Single hepatocellular carcinoma ranging from 3 to 5 cm: radiofrequency ablation or resection?

Abstract

Background. The optimal therapy for hepatocellular carcinoma (HCC) is transplantation. For all those patients not eligible for transplantation (or on the waiting list) among the treatments of choice used more frequently in recent years are resection (RES) and radiofrequency ablation (RFA). RFA is less efficacious for HCC ranging over 3 cm. The aim of this study was to compare RFA to RES in a restricted cohort of patients with a single naive HCC ranging from 3 to 5 cm in size and without end-stage liver disease. Patients and methods. A total of 102 patients who had never been treated before were enrolled. Those patients whose HCC position would have required too much parenchymal loss at RES (central or close to main vascular structures) were treated with RFA (n=60), and the others underwent RES (n=42). The two groups were similar for HCC size and liver disease status. The outcome was considered in terms of overall survival (OS) and disease-free survival (DFS) calculated by the Kaplan–Meier method. Differences among groups were validated by log-rank test. Results. The RES group seemed to present a better long-term OS (91%, 57%, and 43% vs 96%, 53%, and 32% at 1, 3, and 5 years, respectively) and DFS (74%, 35%, and 14% vs 68%, 18%, and 0%, respectively) but there was no statistical significance. Age, gender, virus etiology, HCC size and α-fetoprotein levels did not correlate with survival. Patients with recurrence within the first 12 months after treatment showed a worse long-term survival (p=0.011). Patients in Child-Pugh class B had poor prognoses compared with those in class A (p=0.047). Conclusion. Even if RES seemed to promise better long-term results, in the medium term this difference had no statistical significance. Survival in this series was more closely related to the stage of the underlying liver disease than to treatment (RES/RFA).

Introduction

The incidence of hepatocellular carcinoma (HCC) is still rising, making it one of the most common malignancies worldwide ¹. Up to now, surgical resection (RES) has been considered the first choice treatment ²,³,⁴. Orthotopic liver transplantation (OLT) guarantees the best long-term survival and disease-free survival in selected patients because it removes the tumor as well as the cirrhotic liver ⁵,⁶,⁷. Nevertheless, the shortage of organs makes it possible to treat only a small proportion of patients by OLT. Improvements in imaging and technology have promoted the development of percutaneous techniques such as transcatheter arterial embolization (TACE), percutaneous ethanol injection (PEI), percutaneous acetic injection (PAI), cryo-ablation, and radiofrequency ablation (RFA). Some of these techniques are currently less used (PAI, cryo-ablation) while TACE is mainly employed for advanced HCC ⁸. RFA appears to be one of the most efficacious percutaneous techniques for HCC ablation, thanks to its predictable area of induced necrosis and percentage of induced necrotic cells, checked at the histological examination ⁹,¹⁰,¹¹,¹²,¹³. In the recent past RFA was mainly reserved for patients with compromised hepatic function, who were not eligible for surgery. It was also reserved for patients with smaller tumors, since percutaneous ablative procedures were considered less efficacious for bigger HCCs ¹⁴. However, clinical trials have recently demonstrated that RFA shows a similar long-term outcome to resection in the treatment of small HCCs. From the literature it is not clear how comparable the two techniques are for larger HCCs (>3 cm). The aim of this study was to compare the outcome of RFA and RES in patients at first diagnosis of a single HCC, ranging in size from 3 to 5 cm, and with similar liver function (Child Pugh class A or B).

Materials and methods

Patients

From 1999 to 2006, 283 patients underwent RES and 548 underwent RFA for HCC at our institute. In this retrospective study, only patients with a single nodule ranging from 3 to 5 cm, with no previous treatment for HCC, with absence of extrahepatic tumor, and assigned to Child-Pugh class A or B, were included. The criteria for choosing RFA treatment for a Child-Pugh A/B patient were: a localization of the tumor requiring too much parenchymal loss at RES or patient's refusal of surgery. All patients where reviewed by a multidisciplinary team and treated within 1 month from diagnosis. A cohort of 60 patients in the RFA group and 42 in the RES group was collected. The groups were similar in terms of median age (RES: 67 years, range 28–80 years; RFA: 68 years, range 40–85 years), grade of liver disease calculated as Child-Pugh score (RES: 28 A and 14 B; RFA: 44 A and 16 B), and median tumor size (RES: 40 mm, range 30–50 mm; RFA: 36.5 mm, range 30–50 mm). Of the total cohort, 57 patients suffered from hepatitis C infection (HCV), 22 from hepatitis B infection (HBV), 14 from both infections (HCV + HBV), 1 from alcoholic liver disease (ALD), 1 from HCV + ALD, 1 from primary sclerosing cholangitis, and 6 patients had unknown etiology (Table I). Both surgery and RFA were performed by the same team of surgeons.

Table I. Patients’ clinicopathological characteristics.

Characteristic	RFA (n = 60)	Resection (n = 42)	p value
Age (median and range)	68 (42–85)	67 (28–80)	NS (0.595)‡
Gender			NS (0.977)
Male	47	33
Female	13	9
Etiology of cirrhosis
Hepatitis C virus	32	25	NS (0.534)
Hepatitis B virus	12	10	NS (0.645)
Hepatitis B and C virus	11	3	NS (0.066)†
Other causes	5	4	NS (0.266) †
Liver function (Child-Pugh score)			NS (0.467)
Class A	44	28
Class B	16	14
AFP > 200 (ng/ml)	3	5	NS (0.201)†
Median tumor size (mm)	36.5 (30–50)	40 (30–50)	NS (0.149) ‡

NS, not significant; AFP, α-fetoprotein.

χ², †Fisher's exact test, ‡Mann–Whitney.

Diagnosis of HCC and staging

At ultrasound (US; Toshiba Sono Layer v sal 38b, Aloka Pro Sound SSD 5500), all subjects in the study population presented a single lesion ranging from 3 to 5 cm. In cases with evidence of cirrhosis with α-fetoprotein (AFP) > 200 and/or typical CT or MRI features (rapid contrast enhancement and rapid washout) biopsy was not performed ¹⁵,¹⁶, although in the early years of the study many biopsies had been performed even in these patients. All biopsies were US-guided with a 17 gauge needle from HS Medical (Aprilia, Latina, Italy).

In patient selection only the preoperative imaging staging was considered, disregarding postoperative staging (usually more sensitive thanks to the opportunities for peritoneal inspection, palpation, and intraoperative US), as this was available only for patients in the RES group.

Surgery

All procedures were carried out by open surgery: 31 patients underwent an anatomic segmentectomy, 5 a bisegmentectomy, 5 a wedge resection, and 1 a right hepatectomy. Different techniques were employed to perform the parenchymal transection (clamp-crushing, bipolar coagulation, US dissection, or radiofrequency precoagulation).

RFA

Patients were usually admitted the afternoon before the day of the procedure. All procedures were percutaneous and US-guided. A platelet count ≥40 000 and a PT-INR ≤1.4 were required in our unit at the time of treatment. Only four cases were performed under general anesthesia because of poor patient compliance; all the others were treated under sedation and local anesthesia in the presence of an anesthetist and with continuous non-invasive hemodynamic monitoring. Both subcostal and intercostal approaches were used; the anti-Trendelenburg position was often required and in the intercostal approach the operating table was slightly rotated towards the patient's left side. A radiofrequency generator from Radionics (Tyco, Burlington, Massachusetts, USA; now Valleylab) was employed; 31 patients were treated with a single cooled needle (usually with multiple positioning), 29 patients with a triple cooled needle. Bigger nodules often required multiple insertions. In each insertion up to 200 W for 12 min were given, using the automatic impedance-controlled modality. Needles were extracted at a temperature of ≥85°C. No antibiotic prophylaxis was used.

Follow-up

All patients were followed up by three-phase CT scans performed 40 days after treatment and then once every 12 months. Blood tests (including AFP level, bilirubin, albumin, PT-INR, and creatinine) and US were repeated every 3 months. In cases of doubt at US, a CT scan was performed. In cases of increased AFP with negative US and CT a PET scan was required. Intrahepatic HCC recurrence was considered as either residual vital tumor at the treated site (unclear surgical margins, incomplete RFA treatment) or at a distant site from the primary tumor.

Validation of treatment efficacy and statistical analysis

Categorical variables were compared with the χ² test; continuous variables with Mann–Whitney; a p value < 0.05 was considered statistically significant.

Overall survival (OS) was defined as the interval between treatment and death or the most recent follow-up visit. Disease-free survival (DSF) was defined as the interval between treatment and the date of diagnosis of recurrence.

The cumulative OS and DFS curves were plotted by the Kaplan–Meier method. Comparisons between subgroups by treatment, age, tumor size, Child-Pugh class, etiology of background liver disease, serum AFP level, and time of appearance of recurrence were made by the log-rank test. A value of p<0.05 was considered statistically significant.

Complications were classified in five classes: 1, any deviation from the normal postoperative course without the need for drugs or intervention; 2, requiring specific pharmacological treatment; 3, requiring surgical or radiological intervention; 4, life-threatening complications; 5, death ¹⁷.

Results

The mean±SD duration of follow-up was 27±18.7 months in the RFA group and 31.3±24.3 in the RES group.

There was one in-hospital death in the RES group (1/42 = 2.3%) from bleeding and liver failure after a bisegmentectomy (left-lateral bisebmentectomy, Child-Pugh class B). No death occurred in the RFA group (Table II).

Table II.  Early outcome.

Parameter	RFA	Resection	p value
Hospital mortality	0	1	NS (0.411)†
Complications	6	7	NS (0.987)
Median hospital stay	2 (1–16)	5.5 (3–43)	<0.001‡

χ², †Fisher's exact test, ‡Mann-Whitney.

Two patients had a transient liver failure after RFA, none after surgery. One patient developed metastasis in the thoracic wall at the insertion site of the needle after RFA. Complications are summarized and classified in Table III.

Table III. Complications for each method.

Grade	Description	RFA (n=60)	Resection (n=42)	p value
Grade 1	Wound infection	0	1	0.411
Grade 2	Urine infection	0	2	0.167
	Liver failure	2	0	0.343
Grade 3	Hepatic abscess	1	0	0.588
	Bilioma	0	1	0.411
	Pleural effusion	2	1	0.432
	Cutaneous metastasis	1	0	0.588
Grade 4	Renal failure	0	1	0.411
Grade 5	Intra-abdominal bleeding	0	1	0.411
Total		6	7	0.242

Fisher's exact test was used.

Median hospital stay in the RFA group was 2 (1–16) days, vs 5.5 (3–43) days in the RES group (p<0.001).

At the follow-up CT scan, 40 days after treatment, five patients (5.4%) in the RFA group showed contrast enhancement at the treatment site, considered as incomplete necrosis of the HCC nodule; one patient (2.7%) in the RES group showed residual cancer on the surgical section plane.

The 1-, 3- and 5-year OS in the RES group was 91%, 57%, and 43%, respectively, vs 96%, 53%, and 32% in the RFA group, with no statistical difference between the two groups (p = 0.824) (Figure 1).

Figure 1. .

Overall survival:surgical resection (RES) vs radiofrequency ablation (RFA).

The 1-, 3- and 5-year DFS in the RES group was 74%, 35%, and 14%, respectively, vs 68%, 18%, and 0% in the RFA group, with no statistically significant difference (p=0.283) (Figure 2). The variables gender, age, etiology, size of the nodule, and α-fetoprotein >20 were found to have no effect on OS and DFS.

Figure 2. .

Disease-free survival: surgical resection (RES) vs radiofrequency ablation (RFA).

Patients in Child-Pugh class A showed a better long-term OS (p=0.0467) and a similar DFS (p=0.289) compared to those in class B (Figure 3).

Figure 3. .

Child- Pugh class: A vs B.

Recurrence within the first year after treatment was related to a worse OS (p=0.011) (Figure 4).

Figure 4. .

Patients with early (E) (within first year) vs late (L) recurrence: overall survival.

Discussion

Surgical resection (RES) is usually still considered the most effective treatment in patients with resectable HCC and preserved liver function who are not eligible for transplantation or are on the waiting list ²,³,⁴,⁵,⁶,⁷. According the Barcelona Clinic Liver Cancer (BCLC) strategy for staging and treatment, single small HCC should be treated by RES and RFA should be reserved only for multiple nodules (two or three) or single nodules with increased portal pressure or bilirubin levels ⁴,¹⁸. Furthermore, RFA seemed to be less effective for larger nodules (≥3 cm) compared with RES ¹⁸.

On these bases, in the past many institutes performed RFA only in patients in worse conditions. Thus, there is a mean bias when comparing results of the outcome of RES and RFA.

To the authors’ knowledge there has been only one randomized controlled trial comparing RES to RFA ¹⁹. Even if the present study is not randomized, the results could be of some interest because of the restricted inclusion criteria: the two groups (RFA/RES) were similar for Child-Pugh class; the staging, the treatment, and the follow-up were performed in the same institution by the same team; all patients had a single HCC at first diagnosis; and only medium-sized HCC (3–5 cm) were included. The main difference between the two groups is the tumor localization within the liver, since in the RFA group HCCs were usually situated centrally, and often closer to main vessels. This could have a role when considering the outcomes because it could be associated with a higher vascular invasion, and consequently lead to a worse prognosis for RFA patients ²⁰.

The two populations were similarly matched for the relevant prognostic factors such as number (n = 1) and mean size of the nodes, Child-Pugh score, AFP levels, and etiology of the underlying liver disease.

Both treatments in this series were confirmed to be safe, with only one death occurring in the RES group. Also the complications rate was not different between RES and RFA and was similar to the rates reported by other authors ¹⁹,²⁰,²¹,²². Important prognostic factors were confirmed to be the grade of the underlying liver disease (Child-Pugh class, p=0.047) and the time to recurrence (p=0.011). Even if there was a trend towards a worse DFS after RFA compared with RES, and even if the 5-year survival was also lower in the RFA group (32% vs 43%), there was no statistically significant difference. Comparing data from the present study to those from five other studies on the same topic ¹⁹,²³,²⁴,²⁵,²⁶ (Table IV), the following observations could be made. All papers come from two distinct geographical areas: Italy and the south-east of Asia. Absolute values in terms of OS and DFS are better for patients from Asia. This difference could be explained because patients enrolled in studies from Asia are younger and have better hepatic function, since all are in Child-Pugh class A, indirectly confirming that liver function is one of the most important prognostic factors. The shapes of the OS and DFS curves are similar in most of the studies, confirming a tendency to a better long-term survival after RES. Nevertheless, a statistically significant difference in OS was found in only one paper ²³, but in this paper there was a major selection bias since patients enrolled in the RFA group had more nodules (p=0.0001) and a worse Child-Pugh class (p = 0.0001). In all other comparative studies on this topic ¹⁹,²⁴,²⁵,²⁶, and in the present one, no statistical difference in OS was found between RES and RFA. Perhaps by prolonging the follow-up of the studies this difference might become evident. Perhaps a difference in DFS, that was actually found only in two comparative studies, could be even more evident ²³,²⁶. At present it seems reasonable to conclude that RFA guarantees nearly similar medium-term OS when compared to RES in patients suffering from a 3–5 cm single HCC, but at a lower cost (shorter hospital stay) and less invasiveness for patients. Patients affected by HCC will undergo several treatments during their lives because of a recurrence rate exceeding 70% in 5 years ²⁷,²⁸,²⁹,³⁰; for this reason minimally invasive procedures such as RFA are desirable. Actually, there is also an increasing tendency towards using combinations of therapies (e.g. chemoembolization, portal vein embolization, percutaneous ablation, radiotherapy, etc), as treatment or as neoadjuvant therapies followed by surgical resection, showing promising results ³¹,³².

Table IV. Comparative studies (RES vs RFA) for HCC < 5 cm.

Author	Setting units	Treatment	No. of patients	Mean age	Child-Pugh class A/B	No. of nodes 1/> 1	Node size range (cm)	Mean follow-up	% 1,3-years OS	% 1,3-years DFS	Difference OS (p value)	Difference DFS (p value)
Vivarelli et al. 2004 23	2	RES	79	65	70/9*	66/13*	–	28.9	83, 65	79, 50
Italy		RFA	79	67	43/36*	46/36*	–	15.6	78, 33	60, 20	0.002*	0.001*
Cho et al. 2005 24	1	RES	61	57	61/0	–	<5	21.9	98, 77	71, 36
Korea		RFA	99	58	98/0	–	<5	23	95, 80	36, 30	0.77	0.65
Hong et al. 2005 25	1	RES	93	49*	93/0	93/0	<4	25.5‡	97,.83	75, 54
Korea		RFA	55	59*	55/0	55/0	<4	22.7‡	100, 72	74, 40	0.24	054
Montorsi et al. 2005 26	1	RES	40	67	36/4	40/0	<5	22.4	84, 73	84, 59
Italy		RFA†	58	67	40/18	58/0	<5	25.7	85, 61	67, 32	0.139	0.024*
Chen MS & coll. 2006 19 (RCT)	1	RES	71	49.4	71/0	71/0	≤5	29.2	93, 73	86, 69
China		RFA	90	51.9	90/0	90/0	≤5	27.9	95, 71	85, 64	>0.05	>0.05
Present study 2007	1	RES	42	67	28/14	42/0	3 ≥ 5	31.3	91, 57	74, 35
Italy		RFA	60	68	44/16	60/0	3 ≥ 5	27	96, 53	68, 18	0.824	0.283

RES, resection; RFA, radiofrequency ablation; RCT, randomized controlled trial; OS, overall survival; DFS, disease-free survival.

*Statistical significance (p < 0.05).

†Laparoscopic RFA.

‡Median.

Survival with HCC seems to be most strictly related to the underlying liver disease ³³. This is the main reason for assessing whether there is still a real advantage of surgical resection over less invasive treatments, which may be true only for selected patients.