Статья посвящена сравнительному анализу методов лечения рака печени, осложненного (или являющегося следствием) цирроза печени

Статья посвящена сравнительному анализу методов лечения рака печени, осложненного (или являющегося следствием) цирроза печени. Сравниваются 3 метода – резекция печени, её трансплантация и трансартериальная хемоэмболизация. Учитываются такие факторы, как послеоперационная смертность, развитие сопутствующих осложнений, отдаленная выживаемость. Исследование проводилось с апреля 1998 по май 2007 в Германии, при участии ряда специалистов – онкологов. Опубликована статья в декабре 2008 года.

Georgios C. Sotiropoulos^{1, 2} , Nina Drühe¹, George Sgourakis¹, Ernesto P. Molmenti¹, Susanne Beckebaum¹, Hideo A. Baba³, Gerald Antoch⁴, Philip Hilgard⁵, Arnold Radtke¹, Fuat H. Saner¹, Silvio Nadalin¹, Andreas Paul¹, Massimo Malagó¹, Christoph E. Broelsch¹ and Hauke Lang¹

Georgios C. Sotiropoulos Email: sotiropoulos@ach.klinik.uni-mainz.de Email: georgios.sotiropoulos@uni-essen.de

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths worldwide, accounting for almost half a million annual deaths [1, 2]. Its incidence has increased dramatically in the US as a result of the current hepatitis C epidemic, and is expected to continue to do so over the next 2 decades [1, 3]. Hepatitis B virus, hepatitis C virus, and chronic heavy alcohol use leading to cirrhosis remain the most important predisposing factors [1–4].

Since HCC usually arises in the setting of cirrhosis, several treatment modalities have been developed to better address both pathologies, i.e., malignancy and end-stage liver disease. Surgical procedures, ablative treatments, and systematic therapies have been evaluated in both single- and multi-center trials. Liver resection, transarterial chemoembolization (TACE), and liver transplantation (LT) are the most popular options. Radiofrequency ablation represents another widely applied method in this task. However, because of the known discrepancies among radiological and pathological tumor findings regarding HCC in the end-stage cirrhotic liver [5] and our bad experience concerning treatment with RFA for HCC in cirrhosis, RFA was only sporadically performed in our institution for this indication. As a consequence, these sporadic data were not included for the purposes of this analysis. Although many studies have compared results in various combinations [6–27], little information is available on outcomes as monotherapies at a single center.

The purpose of this study was to evaluate and compare long-term results after liver resection, TACE, and LT when implemented as monotherapies in patients with HCC and cirrhosis at a single European tertiary hepatobiliary specialty center.

Patients and Methods

Between April 1998 and May 2007, a total of 185 consecutive HCC patients with cirrhosis and no prior tumor treatments were admitted to our department. As per our protocol, the standard initial evaluation addressed the exclusion of extrahepatic liver disease and operability/operative risk. Criteria for partial hepatectomy included anatomically resectable disease and adequate reserve liver function. In cases of end-stage liver disease, LT was discussed as a possible therapeutic option. Patients with live donors were evaluated on a case-by-case basis according to their age, severity of liver disease, alpha fetoprotein (AFP) levels, and HCC characteristics at the time of presentation. Milan criteria (single tumor ≤5 cm; two to three tumors, none of them >3 cm; no vascular invasion) [28] were not considered an absolute contraindication in cases of living donor LT [29]. TACE was considered for patients who did not qualify for liver resection, deceased or live donor LT, and was performed at our Institution or elsewhere. Portal vein thrombosis, arterio-portal fistula, extrahepatic disease, and decompensated liver cirrhosis were considered contraindications for TACE. Patients with prior tumor-specific treatments, bridging treatments prior to LT, and re-admission for HCC recurrence were excluded from the study.

Preoperative diagnostic workup included abdominal ultrasound, computed tomography (CT) scan, and chest X-ray. All patients were followed up until 31 August 2007. Data were divided into three groups, according to the therapy received (group 1: liver resection; group 2: TACE; group 3: LT) and evaluated for (1) demographic and preoperative characteristics, (2) operative and intervention-related details, (3) morbidity and mortality, (4) pathologic findings, including number of tumors, tumor size, tumor stage, and grading, (5) outcome as determined by tumor recurrence and survival, and (6) analysis of prognostic factors for overall and disease-free survival.

Statistical methods used included nonparametric Fisher’s exact test (one-tail) for categorical variables, Mann–Whitney U test for quantitative variables, and power analysis (one- and two-sided) for sample size calculations. Life tables were constructed with the Kaplan–Meier product limit method. Gehan’s Wilcoxon test was employed to compare survival. Univariate and multivariable Cox proportional hazard regression analyses were used for predictors of morbidity and mortality. Significance was assigned at 0.05. Statistica release 7 (Statsoft) was used for statistical analysis.

Results

Demographic and Preoperative Characteristics

There were 185 patients, of which 149 were males and 36 females (ratio 4.14:1). Median age was 62.4 years (range 17.7–80.5). The etiology of cirrhosis was hepatitis C virus (HCV) in 36% of cases, hepatitis B virus (HBV) in 17%, alcohol abuse in 21%, cryptogenic in 12%, steatohepatitis in 3%, HCV and alcohol abuse in 2%, HBV and HCV in 2%, and other causes or combinations in 7%. Median Model for End-Stage Liver Disease (MELD) score was 9 (range 2–42). Sixty-two patients met the Milan criteria. The remaining 123 were beyond these criteria. Median AFP level was 19 ng/ml (range 1–64,200 ng/ml).

Sixty-one (n = 61) patients underwent liver resection, 64 TACE, and 60 LT. Among the resection group (n = 61), 20 underwent simple segmentectomies, 14 bisegmentectomies, 8 right hemihepatectomies, 5 left lateral sectionectomies, 5 left hemihepatectomies, 4 atypical resections, 3 trisegmentectomies, 1 right hemihepatectomy plus segment II excision, and 1 right hemihepatectomy plus caudate lobectomy. Among the TACE group (n = 64), 10 patients underwent 1 session, 15 patients 2 sessions, 5 patients 3 sessions, 8 patients 4 sessions, and 24 patients 5–14 sessions. Among the LT group (n = 60), 30 patients underwent living donor LT, 27 deceased donor LT, and 3 split graft LT.

Median ICU and hospital stays for the resection, TACE, and LT groups were 1 and 16 days, 0 and 2 days, and 6 and 29 days, respectively.

Morbidity and Mortality

Morbidity rates for the resection, TACE, and LT groups were 38%, 11%, and 38%, respectively. Resection group complications included biliary leaks (n = 2), abdominal abscesses (n = 2), evisceration (n = 1), pulmonary failure (n = 2), multiorgan failure (n = 5), liver insufficiency (n = 9), massive ascites (n = 1), and peritonitis (n = 1). TACE complications included encephalopathy (n = 1), emesis (n = 1), fever (n = 1), super-infection of the ablated liver requiring CT guided drainage (n = 3), and extensive liver necrosis (n = 1). LT complications included encephalopathy (n = 1), bleeding/hematoma (n = 5), hepatic artery thrombosis (n = 4), bile leaks (n = 2), pulmonary embolism (n = 1), pancreatitis (n = 1), renal insufficiency (n = 1), poor primary function/small-for-size syndrome (n = 2), pneumonia (n = 2), central pontine myelinolysis (n = 2), and abscess (n = 1).

Predictors of complications were defined by general linear models analyses of covariance. Age, MELD score, AFP levels, and tumor diameter constituted continuous variables. Extent of surgery in the resection group, number of sessions in the TACE group, deceased versus living donors in the transplantation group, UICC stage, tumor multifocality, and meeting/exceeding the Milan criteria constituted categorical variables. None of the above variables reached statistical significance as an independent predictor of complications by multivariate analysis (sigma-restricted parameterization). The corresponding model (Test of SS Whole Model versus SS Residual) was not able to predict complications in any of the three groups (resection group, P = 0.109; TACE group, P = 0.231; LT group, P = 0.740).

Thirty-day mortality for the resection, TACE, and LT groups was 23%, 0%, and 8%, respectively. Fourteen resection group patients died within 30 postoperative days: 12 as a result of multiorgan failure, 1 because of aspiration pneumonia, and 1 because of a massive pulmonary embolism. There were five 30-day deaths in the LT group: two patients died secondary to multiorgan failure and one each as a result of massive pulmonary embolism, methicillin-resistant Staphylococcus aureus pneumonia and right ventricular failure. No early deaths were observed in the TACE group. One-tailed Fisher exact P test showed a significant difference in 30-day mortality among the resection and LT groups (P = 0.048).

Median tumor diameter was 4 cm in the resection group (range 1–19 cm), 6 cm (range 3–14 cm) in the TACE group, and 4 cm (range 1–16 cm) in the LT group. HCCs were solitary in 101 (55%) patients (50 in group 1, 22 in group 2, and 29 in group 3). Multifocal tumors were encountered in 66% (n = 42) of TACE cases, 52% (n = 31) of LT instances, and 18% (n = 11) of resection patients. Union International Contre le Cancer (UICC) stages I, II, and >II were encountered in 66 (group 1 = 31, group 2 = 14, group 3 = 21), 50 (group 1 = 12, group 2 = 16, group 3 = 22), and 69 patients (group 1 = 18, group 2 = 34, group 3 = 17), respectively. Microscopic vascular invasion was detected in 17 patients from group 1 and 12 patients from group 3. Evaluation of tumor grading by pathology revealed moderately differentiated HCCs in about half of the cases in the resection and LT group (36/61 and 35/60, respectively). Tumor characteristics are depicted in Table 1.

Comparisons were made among the three groups by Kruskal-Wallis ANOVA and median test for the following parameters: gender (P = 0.0366; all comparisons significant), age (P = 0.0001; resection versus TACE = NS), MELD score (P = 0.0001; resection versus TACE = NS), AFP (P = 0.6769; all comparisons NS), etiology of cirrhosis [P = 0.0024; significant only TACE (HBV/HCV = 10/19) versus LTx (HBV/HCV = 17/30)], Milan criteria (P = 0.988; significant only for TACE versus resection and TACE versus LT), complications (P = 0.989; significant only TACE versus resection and TACE versus LT), ICU stay P = 0.0001 (all comparisons significant), and hospital stay P = 0.0001 (all comparisons significant).

Seventy-one (n = 71) patients are currently alive after a median follow-up of 15.3 months (range: 0.2–144). Twenty-one (n = 21) resection patients are alive, 14 without recurrence, in the resection group. Thirteen (n = 13) patients are alive, three without current evidence of vital tumor, in the TACE group. Thirty-seven (n = 37) recipients are alive, 35 without recurrence, in the LT group. Forty of 185 total patients (22%) were followed for more than 3 years. Five-year overall and recurrence-free survival rates for the entire series (n = 185 patients) were 29% and 32%, respectively (Fig. 1).

Fig. 1 Overall (a) and recurrence-free (b) survival for the entire series (n = 185)

Median survival after resection, TACE, and LT was 11 months (range: 0.2–67.4), 14 months (range: 1.2–143.9), and 23 months (range: 0.3–105), respectively. One, 3-, and 5-year cumulative survival rates after resection, TACE, and LT were 52, 43, 23%; 60, 13, 10%; and 78, 69, 59%, respectively (Table 2). LT recipients had significantly increased survival (Fig. 2) (P = 0.001). Survival differences among resection and TACE (P = 0.754) achieved statistical significance (P = 0.006) only when 30-day mortality was excluded.