; duration, 1–11 days [mean, 5.5 days] vs 1–8 days [mean, 4.0 days]; peak body temperature, 37.4–38.9°C [mean, 38.2°C] vs 37.6– 39.2°C [mean, 38.4°C]). Laboratory data such as serum aspartate aminotransferase levels were also similar (517.8 IU/L ⫾ 100.8 vs 555.0 IU/L ⫾ 825.8). Overlaps like these are quite natural considering both are caused by a hepatic parenchymal necrosis—whether induced by ablation alone or both ablation and infarction. Despite these similarities, abdominal pain was more predominant in our study and might be a distinctive clinical clue suggestive of hepatic infarction rather than postablation syndrome. In addition to postablation syndrome, hepatic abscess is another entity to be differentiated from hepatic infarction. In our 20 patients with
September 2007
JVIR
hepatic infarction, 10 (50%) complained of both fever and abdominal pain. In addition, we found gas collections at CT in six of them. Although these could be the findings of a hepatic abscess, only one patient was found to have a hepatic abscess at clinical examination. Limitations of our study were as follow: First, because our study was retrospective, unknown bias might be involved and standardization for the evaluation of clinical symptoms and laboratory data was not possible. Second, we cannot exclude completely the possibility that another pathologic condition was misinterpreted as an infarction because there was no histologic backup. However, our criteria for hepatic infarction were as good as possible considering the retrospective and CT image analysis– based nature of our study. Third, delineations of the border between the infarction and ablation zone were not always evident. In these cases, it was possible to draw an imaginary border considering both shape of proximal margin of ablation zone and evolution of ablation zone and infarcted area at follow-up CT. This might cause an error in the assessment of volume, but we do not think it was crucial. Last, some clinical features of the patients with hepatic infarction might be attributed to a tissue necrosis caused by ablation itself (ie, postablation syndrome). Despite these limitations, we can conclude that hepatic infarction, as a complication of RFA for HCC, is uncommon. It is, however, more frequent than previously reported, especially in patients with larger tumors and those with a more advanced age. We can manage most cases of hepatic infarction after RFA conservatively. However, surveillance for extensive infarction is needed because this can cause hepatic failure and subsequent mortality. References 1. Choi D, Lim HK, Rhim H, et al. Percutaneous radiofrequency ablation for early-stage hepatocellular carcinoma as a first-line treatment: longterm results and prognostic factors in a large single-institution series. Eur Radiol 2007; 17:684 – 692. 2. Rhim H, Yoon KH, Lee JM, et al. Major complications after radio-frequency thermal ablation of hepatic tumors: spectrum of imaging findings. Radiographics 2003; 23:123–136.
Volume 18
Number 9
3. Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter study. Radiology 2003; 226: 441– 451. 4. de Baere T, Risse O, Kouch V, et al. Adverse events during radiofrequency treatment of 582 hepatic tumors. AJR Am J Roentgenol 2003; 181:695–700. 5. Curley SA, Paolo M, Beaty K, et al. Early and late complications after radiofrequency ablation of malignant liver tumors in 608 patients. Ann Surg 2004; 239:450 – 458. 6. Choi D, Lim HK, Kim MJ, et al. Liver abscess after percutaneous radiofrequency ablation for hepatocellular carcinomas: frequency and risk factors. AJR Am J Roentgenol 2005; 184:1800 – 1807. 7. Kim SH, Lim HK, Choi D, et al. Changes in bile ducts after radiofrequency ablation of hepatocellular carcinoma: frequency and clinical significance. AJR Am J Roentgenol 2004; 183: 1611–1617. 8. Choi D, Lim HK, Kim MJ, et al. Therapeutic efficacy and safety of percutaneous radiofrequency ablation of
Kim et al
9.
10.
11.
12.
13.
14.
hepatocellular carcinoma abutting the gastrointestinal tract. AJR Am J Roentgenol 2004; 183:1417–1424. Jaskolka JD, Asch MR, Kachura JR, et al. Needle tract seeding after radiofrequency ablation of hepatic tumors. J Vasc Interv Radiol 2005; 16:485– 491. Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusion of the Barcellona-2000 EASL conference. J Hepatol 2001; 35:421– 430. Smith GS, Birnbaum BA, Jacobs JE. Hepatic infarction secondary to arterial insufficiency in native livers: CT findings in 10 patients. Radiology 1998; 208: 223–229. Holbert BL, Baron RL, Dodd GD III. Hepatic infarction caused by arterial insufficiency: spectrum and evolution of CT findings. AJR Am J Roentgenol 1996; 166:815– 820. Breiman RS, Beck JW, Korobkin M, et al. Volume determinations using computed tomography. AJR Am J Roentgenol 1982; 138:329 –338. Lev-Toaff AS, Friedman AC, Cohen LM, Radeki PD, Caroline DF. Hepatic infarcts: new observations by CT and sonography. AJR Am J Roentgenol 1987; 149:87–90.
•
1133
15. Goldberg SN, Grassi CJ, Cardella JF, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria. J Vasc Interv Radiol 2005; 16: 765–778. 16. Dodd GD III, Napier D, Schoolfield JD, Hubbard L. Percutaneous radiofrequency ablation of hepatic tumors: postablation syndrome. AJR Am J Roentgenol 2005; 185:51–57. 17. Kanter DM. Hepatic infarction. Arch Intern Med 1965; 115:479 – 481. 18. Temberg JL, Butcher HR. Blood flow relation between hepatic artery and portal vein. Science 1965; 150:1030 –1031. 19. Redman HC, Stewart RR. Arterial collaterals in the liver hilus. Radiology 1970; 94:575–579. 20. Popper HL, Jefferson NC, Necheles H. Liver necrosis following complete interruption of hepatic artery and partial ligation of portal vein. Am J Surg 1953; 86:309 –311. 21. Loeffler L. Factors determining necrosis or survival of liver tissue after ligation of hepatic artery. Arch Pathol 1936; 21:496 –503. 22. Wooling KR, Baggenstoss AH, Weir JF. Infarcts of the liver. Gastroenterology 1951; 17:479 – 493.
PURPOSE: To elucidate the attributes of hepatic infarction after radiofrequency ablation (RFA) of hepatocellular carcinoma (HCC) with an internally cooled electrode. MATERIALS AND METHODS: The authors retrospectively reviewed follow-up computed tomographic (CT) scans (follow-up period, 1– 60.1 months; mean, 15.4 months) in 872 consecutive patients (male:female ratio, 672:200; mean age, 59.5 years) who had undergone 1,120 sessions of RFA for 1,335 HCCs with an internally cooled electrode. Diagnosis of hepatic infarction was made on the basis of CT findings. The authors evaluated the frequency of hepatic infarction, clinical features, initial and follow-up CT findings, accompanied complications, and prognosis. Potential risk factors were evaluated with multiple logistic regression analysis. RESULTS: The frequency of hepatic infarction was 1.8% (20 of 1,120 sessions). Common presenting symptoms were abdominal pain (16 of 20 patients) and fever (11 of 20 patients). All infarctions were found at the first follow-up CT examination. Gas collections were noted in 65% of patients. All lesions showed progressive shrinkage. Accompanied complications were biloma (n ⴝ 2), abscess (n ⴝ 2), and portal vein thrombosis (n ⴝ 1). One patient with a lobar infarction died from hepatic failure. Older age (P ⴝ .048) and larger tumor size (P ⴝ .026) were statistically significant risk factors by multivariate analysis. CONCLUSION: RFA complicated by hepatic infarction is uncommon. Although hepatic infarction can be managed conservatively in most cases, possible extensive involvement should be considered seriously because it may cause mortality. J Vasc Interv Radiol 2007; 18:1126 –1133 Abbreviations:
HCC ⫽ hepatocellular carcinoma, RFA ⫽ radiofrequency ablation, TACE ⫽ transcatheter arterial chemoembolization
RADIOFREQUENCY ablation (RFA) for hepatocellular carcinoma (HCC) is accepted as a complement or even alternative to surgical resection because of its therapeutic efficacy and safety (1). Although RFA of the liver is known to be safe, there is a spectrum
From the Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 IIwon-dong, Kangnam-ku, Seoul 135-710, Korea. Received January 11, 2007; final revision received May 28, 2007; accepted June 1, 2007. Address correspondence to H.R.; E-mail: [email protected] None of the authors have identified a conflict of interest. © SIR, 2007 DOI: 10.1016/j.jvir.2007.06.005
1126
of major and minor complications associated with it that has an incidence of 2.4%–10.6% according to recent large series studies (2–5). Important complications include hemorrhage, hepatic abscess, biliary tree injury, bowel injury, and cancer seeding along the electrode tract. These complications have been studied by many investigators. The attributes of these problems are well established (2–9). Infarction of the hepatic parenchyma as a complication of RFA has been known to be a less common complication than those listed above; its incidence has been reported to be 0%– 0.07% (2–5). Previous studies, however, have only included cases with substantial clinical problems. Therefore, a considerable number of cases
with hepatic infarction that may cause an abnormality only seen at imaging or mild signs and/or symptoms might have been overlooked or misclassified as something else (eg, side effects or manifestations of postablation syndrome). The purpose of our study was to elucidate the frequency, clinical features, risk factors, findings at longterm follow-up computed tomography (CT), and prognosis of hepatic infarction caused by RFA of HCC with an internally cooled electrode.
MATERIALS AND METHODS Patients and Tumors Written informed consent for treatment and data review for research purposes was obtained from all pa-
Volume 18
Number 9
tients at therapy. Institutional review board approval was not required for this retrospective clinical study. From July 2000 to April 2006, we performed a total of 1,120 sessions of RFA in 872 patients with 1,335 HCCs by using internally cooled electrodes. The study population consisted of 672 men and 200 women aged 24 –91 years (mean age, 59.5 years). Clinical characteristics of the study patients and tumors are summarized in Table 1. The diagnosis of HCC was confirmed with percutaneous needle biopsy in 231 tumors from 178 patients. The remaining 1,104 tumors from 694 patients were considered to be HCC on the basis of clinical criteria from the Barcelona-2000 European Association for the Study of the Liver (EASL) conference (10). Of the 1,335 tumors, 741 (55.5%) in 478 patients were residual or recurrent lesions after surgical resection (186 tumors in 124 patients), transcatheter arterial chemoembolization (TACE) (510 tumors in 312 patients), or previous RFA (352 tumors in 224 patients). The remaining 594 tumors (44.5%) in 476 patients were newly detected. RFA Procedures All procedures were performed on an inpatient basis. Patients were discharged 1 day after the procedure if there was no subsequent problem. RFA was performed percutaneously when all of the following criteria were met: (a) the patient had a single HCC 5 cm or less in maximal diameter, (b) the patient had a multinodular HCC (ⱕ3) with each tumor measuring up to 3 cm in maximal diameter, (c) the patient had Child-Pugh class A or B, (d) tumors could be visualized at ultrasonography (US) and were accessible with a percutaneous route, (e) the prothrombin time ratio was greater than 50% (prothrombin time with international normalized ratio ⬍1.7), and (f) the platelet count was greater than 70,000 cells/mm3 (70 cells ⫻ 109/L). In cases in which the size and/or number of tumors were beyond these criteria or when the location of the tumor was inappropriate for a US-guided percutaneous approach, intraoperative procedures after open laparotomy were adopted. We performed RFA percutaneously under US guidance with use of intra-
Kim et al
•
1127
Table 1 Clinical Characteristics of Patients with HCC Treated with RFA by using an Internally Cooled Electrode No. of Patients (n ⫽ 872)
Clinical Features Type of liver disease Liver cirrhosis Child-Pugh classification Class A Class B Chronic hepatitis No disease Healthy carrier of hepatitis type B virus Cause of liver disease (n ⫽ 841) Viral hepatitis Type B Type C Others Alcohol abuse Miscellaneous
765 574 191 76 31 15 801 613 174 14 33 7
Note.—Each patient had 1– 6 tumors (mean, 1.27). The maximal tumor size at US was 0.5– 6.5 cm (range, 2.17 cm).
venous sedation and local anesthesia in 1,005 sessions for 1,163 tumors. We ablated the tumors after laparotomy under US guidance with general anesthesia in 113 sessions for 170 tumors. For two cases, we adopted laparoscopic guidance due to abutted colonic loops. We used an internally cooled electrode system (Cool-tip RF system; Valleylab, Boulder, Colo) for all tumors. Three different types of electrodes were used: a single 17-gauge straight electrode with a 2- and 3-cm active tip, and a cluster of three electrodes with a 2.5-cm active tip mounted on a common handle. We used a single 3-cm electrode for 980 tumors, a single 2-cm electrode for 159 tumors, a cluster electrode for 183 tumors, both a single 3-cm and a cluster electrode for 11 tumors, and both a single 3- and 2-cm electrode for two tumors. Our strategy for complete necrosis of the tumor was to ablate at least 0.5 cm of normal hepatic parenchyma surrounding the tumor for a safety margin. Therefore, for tumors larger than 2.5 cm in diameter (n ⫽ 184), we adopted a multiple overlapping technique (n ⫽ 71) and/or used a cluster electrode (n ⫽ 117). A multiple overlapping technique was also adopted for tumors less than or equal to 2.5 cm if the operator determined it to be appropriate. This technique was adopted for 136 tumors in 65 patients (2.21 times ⫾ 0.64; range, 2– 6 times). Total
ablation time varied (13.2 minutes ⫾ 5.6; range, 3–75 minutes). Follow-up after RFA For early evaluation of the therapeutic response or possible immediate complications after the percutaneous procedure, contrast medium– enhanced US (n ⫽ 140), CT (n ⫽ 693), or both (n ⫽ 172) were performed within 24 hours. An immediate examination was omitted in cases with an intraoperative procedure (n ⫽ 115). All patients were followed-up with contrast-enhanced three-phase CT 1 month later and every 3 months thereafter. However, the interval of follow-up was possibly changed to 2–6 months according to the opinion of the clinicians. Contrast-enhanced CT was performed with one of five helical scanners (HiSpeed CT/i, LightSpeed QX/i, LightSpeed Ultra, or LightSpeed 16, GE Healthcare, Milwaukee, Wis; Brilliance 40, Philips Medical Systems, Best, the Netherlands). A total of 120 mL of nonionic contrast medium (iopromide; Ultravist 300 [300 mg of iodine per milliliter], Schering, Berlin, Germany) was administered intravenously with an automatic injector (OP 100; Medrad, Indianola, Pa) at a rate of 3 mL/sec. Images were obtained 25– 35, 60 –70, and 180 seconds after the initiation of contrast medium injection, representing the hepatic arterial, portal venous, and equilibrium phases,
1128
•
Hepatic Infarction after RF Ablation of HCC
September 2007
JVIR
Figure 1. Images in a 62-year-old woman treated with percutaneous RFA by using an internally cooled electrode (3 cm, single type) for 10 minutes. (a) CT scan obtained in the portal phase before treatment shows a lowattenuation HCC (arrows) in segment VII of the liver. (b–d) Triple-phase contrast-enhanced CT scans obtained 2 hours after RFA in the (b) hepatic arterial, (c) portal venous, and (d) equilibrium phases show a newly developed, wedge-shaped nonenhanced area peripheral to the ablation zone (solid arrows) with its base on the hepatic capsule, a finding that is representative of infarcted hepatic parenchyma (open arrows), in segment VII of the liver. The infarcted area contains gas collections (arrowheads) of a curvilinear configuration. The patient complained of severe abdominal pain and a febrile sensation and was treated with intravenous pethidine hydrochloride and antibiotics for 10 days. (e) CT scan obtained in the equilibrium phase 11 months after treatment demonstrates prominent atrophy of the infarcted hepatic parenchyma (open arrows) compared with a relatively mild degree of shrinkage of the ablation zone (solid arrows).
respectively. By using a single-detector helical CT unit, we obtained images in the craniocaudal direction with 7-mm-thick sections and a 7-mm interval. For multi-detector CT, we used 2.5–5.0-mm-thick sections and 2.5–5.0-mm intervals. We conducted 46 sessions of additional RFA because an unablated residual tumor was found at immediate
follow-up. If it was not possible to perform additional RFA due to poor delineation or new concurrent multiple lesions, we preferred to use TACE for repeat procedures. Imaging and Clinical Data Analysis We retrospectively reviewed follow-up CT scans obtained immediately
(n ⫽ 693) and/or 1 month (n ⫽ 1,074) after all RFA sessions. We focused on the detection of lesions suggestive of hepatic parenchymal infarction, defined as a newly developed, wedge-shaped, round or oval, consistently nonenhanced lesion peripheral to the ablation zone with its base on the hepatic capsule at all three phases of contrast-enhanced CT (Fig 1) (11, 12).
Volume 18
Number 9
Kim et al
•
1129
Table 2 Comparison of Potential Risk Factors between Groups with and without Complicated Hepatic Infarction after RFA for HCC and Results of Multivariate Analysis Variable Age (y) Underlying liver disease No disease Chronic hepatitis Liver cirrhosis Child-Pugh classification Class A Class B History of TACE Mean size of HCC (cm) Type of radiofrequency electrode† Single, 2 cm Single, 3 cm Cluster Adoption of overlapping ablation technique Mean ablation time (min)
Patients without Infarction (n ⫽ 852)
Patients with Infarction (n ⫽ 20)
58.7 ⫾ 10.0
64.1 ⫾ 8.4
30 (3.5) 74 (8.7) 748 (87.8)
1 (5) 2 (10) 17 (85)
557 (65.4) 188 (22.1) 306 (35.9) 2.2 ⫾ 0.9
17 (85) 3 (15) 6 (30) 2.8 ⫾ 1.1
157 (11.9)‡ 970 (73.8)‡ 188 (14.3)‡ 136 (10.3)‡ 13.2 ⫾ 5.5
2 (10) 12 (60) 6 (30) 6 (30) 16.2 ⫾ 8.7
P Value .048* .616
.924 .635 .026* .994
.141 .657
Note.—Numbers in parentheses are percentages. * Statistically significant with the multiple logistic regression model. † In cases in which multiple types of electrode were used, the electrode that made the largest ablation area was used for data analysis. ‡ Data were determined from 1,315 HCCs.
In cases in which hepatic infarction was detected initially, we examined the following data associated with imaging features at CT: (a) absolute volume (in cubic centimeters) by using a tool to measure area and the summation-of-areas technique at the portal venous phase on a picture archiving and communications system (Centricity; GE Healthcare) with a 2,000 ⫻ 2,000 pixels monitor (13), (b) relative volume compared with that of the hepatic segment in which the lesion had developed (subsegmental, segmental, and suprasegmental) by using visual assessment, (c) thrombosis in the large portal vein up to a second-order branch, (d) shape and location of gas collection in the infarcted and/or ablated area, and (e) other associated complications. We also evaluated the following imaging features at follow-up CT: (a) the period during which hepatic infarction was present (disappearance of the lesion was diagnosed when the calculated volume was less than 1 cm3), (b) change in the size of hepatic infarction, (c) change in the pattern of contrast enhancement, and (d) evolution of capsular retraction. Data related to RFA included (a) type of electrode used, (b) duration of radiofrequency energy application, (c)
approach method (percutaneous vs intraoperative), and (d) whether overlapping ablation was adopted or not. Clinical features evaluated included (a) the cause and status of underlying liver disease, (b) the size and location of the index tumor, (c) Child-Pugh classification, (d) history of TACE, (e) symptoms and signs after the procedure and their durations based on the nursing medical record, (f) changes in laboratory findings (aspartate aminotransferase, alanine aminotransferase, and total bilirubin levels) within 3 days of the procedure, (g) management, (h) duration of hospitalization, and (i) prognosis. Statistical Analysis for Risk Factors Several data were evaluated as potential risk factors (Table 2). Multiple logistic regression models were used as a multivariate analysis to determine which variables were independently significant. Data based on patients and tumors were analyzed separately. A P value of less than .05 was considered statistically significant. Data analyses were performed by using software (SPSS for Windows, version 11.0; SPSS, Chicago, Ill).
RESULTS Frequency and Clinical Data of Hepatic Infarction Hepatic infarctions were found in 20 patients (male:female ratio, 15:5; mean age, 64.1 years; age range, 50 – 80 years; 1.8% per session [20/1120]; 1.5% per tumor [20/1335]). Seventeen of the 20 patients had one treated tumor and three patients had two treated tumors. The tumors in those 20 patients measured 1.2–5.5 cm (mean, 2.8 cm). Eight of the 20 patients had previously undergone treatment for HCC (TACE, n ⫽ 4; TACE and RFA, n ⫽ 2; surgical resection, n ⫽ 1; surgical resection and RFA, n ⫽ 1). The time between the last treatment and RFA was 41–371 days (mean, 141 days). None of the patients was found to have evidence of hepatic infarction at CT performed before RFA. The procedure was performed percutaneously in 17 patients and intraoperatively in three. One of the intraoperative procedures was performed with the Pringle maneuver. Other important data are described in Table 2. Follow-up CT Findings All hepatic infarctions were found at the first follow-up CT, which was
1130
•
Hepatic Infarction after RF Ablation of HCC
performed 2 hours–52 days (mean, 10.1 days) after RFA. Immediate follow-up CT scans obtained within 24 hours of the procedure were available in only 12 patients because CT was replaced with contrast-enhanced US (n ⫽ 5) or omitted due to recent surgery (n ⫽ 3). The calculated volume for hepatic infarction at their initial manifestations was 8.6 – 679.1 cm3 (mean ⫾ standard deviation, 87.1 cm3 ⫾ 153.5). The hepatic infarction was classified as subsegmental in 11 patients (55%), segmental in seven (35%), and suprasegmental in two (10%). At the initial CT examination, gas collections were noted in 13 of the 20 patients (65%) in the area of the ablation zone or infarction. Nine of the 13 gas collections (69%) were estimated to be in the branches of the hepatic artery or portal vein because of a curvilinear shape and concordant orientation with the vessels. Follow-up CT was performed one to 24 times (mean, 7.1 times) for 1– 60.1 months (mean, 15.4 months) after the diagnosis of infarction. All lesions showed progressive shrinkage (Fig 1). All infarctions in 13 patients who were followed-up for more than 6 months disappeared with only a residual scar. In 12 of those 13 patients (92%), this was accompanied by progressive retraction of the hepatic capsule. In three patients (15%), the contrast enhancement pattern was changed. Contrast enhancement at the equilibrium phase was seen at the second follow-up CT (1– 4 months after initial presentation), which is a finding suggestive of incomplete reperfusion (Fig 2). Four of the 20 patients with hepatic infarction (20%) had complications: one each had biloma at the ablation zone, abscess at the infarcted area (Fig 3), both biloma at the ablation zone and abscess at the infarcted area, and extensive portal vein thrombosis. Clinical Manifestations Clinically, 17 of the 20 patients with hepatic infarction (85%) complained of symptoms. Abdominal pain was most common and occurred in 16 of the 17 patients (94%). Analgesics (acetaminophen and/or pethidine hydrochloride) were used in seven of those 16 patients (44%). Fever was the second most common symptom, occurring in 11 of the 17 patients (65%). The peak body temperature in those patients ranged
from 37.6 to 39.2°C (mean, 38.4°C). Other symptoms were chills (n ⫽ 3, 18%), nausea (n ⫽ 2, 12%), abdominal distention (n ⫽ 1, 5.9%), right shoulder pain (n ⫽ 1, 5.9%), and generalized itching (n ⫽ 1, 5.9%). Laboratory data obtained within 3 days of the procedure were available in 18 patients. The mean serum aspartate aminotransferase/alanine aminotransferase level was 555.0 IU/L ⫾ 825.8/541.0 IU/L ⫾ 863.7 (preoperative values, 59.4 IU/L ⫾ 41.0/64.6 IU/L ⫾ 86.8). The mean serum total bilirubin level was 5.5 mg/dL ⫾ 11.8 (94 mol/L) (preoperative value, 0.8 mg/dL ⫾ 0.7). Eight patients (47.1%) who underwent the percutaneous procedure were discharged after 1 overnight stay without substantial symptoms. Two patients (66.7%) who underwent an intraoperative procedure were discharged 7 and 9 days after RFA, respectively. The duration of the hospital stay of the remaining 10 patients (50%) was increased. A 65-year-old man who underwent intraoperative RFA with a cluster-type electrode for 12 minutes for HCC that had been treated partially with TACE had a large infarction (volume, 679.1 cm3) involving the right hepatic lobe with portal venous thrombosis from the main to the segmental branches. He refused liver transplantation. Despite intensive care, he died 36 days later due to hepatic failure. Therefore, the mortality rate of patients with hepatic infarction as a complication of RFA was 5% (one of 20 patients). Risk Factor Analysis Among the eight potential risk factors evaluated, old age (P ⫽ .048) and large tumor size (P ⫽ .026) turned out to be independently significant risk factors for developing hepatic infarction after RFA for HCC according to multiple logistic regression analysis (Table 2).
DISCUSSION Previous studies have described the CT appearance of hepatic infarction (11,12,14). One of two recent reports (11) described hepatic infarction as a lowattenuation, peripheral, and wedgeshaped lesion at CT. The other report (12) described three patterns: (a) peripheral, wedge-shaped; (b) central or peripheral, round or oval; and (c)
September 2007
JVIR
irregular and parallel to bile duct. In both studies, proof of the hepatic infarction had been made with histopathologic, angiographic, or clinical methods. In our study, we used the imaging definition of hepatic infarction as a newly developed, wedge-shaped, round or oval, nonenhanced lesion peripheral to the ablation zone with its base on the hepatic capsule at all three phases of contrast-enhanced CT. Recently, several multi-institutional, large series studies (2–5) have been performed to evaluate the complications of RFA for hepatic tumors. Hepatic infarction was so uncommon that the incidence was reported to be no more than 0%– 0.07%. There were only two cases of hepatic infarction noted after RFA in 4,379 patients with 6,524 hepatic tumors. Overall, the incidence was only 0.045%. Conversely, 20 cases of hepatic infarction were found in 1,120 sessions (1.8%) of RFA for 872 patients (2.3%) with 1,335 HCCs (1.5%) in our study; this frequency was higher than that reported in previous studies. We believe that this discrepancy might be caused by the method of identifying cases. We reviewed all CT scans obtained after RFA to search for lesions suggestive of hepatic infarction irrespective of clinical symptoms. However, previous studies might have included only cases that caused a substantial clinical problem or a noticeable change on imaging studies. According to standardization of terminology and reporting criteria developed the International Working Group on Image Guided Tumor Ablation (15), only 11 patients in our study actually could be classified as having a major complication due to a lengthening of the hospital stay (n ⫽ 10) and/or an increase in the level of care (n ⫽ 7). Moreover, some patients classified as having a major complication might even be regarded as merely having postablation syndrome clinically (16), in which similar symptoms and laboratory changes can occur. Therefore, previous studies might have overlooked asymptomatic cases, cases with minor symptoms, or cases with similar features to postablation syndrome; therefore, the real incidence must be higher than reported previously. Because the liver has a dual blood supply from the hepatic artery and
Volume 18
Number 9
Kim et al
•
1131
Figure 2. Images in a 50-year-old man treated with percutaneous RFA by using an internally cooled electrode (2 cm, single type) for 8 minutes. (a– c) CT scans obtained in the (a) arterial, (b) portal venous, and (c) equilibrium phases 2 hours after RFA show a nonenhanced RFA zone (solid arrow) and complicated hepatic infarction (open arrows) in segment VI of the liver. The patient complained of tolerable abdominal pain and a mild febrile sensation and was discharged 2 days later after undergoing only conservative treatment. (d–f) Follow-up CT scans obtained 1 month after treatment demonstrate atrophic change of the low-attenuation infarcted area (open arrows) with mild capsular retraction at the hepatic arterial (d) and portal venous (e) phases. The equilibrium phase scan (f), however, shows reperfusion of the previously infarcted lesion. RFA zone (solid arrow) showed a relatively small change. This lesion disappeared completely at 4-month follow-up CT, with only minimal residual scar and further capsular retraction.
portal vein as well as an ability to develop extensive collateral pathways, hepatic infarction is generally thought
to be relatively rare (17–19). When hepatic infarction occurs, however, it is known to result from an insult either
to the hepatic arterial supply or to both the hepatic arterial and portal venous systems, but not to the portal
1132
•
Hepatic Infarction after RF Ablation of HCC
Figure 3. Images in a 56-year-old man treated with percutaneous RFA by using two types of internally cooled electrode (both the cluster and single 3-cm type) for 8 minutes. This patient complained of mild abdominal pain and high fever (39.3°C) 2 days after treatment. (a) CT scan obtained during the portal venous phase immediately after RFA shows the typical RFA zone (solid arrow) and complicated hepatic infarction (open arrows). (b) Follow-up CT scan obtained 4 days later shows that multifocal, small, low-enhanced lesions with peripheral rim enhancement suggestive of hepatic microabscesses (arrowheads) had developed at the infarcted area. Purulent aspirate from the lesion confirmed the diagnosis (arrow: RFA zone). Symptoms were relieved after a 9-day course of intravenous antibiotic therapy.
venous supply alone (20 –22). Therefore, the hepatic artery appears to have a key role in the development of hepatic infarction. This concept might help explain the risk factors that were proved to be significant in our current study. Older patients have age-related reduced status of their arterial system; therefore, the hepatic artery in older patients is more prone to insult by the same stimulation (ie, hyperthermia) than that of younger patients. In cases with larger tumors, the chance of vascular injury increases with the ablation of a larger area. Although other factors (eg. total ablation time and history of TACE) may be related to an increased risk of arterial injury, they were proved not to be independently significant with multivariate analysis. The clinical features associated with hepatic infarction, especially in mild cases, overlap those of postablation syndrome. According to a recent prospective study (16), fever and malaise were the most common symp-
toms of postablation syndrome; abdominal pain and fever were most the common symptoms in our patients. As for fever, patterns are very similar (postablation syndrome vs hepatic infarction: onset, 1–9 days [mean, 3.0 days] vs 1–3 days [mean, 2.0 days]; duration, 1–11 days [mean, 5.5 days] vs 1–8 days [mean, 4.0 days]; peak body temperature, 37.4–38.9°C [mean, 38.2°C] vs 37.6– 39.2°C [mean, 38.4°C]). Laboratory data such as serum aspartate aminotransferase levels were also similar (517.8 IU/L ⫾ 100.8 vs 555.0 IU/L ⫾ 825.8). Overlaps like these are quite natural considering both are caused by a hepatic parenchymal necrosis—whether induced by ablation alone or both ablation and infarction. Despite these similarities, abdominal pain was more predominant in our study and might be a distinctive clinical clue suggestive of hepatic infarction rather than postablation syndrome. In addition to postablation syndrome, hepatic abscess is another entity to be differentiated from hepatic infarction. In our 20 patients with
September 2007
JVIR
hepatic infarction, 10 (50%) complained of both fever and abdominal pain. In addition, we found gas collections at CT in six of them. Although these could be the findings of a hepatic abscess, only one patient was found to have a hepatic abscess at clinical examination. Limitations of our study were as follow: First, because our study was retrospective, unknown bias might be involved and standardization for the evaluation of clinical symptoms and laboratory data was not possible. Second, we cannot exclude completely the possibility that another pathologic condition was misinterpreted as an infarction because there was no histologic backup. However, our criteria for hepatic infarction were as good as possible considering the retrospective and CT image analysis– based nature of our study. Third, delineations of the border between the infarction and ablation zone were not always evident. In these cases, it was possible to draw an imaginary border considering both shape of proximal margin of ablation zone and evolution of ablation zone and infarcted area at follow-up CT. This might cause an error in the assessment of volume, but we do not think it was crucial. Last, some clinical features of the patients with hepatic infarction might be attributed to a tissue necrosis caused by ablation itself (ie, postablation syndrome). Despite these limitations, we can conclude that hepatic infarction, as a complication of RFA for HCC, is uncommon. It is, however, more frequent than previously reported, especially in patients with larger tumors and those with a more advanced age. We can manage most cases of hepatic infarction after RFA conservatively. However, surveillance for extensive infarction is needed because this can cause hepatic failure and subsequent mortality. References 1. Choi D, Lim HK, Rhim H, et al. Percutaneous radiofrequency ablation for early-stage hepatocellular carcinoma as a first-line treatment: longterm results and prognostic factors in a large single-institution series. Eur Radiol 2007; 17:684 – 692. 2. Rhim H, Yoon KH, Lee JM, et al. Major complications after radio-frequency thermal ablation of hepatic tumors: spectrum of imaging findings. Radiographics 2003; 23:123–136.
Volume 18
Number 9
3. Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter study. Radiology 2003; 226: 441– 451. 4. de Baere T, Risse O, Kouch V, et al. Adverse events during radiofrequency treatment of 582 hepatic tumors. AJR Am J Roentgenol 2003; 181:695–700. 5. Curley SA, Paolo M, Beaty K, et al. Early and late complications after radiofrequency ablation of malignant liver tumors in 608 patients. Ann Surg 2004; 239:450 – 458. 6. Choi D, Lim HK, Kim MJ, et al. Liver abscess after percutaneous radiofrequency ablation for hepatocellular carcinomas: frequency and risk factors. AJR Am J Roentgenol 2005; 184:1800 – 1807. 7. Kim SH, Lim HK, Choi D, et al. Changes in bile ducts after radiofrequency ablation of hepatocellular carcinoma: frequency and clinical significance. AJR Am J Roentgenol 2004; 183: 1611–1617. 8. Choi D, Lim HK, Kim MJ, et al. Therapeutic efficacy and safety of percutaneous radiofrequency ablation of
Kim et al
9.
10.
11.
12.
13.
14.
hepatocellular carcinoma abutting the gastrointestinal tract. AJR Am J Roentgenol 2004; 183:1417–1424. Jaskolka JD, Asch MR, Kachura JR, et al. Needle tract seeding after radiofrequency ablation of hepatic tumors. J Vasc Interv Radiol 2005; 16:485– 491. Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusion of the Barcellona-2000 EASL conference. J Hepatol 2001; 35:421– 430. Smith GS, Birnbaum BA, Jacobs JE. Hepatic infarction secondary to arterial insufficiency in native livers: CT findings in 10 patients. Radiology 1998; 208: 223–229. Holbert BL, Baron RL, Dodd GD III. Hepatic infarction caused by arterial insufficiency: spectrum and evolution of CT findings. AJR Am J Roentgenol 1996; 166:815– 820. Breiman RS, Beck JW, Korobkin M, et al. Volume determinations using computed tomography. AJR Am J Roentgenol 1982; 138:329 –338. Lev-Toaff AS, Friedman AC, Cohen LM, Radeki PD, Caroline DF. Hepatic infarcts: new observations by CT and sonography. AJR Am J Roentgenol 1987; 149:87–90.
•
1133
15. Goldberg SN, Grassi CJ, Cardella JF, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria. J Vasc Interv Radiol 2005; 16: 765–778. 16. Dodd GD III, Napier D, Schoolfield JD, Hubbard L. Percutaneous radiofrequency ablation of hepatic tumors: postablation syndrome. AJR Am J Roentgenol 2005; 185:51–57. 17. Kanter DM. Hepatic infarction. Arch Intern Med 1965; 115:479 – 481. 18. Temberg JL, Butcher HR. Blood flow relation between hepatic artery and portal vein. Science 1965; 150:1030 –1031. 19. Redman HC, Stewart RR. Arterial collaterals in the liver hilus. Radiology 1970; 94:575–579. 20. Popper HL, Jefferson NC, Necheles H. Liver necrosis following complete interruption of hepatic artery and partial ligation of portal vein. Am J Surg 1953; 86:309 –311. 21. Loeffler L. Factors determining necrosis or survival of liver tissue after ligation of hepatic artery. Arch Pathol 1936; 21:496 –503. 22. Wooling KR, Baggenstoss AH, Weir JF. Infarcts of the liver. Gastroenterology 1951; 17:479 – 493.