Longterm survival outcomes of patients undergoing treatment with radiofrequency ablation for hepatocellular carcinoma and metastatic colorectal cancer liver tumors

Abstract

Background

We aim to investigate long-term survival outcomes in patients undergoing radiofrequency ablation (RFA), based on our longitudinal 5 and 10 year follow-up data.

Methods

All patients who underwent RFA for hepatocellular carcinoma (HCC) and colorectal liver metastasis (CLM) between 1999 and 2010.

Results

320 patients were included with oncologic diagnoses of HCC in 122 (38.1%) and CLM in 198 (61.9%). The majority of patients had a single tumor ablation (71% RFA 1 lesion). Minimum 5 year follow-up information was available in 89% patients, with a median follow-up of 115.3 months. In patients with HCC, disease eventually recurred in 73 (64%) patients. In patients with CLM, disease recurrence was ultimately seen in 143 (84.1%) patients. In the HCC group, the 5- and 10-year overall survivals were 38.5% and 23.4%, while in the CLM group, the 5- and 10-year overall survivals were 27.6% and 15%, respectively.

Conclusions

The use of RFA as a part of treatment strategy for primary and metastatic liver tumors imparts 10-year overall survivals of >23% and 15%, respectively. This study indicates that long-term survival is possible with RFA treatment.

Introduction

For patients with liver metastases from colorectal cancer and neuroendocrine tumors, complete surgical resection is currently the only curative option. Colorectal cancer is the third most common cancer worldwide and it is ranked the second most frequent cause of cancer associated mortality in the industrialized countries.¹ Around 50% of colorectal cancer patients eventually develop liver metastases.² Studies have demonstrated that patients with untreated colorectal liver metastatic lesions have 31% survival rate at 1 year, 7.9% at 2 years, 2.6% at 3 years and 0.9% at 4 years.^{3, 4, 5} The addition of chemotherapy regiments improves median survival from 6-12 months–20 months.⁴ In the 1990's, irinotecan (FOLFIRI) and oxaliplatin (FOLFOX)-based regiments further improved response rate up to 50%.^{3, 6} Despite improvement in chemotherapy regiments and multimodality approach in the last decade, patients with metastatic colorectal cancer continue to have a poor prognosis with medial survival of approximately 21 months.³ Therefore, parenchymal sparing surgical resection with negative margins remains the gold standard for operable patients with metastatic colorectal cancer to the liver.

In patients with hepatocellular carcinoma (HCC), liver transplantation provides the best long term overall survival and disease-free survival, as it removes both the tumors and the cirrhotic background liver.⁷ Due to modern advances in liver transplantation, recurrence rate for patients with HCC after liver transplantation is 16%, with a median time from liver transplantation to HCC recurrence of 13 months.⁸ Liver transplantation for HCC, however, is limited by strict eligibility criteria, high costs, and a limited availability of donor organs. The drop out rate for patients with HCC awaiting liver transplantation in contemporary series is 14.5%, which is mainly secondary to tumor progression or hepatic decompensation.⁹ While the results with liver resection and hepatic transplantation are encouraging, approximately 70–80% of patients presenting with HCC and CLM are not resection or transplant candidates due to extent of their disease, presence of vascular invasion, or inadequate hepatic reserve. For the majority of patients with HCC or CLM who are not transplant or resection candidates, a number of treatment options including systemic chemotherapy, hepatic-artery directed modalities, isolated hepatic perfusion, external beam radiation therapy, and ablative techniques are currently available.

An obvious conundrum is that surgery and liver transplantation afford patients with CLM and HCC the best chance for a cure, but the vast majority of these patients are not candidates. Several local ablative treatments which include transarterial chemoembolization, percutaneous ethanol injection, and RFA were then developed. RFA particularly have emerged as a promising adjunct in the treatment of CLM and HCC in the past decade due to its safety, efficacy, and ability to provide more consistent results in local tumor control, especially in patients with limited hepatic reserve.^{10, 11} While there have been several studies evaluating 5-year survival in small cohorts of patients undergoing RFA as part of their treatment strategy for HCC or CLM, there are only limited data on 10-year survival, which is a time interval that we believe is equivocal with cure.^{3, 12, 13, 14, 15} In this study, we sought to determine whether long-term overall survival at 10 years is obtainable in patients, who undergo RFA as part of their treatment for primary and metastatic colorectal hepatic malignancies.

Materials and methods

After Institutional Review Board approval, a single-institution review of all patients (n = 320), who underwent RFA as part of their treatment for hepatocellular carcinoma and colorectal liver metastasis at the University of Pittsburgh Medical Center between 1999 and 2010 was performed. Patients who underwent RFA for treatment of other hepatic lesions were excluded from the study. Data were retrospectively analyzed from a prospective institutional hepatic cancer registry. The primary endpoint of the study are long term 10-year overall survival outcome and tumor recurrence, following RFA for treatment of HCC and CLM.

Treatment planning for patients with hepatic malignancy is carried out under the direction of a multidisciplinary liver tumor board, which is comprised of hepatobiliary surgeons, transplant surgeons, medical oncologists, hepatologists, and hepatobiliary radiologists. The diagnosis of HCC was based on the criteria defined in the practice guidelines of the European Association for the Study of Liver (EASL) or the American Association for the Study of Liver Disease (AASLD). Preoperative scans using either triphasic helical axial computed tomography (CT) or liver magnetic resonance imaging (MRI) were obtained within 1 month before the ablations to allow for the most accurate planning. All patients were initially evaluated for resection, ablation, transplantation, or combination of treatments, based on their extent of disease, tumor characteristics, degree of medical comorbidities, and pre-existing liver reserve. All RFA procedures were performed in the operating room under general anesthesia. Microwave technology was not used. Laparoscopic, open, and percutaneous approaches were all utilized and their applications were at the discretion of the treating surgeon. Laparoscopic ablation was normally performed via three laparoscopic ports under intraoperative ultrasound guidance to ensure proper placement of RFA probe. Tissue biopsy was routinely obtained from the tumors prior to ablation. Patients were normally admitted to the floor postoperatively.

All patients were evaluated in our liver cancer clinic according to imaging studies (CT scan or MRI) and tumor biomarkers [alpha-fetoprotein (AFP) for HCC and carcinoembryonic antigen (CEA) for CLM] before ablation and at intervals of 1–3 month post-ablation. Ablation success was defined according to the appearance of an area equal to or larger than that of the original tumor, without evidence of contrast enhancement. Disease-free interval was defined as an interval between the date of RFA until radiographic documentation of tumor recurrence at the treated RFA site and/or presence of new tumor elsewhere (intra or extrahepatic locations). Overall survival was defined as survival between the date of RFA and the date of death, as determined by medical record review and confirmed with a query of the Social Security Death Index. Statistical analysis was performed using SPSS version 23 for Windows (SPSS, Inc., Chicago, IL, USA). Comparison between 2 groups was performed using T-test. Disease-free and overall survival outcomes were analyzed using the Kaplan–Meier method (log-rank). In all comparisons, a p-value of <.05 was considered statistically significant.

Results

Between 1999 and 2010, 320 patients (laparoscopic: 102, open: 212, percutaneous: 6) underwent RFA for HCC and CLM at the University of Pittsburgh Liver Cancer Center. The mean overall age was 65 (range: 33–92) years, predominantly male (63.4%), and white (89.4%). Oncologic diagnoses included HCC in 122 patients (38.1%) and CLM in 198 patients (61.9%). Demographic characteristics of both groups are summarized in Table 1. RFA alone was performed in 37.8% of all patients, while 62.2% underwent combination of RFA and liver resection. In patients with HCC, vast majority of the radiofrequency ablations were performed for a single hepatic lesion (90.2%) via laparoscopic approach (73.8%). Only 1.6% of HCC patients underwent RFA for 3 or more liver lesions. Simultaneous liver resection was performed in 38.7% of HCC patients, mostly laparoscopically. In these patients subgroup, nonanatomic wedge liver resection was the most commonly performed procedure (75%), followed by anatomic resection of 2 or more segments (20%), and others (5%). Approximately 10% of HCC patients were found to have noncirrhotic background liver. In HCC patients with liver cirrhosis, alcoholism was the most common cause (35%), which is followed by nonalcoholic steatohepatitis (20%), hepatitis B/C (30%), cryptogenic (10%), and others (5%).

Table 1.

Patient demographics

Variables	HCC (n = 122)	CLM (n = 198)
Age (years)	65.7 (range 33–92)	64.7 (range 33–90)
Gender (Male:Female)	84:38	119:79
Race (n)
a. White	98 (80.3%)	188 (94.9%)
b. Black	18 (14.8%)	7 (3.5%)
c. Others	6 (4.9%)	3 (1.6%)
RFA site (n)
a. Single lesion	110 (90.2%)	117 (59.1%)
b. 2 lesions	10 (8.2%)	40 (20.2%)
c. ≥ 3 lesions	2 (1.6%)	41 (20.7%)
RFA approach (n)
a. Laparoscopic	90 (73.8%)	12 (6.1%)
b. Open	30 (24.6%)	182 (91.9%)
c. Percutaneous	2 (1.6%)	4 (2%)
AFP (ng/ml)	36.6 (range 1.7–540)	Not applicable
CEA (ng/ml)	Not applicable	16.1 (range 0.5–636)
Total bilirubin (mg/dL)	1.1 (range 0.2–4.1)	0.6 (range 0.1–2.4)
Albumin (g/dL)	3.7 (range 2.3–5.3)	4 (range 2.1–4.9)
INR	1.2 (range 0.9–1.9)	1.1 (range 0.8–3.3)
Platelet (1000/mm3)	119 (range 13–341)	210 (range 72–703)
BUN (mg/dL)	14.9 (range 5–35)	15.4 (range 3–80)
Creatinine (mg/dL)	1 (range 0.5–3.9)	1 (range 0.5–8.4)
Child-pugh classification
a. Child A	102 (83.6%)	196 (99%)
b. Child B	20 (16.4%)	2 (1%)
Average MELD score	10 (range 6–34)	8 (range 6–31)

In patients with CLM, single tumor ablation was performed in approximately 60% of patients, with most ablations were accomplished using open approach (91.9%). Simultaneous liver resections during RFA were performed in 77.2% of patients with CLM open right hepatic lobectomy (35%), open left hepatic lobectomy (30%), and open nonanatomical wedge liver resections (30%), others (5%).

Minimum 5 year follow-up information was available in 89% patients, with a median follow-up of 115.3 months. Of the 114 patients with HCC in whom follow-up information were obtainable, disease eventually recurred in 73 (64%) patients [RFA site recurrence = 16.7%, new hepatic lesion(s) = 45.6%, extrahepatic recurrence = 7.9%]. A median disease free interval of 13 months was seen in the HCC group. Of the 170 patients with CLM in whom follow-up information were obtainable, disease recurrence was ultimately seen in 143 (84.1%) patients [RFA site recurrence = 7.6%, new hepatic lesion(s) = 58.8%, extrahepatic recurrence = 44.7%]. A slightly shorter median disease free interval of 12.2 months was seen in the CLM group (p-value > .05). A detailed breakdown of the tumor recurrence characteristics for both HCC and CLM is displayed in Table 2.

Table 2.

HCC and CLM recurrence characteristics after RFA

Clinical findings after RFA	HCC (n = 122)	CLM (n = 198)
No evidence of disease	41 (33.6%)	27 (13.65)
Extrahepatic recurrence	7 (5.7%)	32 (16.2%)
New hepatic lesion (s)	47 (38.5%)	59 (29.85)
RFA site recurrence	14 (11.5%)	8 (4%)
RFA site recurrence + Extrahepatic lesion (s)	N/A	3 (1.5%)
RFA site recurrence + New hepatic lesion (s)	3 (2.5%)	N/A
RFA site recurrence + New hepatic lesion (s) + Extrahepatic lesion (s)	2 (1.6%)	2 (1%)
Loss to follow-up	8 (6.6%)	28 (14.1%)

The 10-year disease free survivals for HCC and CLM groups were 35.6% and 16.4%, respectively [(p-value = .001) Graph 1]. In patients with HCC, the 5- and 10-year overall survivals were 38.5% and 23.4%, while in patients with CLM, the 5- and 10-year overall survivals were 27.6% and 15%, respectively [(p-value = .052) Graph 2].

Graph 1.

Disease free survivals after RFA

Graph 2.

Overall survivals after RFA

Discussion

Surgical resection remains the gold standard of curative therapy in patients with liver cancer, both primary and metastatic. Due to the fact that significant percentages of these patients are not resection candidates, alternative treatment modalities play important roles. Advances in tumor biology and technology have led to new treatment options for unresectable liver tumors. National Institute of Health (NIH) guidelines recommend systemic chemotherapy for stage 3 and stage 4 colorectal cancer, whereas in patients in stage 2, administration of chemotherapy should be individualized. According to the American Joint Committee on Cancer, 20% stage 2 and 50% stage 3 patients will progress to have liver metastasis.¹⁶ In addition to regional and systemic chemotherapy, RFA has been used as an alternative to liver resection in recent decade with documented safety and efficacy in primary and metastatic hepatic malignancies. In clinical practice, patient's physical status, tumor size, tumor location, underlying hepatic reserve, and surgeon's expertise influence surgical decision on whether to offer RFA, liver resection, or a combination of both. Despite a randomized trial concluded that liver resection may provide better survival and lower local recurrence than RFA for HCC, data from a North American multicenter study suggested a survival benefit of 2.5 years with RFA for metastatic colorectal cancer.^{17, 18} While there are documented 5-year survival rates with RFA, there are only few studies to date evaluating the long-term efficacy (i.e. 10-year survival).^{7, 13, 14}

Radiofrequency ablation (RFA) utilizes a high-frequency alternating current to generate frictional energy and heat within tissues. When applied to a tumor, the resultant temperature increase produces protein denaturation, cell death, and coagulation necrosis. RFA was initially approved for general tissue ablation by the Food and Drug Administration in 1996 and for the treatment of unresectable hepatic metastases in 2000. Since the first-generation RFA electrodes were unipolar and produced a focused cylinder of ablation, early ablations were limited to tumors smaller than 2.0 cm. As the technology evolved to include multiarray electrodes with multiple deploying tines, ablations fields up to 5 cm can be generated through overlapping and sequential deployments. RFA can be safely and effectively performed via percutaneous, laparoscopic, and open techniques. The decision as to which is the best approach is based on patients functional status, underlying hepatic reserve, history of prior abdominal procedures, concurrent abdominal operations (i.e. liver resections or colorectal resections), and anatomical distribution of tumor burden. Laparoscopic approach in liver ablation offers several advantages when compared to open or percutaneous approaches. Laparoscopic approach affords the ability to identify and treat lesions located at the dome of the liver, peripherally in the liver, or in close proximity to other abdominal organs, while staying minimally invasively. Intraoperative liver examination by laparoscopic ultrasound is another advantage of this approach. Lo et al. reported that in 14% of patients, other liver lesions located in other liver segments were able to be identified during laparoscopy and subsequently treated.¹⁹ This allows a better oncologic staging and a more complete disease treatment. In this study, we used the opportunity during laparoscopic RFA to concomitantly performed laparoscopic liver resection for other liver lesions (found pre-or intraoperatively) in approximately a third of our HCC patients. Ability to ensure liver hemostasis via direct visualization of the liver surface is unique to the surgically delivered RFA. Only less than 2% of patients in our series underwent percutaneous RFA. We only utilize percutaneous approach via subcostal or lower intercostal puncture in patients who have undergone multiple prior liver resections or when significant technical challenges are anticipated. Difficult transabdominal sonographic visualization and absence of access window limit the use of percutaneous technique.

Radiofrequency ablation technology comes with several limitations and specific risks that should be considered prior to application. The liver lesions must be visible by ultrasound and should be away from large blood vessels to avoid the heat sink effect, which may hinder complete tumor ablation. Residual unablated tissue is a relatively common occurrence after laparoscopic RFA in patients with HCC lesions larger than 3 cm in diameter.^{20, 21} Therefore, liver ablation must be cautiously considered for liver tumors larger than 3 cm, due to the higher probability of incomplete ablation. Additionally, care must be taken to avoid thermal injury to the major bile ducts, especially when RFA is used in close proximity to the hepatic hilum. Tohme et al. reported that microsatellite disease next to the primary liver tumor was detected in final pathology (but not on preoperative imaging) in about 16% of patients after liver resections.⁷ This finding suggested that liver resection is superior to the RFA because it allows for removal of tissues adjacent to the tumor that might harbor microsatellite disease or venous tumor thrombi. Even though this theoretically should result in lower frequency of intrahepatic recurrences, their study did not find any significant difference in recurrence between the liver resection and RFA groups.⁷

In our experience, the majority of patients undergoing RFA were diagnosed with hepatic lesions which were deemed unresectable secondary to either extent of disease, underlying liver insufficiency, or presence of significant medical comorbidities. It is our practice to routinely perform intraoperative ultrasound-guided core biopsy of all liver lesions prior to RFA. This provides a diagnosis and allows for accurate 3-dimensional targeting of lesions in real-time, prior to disruption by larger RFA probes. The majority of the patients (61.9%) in this study carried a diagnosis of metastatic colorectal cancer, which is concordant with the fact that the majority of patients referred to our practice have metastatic colon cancer. With increasing efficacy of systemic chemotherapeutic regimens, we are seeing an increasing number of colorectal cancer patients with extensive bilobar disease, in whom complete surgical resection is not feasible. In patients with colorectal metastases, complete surgical resection offers the best chance for cure and that is the primary goal of every treatment, whenever possible. Our increasing experience with patients with extensive unresectable disease has led to the use of RFA as an adjunct to liver resection. A similar approach applies for patients with metastatic neuroendocrine cancer.

As for HCC in the setting of chronic liver disease, all patients within Milan criteria are referred for liver transplant evaluation. Once they are listed for liver transplantation, we utilize RFA as a bridge therapy while waiting for organ availability. In patients with disease outside Milan criteria, we employ RFA with aim to either downstage their disease or as a mean of delaying progression and prolonging survivals. In our current study, 32 (26.2%) of HCC patients were successfully downstaged using RFA and subsequently underwent liver transplantation. At the University of Pittsburgh Liver Cancer Center, approximately 2/3 of patients present with multifocal disease and approximately half of them underwent RFA as an adjunct to surgical resection to clear the liver of unresectable lesions. For unresectable advanced hepatocellular carcinoma, Nexavar^® (Sorafenib) treatment was offered in about 15% of HCC patients.

Local tumor recurrence is defined as reappearance of enhancing tissue within and around the ablation zone, which is secondary to presence of residual viable or unablated tumor in patients previously considered to have a complete ablation. On postoperative imaging after ablations, we evaluate the site(s) of prior RFA for signs of local tumor recurrence, presence of new intrahepatic lesions, and/or extrahepatic metastasis. A reported cumulative rate of tumor recurrence after RFA in patients with HCC is 60–80%.^{22, 23} With a median follow up of nearly 10 years, local RFA site recurrence was found in 15.6% of our HCC patients. In this study, we initially used an RF 2000^® Radiofrequency Generator for a year, before updating our generator to the newer version (higher power) of RF 3000^® Radiofrequency Generator (Boston Scientific, Natick, MA, USA). We did achieve complete roll-off with every ablations. We did not observe any difference in local failure rates between ablations performed using the older and the newer RFA generators. Almost 40% of patients developed new hepatic lesion(s) on their post ablation surveillance imaging, while less than 10% experienced extrahepatic tumor recurrence. Because HCC most commonly arises in livers with background cirrhosis or chronic inflammation, a significant number of patients develops new primary liver cancer after successful ablative treatments. Local tumor recurrence maybe caused by incomplete tumor/tissue ablation, presence of tumor venous invasion, and/or presence of satellite nodules in the adjacent liver parenchyma. Santambrogio et al. reported a 15% rate of local tumor recurrence/progression within 3 years of the RFA treatments. Based on a subset analysis, approximately 35% patients recurred locally within 12 months post ablation.¹² It is often difficult to determine whether the development of a new HCC lesion is due to local recurrence after ablation versus multicentric tumor genesis. HCC recurrence tends to be multicentric, away from the ablation area, which is related to the carcinogenesis potential of hepatocytes affected by chronic inflammation and cirrhosis.

In patients with CLM, local RFA site recurrence was less than 10%. Almost 40% patients, however, developed extrahepatic metastasis, which confirms a more systemic nature of the colorectal cancer spread. Common extrahepatic sites included aortocaval or periportal enlarged nodes, lungs, peritoneum, abdominal wall, and skeletal bones. Lung metastasis varied from a single nodule to the more commonly seen multiple bilateral pulmonary nodules. Due to the colonic mesenteric venous drainage via the liver, half of patients in our CLM group developed new intrahepatic tumor(s). Neoadjuvant chemotherapy was given preoperatively in majority of our CLM patients, nearly 80%. Half of these patients had received chemotherapy prior to their referral to our liver cancer center. Most commonly used regiments are Folfox/Folfiri ± Avastin. CLM patients are typically given two months of neoadjuvant chemotherapy, followed by liver surgery. Last dose of chemotherapy and/or Avastin must be at least one month before any liver operation. Postoperatively, the patients typically receive four more months of chemotherapy. In terms of postoperative follow-up schedule, all of our patients are seen in about 2 weeks postoperatively for their initial post-RFA visit. They are then followed every 3 months with a CT scan triphasic liver/MRI for one year. If there are no signs of tumor recurrence, they are then seen every six months for the next four years. Beyond five years postoperatively without tumor recurrence, the patients are followed annually. In the subgroup of patients had RFA local recurrence, approximately 35% of them underwent another RF reablation. We did not have any patients who required reablation more than once. Transarterial chemoembolization (TACE) with Cisplatin was used in 9 (7.3%) of our HCC patients who unfortunately developed multifocal bilobar intrahepatic tumor recurrence after RFA.

Liver function and alpha fetoprotein level strongly influence survivals in patients with HCC.²⁴ Berber at al reported that patients with a carcinoembryonic antigen (CEA) less than 200 ng/mL has an improved overall survival compared with those with a CEA more than 200.¹⁷ Furthermore, patients with dominant lesion less than 3 cm in diameter had a median survival of 38 versus 34 months for lesions 3–5 cm, and 21 months for lesions greater than 5 cm (P = .03). Survival approached statistical significance for patients with one to three tumors versus more than three tumors (29 v 22 months, P = .09). Gender, age, colon versus rectal primary, nodal status at the time of diagnosis, metachronous versus synchronous, bilobar versus unilobar, pretreatment chemotherapy and presence of extrahepatic disease did not affect survival.¹⁷

Over the past decade, several studies have assessed the safety and efficacy of RFA in HCC and CLM for a number of different primary tumors. Babawale et al. reported 29% overall survival at 5 years after RFA for colorectal liver metastasis, which is comparable to our result of 27.6% overall survival at 5 years.¹⁵ Sorensen et al. and Solbiate et al., however, reported higher a 5 year overall survival of 44% and 47.8%, respectively.^{13, 15, 25} In an European study of 426 patients who underwent RFA for HCC, Santambrogio et al. reported 34% overall survival at 5 years.¹² The median follow up interval in their study was 37.2 months (range 2–193 months). When compared with this European study, our finding of 38.5% five year overall survival in HCC group compares favorably, especially with our significantly longer median follow up of 115.3 months.

Even though several publications have described 10 year experience of using RFA for treatment of colorectal liver metastasis, hepatocellular carcinoma, and neuroendocrine liver metastasis, there are only a few studies with actual 10 year follow-up data.^{26, 27, 28} Most of them are single institution in nature. Zhang et al. reported their 10 year disease-free and overall survivals of 15% and 33% in HCC patients, respectively. Patients enrolled in this study have small ≤3 cm HCC lesions in a background of hepatitis B infection, which is common among the Chinese population. When compared to our HCC patients, a higher 10 year disease free survival (35.6% versus 15%) but lower 10 year overall survival (23.4% vs 33%) were found. Incidence, severity, and associated mortality of liver cirrhosis caused by different hepatitis virus types, as well as dissimilar size of lesions treated, may explain the difference in survival outcomes. Another study with 10 year follow up by Solbiati et al. described an overall 10 year survival of 18% after RFA for colorectal liver metastasis.¹³ A comparable 15% overall survival after RFA in patients with colorectal liver metastasis at 10 year mark was seen in our study.

Conclusions

This study provides convincing data that long-term survival is possible with RFA treatment, especially when combined with hepatic resection for multifocal tumors. This study shows that, patients who undergo RFA as part of their hepatocellular and colorectal metastatic treatments exhibit 10-year overall survival rates of >23% and 15%, respectively. Therefore, selective application of this modality has the potential to improve patient survival in the appropriate setting.

Conflict of interest

None declared.