Ultrasound-guided percutaneous ethanol injection under general anesthesia for the treatment of hepatocellular carcinoma on cirrhosis: long-term results in 268 patients

European Journal of Ultrasound 12 (2000) 145 – 154 www.elsevier.com/locate/ejultrasou

Clinical Science: Original Paper

Ultrasound-guided percutaneous ethanol injection under general anesthesia for the treatment of hepatocellular carcinoma on cirrhosis: long-term results in 268 patients Antonio Giorgio a,*, Luciano Tarantino a, Giorgio de Stefano a, Anna Perrotta a, Vincenza Aloisio a, Luca del Viscovo b, Alfredo Alaia c, Gennaro Lettieri a a

Inter6entional Ultrasound Ser6ice, ‘D. Cotugno’ Hospital, Via Quagliariello 54, 80131 Naples, Italy b Department of Radiology, 2nd Uni6ersity of Naples, Naples, Italy c Department of Anesthesiology, ‘D. Cotugno’ Hospital, Via Quagliariello 54, 80131 Naples, Italy Received 29 June 2000; received in revised form 29 August 2000; accepted 5 September 2000

Abstract Objecti6e: Percutaneous ethanol injection (PEI) under general anesthesia (One-shot PEI) is a therapy for large and multiple hepatocellular carcinoma (HCC) by the injection of a large amount of ethanol into the tumor. We report our results with 5-year survival rates in patients with HCC on cirrhosis treated with One-shot PEI. Patients and methods: From October 1992 to March 1998, 268 cirrhotic patients (age 42 – 82 years; 191 males; 95 Child-Pugh’s A class, 150 B and 23 C class of cirrhosis) with 515 HCC nodules underwent One-shot PEI. Diameter of HCC nodules ranged from 0.6 to 14 cm (mean 5.02 9 2.2 cm; median: 4 cm). One hundred and thirty-eight patients had a single nodule (range 3.2–14 cm; mean 5.692.1 cm), 130 had multiple nodules, up to six nodules (mean 2.9 nodules) (range 0.6–11 cm; mean 4.89 2.1 cm) Results: CT showed complete necrosis in 357/506 nodules (70%). Five patients (1.8%) with nine nodules died as a result of the procedure (variceal bleeding in three cases, liver failure in one and hemoperitoneum in one). The overall survival rates were 93, 83, 74, 65 and 59% at 1, 2, 3, 4 and 5 years, respectively. Survival rates were 90, 84, 82 and 82% at 12, 24, 36 and 48 months, respectively, in patients with a single nodule 5 5 cm, and 97, 71, 59, 59 and 59% at 12, 24, 36, 48 and 60 months, respectively, in patients with single nodule \ 5 cm. Patients with multiple nodules had survival rates of 97, 89, 75, 60 and 60% at 12, 24, 36, 48 and 60 months, respectively. Conclusion: PEI of large and multiple HCC showed survivals similar to conventional PEI for patients with smaller tumors. © 2000 Elsevier Science Ireland Ltd. All rights reserved.

* Corresponding author. Present address: Viale Colli Aminei 491, 80131 Naples, Italy. Tel.: +39-081-7435191; fax: + 39-0815908278. E-mail address: [email protected] (A. Giorgio). 0929-8266/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 9 2 9 - 8 2 6 6 ( 0 0 ) 0 0 1 1 3 - 0

146

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

Keywords: Liver neoplasm; Cirrhosis; Hepatocellular carcinoma; Ultrasound; Interventional procedure; Percutaneous ablation; Ethanol injection

1. Introduction Percutaneous ethanol injection (PEI) under ultrasonographic (US) guidance is now widely used in the treatment of Hepatocellular Carcinoma (HCC) on cirrhosis (Livraghi et al., 1986; Seky et al., 1989; Ebara et al., 1990; Giorgio et al., 1992; Tanikawa, 1992; Vilana et al., 1992; Shiina et al., 1993; Kotoh et al., 1994; Ondera et al., 1994; Livraghi et al., 1995a,b). This therapeutic modality has gained great popularity because of its safety, simplicity, cheapness and repeatability (Livraghi et al., 1995b; Di Stasi et al., 1997). The best results are obtained on HCC nodules up to 5 cm in diameter in which pathological studies have shown up to 100% of necrosis (Shiina et al., 1991, 1993; Troisi et al., 1998; Veltri et al., 1998). PEI is performed in multiple sessions and usually in the outpatient clinic. Nevertheless, there are two main draw-backs with this ‘conventional’ procedure: pain (especially when treated nodules are superficially located) and the high number of sessions in the treatment of nodules \ 2 cm. Therefore, in 1993 Livraghi and co-workers proposed a new technique, called ‘One-Shot PEI’, for the treatment of large or multiple nodules by injecting large amounts of ethanol into the tumor, in a ‘single session’ with the patient under general anesthesia (Livraghi et al., 1993). In a previous study, we confirmed that PEI under general anesthesia (PEI u.g.a.) is an effective and feasible procedure, even if more aggressive than conventional PEI (Giorgio et al., 1996). In 1998, we reported the 3-year survival in the first 112 patients (Giorgio et al., 1998). In this study we report the 5-year results of HCC treatment with PEI u.g.a. in a larger series of cirrhotic patients.

2. Materials and methods

2.1. Patient selection A Specialized Service for US guided interventional procedures has been operating at our Institution since 1978 and we have performed percutaneous treatment of liver tumors since January 1987. From October 1992 to March 1998, 306 patients with HCC on cirrhosis, coming from many institutions in the Naples area as well as other areas of southern Italy, were referred to our Interventional US Service in order to be treated with PEI u.g.a. Most of the patients (144 cases = 47%) had both the diagnosis of liver cirrhosis and of focal liver lesion at the first US examination which had been performed because clinical and laboratory data suggested a diffuse chronic liver disease. A low number (68 cases= 22%) of the patients were aware of the chronic liver disease and regularly underwent US and AFP screening, while 94/306 patients (31%) were irregularly controlled with US for the underlying cirrhosis without following any screening schedule. Two hundred and sixty-eight patients were admitted to the procedure, while 38 patients were advised to have other treatments (conventional PEI, surgery, trans-arterial chemoembolization, hormonal treatment). The patients were admitted to PEI u.g.a. when one or more of the following criteria were present (Giorgio et al., 1996): 1. Refusal of surgery (157 patients). 2. Difficult surgical approach (e.g. nodules located near to major vascular structures or postero-superior subdiaphragmatic segments) (87 patients). 3. Presence of a single nodule \ 3 cm (138 patients); 4. Presence of two or more nodules (up to six) with at least one \3 cm (130 patients).

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154 Table 1 Clinical features of the patients in our series Range Age (years) No. of nodules Diam. of nodules (cm) Volume of nodules (cc) Serum AFP (ng/ml)

42–82 1–6 1.6–14 0.12–480 1.2–16 000 Patients

Child-Pugh classification A B C

95 150 23

Etiology of the cirrhosis HbsAg+ HCVAb+ HbsAg+–HCVAb+ Cryptogenic

29 196 35 8

Mean 63 2.9 5.02 52 Mean Pugh score 5.8 7.6 10.2

In all treated cases there were overlapping results of US and CT. The presence of ascites (mild – moderate in 23 patients) or partial neoplastic thrombosis of a single main branch of portal vein (less than 1.5 cm in length for left portal vein in seven patients, and less than 2.5 cm for right portal vein in 11 patients) or of a segmental portal vessel (13 patients) were not considered absolute contraindications for procedure. All patients underwent abdominal CT and chest X-ray that did not show extrahepatic spread of the tumor. The 268 patients treated with PEI u.g.a. (191 males) had 515 HCC nodules. Clinical and US data, as well as, etiology and stage of the underlying cirrhosis in all patients are reported in Table 1. Diagnosis was established with fine needle biopsy under sonographic guidance in 233 patients. Thirty-five patients had serum AFP\ 400 ng/ml and hypervascular hepatic nodules in early arterial phase at enhanced imaging techniques (CT in 28 patients, MR in five and angiography in two). AFP levels (range 1.2 – 16 000 ng/ml) were B 20 ng/ml in 35 patients, between 21 and 200 ng/ml in 179 and \ 200 ng/ml in the others.

147

2.2. Procedure Before the procedure, all patients underwent Dynamic or Spiral CT with contrast medium (Fig. 1A), ECG, chest X-ray and routine blood tests. The range and the mean, respectively, of the main laboratory data of patients before treatment are reported in Table 2. All patients had undergone Oesophago-Gastroduodenoscopy from 1 week to 6 months before procedure. Abdominal US examination was performed the day before and the day after the procedure. Informed consent was obtained from all patients. PEI u.g.a. was carried out as reported elsewhere (Giorgio et al., 1996). All procedures required the presence of the same operator (A.G.) together with a physician with experience in abdominal US (L.T., G.d.S., N.M.), an anesthesiologist and a nurse. General anesthesia, without endotracheal intubation, was delivered as follows: premedication with atropine 0.5 mg i.v. and induction with propofol 9–12 mg/kg per h plus fentanyl 50 mg i.v. Propofol was chosen because of its rapid hepatic elimination and lack of relevant side effects. Under continuous real-time ultrasound guidance using a 3.5 MHz convex probe, a 22-gauge, 20-cm long needle (Ecoject, HS, Tokyo, Japan) was inserted into the deepest portion of the nodule (Fig. 1B) and 95% sterile ethanol was slowly injected into the lesion. Real-time US scanning during the injection allowed the evaluation of the intranodular diffusion of the ethanol assessed by the characteristically intense hyperechogenicity induced by the alcohol at the site of the injection (Fig. 1C). The needle was then partially withdrawn and the more superficial parts of the nodule were injected. After that, the needle was completely withdrawn, and another puncture with a new needle was performed to inject other parts of the lesion that had not yet been perfused by the ethanol (Fig. 1C,D). Multiple punctures (1–8 per nodule; median: 3) were carried out until the nodule appeared homogeneously hyperechoic (Fig. 1E). We considered this appearance the US sign of good ethanol distribution throughout the liver lesion. The volume of the nodule, calculated by the US equipment with the ellipsoid method (Giorgio et

148

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

Fig. 1.

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

149

Table 2 Main laboratory tests of 268 patients submitted to One Shot PEI before and after procedure and mean time of recoverya Test

ALT Total bilirubin PT% Creatininaemia

Normal values

5–60 0.2–1.2 70–100 0.6–1.4

Before treatment

After treatment

Range

Mean

9S.D.

Variation

TR

37–241 0.6–5.9 42–100 0.6–2.3

106 1.6 67.4 1.4

57 1.4 16.4 0.6

123 952 410 9 28 −159 9 46 924

12.1 9 4.4 12.8 9 5.5 7.1 92.8 3.5 9 1.2

a

PT, protrombin time, percentage of activity; S.D., standard deviation; variation, percentage of increase or decrease in values (mean 9S.D.); TR, time of recovery (in days): mean 9S.D.

al., 1996) was considered the scheduled volume of alcohol to be injected. However, the procedure was always terminated when all portions of the nodule had been apparently perfused by ethanol through the multiple punctures, so that, for nodules \2 cm, the injected ethanol was never greater than the volume of the nodule in all cases (50–90% of the volume of the nodule). For very small nodules, the amount of injected ethanol was sometimes higher than the volume of the nodule because some of the ethanol escaped from the lesion during the injection. As our experience increased and more procedures were performed, we learned that major complications occurred when a large amount of ethanol (exceeding 60 ml) was injected (see Section 3.4). Therefore, from January 1995, we decided to schedule two or more sessions, under general anesthesia performed every 2 weeks, when treating large nodules (\ 8 cm). In the first session, we injected a maximum of 60 ml of ethanol and the perfused areas of the nodule were recorded on tape. Before the start of the subsequent session, the nodules were evaluated by US to assess the changes in echogenicity of part or whole of the nodule due to the effect of the ethanol injected in the previous PEI sessions.

In the subsequent sessions, the ethanol was injected until all parts of the tumor had been perfused. Usually patients were discharged from the hospital the day after the procedure.

2.3. E6aluation of therapeutic efficacy of the PEI u.g.a. The evaluation of the efficacy of the treatment was performed by: (1) CT with i.v. contrast medium carried out from 1 to 4 weeks following the end of the procedure (Livraghi et al., 1993; Giorgio et al., 1996); (2) follow-up by US of the size and echogenicity of the treated lesions and (3) analysis of the patients’ survivals. One hundred and sixty-seven patients underwent Dynamic ‘single slice’ conventional CT (CT Pace, General Electrics, Milwaukee, WI) and 101 patients underwent three-phase Spiral CT (Synergy Power, General Electric, Milwaukee, WI). Tumor necrosis was considered complete when CT scans showed no areas of enhancement within the lesion in the arterial phase (Fig. 1F), otherwise the response was rated as incomplete (Livraghi et al., 1993; Giorgio et al., 1996).

Fig. 1. (A) Enhanced CT scan showing a large hepatocellular carcinoma (arrows) involving V, VI, VII and VIII segment of the liver, before Percutaneous Ethanol Injection under general anesthesia. (B – E) US scan of the tumor during the treatment. (B) The needle (arrowheads) is inserted into the deep medial portion of the tumor (T) and injection of ethanol is started. The injected ethanol can be seen as a hyperechoic area in the nodule (arrows). (C) After withdrawing the needle, a second needle (N) is inserted into another portion of the nodule (arrows), not yet perfused by the ethanol. (D) Then, ethanol (arrows) is injected again into the deep and superficial portion of the nodule along the needle path. (E) After treatment the nodule appears homogeneously hyperechoic, the US sign of good ethanol distribution throughout the liver lesion. (F) Post-treatment dynamic CT scan of the tumor shows the almost complete necrosis of the large lesion.

150

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

All patients were followed-up every 2 months with clinical examination, and abdominal US as well as with liver function tests and AFP for the first year and every 3 months thereafter. US studies were carried out in order to evaluate changes in treated lesion volume and to detect new nodules (recurrence) and portal thrombosis. If US demonstrated recurrence, a CT exam was performed. Survival rates were calculated, for 263 patients, by using the Kaplan – Meier method and the generalized Wilcoxon test was used for the statistical analysis (Kaplan and Meyer, 1958; Gehan, 1965).

3. Results

3.1. Outcome of the treatment (short-term e6aluation) Injected ethanol ranged between 16 and 120 ml per session (mean: 32 cc). The total amount of ethanol injected per patient ranged between 16 and 180 ml (mean: 48 ml). The injected ethanol per nodule ranged from 16 to 180 ml (mean: 39 ml). Two hundred and forty-five patients underwent one session under general anesthesia; 19 had two sessions and four had three sessions. The number of needle punctures ranged from 2 to 12 (mean: 5). The duration of PEI u.g.a. was never more than 50 min (range 15 – 50 min). Five hundred and six out of 515 nodules underwent CT examinations after completion of procedure; 357/506 (70%) HCC lesions were completely necrotic on imaging. Five patients with nine nodules died as a result of the procedure (see later) and did not undergo CT exam. On post-procedure CT examination, necrosis was complete in 92% of nodules of up to 6 cm; 81% of nodules \6 and B 8 cm showed no enhanced areas in arteriographic phase and 60% of nodules \ 8 cm appeared completely necrotic. One hundred and thirty-four out 149 of the incompletely necrotic nodules at post-treatment CT, were successfully retreated by conventional PEI or PEI u.g.a. (mean 2.8 days).

3.2. US follow-up (medium-term e6aluation of the efficacy) During the follow-up, in 15 cases (5.7%) local recurrences were observed at US. In these patients the pre-treatment diameter of the nodules (all completely necrotic at post-treatment CT) ranged from 3 to 7 cm and recurrence was not apparently related to the size of the nodule. In all the remaining cases the treated lesions shrunk or rarely disappeared and showed changes of pattern (hypoechogenicity and/or calcifications) at US follow-up examination. One hundred and fifty-three (58%) patients showed hepatic recurrences (e.g. presence of new neoplastic nodules in areas distant from treated ones). In 102 cases there were new single lesions; in the remaining cases, two or more nodules were found. Recurrences were treated with conventional PEI in 107 cases, PEI u.g.a. in 33 cases and Trans-Arterial-Chemo-Embolization in 10 cases.

3.3. Patients’ sur6i6al (long-term e6aluation of the efficacy) The follow-up ranged between 6 and 72 months (mean 29.4 918.2; median 24 months). During the study, 18 patients dropped out of the followup after a mean period of 34.3 months. All the remaining patients were regularly in follow-up at the end of the study. All survival rates with confidence intervals are reported in Table 3. Figs. 2 and 3 show 5-year survival curve in all patients and cumulative survival curves stratified according to Child Classes, respectively. The overall survival rates were 93, 83, 74, 65 and 59% at 1, 2, 3, 4, and 5 years, respectively (Fig. 2). According to Child-Pugh’s classification (Fig. 3), survival rates were 98, 88, 79, 70 and 70% in Child-Pugh A patients at 1, 2, 3, 4 and 5 years, respectively; 94, 85, 76, 67, and 54% at 1, 2, 3, 4,and 5 years, respectively, in Child-Pugh B patients; 60 and 24% at 1 and 2 years, respectively in Child-Pugh C patients. The difference between survival rates of patients in Child-Pugh A and B classes compared with patients in Child-Pugh C

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

151

Table 3 Survival rates of our series according to Child-Pugh’s Classes and size of the tumorsa Child-Pugh’s Class

1 year

2 years

3 years

4 years

5 years

SR

(CI)

SR

(CI)

SR

(CI)

SR

A B C

98 94 60

(100–94) (98–90) (95–25)

88 85 24

(99–77) (92–78) (84–0)

79 76 –

(96–62) (87–65)

70 67 –

(94–46) (87–47)

70 54 –

Single nodule B5 cm Single nodule \5 cm Multiple nodules

90 97 97

(97–83) (100–91) (100–94)

84 71 89

(95–73) (92–50) (96–82)

82 59 75

(98–66) (92–26) (87–63)

82 59 60

(100–51) (99–19) (79–41)

– 59 60

(100–10) (83–37)

Overall

93

(97–89)

83

(90–76)

74

(84–64)

65

(81–49)

59

(81–37)

a

(CI)

SR

(CI) 96–44) (87–21)

SR, survival rate (percentage); CI, 95% confidence interval (percentage).

was statistically significant (P B0.01). Differences between Child-Pugh A and Child-Pugh B patients were not statistically significant. According to the size and number of nodules, survival rate was 90, 84, 82, and 82% at 12, 24, 36 and 48 months, respectively, in patients with a single nodule of up to 5 cm and 97, 71, 59, 59 and 59% at 12, 24, 36, 48 and 60 months, respectively, in patients with single nodule \ 5 cm. Patients with multiple nodules had survival rates of 97, 89, 75, 60, and 60% at 12, 24, 36, 48 and 60 months, respectively.

3.4. Complications No complications occurred related to general anesthesia. Table 2 reports the variations of laboratory tests values after procedure and the mean time of recovery. Five patients (1.8%) died between 7 h and 10 days after the procedure. One patient with four nodules (diameter range 1.7 – 6 cm; total injected ethanol: 90 ml) died from hemoperitoneum; three patients (all with single nodule [range 6.5–8 cm]) died from haemorrhage from rupture of oesophageal varices (injected ethanol range 70–90 ml); one patient (single nodule, diameter 7 cm — injected ethanol 70 ml) died from acute liver failure. All the deaths occurred before January 1995. Other major complications were the following: two cases of hemoperitoneum recovered with-

out blood transfusions; two patients had acute tubular necrosis, and recovered within 1 week with medical therapy, while eight patients who showed decompensation of liver cirrhosis recovered with normal medical therapy. After treatment, US showed the presence of mild–moderate ascites in 22 patients without US detectable ascites before the procedure and an increase in amount of fluid in 11 out of 23 patients who had ascites before procedure. In all of them, ascites recovered within 7–11 days with medical treatment. Other minor complications were pain and/or fever which occurred in the majority of patients and were treated with common analgesics and paracetamol, respectively. In three cases, abscess formation in the treated nodules was observed 7–12 months after procedure. All patients were treated with US-guided percutaneous needle aspiration and in one case no

Fig. 2. Overall 5-year survival curve of our series.

152

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

Fig. 3. Five-year survival curve of our series according to Child-Pugh’s class.

focal liver lesion was detectable 4 years later. The hospital stay ranged from 2 to 19 days.

4. Discussion Hepatocellular carcinoma is one of the world’s most common malignancies (Sherman, 1995). Generally, prognosis is poor because of the disease’s highly malignant nature and because it is often associated with cirrhosis (Sherman, 1995). Before 1986, liver resection was considered the treatment of choice and Trans Arterial Chemo Embolization was employed in cases of unresectable tumor. The introduction of PEI in 1986 by Livraghi et al. (1986) introduced a new therapeutic tool for HCC. To shorten the long treatment times of multisession conventional PEI, Livraghi et al. (1993) proposed the ‘single session PEI’ with the patient in general anesthesia in order to inject in a single procedure as much ethanol as was needed to obtain complete necrosis of the lesion. Radiofrequency (RF) thermal ablation is a new, promising technique for percutaneous treatment of HCC, and good results have been reported for the treatment of small HCC (Livraghi et al., 1999) with this technique. Although at present there are no controlled, prospective, randomized trials, long-term results with PEI of HCC nodules up to 5 cm in diameter are similar, if not better, than surgery (Shiina et al., 1993; Livraghi et al., 1995a,b; Ryu et al., 1997; Colella et al., 1998). In the last few years, increasing AFP and US screening in cirrhotic

patients has allowed the detection of nodules of small dimensions. However, in spite of US and AFP screening large or multiple nodules are not infrequently discovered at first presentation. For this reason it is also necessary to treat this type of tumor with minimally invasive therapy, as it is certainly not eligible for surgery nor for liver transplantation. Although this latter therapeutic modality has given the best results in terms of survival (Colella et al., 1998), up to now there have been very strict criteria for admission and organs are not available for the great number of patients who need liver transplantation (Philosophe et al., 1998). Therefore HCC cirrhotics who are not good candidates for liver transplantation have many different treatment options: surgery, Trans-Arterial-Chemo-Embolization, conventional PEI, ‘single- or multiple-session’ PEI u.g.a. and RF. In our study the 1-, 3- and 5-year overall survival rates were 94, 75 and 65%, respectively. These results are similar to those reported with surgical resection for large tumors (Nagasue et al., 1993; Sugioka et al., 1993; Izumi et al., 1994; Kawarada et al., 1994; Kim and Kim, 1994; The Liver Cancer Study Group of Japan, 1994). Moreover they appear similar to, if not better, than two recent Western and Eastern surgical experiences. In fact 3- and 5-year survival rates of the patients who underwent hepatic resection were 57.5 and 40.8%, respectively in the last report of The Liver Cancer Study Group of Japan (1994). In the Italian series of Mazziotti et al. (1998) the 5-year survival rate was 41.3%. Furthermore, our data show that by using PEI u.g.a., patients with large or multiple nodules have similar survival when compared with patients having lesions of smaller diameter treated with conventional PEI. In fact in a recent Italian multicenter study on 746 HCC cirrhotics treated with conventional PEI with multiple sessions technique, survival rates were 98, 79 and 47% at 1, 3 and 5 years, respectively, in Child A patients and 98, 63 and 29% at 1, 3 and 5 years, respectively, in patients in Child B class (Livraghi et al., 1995a,b). In our series, 72% of nodules were necrotic on imaging and most of the remaining nodules with incomplete necrosis were immediately treated with

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

conventional or PEI u.g.a. Consequently, this procedure is virtually able to destroy all the tumor mass or, at any rate, considerably slow down the neoplastic growth and therefore probably improves survival. In the last few years, some studies combining PEI and Trans-Arterial-Chemo-Embolization have reported increased survival (Allgaier et al., 1998; Lencioni et al., 1998). Generally, it is reported that nodules B3 cm can be successfully treated by PEI and in nodules \ 3 cm Trans-Arterial-Chemo-Embolization must be combined with PEI (Lencioni et al., 1998). Although comparing different series of cirrhotics is always difficult, our results clearly indicate that PEI u.g.a. alone is able to destroy the majority of nodules of up to 6 cm in diameter thereby avoiding adding another invasive procedure. PEI u.g.a. is more aggressive than traditional PEI. In this study, mortality was 1.8%, compared to 0.1% of the recent Italian survey which collected 1066 cirrhotic patients (Di Stasi et al., 1997). Certainly, the variceal bleeding in the three patients who died is to be attributed to the sudden increase in the portal pressure after the injection of large quantities of ethanol. In fact ethanol acts also inducing thrombosis of the tumor vessels, but it can also induce thrombosis of the small vessels in the adjacent segments of the liver, with a subsequent increase in portal pressure. Thrombosis of intrahepatic small portal and hepatic vessels could have probably determined decreased blood flow and liver anoxia in the patient who died from acute liver failure. Nevertheless, a case of death for massive necrosis of the liver has been reported by Taavitsainen et al. (1993) after injection of only a few milliliters of ethanol and two deaths from hemoperitoneum after conventional PEI have been reported in other series (Shiina et al., 1993; Di Stasi et al., 1997). Our series is primarily composed of patients with cryptogenic and post-viral cirrhosis. Due to the geographic area where our patients come from, this is only an initial study on the treatment of HCC on alcoholic cirrhosis. However, we think that the main issue that can influence the outcome of the treatment is the compensation of the cirrhosis rather than the etiology.

153

Therefore, in the light of our experience, the following conclusions can be reasonably made: 1. PEI u.g.a. avoids multiple treatments over a long period of time associated with conventional PEI; 2. PEI u.g.a. seems effective in inducing necrosis of large and multiple HCCs; 3. The best results are obtained in patients with a single nodule, 3–6 cm in diameter; 4. Survival rates of patients treated with PEI u.g.a. are similar to those of patients treated with conventional PEI for tumors of smaller dimension; 5. PEI u.g.a. is more aggressive than traditional PEI: therefore patients with Child-Pugh C cirrhosis and/or oesophageal varices at risk of bleeding should be excluded; 6. Because of dose-dependent side effects, in case of large nodules, it is preferable to carry out a further session under general anesthesia to avoid major complications, injecting no more than 60 ml of ethanol per session; 7. Before performing PEI u.g.a., experience with conventional PEI is necessary.

References Allgaier HP, Deibert P, Olschewski M, et al. Survival benefit of patients with inoperable hepatocellular carcinoma treated by a combination of transarterial chemoembolization and percutaneous ethanol injection — a single center analysis including 132 patients. Int J Cancer 1998;18:601– 5. Colella G, Bottelli R, De Carlis L, et al. Hepatocellular carcinoma: comparison between liver transplantation, resective surgery, ethanol injection, and chemoembolization. Transpl Int 1998;11(Suppl.1):S193– 6. Di Stasi M, Buscarini L, Livraghi T, et al. Percutaneous ethanol injection in the treatment of hepatocellular carcinoma. Scand J Gastroenterol 1997;32:1168 – 73. Ebara M, Otho M, Sugiura N, et al. Percutaneous ethanol injection therapy for hepatocellular carcinoma: study of 95 patients. J Gastoenterol Hepatol 1990;5:616 – 26. Gehan EA. A generalized Wilcoxon test for comparing arbitrary simple censored samples. Biometrika 1965;52:457–81. Giorgio A, Tarantino L, Francica G, et al. Percutaneous ethanol injection under sonographic guidance of hepatocellular carcinoma in compensated and decompensated cirrhotic patients. J Ultrasound Med 1992;11:587 – 95. Giorgio A, Tarantino L, Francica G, et al. One-shot percutaneous ethanol injection of liver tumors under general

154

A. Giorgio et al. / European Journal of Ultrasound 12 (2000) 145–154

anaesthesia: preliminary data on efficacy and complications. Cardiovasc Intervent Radiol 1996;19:27–31. Giorgio A, Tarantino L, Mariniello N, et al. Percutaneous ethanol injection under general anaesthesia for hepatocellular carcinoma: 3 year survival in 112 patients. Eur J Ultrasound 1998;8:201–8. Izumi R, Shimizu K, Tohru II, et al. Prognostic factor of hepatocellular carcinoma in patients undergoing hepatic resection. Gastroenterology 1994;106:720–7. Kaplan L, Meyer P. Non parametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457 – 81. Kawarada Y, Ito F, Sakurai H, et al. Surgical treatment of hepatocellular carcinoma. Cancer Chemother Pharmacol 1994;33(Suppl.):12–7. Kim ST, Kim KT. Hepatic resections for primary liver cancer. Cancer Chemother Pharmacol 1994;33(Suppl.):18– 23. Kotoh K, Sakai H, Sakamoto S, et al. The effect of percutaneous ethanol injection therapy on small solitary hepatocellular carcinoma is comparable to that of hepatectomy. Am J Gastroenterol 1994;89:194–8. Lencioni R, Paolicchi A, Moretti M, et al. Combined transcatheter arterial chemoembolization and percutaneous ethanol injection for the treatment of large hepatocellular carcinoma: local therapeutic effect and long-term survival rate. Eur Radiol 1998;8:439–44. Livraghi T, Festi D, Monti F, et al. US guided percutaneous alcohol injection of small hepatic and abdominal tumors. Radiology 1986;161:309–12. Livraghi T, Vettori C, Torzilli G, Lazzaroni S, Pellicano` S, Ravasi S. Percutaneous ethanol injection of hepatic tumors: single session therapy under general anaesthesia. Am J Roentgenol 1993;161:1065–9. Livraghi T, Bolondi L, Buscarini L, et al. No treatment, resection and ethanol injection in hepatocellular carcinoma: a retrospective analysis of survival in 391 cirrhotic patients. J Hepatol 1995a;22:522–6. Livraghi T, Giorgio A, Marin G, et al. Hepatocellular carcinoma on cirrhosis in 746 patients: long term results of percutaneous ethanol injection. Radiology 1995b;197:101– 8. Livraghi T, Goldberg SN, Lazzaroni S, et al. Small hepatocellular carcinoma: Treatment with radio-frequency ablation versus ethanol injection. Radiology 1999;210:655 – 9. Mazziotti A, Grazi GL, Cavallari A. Surgical treatment of hepatocellular carcinoma on cirrhosis: a Western experience. Hepatogastroenterology 1998;45(Suppl 3):1281 – 7. Nagasue N, Kohno H, Chang Y, et al. Liver resection of

.

hepatocellular carcinoma: results of 229 consecutive patients during 11 years. Ann Surg 1993;217:375 – 84. Ondera H, Ukai K, Nakano N, Takeda T, et al. Outcomes of 116 patients with hepatocellular carcinoma. Cancer Chemother Pharmacol 1994;33(Suppl.):103. Philosophe B, Greig PD, Hemming AW, et al. Surgical management of hepatocellular carcinoma: resection or transplantation? J Gastrointest Surg 1998;2:21 – 7. Ryu M, Shimamura Y, Kinoshita T, et al. Therapeutic results of resection, transcatheter arterial embolization and percutaneous transhepatic ethanol injection in 3225 patients with hepatocellular carcinoma: a retrospective multicenter study. Jpn Clin Oncol 1997;27:251 – 7. Seky T, Nonaka T, Kubota Y, et al. Ultrasonically guided percutaneous ethanol injection therapy for hepatocellular carcinoma. Am J Gastroenterol 1989;84:1400 – 3. Sherman M. Hepatocellular carcinoma. Gastroenterologist 1995;3:55 – 66. Shiina S, Tagawa K, Unuma T. Percutaneous ethanol injection for hepatocellular carcinoma: a histopathologic study. Cancer 1991;68:1524 – 30. Shiina S, Tagawa K, Niwa Y, et al. Percutaneous ethanol injection therapy for hepatocellular carcinoma: results in 146 patients. Am J Roentgenol 1993;160:1023 – 8. Sugioka A, Tsuzuky T, Kanai T. Post resection prognosis of patients with hepatocellular carcinoma. Surgery 1993;113:612 – 8. Taavitsainen M, Vehmas T, Kauppila R. Fatal necrosis following percutaneous ethanol injection for hepatocellular carcinoma. Abdom Imaging 1993;18:357 – 9. Tanikawa K. Multidisciplinary treatment of hepatocellular carcinoma. In: Tobe T, Kameda H, editors. Primary Liver Cancer in Japan. Tokyo: Springer, 1992:327 – 31. The Liver Cancer Study Group of Japan. Predictive factors for long term prognosis after partial hepatectomy for patients with hepatocellular carcinoma. Cancer 1994; 74: 2772 –2780. Troisi R, Defreyne L, Hesse J, et al. Multimodal treatment for hepatocellular carcinoma on cirrhosis: the role of chemoembolization and alcoholization before liver transplantation. Clin Transplant 1998;12:313 – 9. Veltri A, Grosso M, Martina M, et al. Effect of preoperative radiological treatment of hepatocellular carcinoma before liver transplantation: a retrospective study. Cardiovasc Intervent Radiol 1998;21:393 – 8. Vilana R, Bruix J, Bru C, et al. Tumor size determines the efficacy of percutaneous ethanol injection for the treatment of small hepatocellular carcinoma. Hepatology 1992;16: 353 – 7.