Dr. Krishnamurthy is an Assistant Professor of Clinical Radiology, Dr. Casillas is a Professor of Radiology , and Dr. Latorre is a Visiting Resident from Columbia in the Department of Radiology, University of Miami/Jackson Memorial Hospital, Miami, FL.

The liver is a common site for both primary and secondary malignancies, the most common being hepatocellular carcinoma (HCC) and metastasis from colorectal carcinoma. Hepatocellular carcinoma is the fourth most common cause of cancer-related deaths worldwide, and approximately one million new cases are reported annually. ¹ Mortality is essentially 100% when these tumors are not treated. Surgical resection is the standard of care because it has been shown to provide survival benefits, while systemic chemotherapy and radiotherapy are largely ineffective. ^2,3 However, only 5% to 15% of patients with HCC or hepatic metastasis are candidates for curative surgery. ^2,4,5 The major limiting factors for surgical treatment are too many tumors, large lesions, tumors in unresectable locations, tumors with major vascular invasion, insufficient hepatic reserve from cirrhosis, and underlying medical problems that increase the surgical risk. There is also significant peri-operative morbidity and mortality. The average 5-year survival rate after successful resection for both HCC and metastasis is only 20% to 40%. ^4,5 A considerable number of patients will develop recurrence of tumor, which is usually fatal.

Various minimally invasive techniques have been developed as an alternative to surgery. The most important are percutaneous ethanol injection (PEI), hepatic arterial chemoembolization, and thermal ablation techniques. Percutaneous ethanol injection has been reported to be primarily successful in small HCC and is not very useful for treatment of metastatic tumors. ⁶ Chemoembolization is often reserved for unresectable tumors. ⁷ Recently, thermal ablation techniques have become popular. There are two types--cold (cryoablation), and heating ablation techniques. The latter includes radiofrequency, microwave, laser, and high-intensity focused ultrasound (US) ablations. Cryoablation is the most popular among surgeons with reported success rates similar to resection surgery. ⁸ But this technique is invasive, requiring laparotomy in most cases. Experience with microwave and laser ablations is limited and the high-intensity focused US technique is still in the experimental stage. ^9­11

Radiofrequency (RF) ablation has gained more popularity in the last decade and is claimed to be a very promising technique for both primary and secondary hepatic tumors. ^12,13 Radiofrequency ablation produces controlled coagulation of the tumor by heating the tissue to temperatures above 50ºC. Recent studies have reported favorable survival rates and excellent rates of local tumor control by RF ablation, especially in patients with HCC. ^12-15 In this article, we will discuss the historical background, mechanism of RF ablation, optimal ablation technique, imaging follow-up evaluation, procedural complications, and results. Combination therapy of RF ablation with other techniques is also presented.

Historical perspective

The effect of RF waves was first reported by d'Arsonval in 1891. ¹⁶ He described heating of tissue when the RF waves pass through living tissue. This led to the development of medical diathermy and surgical electrocautery probes in the 1990s. ¹⁷ The electrocautery probe works by desiccation and charring of the tissue at the point of contact when alternating current passes through the patient between the cautery probe and a large grounding pad applied usually to the thigh of the patient. Radiofrequency ablation technique is based on this principle. Physical principle of tissue interaction with RF waves was first described by Organ, ¹⁸ who demonstrated that alternating current causes agitation of ions in the living tissue that results in frictional heat and thermal injury.

Radiofrequency ablation in the liver was first conceptualized by McGahan et al ¹⁹ and Rossie et al ²⁰ in the 1990s. Using alternating current, a grounding pad, and a stock needle insulated to the distal tip (monopolar electrode) and inserted percutaneously, they produced focal thermal ablation at the needle tip deep within the liver. They proposed that RF ablation could be an effective treatment for small hepatic tumors.

Mechanism of RF ablation

An RF ablation system consists of a very high frequency (200 to 1200 KHz) alternating current generator, RF needle (monopolar electrode), grounding pad (which serves as a large dispersive electrode) and the patient must all be connected in series. In this circuit, electric current enters through both the electrodes with the patient as resistor. As the electric current alternates in directions at high frequency, tissue ions that are attempting to follow the direction of the current get agitated. Due to natural high resistivity in the living tissue, ionic agitation produces frictional heat at the immediate vicinity of electrodes. Because the grounding pad has a very large surface area, the electrical resistance is low; hence, the production of frictional heat is concentrated at the needle electrode. Thus, deposition of electromagnetic energy from electric current produces thermal injury. The extent and nature of this injury are dependent on two important factors: temperature and RF application duration. ^21,22 To produce irreversible cell damage, it takes several hours at 45ºC, but it takes only 4 to 5 minutes at 50°C to 55ºC. At temperatures between 60ºC and 100ºC, there is immediate tissue coagulation (due to irreversible damage to mitochondrial and cytosolic enzymes by heat-induced denaturation of proteins). Above 100ºC, tissue simply vaporizes. Therefore, temperatures between 50ºC and 100ºC are ideal for RF ablation.

RF devices

The U.S. Food and Drug Administration has approved RF devices manufactured by three companies--RITA Medical Systems (Mountain View, CA), Radiotherapeutics (Mountain View, CA), and Radionics (Burlington, MA)--for treatment of liver tumors that are not amenable to surgery (Table 1).

RITA device

For the model 70 or model 90 Starburst XL devices (RITA Medical Systems), the ablation protocols are based on the attainment of a target temperature in the tissue. To perform ablation, two grounding pads are applied to the thighs or back of the patient. The tip of the needle is advanced with the prongs retracted, to the desired location, and then the prongs are partially deployed. There is an automated program that will slowly raise the generator power from 25 W to peak within 1 to 1.5 minutes. Full deployment of the prongs is performed when the target temperature (usually 95ºC to 105ºC) is achieved. The device automatically adjusts the power to maintain the target temperature as the tumor is cooked. The ablation is carried out to a preselected time in the treatment protocol that is at least 10 minutes for smaller tumors and >20 minutes for tumors >5 cm in diameter. For satisfactory ablation, the company recommends that after completion of the ablation cycle, the temperature reading from the central electrode must be >50ºC for at least 1 minute. The needle tract is ablated with a tract ablation program, which maintains the temperature >60ºC while removing the needle.

Radiotherapeutics device

The ablation protocols for the LeVeen needle electrode (Radiotherapeutics) are based on tissue impedance, unlike the target temperature used in the RITA devices. Two grounding pads are applied to the patient's thighs. The needle is inserted to the desired location, and the prongs are deployed to full length. The generator is initially set to 30 W and is then increased by 10 W steps every minute until the peak power of 90 W is attained. At this power, ablation is continued for 15 minutes or until the high impedance stops current flow. If the device did not impede out in the first 15 minutes, then the generator is switched off for 30 seconds and then restarted at maximum power and run until high impedance again stops further ablation or until 15 minutes has elapsed. If the ablation stops due to high impedance during the first cycle, the machine is switched off for 30 seconds and restarted at 70% of maximum power and run for 15 minutes or until high impedance again stops the ablation process.

Radionics device

The ablation protocols for the Cool-tip RF electrode (Radionics) are also based on tissue impedance. The needle shaft has two internal channels connected at the tip, allowing perfusion of chilled water. The internal cooling of the needle prevents desiccation and charring at the needle tip, allowing better dis-sipation of the frictional heat, which increases the size of the thermal injury. ²³ Recently, a clustered three-needle electrode system has been introduced, which has three similar internally cooled needles arranged in a parallel, triangular manner and attached to a common hub. This arrangement is shown to produce a larger volume of ablation than a single-electrode system produces. ²⁴

Four grounding pads are applied on the patient's thighs. The single or cluster needle is inserted to the desired location and then connected to the generator and perfusion pump that circulates sterile chilled water. The automated program in the generator increases the power to a peak of 200 W within 1 minute. The power is maintained until the impedance rises more than 20 (omega), then the power is reduced to 10 W for 15 seconds. The power is increased once more to maximal power until the impedance rises again above 20 (omega). This cycle is repeated for a total of 12 minutes.

Preprocedural evaluation

Complete evaluation of the patient, including clinical history and directed physical examination, is required. Table 2 gives a summary of the required preprocedural investigations and their significance. Any active infection, uncorrectable coagulopathy, untreatable extrahepatic tumor spread, and extreme debility are contraindications. Histopathologic confirmation of malignancy by biopsy is essential in all patients. Written and informed consent should be obtained.

Imaging studies form the basis for treatment. Triple-phase computed tomography (CT) scanning or contrast-enhanced magnetic resonance imaging (MRI) scanning is usually required. The number of lesions, their size, location, relationship to adjacent viscera (eg, intestines, diaphragm, gallbladder), and proximity to major intrahepatic vessels (eg, hepatic and portal veins) should be carefully recorded. Ideal tumors for RF ablation are <3 cm in diameter, completely surrounded by hepatic parenchyma, >=1 cm deep to the liver capsule, >=2 cm away from large hepatic or portal veins, and <5 in number. Patients with too much of a tumor burden (>5 cm in diameter and >5 lesions) are not good candidates. Extrahepatic tumor spread should be evaluated with a bone scan or chest CT scan, if clinically appropriate.

RF ablation procedure

The RF ablation can be performed by three different approaches, including percutaneous, intraoperative, and laparoscopic routes. The percutaneous approach is preferred because it is least invasive, is associated with minimal morbidity, has a lower cost, can be performed on an outpatient basis, and can be repeated as necessary to treat recurrent disease. The intraoperative and laparoscopic approaches are useful in certain special circumstances, which will be discussed separately.

The patient should fast for 6 to 8 hours prior to the procedure. A large peripheral intravenous (IV) access is established, and the patient is monitored to track cardiac and respiratory rate, blood pressure, and peripheral oxygen saturation. Although there is no consensus about routine administration of antibiotics, we administer a single IV dose of ampicillin and cephalosporin, starting 30 minutes to 1 hour prior to the procedure. We continue oral cephalosporin for 5 to 7 days after the procedure.

The procedure can be performed under IV conscious sedation or general anesthesia, depending on the patient and operator preferences. A preanesthetic evaluation should be performed. Fentanyl citrate and midazolam hydrochloride are used for IV conscious sedation. Deeper sedation can be obtained with IV propofol infusion. General anesthesia is highly recommended if the ablation procedure is expected to be extensive and last >=3 hours.

The actual needle insertion can be performed under US or CT guidance, depending on which shows the tumor best and the operator preference (Figure 1). In general, US is the most commonly used modality because it provides good real-time guidance for needle placement, is less expensive, and is quicker to perform. ^25­27 However, during the ablation process, echogenic bubbles form that can limit visibility, making it difficult to assess the completeness or the extent of the coagulation necrosis. CT is useful in monitoring the ablation process since the perimeter of the ablated area can be better visualized.

The goal of treatment is to obtain complete tumor destruction. If the RF ablation were to be as successful as surgical resection, the ablation should achieve a reasonably good tumor-free margin (at least 1 cm, preferably 2 cm). ²⁸ Therefore, ablation should extend to a cuff of at least 1 cm of normal hepatic parenchyma. Current RF devices can produce a 3-cm sphere of ablation. Therefore, they are ideal to ablate tumors <1 cm in diameter (subtracting 2 cm from the 3-cm diameter for the 360š 1-cm tumor-free margin). ¹³ Technical strategy of larger tumor ablation is described below.

Ablation of lesions under special circumstances

Larger lesions

When the tumor is >=3 cm, single ablation is not sufficient, thus increasing the concern for under treatment and tumor recurrence. The difficulty in achieving complete ablation of larger tumors is not only because of the tumor volume, but also because of poor local electrical and heat conductivity within the tumor. ²⁹ The latter is attributed to the heterogeneity of tumor caused by fibrosis and calcifications. Given these factors, the strategy to obtain larger ablation includes overlapping of ablations, improving electrical conductivity within the tumor by intratumoral saline injection, and modifications of electrode design.

Overlapping ablation technique-- It is logical to expect that overlapping ablations can produce larger ablation volumes. After a single-ablation sphere, 6 overlapping ablations (4 in x-y plane and 2 in the z-axis) produce, geometrically, the next largest ablation. This approach will produce a composite sphere with internal diameter of only 3.75 cm (just 1.25 times the single-ablation sphere). This is suitable for tumors <1.75 cm in diameter (subtracting 2 cm from the 3.75 cm diameter for the 360š 1-cm tumor-free margin). Fourteen overlapping ablations is the next step, and the composite sphere produced will have a diameter of 5 cm (1.7 times that of the single-ablation sphere). This is suitable for a 3-cm diameter tumor (subtracting 2 cm from the 5-cm diameter for the 360š 1-cm tumor-free margin). In actual practice, it is very difficult (almost impossible) to reproduce this model. To overcome this problem, the concept of overlapping cylinder strategy has been advocated. ³⁰ In this technique, spheres of ablation are overlapped linearly to create cylinders, and the cylinders are then overlapped. Geometrically, this model may produce imperfect coverage, but it can be performed practically with greater ease and success.

Improving electrical conductivity within the tumor­­ Intratumoral saline injection prior to or during RF ablation improves electrical conductivity within the tumor. This causes greater deposition of the RF energy, leading to increased tumor coagulation. ³¹ It is of interest that neither the concentration nor the volume of the saline solution correlates linearly with the size of the RF thermal injury. Therefore, optimal saline concentration and volume need to be titrated carefully for each of the different devices and
different tumors.

Modifications of electrode design ­­ Electrode modifications to ablate larger tumors include clustered array electrodes, bipolar electrodes, and pulsed RF deposition systems. ^24,32,33 Clustered array electrodes (such as the Radionics model) represent an extension of the same concept as the single internally cooled electrode described earlier. Goldberg et al ²⁴ demonstrated that the sphere of coagulation necrosis produced by a cluster of electrodes placed <1 cm apart is greater than that created by a composite of individual electrodes. Their in vivo experiments on liver ablation demonstrated that cluster electrodes placed 0.5 cm apart produced an ablation sphere of 3.1 cm versus 1.8 cm produced by conventional single electrodes under otherwise similar conditions. ²⁴ Their initial clinical experience in 10 patients with colorectal metastasis showed ablation injury of 5.3 cm ± 0.6 using a clustered electrode with a single 12- to 15-minute ablation.

A bipolar electrode has a different design. There is no grounding pad; instead, there is an active electrode and a closely placed grounding electrode. The heat is generated not only around the active electrode, but also around the grounding electrode and in the space between the two. This is in contrast to the monopolar electrode, where heat is generated only at the active electrode. Early clinical experience demonstrates that bipolar needles produce a larger coagulation volume of 3-cm diameter by a single application alone. ³² Absence of a grounding pad eliminates the risk of grounding pad burns also.

Pulse RF ablation is another strategy aimed at increasing volume of coagulation by increasing the RF energy deposition. ³³ In this technique, higher energy deposition is alternated with lower energy deposition. During periods of low energy deposition, the tissue around the electrode cools down, allowing for even higher energy deposition during the next cycle of ablation. This method allows for deeper heat penetration, creating a larger ablation zone. The experience is limited, but seems a promising technique for treating larger lesions.

Tumors at difficult locations

Tumors that are subcapsular or at the liver hilum; and those adjacent to the gallbladder, colon, intestines, and diaphragm are of special concern because of the difficulty in achieving complete ablation with a good tumor-free margin and increased risk of complication. In the case of subcapsular tumors, if the tumor involves the capsule, the latter should also be ablated with the tumor (Figure 2). Careful selection of the needle configuration is important. If the tumor is small and rounded, an expandable needle system (such as the RITA and Radiotherapeutics models) is useful because they can produce more spherical ablation injury compared with the straight Cool-tip needles (Radionics). For tumors that are stranded at the angles of liver, such as in the inferior angle of right lobe or in lateral segment of left lobe, a straight cooled-tip needle would be more useful since it can be fully deployed for ablation. The prongs of the expandable needle systems often penetrate the liver capsule, increasing the risk of bleeding when they are fully deployed. If the expandable needle has to be used, then the prongs should be partially deployed and ablation has to be performed multiple times to cover the whole tumor.

Tumors at the hepatic hilum present another distinct problem because of their proximity to main vessels, eg, the portal vein and hepatic artery. These large vessels produce what is known as perfusion-mediated tissue cooling or the heat-sink effect. ³⁴ The rapid flow of blood in these vessels washes out the heat quickly, preventing build-up of optimal temperature to induce coagulation necrosis in the periphery of tumor adjacent to the vessel. In pig liver models, Hansen et al ³⁵ showed that vessels >3 mm in diameter prevent complete ablation of liver tissue. Studies have shown that reducing hepatic perfusion by mechanical or pharmacologic methods improves coagulation in the tumor. ^34,36 This can be produced at laparotomy or during the laparoscopic-guided approach by clamping the hepatic artery and portal vein in the hepatoduodenal ligament (Pringle maneuver). Although clamping can be applied safely for up to 1 hour, there is increased risk of liver ischemia and related complications, such as liver failure and biliary strictures. Also, resorting to open procedures would offset the potential benefits of a minimally invasive percutaneous approach. Alternatively, angiographic balloon occlusion of a hepatic artery has been used during percutaneous procedures. ³⁷ Hepatic artery embolization with temporary embolic agents, such as gel foam has also been reported. However, the practical utility and the efficacy of these additional maneuvers is questionable because of the invasive nature of the techniques and the dual blood supply of the liver. Other techniques for reducing blood flow, eg,
pharmacologic modulation and antiangiogenesis therapy, are largely experimental at this time. ³⁶

There is another difficulty in targeting tumors at the hepatic hilum, since the tumor is located in the space between large vessels, which is small and narrow. The expandable needle electrodes should be partially deployed; otherwise they may penetrate the large vessels, resulting in internal bleeding or pseudoaneurysm formation. This usually necessitates multiple ablations. Straight-tip internally cooled needles are advantageous in this location because they can be deployed fully and there is little concern for vessel penetration. They also have better tumor coagulation due to internal cooling design, as discussed earlier.

The tumors located close to the diaphragm or other viscera, such as the gallbladder and intestines, are also difficult to treat. Diaphragmatic thermal injury can cause severe, persistent pain and breathing difficulty, leading to lower lung atelectasis and pleural effusion. The chronic pain may require long-term analgesics. Damage to viscera can cause cholecystitis or ischemic bowel injury. Therefore, treatment of tumors located at these critical locations may be better performed using open surgical procedures, as will be discussed in the following section.

Laparotomy and laparoscopic-guided ablations

Tumors located adjacent to the dia-phragm, gallbladder, or bowel can be treated more easily during an open procedure because the liver and the adjacent organs can be separated or mobilized, thus avoiding potential injury. ^38,39 These techniques also allow for more accurate tumor staging, like detection of peritoneal or surface implants and lymphadenopathy. They provide an opportunity to perform intraoperative US that can detect unsuspected tumor deposits (Figure 3). These findings can avoid an unnecessary local tumor treatment. ^39,40 Both of these approaches also allow for performance of the Pringle maneuver, which can increase the extent of tumor coagulation. But the open techniques are invasive and more costly, and are associated with higher morbidity and mortality than percutaneous approaches.

Monitoring of coagulation during RF ablation

Currently, imaging techniques (US, CT, and MR) are standard for monitoring coagulation. ⁴¹ Of these, US is the most commonly used modality and the ablation is seen as an echogenic area due to liberation of nitrogen gas during the coagulation (Figure 4). But US is unsuitable for larger lesions or when multiple ablations are planned because the echogenicity from the initial ablation can obscure visibility during subsequent needle placement. We use a combination of US and CT in these circumstances for intraprocedural monitoring.

Both noncontrast and contrast-enhanced US, CT, and MR scans are used for assessing the adequacy of ablation. In general, noncontrast imaging findings are unreliable parameters for predicting the adequacy of ablation. ^29,41 The ablated area appears hyperechoic on US or hypodense on CT scan. In experimental porcine liver models, some authors have reported the appearance of decreased hypoattenuation on noncontrast CT as a sensitive indicator of complete ablation. ⁴² But more important morphologic features to be determined are the size and margin (the interface between normal liver tissue and the ablation defect) of the defect. After complete ablation, the defect should be centered on the tumor location, it should beinitially larger than the tumor (which ensures desired tumor-free margins), and the interface should be sharp, smooth, and devoid of nodularity. ¹³ When blood-pool contrast agents are used (iodinated contrast in CT and gadolinium contrast in MR), there should be complete lack of tumor enhancement with thin, smooth rim enhancement at the interface. ⁴¹ The rim represents ablation-induced hyperemia (Figure 5). Irregularity and nodularity suggests incomplete ablation with residual tumor. Contrast-enhanced power and color Doppler US has also been used for monitoring of coagulation. Ultrasound contrast agents allow for better differentiation between perfused and nonperfused tissue, which indirectly represent viable and nonviable (coagulated) tissue. ^43,44 Therefore, adequacy of ablation can potentially be predicted in real-time during the procedure. ^45-47 The lack of general availability and operator experience are limiting factors for universal use of US contrast agents.

Direct documentation of tissue temperature (using additional remote temperature sensors during and immediately after ablation) has also been used to predict adequacy of ablation. ^33,48 Temperature-sensing electrodes are placed under US/CT guidance in the periphery of the tumor or cuff of normal tissue. Temperatures in these electrodes >55ºC indicates adequacy of tumor ablation. However, this technique is not fully optimized and reliable systems have yet to become available for general use.

Immediate postprocedural care

The patient needs intensive monitoring with frequent recording of vital signs for any immediate complications like internal bleeding. Patients usually experience significant pain and nausea and require IV narcotic analgesics, eg, morphine sulphate and antiemetics. Once stabilized, the patient can be moved to a regular floor bed with routine vital monitoring. Oral fluids and regular diet can be resumed as tolerated by the patient. Patients usually require oral analgesics and we prescribe a combination of acetaminophen and hydrocodone bitartrate. Patients can be discharged home the next day. We prescribe 5 days of oral cephalosporin tablets. If significant pain persists, oral hydrocodone bitartrate and acetaminophen tablets are continued for 3 to 5 days.

Long-term follow-up

The goal of long-term follow-up is to detect untreated residual and recurrent tumor when it is small and potentially treatable (Figure 6). Patients are also followed to identify development of any extrahepatic disease. Any one or a combination of the imaging modalities (CT, MRI, and/or US) may be used for follow-up. The choice of modality is based on patient concerns (eg, contrast-media­related allergy, renal status, and clinical availability) and operator experience. We recommend that follow-up should be performed with the same imaging performed prior to treatment, as this will allow for accurate comparison of findings.

On CT scans, complete ablation is seen as a low-attenuation area devoid of enhancement or nodules, as described earlier. Hepatic arterial-phase images are useful in evaluating hypervascular HCC and portal venous-phase images are useful in evaluating metastases. On MR imaging, the ablated tissue demonstrates low signal and there is lack of enhancement in dynamic contrast-enhanced sequences. Bright signal on T2-weighted images and nodular contrast enhancement are suggestive of viable tumor tissue. In the early postoperative period, there is contrast enhancement of the rim or interface. The finding usually persists for a few weeks, usually disappearing by 4 weeks on CT and possibly persisting for few months on contrast-enhanced MR. Another important finding of complete ablation is that the size of the defect remains stable or decreases over time. Any increase in size is suspicious for tumor recurrence.

The follow-up sequence most commonly used is to obtain scans immediately after the procedure (within 24 hours) followed by scans at 1 month, 3 months, and then every 3 months. It is also useful to obtain serum alpha-fetoprotein or carcinoembryonic antigen, as appropriate, at 3 monthly intervals in HCC and colorectal metastasis. Ideally, the follow-up should be continued indefinitely since tumors are known to recur even after months.

Two other imaging techniques used for follow-up need special mention: contrast-enhanced US and functional positron emission tomography (PET) imaging. ^43,49 Unenhanced color and power Doppler US do not reliably detect residual or recurrent tumor. ^41,50 However, combining US contrast agents with the Doppler technique improves detection of residual or recurrent tumor. ^43,44 Recent studies suggest that the addition of harmonic imaging to contrast-enhanced Doppler techniques can improve accuracy levels comparable to that of triple-phase CT scanning. ^46,47,51 With 18-fluorodeoxyglucose (18-FDG)-labeled PET scanning, abnormal increased activity after ablation has been reported to represent residual or recurrent tumor. ⁴⁹ However, the experience is still limited, and this modality cannot be recommended for routine follow up.

Percutaneous biopsy is another potential tool in diagnosing tumor recurrence. Imaging findings are usually sufficient to diagnose residual or recurrent tumor. If there is any discrepancy between clinical and imaging findings or if the imaging findings are not diagnostic, the suspicious area should be biopsied. Biopsy is not routinely recommended because of its limitations, such as sampling error and limited experience in interpretation of biopsy specimens with heat-induced damage. Thus, a negative biopsy does not guarantee complete treatment. ⁵² Similarly, imaging techniques also have limitations. The resolution of the current scanners is only 2 to 3 mm, making it difficult to diagnose microscopic residual or recurrent tumor, and to differentiate a small recurrence from ablation-induced hyperemia in the early postprocedural period. ⁵³ Therefore, optimal follow-up protocol will include anatomic and functional imaging, and image-guided biopsy when the imaging findings are not definite or when a discrepancy exists between clinical and imaging findings.

Complications of RF ablation

The incidence of clinically relevant complications following percutaneous RF ablation is low (2% to 7%). ^53,54 Although significant pain requiring IV narcotics is universal in all patients immediately after the procedure, persistent severe pain lasting more than 1 week is very rare. Ablation-induced injury to the diaphragm, gallbladder, intestines, liver capsule, and major portal veins are some of the causes of extended pain. These patients should be evaluated carefully, and any additional causes for acute abdominal pain should also be excluded.

Tumor ablation syndrome is another complication that results from the release of mediators of the inflammatory response due to ablation-induced cell death. The severity of the syndrome is dependent on the volume of tissue ablated. Smaller ablations result in low-grade fever (<38.8°C), malaise and leukocytosis starting about 3 to 4 days after the procedure and lasting for 1 week to 10 days. Larger ablations result in more severe symptoms of high-grade fever, nausea, vomiting, and lethargy. The symptoms can begin as early as day 1 and persist for 2 to 3 weeks. The treatment for ablation syndrome is primarily supportive, including antipyretics and hydration. The most important differential diagnosis is abscess at the site of ablation or septicemia. If infection is suspected, blood cultures should be performed and IV broad-spectrum antibiotics given. CT scanning may demonstrate gas at the ablation site, which need not necessarily signify abscess, since gas may be seen routinely following ablation without infection. Therefore, drainage of the ablation cavity that is suspected to be an abscess is purely a clinical decision.

Other reported complications are rare and include bleeding, injury to the bile duct (resulting stricture or biliary fistula) (Figure 7), pleural and lung injury causing pneumothorax and pleural effusion, gallbladder injury causing cholecystitis, and colon and intestinal injury causing ischemic bowel. Rarely, tumor seeding in the tract of the needle may be seen (Figure 8). But proper tract ablation can eliminate this complication (Figure 9). Vascular complications are hepatic artery pseudoaneurysm and arteriovenous fistula. Grounding-pad­related burns are also reported occasionally. ¹² Procedure-related mortality is very rare (<1%). ²⁵

Complications are also seen when general anesthesia is used. Open procedures (RF ablation at laparotomy and laparoscopy) have additional increased morbidity and mortality related to abdominal surgery and increased length of hospitalization.

Outcome analysis

Radiofrequency ablation is being increasingly performed worldwide. Currently, it is predominantly performed for treatment of hepatic tumors that are not amenable to resection. The most common tumors treated by RF ablation are HCC and colorectal metastases. Results of long-term survival rates from some of the larger studies have become available recently. ^25,26,54-56 These studies show that the most important determinant for achieving complete ablation is the size of the tumor. In general, technically successful complete ablation is possible in 90% of tumors <2.5 cm in diameter, in approximately 70% to 90% of tumors with 2.5- to 3.5-cm diameter, in 50% to 70% of tumors with 3.5 to 5.0 cm in diameter, and <50% of tumors >5.0 cm in diameter. ^25,29,56-59 Tumor type is also shown to affect the outcome with better success reported in treating HCC and breast metastases compared with colorectal metastases. ^25,57,60 Studies have also shown that tumor histology directly affects the success rate with well-differentiated HCC having better results of ablation than infiltrating lesions.

Radiofrequency ablation of HCC has been documented to provide survival benefits, particularly for smaller lesions (<3.5 cm). ^14,26,27 Disease-free survival on long-term follow-up, as reported by larger studies are as follows: 64% at 23-month and 71% at 12-month mean follow-up, according to Rossie et al ^14,27 ; 67% at 15-month mean follow-up, according to Francica et al ³³ ; and 92% at 1-year and 60% at 3-year follow-up, according to Iannitti et al. ⁵⁴

Radiofrequency ablation for liver metastases is promising but is less effective than it is for HCC. ^57,58,61 Colorectal metastasis is the most common metastasis treated by RF ablation. Disease-free survival has generally been reported to be lower than that reported for HCC. However, two recent studies show comparable results with disease-free survival of 87% to 93% at follow-up at 12 months, 62% to 77% at 24 months, and 41% to 50% at 36 months. ^54,61

Although hepatic resection is the standard of care for hepatic malignancy, major hepatic resection is associated with significant morbidity and mortality (2% to 10%). ⁶² Radiofrequency ablation, in comparison, is a minimally invasive and safe procedure with lower morbidity and mortality (<2%). ^53,54

Comparison with other minimally invasive techniques

Radiofrequency ablation is only one of many minimally invasive techniques, and other techniques include PEI, chemoembolization, and microwave, laser, or cryo-ablations. There are at least two studies comparing the outcome of RF ablation with PEI. ^56,63 They showed that complete tumor necrosis was better (90% versus 80% to 85%) and local tumor recurrence was lower (2% versus 7%) with RF ablation when compared with PEI. Another advantage was the fewer number of sessions (1.2 to 1.3 versus 3.3 to 4.8) required with the RF ablation technique. However, there is paucity of literature providing a direct comparison between RF ablation and other techniques.

Combination therapy of hepatic tumors­­The future of RF ablation?

Complete ablation of hepatic tumor depends not only on a perfect technique, but also on the biology of the tumor itself. There are strategies for combining RF ablation with other minimally invasive techniques, eg, intratumoral chemotherapeutic drug injection and PEI, which can alter the biology of the tumor such that tumor cells become more susceptible to heat.

It has been documented that the chemo-therapeutic agents and hyperthermia (42ºC to 45ºC) have synergetic effects. ^21,22
Goldberg et al ⁶⁴ demonstrated in an ex-perimental rat model (R3230 mammary adenocarcinoma) that intratumoral injection of doxorubicin increases the volume of coagulation necrosis by RF ablation. The ablation volume was significantly greater than the volume with RF therapy or doxorubicin therapy alone. The extent of coagulation was dependent on both the dose and timing of the doxorubicin, with the greatest coagulation achieved when doxorubicin was administered within 30 minutes of RF ablation. In another experiment (R3230 mammary adenocarcinoma, breast tumor model in rats), Goldberg et al ⁶⁵ showed that combined therapy consisting of PEI immediately followed by RF ablation showed significant increase in the extent of coagulation necrosis compared with either treatment performed alone. These combined strategies open a new venue for achieving complete tumor ablation of smaller tumors and may make ablation of larger tumors realistic in the future. However, further studies are necessary to determine which combinations are more effective and if they improve long-term outcome.

Conclusion

Current literature shows that RF ablation is a feasible and safe procedure for the treatment of primary and secondary liver tumors. Although surgical resection continues to be the standard treatment option, RF ablation is shown to be an effective alternative for tumors deemed not resectable. With current technology, RF ablation is a very effective treatment for small tumors, particularly HCC lesions. Recent advances in the device technology, including newer electrode designs, methods of increasing tissue susceptibility to heat, and combining with other treatment modalities, eg, chemotherapy, may make RF ablation a realistic treatment option for larger tumors, and may improve disease-free survival in these patients. In view of multiple, invasive, and minimally invasive treatment options available, further research and well-conducted, randomized, multicenter trials are required to determine the proper role of RF ablation in treatment of hepatic tumors.