RFA can be performed percutaneously, or during laparoscopic or open surgery. There are advantages and disadvantages to each, and the approach will depend on the condition of the patient, tumor characteristics such as location, size, number and growth pattern, and experience and preference of the provider[8]. There is insufficient evidence as of date indicating which delivery method is the preferred due to a lack of randomized control trials and varying patient and tumor characteristics between single technique studies. In a study comparing open, laparoscopic, and percutaneous approaches for liver tumors, there was no difference in mortality, major complications, or overall survival; but open compared to percutaneous approach resulted in improved disease free survival and decreased local tumor recurrence[9].

The percutaneous approach has the advantage of being performed under conscious or deep sedation, providing an option for patients who are higher surgical risk. This can usually be done as an outpatient or with a very short hospital stay, and can be performed multiple times if needed. The percutaneous approach can also be performed under anesthesia. Other advantages of this technique are the use of sonographic, computed tomography (CT) or magnetic resonance imaging (MRI) to guide precise electrode placement. At the same setting, contrast enhanced sonography or contrast enhanced CT can be done during the procedure to check for adequacy of ablation. Disadvantages of the percutaneous technique are lack of visualization of small surface tumors or deeper tumors which can be better identified with the open technique. Percutaneous RFA has shown excellent results for small < 3 cm neoplasms in the liver, lung or kidney. However, higher local recurrence has been shown with the percutaneous approach for larger tumors[10] and tumors in close proximity to major vessels, such as the portal vein.

Open RFA allows for better visualization and the ability to manipulate adjacent structures. It has the advantage of being able to detect occult metastatic disease with use of intra-operative ultrasound (US) and allows for treatment within a greater anatomic range. With hepatic RFA, another advantage is the ability to occlude portal inflow (Pringle maneuver) which, as described above, reduces heat dissipation and, therefore, increases the volume of tissue ablated. This technique is particularly valuable when tumors are located in proximity to vascular structures.

Laparoscopic RFA combines many of the benefits of both the percutaneous and open approaches. It is minimally invasive with less morbidity of a large incision while still allowing better visualization of the tumor and of adjacent structures to optimize staging. Pneumoperitoneum may also work in a similar manner to the Pringle maneuver and decrease the heat sink effect in tumors in proximity to large vessels by decreasing portal flow[11]. It also allows resection or displacement of structures adjacent to tumors that cannot be performed with the percutaneous technique.

Imaging plays an important role in the diagnosis and localization of the tumor, in real-time monitoring of the ablation zone, in assessment of tissue response to RFA therapy, and finally in patient follow-up. The RF probe is usually placed under CT or US guidance, and the RFA procedure monitored with real-time US. Ablation zones are seen on US as hyperechogenic areas which represent microbubbles created from the vaporization of interstitial fluid from ablated tissue. However, these hyperechogenic areas do not completely parallel the ablated zone. To determine the extent of necrosis following RFA in countries outside of the United States, US contrast is used at the time of the procedure to check for complete ablation and whether re-treatment is needed at the setting[12]. In the United States, a follow-up contrast-enhanced CT or MR is typically used, with successfully ablated areas failing to enhance. A thin enhancing rim representing either inflammation or hemorrhagic granulation tissue may surround the ablated zone for several weeks following treatment[13]. Follow-up may be done with CT, MRI or positron emission tomography scan, depending on the type, size and location of tumor.

The goal of RFA is usually to ablate 1 cm margin of normal tissue surrounding the tumor on all sides[8,14,15]. This surgical margin is necessary because of the difficulty of accurately determining the extent of the coagulation zone, and because of the possibility of microscopic malignancy surrounding the gross tumor[8]. Exceptions to the 1 cm margin rule may include organs such as the kidney, in which preservation of normal renal parenchyma would be a priority, or when tumor debulking for palliation or relief of neuroendocrine symptoms is the goal of treatment or when surrounding vital structures limit the extent of ablation.

RFA has been shown to be a relatively safe procedure, with mortality between 0.3% and 0.8% and morbidity 2% and 10%[16,17]. Complications include post-procedural pain, post-RFA syndrome with fever and flu-like symptoms that usually resolves within the first 24 h, skin burns from improperly placed grounding pads, thermal injury to adjacent structures, bleeding, secondary infection, and tumor seeding, which can be prevented by cauterization of the needle tract on withdrawal of the probe[8].

The most extensive body of literature on RFA for the treatment of solid tumors involves its use with hepatic malignancies, both primary and metastatic. Currently, RFA is considered a first line treatment modality for local control of hepatocellular carcinoma in patients with Child-Pugh B or higher cirrhosis where resection would have a higher associated mortality. It is indicated in patients with 3 or fewer tumors that are 3 cm or smaller (Milan criteria)[18]. It has recently been shown to be superior to percutaneous ethanol injection with regards to survival and local recurrence[19]. It’s equivalence to surgical resection in patients who satisfy the Milan criteria remains controversial. A prospective randomized trial and a large retrospective analysis comparing local ablative techniques with surgical resection for patients with small solitary tumors, stage T1, found no difference in overall survival between RFA and surgical resection[20-22]. Smaller observational studies have demonstrated similar results[11]. A meta-analysis comparing RFA to hepatic resection in all subsets of patients found improved 3 and 5 year overall and disease free survival and decreased local recurrence in patients who underwent hepatic resection[23]. However, in patients with tumors smaller than 3 cm, the overall survival was comparable. In patients with larger tumors (> 3 cm), the combination of chemoembolization with RFA has been demonstrated to be superior to RFA alone in improving survival[24,25]. This is based on the hypothesis that RFA results in a zone of inflammation that can then be strategically used for targeted delivery of chemotherapeutic agents via chemoembolization.

The majority of the literature regarding hepatic metastases comes from single arm, retrospective or prospective studies evaluating RFA for treatment of unresectable colorectal metastases. In such studies, hepatic resection is superior to both RFA alone or combination of RFA with hepatic resection in regards to local recurrence and overall survival[26]. However, during open resection, additional tumors may be detected on the liver surface or deep metastases may be seen with intra-operative US. These additional lesions can be resected or treated with intra-operative RFA. Randomized control trials directly comparing RFA to hepatic resection for resectable disease have yet to be performed.

There is considerable overlapping variability in the 5 year survival and the local recurrence rates due to differences in definition of local recurrence, inclusion criteria for unresectability, extent of extrahepatic disease, and patient and tumor characteristics between the studies. Local recurrence rates varied between 9% and 40% and 5 year overall survival varied between 18% and 30% (Table 2). The best outcomes were in patients with solitary tumors less than 3 cm and slightly less in patients with 3 or fewer tumors less than 3 cm[27]. Local recurrence was significantly larger in patients with tumors between 3-5 cm[20]. Retrospective studies comparing hepatic resection to RFA for patients who were potentially resectable but poor candidates for surgery due to co-morbidities or refusal, demonstrated decreased local recurrence and improved overall survival with hepatic resection[28]. Therefore it is evident that surgical resection remains the gold standard; but for those who are not candidates for surgery, an alternative such as RFA is valuable.

RFA is increasingly being applied to malignant lung nodules for local control as well as for palliation as its feasibility and efficacy is becoming more established in the literature. Surgical resection remains the gold standard for curative treatment of primary lung cancers and malignant metastasis. However, only about 30% of patients with primary lung cancer are eligible for surgery at the time of diagnosis due to poor functional status and chronic obstructive pulmonary disease[29]. In patients with pulmonary metastasis, multiple lesions and advanced stage usually precludes curative surgical resection.

Currently, there is insufficient evidence to prove that RFA is comparable to surgical resection. There are currently no prospective randomized controlled trials comparing RFA with standard surgical treatment options in patients with malignant lung nodules. Data is limited to case series with differences in number of primary and secondary lung lesions, criteria for unresectability, number of prior resections, history of prior radiation therapy, differences in follow-up protocols, and criteria for determining extent of response to RFA treatment.

However, a small matched case series of 22 patients comparing RFA to resection in patients with stage I non-small cell lung cancer (NSCLC) demonstrated comparable survival in RFA patients at 1, 2, and 5 years[30]. The RAPTURE study, a large prospective multicenter single arm trial, using RFA in patients with early stage NSCLC or lung metastases demonstrated 1 and 2 year overall survival rates of 70% and 48% respectively in patients with primary lung tumors, and 89% and 66% 1 and 2 year overall survival in patients with colorectal metastases. The cancer specific survival was higher in both groups; 92% and 73% at 1 and 2 years in the NSCLC cohort and 91% and 68% in the cohort with colorectal metastases.

The 1, 2, 3 year overall survival for patients with early stage primary lung cancer treated with RFA varies from 70% to 90%, 48% to 84%, 25% to 74%, respectively (Table 3). This is comparable to the 1, 3, 5 year overall survival of patients who undergo lobectomy or segmental resection for early stage lung cancer[31-34]. In most studies that compare outcomes based on size of tumor ablated, patients with tumors smaller than 3 cm had longer median progression free intervals and overall survival[35].

The median procedure related morbidity and mortality are 37.5% and 0% respectively[36]. The majority of complications from thoracic RFA are minor with the most frequently encountered being pneumothorax and pleural effusions (4.5%-61%) and hemoptysis. Others include pain, fever and pneumonia. Despite the high incidence of pneumothorax, only a minority, 11%, require pleural drainage[36]. The incidence of pneumothorax increases as the number of lesions ablated[37].

The role of RFA in breast cancer is still emerging. There is a growing trend towards breast conservation techniques that minimize scarring, breast deformity, and improve overall post procedure cosmesis. Several small single institution studies have established the feasibility of RFA and outlined potential complications (Table 4). In majority of these studies, RFA was followed by lumpectomy or mastectomy, either immediately or in a delayed fashion. The procedure was done under local or general anesthesia depending on whether resection was delayed or followed immediately after RFA, respectively. Response was assessed by pre- and post-procedural MRI which correlated better with pathologic response than US[38]. HE staining, immunohistochemistry with CK 18/8, or nicotinamide adenine dinucleotide-diaphorase cell viability assay were used to assess histopathologic response. There are several studies that have reported HE staining maybe inadequate to assess histopathologic response since it gives a broad spectrum of necrosis and that techniques that assess cell viability are better[38,39]. Complete coagulative necrosis was achieved in 80%-100% of the patients, with skin burn being the most common complication in a very small subset of patients.

Patient selection criteria were strict, including mostly patients with invasive tumors less than 2 cm in size; a few studies had a small portion of patient with non-invasive tumors. The presence of extensive intraductal component was also a relative contraindication to RFA. In addition, estrogen and progesterone receptor status, her 2 status, grade, histology, and need for chemotherapy had to be known prior to RFA since no residual tumor cells would be available post-procedure if 100% successful. Superficial tumors within 1 cm of the skin are a relative contraindication as well, due to increased risk for skin burns. Various strategies to minimize skin burns have been employed in the studies including cooling the breast with sterile ice packs and subcutaneous injection of sterile saline or a high resistance solution to displace the tumor away from the skin. In addition, preoperative chemotherapy is a contraindication since it can lead to an underestimation of tumor size and leave occult foci of residual carcinoma[40].

There are currently no studies comparing RFA to surgical resection, and no long term studies depicting local recurrence rates or survival in patients who receive RFA instead of surgical resection. Very few studies have evaluated RFA as an alternative to surgical resection. Oura et al[41] reported their experience treating 52 patients, with a mean tumor size of 1.3 cm (range 0.5-2.0 cm), with RFA following sentinel node biopsy. There was no local-regional or distant recurrence after an average 15-mo follow-up (range 6-30 mo).

Patient response to RFA has been favorable. Oura et al[41] retrospectively evaluated cosmetic results, which were found to be excellent in 43 patients (83%), good in 6 patients (12%) and fair in 3 patients (6%). The authors found that a major factor leading to poor cosmesis was mass formation at the site of RFA, especially in women with small breasts. This can lead to increased patient anxiety as well.

Progress in the application of RFA for breast tumors is at present hampered by our ability to accurately judge the margin status which is a critical variable in local recurrence rate. Evolution in imaging technology will foster such advancements. Nonetheless, as more breast cancers are being diagnosed at a smaller size, a focused image-guided ablation can minimize destruction of normal breast tissue and thus may positively impact cosmesis.

As with other solid tumors, RFA is increasingly being applied for the therapy for renal tumors as less invasive and nephron-sparing techniques, including partial nephrectomy and laparoscopic nephrectomy, have proven to have comparable 5-year and disease-free survival[42].

Currently, RFA as primary treatment for renal malignancy is limited in study to a select group of patients with early T1a disease or for whom surgical resection is not an option. These include patients with only one kidney, multifocal disease, Von Hippel Lindau, limited renal function, elderly patients or patients with comorbidities that are poor candidates for surgery[8,43-45]. Contraindications include a life expectancy less than one year, the presence of distant metastases, tumors > 5 cm, or tumors in the hilum or central collecting system. Studies have consistently shown 91%-97% complete first ablation success for small (< 3-4 cm), exophytic, peripherally located tumors (Table 5). This is due to the fact that peripherally located tumors are surrounded by peri-renal fat that provides insulation, allowing the high temperatures necessary for successful ablation to be achieved. Conversely, hilar blood flow creates a heat-sink effect making treatment of central tumors more challenging. The recurrence free survival varies from 79%-91% in biopsy proven renal cell cancers, while the 3 and 5 year cancer specific survival ranges from 95%-100% in the few long term studies.

RFA has been long proven efficacious for the treatment of osteoid osteomas. It is performed in patients with typical clinical and radiographic characteristics of an osteoid osteoma (radiolucent nidus surrounded by reactive sclerosis) for treatment of bone pain. It is successful initially in 73%-98% of patients with 92%-100% secondary success rates and majority of patients experiencing pain relief within the first 1-2 wk of treatment[46-49]. Complication rates are minimal with skin necrosis and burns being the most common. It has been demonstrated to be comparable to surgical resection with regards to recurrence[50]. RFA has also been described in case reports for the treatment of other benign bone tumors.

More recently, RFA has been applied as a palliative modality for the treatment of painful metastatic bone lesions. External beam radiation remains the gold standard for treatment of localized bone pain from a metastatic focus. However, 20%-30% of patients don’t respond and are recalcitrant to pharmacotherapy[51,52]. In addition, patients previously irradiated at a recurrent site, may not be eligible for repeat radiation therapy. Ninety percent to ninety-five percent of patients treated with RFA experience a clinically significant reduction in pain that can been seen within the first week of treatment lasting up to 24 wk[52,53]. Complication rates are minimal and can vary from bleeding, pathologic fractures, skin and muscle burns and damage to adjacent neurovascular structures[46].