World J Surg DOI 10.1007/s00268-014-2502-6
Impact of Spontaneous Hepatocellular Carcinoma Rupture on Recurrence Pattern and Long-term Surgical Outcomes after Partial Hepatectomy Hyung Soon Lee • Gi Hong Choi • Dae Ryong Kang • Kwang-Hyub Han Sang Hoon Ahn • Do Young Kim • Jun Yong Park • Seung Up Kim • Jin Sub Choi
•
Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background The clinical significance of spontaneous hepatocellular carcinoma (HCC) rupture association with recurrence pattern and long-term surgical outcomes remains under debate. We investigated the impact of spontaneous HCC rupture on recurrence pattern and longterm surgical outcomes after partial hepatectomy. Methods From 2000 to 2012, 119 patients with diagnosed ruptured HCC were reviewed. To compare outcomes between staged hepatectomy in spontaneously ruptured HCC and hepatectomy in non-ruptured HCC, we performed propensity score-matching to adjust for significant differences in patient characteristics. Overall survival, disease-free survival, and recurrence pattern were compared between the matched groups. Results Forty-four patients with newly diagnosed ruptured HCC and Child A class were initially treated with transcatheter arterial embolization for hemostasis. Three patients underwent emergency laparotomy, 18 underwent
staged hepatectomy, and 23 received transarterial chemoembolization (TACE) alone after transcatheter arterial embolization. Among the 23 patients treated with TACE alone, 10 had resectable tumors. The staged hepatectomy group shows significantly higher overall survival with TACE alone than the resectable tumor group (P \ 0.001). After propensity score-matching, overall survival, diseasefree survival, and recurrence pattern were not significantly different between the ruptured HCC with staged hepatectomy group and the non-ruptured HCC with hepatectomy group. Peritoneal recurrence rates were similar at 14.3 % versus 10.0 %, respectively (P = 0.632). Conclusions Patients with spontaneously ruptured HCC with staged hepatectomy show comparable long-term survival and recurrence pattern as patients with non-ruptured HCC having similar tumor characteristics and liver functional status. Thus, spontaneous HCC rupture may not increase peritoneal recurrence and decrease long-term survival after partial hepatectomy.
H. S. Lee G. H. Choi J. S. Choi (&) Department of Surgery, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea e-mail: [email protected]
Introduction
H. S. Lee e-mail: [email protected] D. R. Kang Biostatistics Collaboration Unit, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea K.-H. Han S. H. Ahn D. Y. Kim J. Y. Park S. U. Kim Department of Internal Medicine, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea
Spontaneous tumor rupture is a life-threatening complication of hepatocellular carcinoma (HCC). Mortality rates of spontaneously ruptured HCC remain high (25–75 %) [1], and clinicians often consider hepatectomy for treating spontaneous HCC rupture as futile because ruptured HCC is classified as T4 stage disease and evokes peritoneal metastasis risk [2, 3]. Although previous studies suggest that emergency hepatectomy is a feasible intervention for spontaneously ruptured HCC [4, 5], surgeons often feel burdened by the prospect of performing emergency hepatectomy. This is because detailed information on the HCC and liver
123
World J Surg
functional reserve after rupture is largely unknown, and liver function becomes aggravated by hemorrhagic shock. Conversely, transarterial embolization (TAE) is highly effective in achieving hemostasis during the acute phase of ruptured HCC, and it provides time to evaluate liver functional reserve and tumor status [6, 7]. Therefore, a recent shift in surgical management trends is away from emergency hepatectomy and toward TAE followed by staged hepatectomy.[8] Although staged hepatectomy for spontaneously ruptured HCC still shows poor 5-year overall survival (OS) rates (15–21 %) and disease-free survival (DFS) rates (0–50 %) [9–12] some staged hepatectomy-treated patients demonstrate prolonged OS [5, 8, 13]. The clinical significance of spontaneously ruptured HCC on recurrence pattern and long-term surgical outcomes after partial hepatectomy remains controversial. Thus, we investigated the impact of spontaneous tumor rupture on recurrence pattern after hepatectomy and long-term surgical outcomes in HCC patients.
Methods From 2000 to 2012, 119 patients diagnosed with ruptured HCC at Severance Hospital, Yonsei University College of Medicine, were retrospectively reviewed. Ruptured HCC was suspected when patients presented with sudden-onset abdominal pain, abdominal distention, and shock. The diagnosis was confirmed by dynamic contrast-enhanced abdominal computed tomography (CT) with typical findings of a ruptured HCC, including active extravasation of the contrast material, HCC with a surrounding perihepatic hematoma showing high attenuation, protrusion of the contour, and focal discontinuity of the hepatic surface [14]. Although abdominal paracentesis was occasionally performed, confirmation of blood-stained ascites was not an indispensable criterion for the diagnosis of ruptured HCC. In hemodynamically unstable patients or patients with continuous intra-abdominal hemorrhage, TAE was the first intervention. Tumor location, active bleeding site, and portal vein patency were determined during TAE performed using 5 mL iodized-oil contrast medium (Lipiodol; Laboratoire Guerbet, Aulnay-sous-Bois, France) and feeding artery embolization using gelatin sponge particles (Cutanplast; Mascia Bruneili Spa, Milano, Italy). If TAE failed, emergency laparotomy was performed. After their hemodynamic status was stabilized, patients underwent a complete clinical evaluation to allow us to formulate a definitive HCC treatment plan. Tha evaluation included serology, Child-Pugh grading, abdominal CT, hepatic angiography, and an indocyanine green retention rate at 15 min [15]. Based on test results, we assessed the
123
feasibility of hepatectomy. In surgical candidates, we conducted elective hepatectomy; resection type was decided by liver functional reserve and patient performance status. Major resection involved three or more anatomical segments, and minor resection was performed on tumors located at the liver periphery or when hepatic reserve was insufficient. Hematoma evacuation and peritoneal lavage were performed in all staged hepatectomies. Peritoneal lavage was performed to eradicate tumor cells that had dispersed into the peritoneal cavity during rupture [14]. In patients for whom surgery was not available, we performed transarterial chemoembolization (TACE) 2–4 months after TAE. Patients with staged hepatectomy were seen at followup 1 month after hepatectomy and every 3 months thereafter. Dynamic CT was performed at each appointment, and laboratory studies were use to screen for serum alpha-fetoprotein and protein induced by vitamin K absence or antagonist-II levels. Hepatocellular carcinoma recurrence was defined by newly developed lesions detected by CT. Patients with suspected recurrence underwent magnetic resonance imaging or hepatic angiography for confirmation. Postoperative complication analysis employed the Clavien–Dindo classification [16]. Wound problems were defined as surgical site infections (superficial or deep infection, wound abscess) or disrupted sutured tissue (wound, fascia). Pleural effusion, ileus, and fluid accumulation in the operating field were radiologically diagnosed. Acute renal failure was defined by abrupt and reversible declines in renal filtration function. To compare outcomes between staged hepatectomy in spontaneously ruptured HCC and hepatectomy in nonruptured HCC, we performed propensity score-matching to adjust for significant differences in patient characteristics [17, 18]. Between January 2000 and December 2007, 601 patients underwent hepatectomy for non-ruptured HCC at our institution; 12 (2.0 %) patients with perioperative mortality and 6 (1.0 %) patients with Child-Pugh class B disease were excluded. We also excluded pathologic T1-stage HCCs (n = 257) because there were no T1 tumors in patients with spontaneously ruptured HCCs who underwent staged hepatectomy. We conducted propensity score-matching on the remaining 326 non-ruptured HCC patients to select the control group. Statistical Analysis The propensity scores were estimated by multiple logistic regression analysis. Regression analysis predicted the probability that each patient would be treated based on nine covariables: age, sex, etiology of cirrhosis, microvascular invasion, gross vascular invasion, the presence of
World J Surg
cirrhosis, tumor size, number, and stage. Regarding of cirrhosis etiology, the hepatitis B virus (HBV) group was defined as being positive for the HBV surface antigen, and the hepatitis C group (HCV) was defined as being positive for the anti-HCV antibody. Tumor stage, size, number, the presence of cirrhosis, and vascular invasion were determined by pathology. Tumor staging was determined based on the American Joint Committee on Cancer (AJCC), 7th Edition, tumor-node-metastasis (TNM) classification [2]. However, the tumor stage of ruptured HCC was assessed by the tumor itself, excluding the rupture factor. Using these nine covariables in the regression analysis, a propensity score was calculated for each patient. The discrimination and calibration abilities of the propensity score models were assessed with the C-statistic and the HosmerLemeshow statistic. The model was then used to obtain a one-to-two match for the ruptured HCC with staged hepatectomy group and the non-ruptured HCC with hepatectomy group. Finally, each patient with ruptured HCC and a staged hepatectomy was matched to one patient with nonruptured HCC and a hepatectomy, and the identical propensity score. After the propensity score-matched sample was formed, we assessed baseline variable balance between the two propensity-matched cohort groups. The effect of HCC rupture on the outcomes and its statistical significance were estimated using appropriate statistical methods for matched data. Continuous variables were compared with the paired t-test or the Wilcoxon signed rank test, as appropriate, and categorical variables were compared with the McNemar’s or marginal homogeneity test, as appropriate. Data are presented as patient numbers (percentages) or as median value (range). Categorical variables were analyzed with the chi-square test or Fisher’s exact test. Continuous variable were analyzed with the Mann–Whitney U-test. Survival rates were calculated by the Kaplan–Meier method, and differences between the groups were assessed with the log rank test. Overall survival was calculated as the time from hepatectomy until death from any cause, or until the observation period conclusion. Disease-free survival was the time from hepatectomy until recurrent disease detection, or until the observation period conclusion without recurrence. To identify prognostic factors for OS in the staged hepatectomy group and the TACE-alone with resectable tumor group, Cox-regression analysis was performed. Variables found to be significant prognostic factors by univariate analysis were considered in the multivariate analysis. Statistical analyses were performed with SAS (version 9.2; SAS Institute Inc, Cary, NC). P-values \ 0.05 were considered statistically significant. This study was approved by the Institutional Review Board of Yonsei
University College of Medicine (Seoul, Republic of Korea).
Results Patient Characteristics Patients with ruptured HCC were managed according to the flow chart presented in Fig. 1. Of the 119 patients diagnosed with ruptured HCC, 53 patients had newly diagnosed ruptured HCCs. Among these 53 newly diagnosed ruptured HCC patients, 3 patients had concomitant extrahepatic metastasis and 6 patients had Child-Pugh grade [ B disease; these 9 patients were managed conservatively. The remaining 44 patients with Child-Pugh grade A were initially treated with TAE for hemostasis. Shock was present in 23/53 patients (52.2 %) on admission, and 3 patients underwent emergency laparotomy due to hemostasis failure after TAE (bleeder ligation in 2, tumor enucleation in 1 patient). Among the three emergency laparotomy patients, one died of liver failure 1 month postoperatively, and one died of tumor rebleeding at postoperative day 3. After TAE, 18 patients underwent staged hepatectomy, and 23 underwent TACE alone. Among those 23 patients, 10 patients were seen on retrospective data review to have had resectable tumors. The reasons for TACE-alone treatment in the 10 patients with resectable tumor were: patient refusal in 8 and comorbid disease in 2. The median duration between TAE and staged hepatectomy was 23 days (range: 9–375 days). Major hepatectomy was conducted in 7 patients (38.9 %): right hepatectomy in 3, left hepatectomy in 3, and central bisectionectomy in 1. Minor hepatectomy was performed in 11 patients (61.1 %): lateral sectionectomy was performed in 2 and segmentectomy performed in 9. The median follow-up duration after staged hepatectomy was 68 months (range: 4–160 months). Staged Hepatectomy Complications Eight staged hepatectomy patients (44 %) experienced complications. Grade I complications developed in 6 patients and were managed conservatively: pleural effusion developed in 3, wound problems in 2, and ileus in 1. Grade II complications developed in two patients: fluid accumulation at the resected liver surface in 1 and acute renal failure in 1. The patient with fluid collection developed fever at postoperative day 6 and recovered after antibiotic optimization. The renal failure patient received intermittent diuretic therapy and recovered normal renal filtration rates.
123
World J Surg Fig. 1 Patient selection schematic
group showed significantly higher OS than the TACEalone group (P \ 0.001). Prognostic Factors for OS in the Staged Hepatectomy Group and the TACE-alone with Resectable Tumor Group
Fig. 2 Comparison of overall survival between the staged hepatectomy group and the transarterial chemoembolization (TACE)-alone with resectable tumor group
There were no complications above type IIIa; nor were there any postoperative deaths. Comparison of OS between the Staged Hepatectomy Group and the TACE-alone with Resectable Tumor Group The 1-, 3-, and 5-year OS rates were, respectively, 40.0 %, 0 %, and 0 % in TACE-alone with resectable tumor patients, and 88.5 %, 64.4 %, and 55.2 % in the staged hepatectomy patients (Fig. 2). The staged hepatectomy
123
The patient characteristics at HCC rupture diagnosis in both groups are summarized in Table 1. Patients in the TACEalone group had significantly higher total bilirubin levels and international normalized ratios than patients in the staged hepatectomy group. In addition, tumors were significantly larger in the TACE-alone group than in the staged hepatectomy group. On univariate analysis, tumor size and type of treatment were found to be significantly associated with OS (Table 2). However, in multivariate analysis, type of treatment was the only independent prognostic factor for OS. Comparison of Long-term Surgical Outcomes between Non-ruptured HCC Patients and Ruptured HCC Patients As noted earlier, propensity score-matching considered nine variables (Table 3). The two groups showed significant differences in tumor size and staging, as well as microvascular and gross vascular invasion, before matching. These differences between groups were eliminated after matching. Before matching, 1-, 3-, and 5-year OS rates were, respectively, 90.3 %, 68.7 %, and 57.2 % in non-ruptured HCC patients and 88.5 %, 64.4 %, and 55.2 % in staged hepatectomy patients. The 1-, 3-, and 5-year DFS rates were,
World J Surg Table 1 Patient characteristics at diagnosis of tumor rupture in the staged hepatectomy group versus the TACE-alone with resectable tumor group Variables
Staged hepatectomy (n = 18)
Age, years Male gender
53.0 (40–78) a
HBsAg positivitya Shock on presentationa
TACE-alone (n = 10) 66.5 (36–79)
P value 0.072
15 (83.3 %)
9 (90.0 %)
16 (88.9 %)
8 (80.0 %)
0.520
5 (50 %)
0.778
8 (44.4 %)
0.629
Hemoglobin, g/dL
11.4 (8.0–16.0)
10.0 (7.6–11.2)
0.002
Creatinine, mg/dL
0.9 (0.5–2.7)
1.0 (0.5–1.7)
0.796
AST, IU/L ALT, IU/L Platelets (/mm3) Total bilirubin, mg/dL INR ICG R15 (%)
49.5 (15–9) 45.5 (5–1) 170,000 (53,000–467,000)
82.0 (1–6) 55.5 (7–1) 161,500 (76,000–249,000)
0.121 0.245 0.796
0.8 (0.3–2.2)
1.2 (0.5–2.4)
0.041
1.07 (0.86–1.47)
1.27 (1.04–1.77)
0.014 0.159
7.0 (4.0–31.5)
9.0 (5.5–31.2)
Alphafetoprotein, IU/mL
106.8 (1.1–83,000)
285.0 (1.8–83,000)
0.356
PIVKA-II, mAU/mL
806.5 (13–20,498)
479.0 (40–2,000)
0.900
Tumor size, cm Tumor number
6.1 (3.0–15.0) 1 (1–3)
8.0 (6.0–13.0) 1 (1–3)
0.045 0.944
Values in parentheses are median with range unless indicated otherwise a
Values are number with percentages
HCC Hepatocellular carcinoma, TACE Transarterial chemoembolization, HBsAg Hepatitis B virus surface antigen, AST Aspartate aminotransferase, ALT Alanine aminotransferase, INR International normalized ratio, ICG R15 Indocyanine green retention rate at 15 min, AFP Alphafetoprotein, PIVKA-II Protein induced by vitamin K absence or antagonist-II
respectively, 64.3 %, 42.8 %, and 36.5 % in non-ruptured HCC patients and 54.9 %, 39.2 %, and 7.8 % in staged hepatectomy patients. The DFS rates were significantly higher in the non-ruptured HCC group than in the staged hepatectomy group (P = 0.019), although OS rates were not significantly different between the two groups (Fig. 3a). After matching, 1-, 3-, and 5-year OS rates were, respectively, 89.0 %, 55.1 %, and 46.5 % in the matched control group and 88.5 %, 64.4 %, and 55.2 % in the staged hepatectomy group. The 1-, 3-, and 5-year DFS rates were, respectively, 55.8 %, 35.5 %, and 35.5 % in the non-ruptured HCC group and 54.9 %, 39.2 %, and 7.8 % in the staged hepatectomy group. No significant intergroup differences in OS and DFS rates were observed between the two matched groups (Fig. 3b). In the staged hepatectomy group, HCC recurrence developed in 14/18 patients (77.8 %), and the peritoneal recurrence rate was 14.3 % (Table 4). In the matched control group, HCC recurrence developed in 20/37 patients (54.1 %), and the peritoneal recurrence rate was 10.0 %. There was no significant difference in recurrence pattern between the two matched groups.
Discussion Comparing surgical outcomes between spontaneously ruptured HCC and non-ruptured HCC is difficult because
of different patient characteristics at diagnosis. Most patients with spontaneously ruptured HCC had higher tumor stage and poorer liver function than non-ruptured HCC patients. Thus, hepatectomy in spontaneously ruptured HCC show lower survival outcomes than hepatectomy in non-ruptured HCC. Liu et al [10] demonstrated that hepatectomy in ruptured HCC shows lower OS and DFS rates than hepatectomy in non-ruptured HCC. Yeh et al [5] reported no significant difference in OS rates after hepatectomy between ruptured HCC and non-ruptured HCC, although non-ruptured HCC show significantly better DFS rates. However, previous studies did not adjust for patient characteristics that affect prognosis after surgery, and required recalibration of patient characteristics between ruptured HCC and non-ruptured HCC groups. Mizuno et al [9] compared survival rates of patients with ruptured and non-ruptured HCC based on the same background factors, and they found no significant differences between groups in OS and DFS. However, this may not be an accurate comparison between the groups because ruptured HCC patients were compared with non-ruptured HCC patients with stage IVa disease (AJCC TNM classification T4N0/1M0). Although tumor rupture is considered T4 stage, most resectable ruptured HCCs are small or mediumsized solitary tumors. Likewise, when the factor of tumor rupture is excluded, there were no T4 tumors in ruptured HCCs treated with staged hepatectomy in our study.
123
World J Surg Table 2 Cox regression analysis of prognostic factor for overall survival in the staged hepatectomy group and the TACE-alone with resectable tumor group (n = 28)
Variable
Category
Overall survival Hazard ratio
95 % Confidence interval
P value
1.035
0.996–1.076
0.076
Univariate analysis Age (1-year increase) Gender
Female vs male
1.762
0.400–7.762
0.454
HBsAg positivity
No vs yes
1.524
0.345–6.727
0.578
Shock on presentation
No vs yes
0.732
0.258–2.075
0.557
Hemoglobin B 10 g/dL
No vs yes
1.944
0.701–1.944
0.202
Creatinine [ 1.2 mg/dL Serum platelet B 100,000 mm3
No vs yes No vs yes
1.100 0.995
0.245–4.950 0.324–3.059
0.901 0.993
Total bilirubin [ 1.1 mg/dL
No vs yes
1.835
0.721–4.667
0.203
INR [ 1.4
No vs yes
1.680
0.462–6.104
0.431
ICG R15 [ 10 %
No vs yes
AFP (continuous), IU/mL
2.183
0.645–7.382
0.209
1.000
0.998–1.002
0.906
PIVKA-II (continuous), mAU/mL
1.000
0.976–1.024
0.981
Tumor size (continuous), cm
1.177
1.012–1.368
0.034
Tumor number
Solitary vs multiple
1.279
0.493–3.316
0.613
Type of treatment
Staged hepatectomy vs TACE-alone
8.635
2.172–34.336
0.002
1.169 7.876
0.983–1.390 1.939–31.996
0.077 0.004
Multivariate analysis Tumor size (continuous), cm Type of treatment
Staged hepatectomy vs TACE-alone
Table 3 Baseline variables in the non-ruptured HCC with hepatectomy group versus the ruptured HCC with staged hepatectomy group Variable
Category
Before matching Non-ruptured HCC (n = 326)
a
After matching Staged hepatectomy (n = 18)
P value
Non-ruptured HCC (n = 37)
Staged hepatectomy (n = 18)
P value
53 (22–72)
53 (40–78)
0.647
52 (32–71)
53 (40–78)
0.831
254 (77.9 %)
15 (83.3 %)
0.588
28 (75.7 %)
15 (83.3 %)
0.731
HBV
259 (79.4 %)
16 (88.9 %)
0.543
31 (83.8 %)
16 (88.9 %)
1.000
HCV
13 (4.0 %)
0 (0 %)
0 (0 %)
0 (0 %)
No hepatitis
54 (16.6 %)
2 (11.1 %)
6 (16.2 %)
2 (11.1 %)
Tumor size, cma
4.5 (0.9–18.0)
6.1 (3.0–15.0)
0.047
6.5 (0.9–18.0)
6.1 (3.0–15.0)
0.905
Multiple tumors
108 (33.1 %)
6 (33.3 %)
0.986
20 (54.1 %)
6 (33.3 %)
0.149
Cirrhosis
176 (54 %)
11 (61.1 %)
0.543
21 (56.8 %)
11 (61.1 %)
0.759 0.505
Age, years
Male gender Etiology
Gross vascular invasion
44 (13.5 %)
8 (44.4 %)
0.000
13 (35.1 %)
8 (44.4 %)
Microvascular invasion
294 (90.2 %)
12 (66.7 %)
0.002
26 (70.7 %)
12 (66.7 %)
0.786
II III
273 (83.7 %) 47 (14.4 %)
4 (22.2 %) 14 (77.8 %)
0.000
9 (24.3 %) 28 (75.7 %)
4 (22.2 %) 14 (77.8 %)
0.403
IV
6 (1.8 %)
Tumor stage
0 (0 %)
Values in parentheses are percentages unless indicated otherwise a
Values are median (range)
HCC Hepatocellular carcinoma, HBV Hepatitis B virus, HCV Hepatitis C virus
123
0 (0 %)
0 (0 %)
World J Surg Fig. 3 Comparison of diseasefree survival (DFS) and overall survival (OS) between staged hepatectomy in the ruptured hepatocellular carcinoma (HCC) group and hepatectomy in the non-ruptured HCC group
Table 4 Recurrence pattern in the staged hepatectomy group versus the matched control group Recurrence pattern
Group
P value
Staged hepatectomy (n = 14)
Matched controls (n = 20)
Intrahepatic
7 (50.0 %)
10 (50.0 %)
Lung
2 (14.3 %)
6 (30.0 %)
Peritoneal
2 (14.3 %)
2 (10.0 %)
Lymph node
3 (21.4 %)
2 (10.0 %)
0.632
Therefore it seems optimal that ruptured HCCs were compared to non-ruptured HCCs with similar tumor characteristics excluding the rupture factor. In this regard, our study redefined tumor characteristics by excluding the factor of tumor rupture, and the patients were restricted to those characterized as Child-Pugh class A. We performed propensity score-matching to overcome differences in patient characteristics that affect postoperative prognosis between the ruptured HCC and non-ruptured HCC groups. After matching, ruptured HCC showed no significant differences in OS and DFS than non-ruptured HCC. Thus, ruptured HCC patients may have long-term survival
comparable to those with non-ruptured HCC, with similar tumor characteristics and liver functional status. There are many reports of peritoneal recurrence after spontaneous HCC rupture [19–22], and some reports insist that spontaneous HCC rupture evokes peritoneal metastasis [3]. However, it remains debatable whether spontaneous rupture of HCC increases peritoneal recurrence rates and decreases positive long-term surgical outcomes. Lin et al [23] reported that most patients with peritoneal metastasis after hepatectomy exhibit no prior evidence of overt HCC rupture. Mizuno et al [9] demonstrated that the recurrence pattern after hepatectomy is similar between ruptured HCC and non-ruptured HCC. Our results also demonstrate no difference in recurrence pattern between ruptured HCC patients and matched non-ruptured HCC controls. Peritoneal lavage is important in reducing peritoneal recurrence rates; however, tumor control by staged hepatectomy may play a larger role. Cheng et al [11] assessed the AJCC 7 staging system validity for resectable HCC prognosis, and tumor rupture does not influence survival after hepatectomy in stage III patients. Kwak et al [24] reported that peritoneal metastasis is not an independent prognostic factor of HCC, and maintenance of
123
World J Surg
favorable hepatic function and controlling intrahepatic HCC are more important prognostic factors. These findings are consistent with our observations that patients in the TACE-alone with resectable tumor group had significantly lower OS rates than those in the staged hepatectomy group. The TACE-alone group had Child-Pugh grade A liver function, and the tumor size was not a crucial factor for resectability. Nonetheless, poor OS rates in the TACE-alone group might be attributed in part to lower baseline liver function and a higher proportion of larger tumors. However, in multivariate analysis of prognostic factors for OS, type of treatment was identified as the only independent prognostic factor. This finding suggests that surgical resection may be the most important factor for long-term survival in patients with ruptured HCC. Consequently, in patients with preserved liver function and resectable tumors, staged hepatectomy should be considered as a potential approach for achieving tumor control and long-term survival. Recently, Hai et al [4] advocated emergency hepatectomy for achieving both hemostasis and definitive treatment in a single operation; mortality rates were 5.8 %. However, morbidity rates were not described and lower mortality rates resulted because emergency hepatectomy was only performed in patients with peripherally located tumors. Ruptured HCC patients have acutely damaged livers and abnormal coagulation profiles [10, 25, 26]. Therefore, emergency hepatectomy may diminish residual hepatic function, leading to postoperative bleeding and liver failure. Lai et al [27] reported that emergency hepatectomy in spontaneously ruptured HCC shows a 71.4 % mortality rate, and hepatic artery ligation has 76.6 % mortality. Vergara et al [28] reported that mortality and morbidity rates after emergency hepatectomy in spontaneously ruptured HCC were 16.5 and 50 %, respectively. In our study, one patient died of postoperative liver failure and one patient died of postoperative bleeding among the three emergency laparotomy patients. Thus, emergency surgery for ruptured HCC may be associated with high morbidity and mortality. We observed no complications above type IIIa after staged hepatectomy. Full preoperative clinical evaluations indicated major hepatectomy for seven patients (38.9 %). Thus, staged hepatectomy is facilitated by full preoperative clinical evaluation, thereby allowing safe and curative treatment in ruptured HCC versus emergency hepatectomy. This study has several limitations. It was performed in single center with a small sample size. In addition, it was designed retrospectively, so selection bias is possible. Although propensity score-matching reduced potential selection bias, the possibility of data loss still exists in the matching process [17, 29].
123
In summary, spontaneously ruptured HCC with staged hepatectomy shows comparable long-term survival and recurrence pattern versus non-ruptured HCC, with similar tumor characteristics and liver functional status. Spontaneous HCC rupture may not increase peritoneal recurrence rates and decrease long-term survival after hepatectomy. Additionally, spontaneously ruptured HCC may need staging reappraisal according to its own tumor characteristics and excluding the rupture factor, when patients have resectable tumors and preserved liver function.
Conflict of interest No author has any conflict of interest to declare, financial or otherwise. Funding
None
References 1. Lai EC, Lau WY (2006) Spontaneous rupture of hepatocellular carcinoma: a systematic review. Arch Surg 141:191–198 2. Edge SB, Byrd DR, Compton CC et al (2009) AJCC cancer staging manual, 7th edn. Springer, New York 3. Sonoda T, Kanematsu T, Takenaka K et al (1989) Ruptured hepatocellular carcinoma evokes risk of implanted metastases. J Surg Oncol 41:183–186 4. Hai L, Yong-Hong P, Yong F et al (2005) One-stage liver resection for spontaneous rupture of hepatocellular carcinoma. World J Surg 29:1316–1318. doi:10.1007/s11805-007-0322-2 5. Yeh CN, Lee WC, Jeng LB et al (2002) Spontaneous tumour rupture and prognosis in patients with hepatocellular carcinoma. Br J Surg 89:1125–1129 6. Leung CS, Tang CN, Fung KH et al (2002) A retrospective review of transcatheter hepatic arterial embolisation for ruptured hepatocellular carcinoma. J R Coll Surg Edinb 47:685–688 7. Kim JY, Lee JS, Oh DH et al (2012) Transcatheter arterial chemoembolization confers survival benefit in patients with a spontaneously ruptured hepatocellular carcinoma. Eur J Gastroenterol Hepatol 24:640–645 8. Buczkowski AK, Kim PT, Ho SG et al (2006) Multidisciplinary management of ruptured hepatocellular carcinoma. J Gastrointest Surg 10:379–386 9. Mizuno S, Yamagiwa K, Ogawa T et al (2004) Are the results of surgical treatment of hepatocellular carcinoma poor if the tumor has spontaneously ruptured? Scand J Gastroenterol 39:567–570 10. Liu CL, Fan ST, Lo CM et al (2001) Management of spontaneous rupture of hepatocellular carcinoma: single-center experience. J Clin Oncol 19:3725–3732 11. Cheng CH, Lee CF, Wu TH et al (2011) Evaluation of the new AJCC staging system for resectable hepatocellular carcinoma. World J Surg Oncol 9:114 12. Yoshida H, Onda M, Tajiri T et al (1999) Treatment of spontaneous ruptured hepatocellular carcinoma. Hepatogastroenterology 46:2451–2453 13. Shimada R, Imamura H, Makuuchi M et al (1998) Staged hepatectomy after emergency transcatheter arterial embolization for ruptured hepatocellular carcinoma. Surgery 124:526–535 14. Zhu LX, Wang GS, Fan ST (1996) Spontaneous rupture of hepatocellular carcinoma. Br J Surg 83:602–607
World J Surg 15. Choi GH, Park JY, Hwang HK et al (2011) Predictive factors for long-term survival in patients with clinically significant portal hypertension following resection of hepatocellular carcinoma. Liver Int 31:485–493 16. Clavien PA, Barkun J, de Oliveira ML et al (2009) The ClavienDindo classification of surgical complications: five-year experience. Ann Surg 250:187–196 17. D’Agostino RB Jr (1998) Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 17:2265–2281 18. Braitman LE, Rosenbaum PR (2002) Rare outcomes, common treatments: analytic strategies using propensity scores. Ann Intern Med 137:693–695 19. Yunoki Y, Takeuchi H, Makino Y et al (1999) Intraperitoneal seeding of ruptured hepatocellular carcinoma: case report. Abdom Imaging 24:398–400 20. Okano J, Shiota G, Horie Y et al (1996) Rupture of metastatic nodule on the peritoneal surface secondary to hepatocellular carcinoma. Intern Med 35:783–784 21. Eriguchi N, Aoyagi S, Okuda K et al (2000) Successful surgical treatment for implanted intraperitoneal metastases of hepatocellular carcinoma. J Hepatobiliary Pancreat Surg 7:520–523 22. Lin CC, Chen CH, Tsang YM et al (2006) Diffuse intraperitoneal metastasis after spontaneous rupture of hepatocellular carcinoma. J Formos Med Assoc 105:577–582
23. Lin CC, Liang HP, Lee HS et al (2009) Clinical manifestations and survival of hepatocellular carcinoma patients with peritoneal metastasis. J Gastroenterol Hepatol 24:815–820 24. Kwak MS, Lee JH, Yoon JH et al (2012) Risk factors, clinical features, and prognosis of the hepatocellular carcinoma with peritoneal metastasis. Dig Dis Sci 57:813–819 25. Miyamoto M, Sudo T, Kuyama T (1991) Spontaneous rupture of hepatocellular carcinoma: a review of 172 Japanese cases. Am J Gastroenterol 86:67–71 26. Hsueh KC, Fan HL, Chen TW et al (2012) Management of spontaneously ruptured hepatocellular carcinoma and hemoperitoneum manifested as acute abdomen in the emergency room. World J Surg 36:2670–2676. doi:10.1007/s00268-012-1734-6 27. Lai EC, Wu KM, Choi TK et al (1989) Spontaneous ruptured hepatocellular carcinoma. an appraisal of surgical treatment. Ann Surg 210:24–28 28. Vergara V, Muratore A, Bouzari H et al (2000) Spontaneous rupture of hepatocelluar carcinoma: surgical resection and longterm survival. Eur J Surg Oncol 26:770–772 29. Austin PC, Grootendorst P, Anderson GM (2007) A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: a Monte Carlo study. Stat Med 26:734–753
123
Impact of Spontaneous Hepatocellular Carcinoma Rupture on Recurrence Pattern and Long-term Surgical Outcomes after Partial Hepatectomy Hyung Soon Lee • Gi Hong Choi • Dae Ryong Kang • Kwang-Hyub Han Sang Hoon Ahn • Do Young Kim • Jun Yong Park • Seung Up Kim • Jin Sub Choi
•
Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background The clinical significance of spontaneous hepatocellular carcinoma (HCC) rupture association with recurrence pattern and long-term surgical outcomes remains under debate. We investigated the impact of spontaneous HCC rupture on recurrence pattern and longterm surgical outcomes after partial hepatectomy. Methods From 2000 to 2012, 119 patients with diagnosed ruptured HCC were reviewed. To compare outcomes between staged hepatectomy in spontaneously ruptured HCC and hepatectomy in non-ruptured HCC, we performed propensity score-matching to adjust for significant differences in patient characteristics. Overall survival, disease-free survival, and recurrence pattern were compared between the matched groups. Results Forty-four patients with newly diagnosed ruptured HCC and Child A class were initially treated with transcatheter arterial embolization for hemostasis. Three patients underwent emergency laparotomy, 18 underwent
staged hepatectomy, and 23 received transarterial chemoembolization (TACE) alone after transcatheter arterial embolization. Among the 23 patients treated with TACE alone, 10 had resectable tumors. The staged hepatectomy group shows significantly higher overall survival with TACE alone than the resectable tumor group (P \ 0.001). After propensity score-matching, overall survival, diseasefree survival, and recurrence pattern were not significantly different between the ruptured HCC with staged hepatectomy group and the non-ruptured HCC with hepatectomy group. Peritoneal recurrence rates were similar at 14.3 % versus 10.0 %, respectively (P = 0.632). Conclusions Patients with spontaneously ruptured HCC with staged hepatectomy show comparable long-term survival and recurrence pattern as patients with non-ruptured HCC having similar tumor characteristics and liver functional status. Thus, spontaneous HCC rupture may not increase peritoneal recurrence and decrease long-term survival after partial hepatectomy.
H. S. Lee G. H. Choi J. S. Choi (&) Department of Surgery, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea e-mail: [email protected]
Introduction
H. S. Lee e-mail: [email protected] D. R. Kang Biostatistics Collaboration Unit, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea K.-H. Han S. H. Ahn D. Y. Kim J. Y. Park S. U. Kim Department of Internal Medicine, Yonsei University College of Medicine, 250 Seongsanno, Seodaemoon-gu, Seoul 120-752, Republic of Korea
Spontaneous tumor rupture is a life-threatening complication of hepatocellular carcinoma (HCC). Mortality rates of spontaneously ruptured HCC remain high (25–75 %) [1], and clinicians often consider hepatectomy for treating spontaneous HCC rupture as futile because ruptured HCC is classified as T4 stage disease and evokes peritoneal metastasis risk [2, 3]. Although previous studies suggest that emergency hepatectomy is a feasible intervention for spontaneously ruptured HCC [4, 5], surgeons often feel burdened by the prospect of performing emergency hepatectomy. This is because detailed information on the HCC and liver
123
World J Surg
functional reserve after rupture is largely unknown, and liver function becomes aggravated by hemorrhagic shock. Conversely, transarterial embolization (TAE) is highly effective in achieving hemostasis during the acute phase of ruptured HCC, and it provides time to evaluate liver functional reserve and tumor status [6, 7]. Therefore, a recent shift in surgical management trends is away from emergency hepatectomy and toward TAE followed by staged hepatectomy.[8] Although staged hepatectomy for spontaneously ruptured HCC still shows poor 5-year overall survival (OS) rates (15–21 %) and disease-free survival (DFS) rates (0–50 %) [9–12] some staged hepatectomy-treated patients demonstrate prolonged OS [5, 8, 13]. The clinical significance of spontaneously ruptured HCC on recurrence pattern and long-term surgical outcomes after partial hepatectomy remains controversial. Thus, we investigated the impact of spontaneous tumor rupture on recurrence pattern after hepatectomy and long-term surgical outcomes in HCC patients.
Methods From 2000 to 2012, 119 patients diagnosed with ruptured HCC at Severance Hospital, Yonsei University College of Medicine, were retrospectively reviewed. Ruptured HCC was suspected when patients presented with sudden-onset abdominal pain, abdominal distention, and shock. The diagnosis was confirmed by dynamic contrast-enhanced abdominal computed tomography (CT) with typical findings of a ruptured HCC, including active extravasation of the contrast material, HCC with a surrounding perihepatic hematoma showing high attenuation, protrusion of the contour, and focal discontinuity of the hepatic surface [14]. Although abdominal paracentesis was occasionally performed, confirmation of blood-stained ascites was not an indispensable criterion for the diagnosis of ruptured HCC. In hemodynamically unstable patients or patients with continuous intra-abdominal hemorrhage, TAE was the first intervention. Tumor location, active bleeding site, and portal vein patency were determined during TAE performed using 5 mL iodized-oil contrast medium (Lipiodol; Laboratoire Guerbet, Aulnay-sous-Bois, France) and feeding artery embolization using gelatin sponge particles (Cutanplast; Mascia Bruneili Spa, Milano, Italy). If TAE failed, emergency laparotomy was performed. After their hemodynamic status was stabilized, patients underwent a complete clinical evaluation to allow us to formulate a definitive HCC treatment plan. Tha evaluation included serology, Child-Pugh grading, abdominal CT, hepatic angiography, and an indocyanine green retention rate at 15 min [15]. Based on test results, we assessed the
123
feasibility of hepatectomy. In surgical candidates, we conducted elective hepatectomy; resection type was decided by liver functional reserve and patient performance status. Major resection involved three or more anatomical segments, and minor resection was performed on tumors located at the liver periphery or when hepatic reserve was insufficient. Hematoma evacuation and peritoneal lavage were performed in all staged hepatectomies. Peritoneal lavage was performed to eradicate tumor cells that had dispersed into the peritoneal cavity during rupture [14]. In patients for whom surgery was not available, we performed transarterial chemoembolization (TACE) 2–4 months after TAE. Patients with staged hepatectomy were seen at followup 1 month after hepatectomy and every 3 months thereafter. Dynamic CT was performed at each appointment, and laboratory studies were use to screen for serum alpha-fetoprotein and protein induced by vitamin K absence or antagonist-II levels. Hepatocellular carcinoma recurrence was defined by newly developed lesions detected by CT. Patients with suspected recurrence underwent magnetic resonance imaging or hepatic angiography for confirmation. Postoperative complication analysis employed the Clavien–Dindo classification [16]. Wound problems were defined as surgical site infections (superficial or deep infection, wound abscess) or disrupted sutured tissue (wound, fascia). Pleural effusion, ileus, and fluid accumulation in the operating field were radiologically diagnosed. Acute renal failure was defined by abrupt and reversible declines in renal filtration function. To compare outcomes between staged hepatectomy in spontaneously ruptured HCC and hepatectomy in nonruptured HCC, we performed propensity score-matching to adjust for significant differences in patient characteristics [17, 18]. Between January 2000 and December 2007, 601 patients underwent hepatectomy for non-ruptured HCC at our institution; 12 (2.0 %) patients with perioperative mortality and 6 (1.0 %) patients with Child-Pugh class B disease were excluded. We also excluded pathologic T1-stage HCCs (n = 257) because there were no T1 tumors in patients with spontaneously ruptured HCCs who underwent staged hepatectomy. We conducted propensity score-matching on the remaining 326 non-ruptured HCC patients to select the control group. Statistical Analysis The propensity scores were estimated by multiple logistic regression analysis. Regression analysis predicted the probability that each patient would be treated based on nine covariables: age, sex, etiology of cirrhosis, microvascular invasion, gross vascular invasion, the presence of
World J Surg
cirrhosis, tumor size, number, and stage. Regarding of cirrhosis etiology, the hepatitis B virus (HBV) group was defined as being positive for the HBV surface antigen, and the hepatitis C group (HCV) was defined as being positive for the anti-HCV antibody. Tumor stage, size, number, the presence of cirrhosis, and vascular invasion were determined by pathology. Tumor staging was determined based on the American Joint Committee on Cancer (AJCC), 7th Edition, tumor-node-metastasis (TNM) classification [2]. However, the tumor stage of ruptured HCC was assessed by the tumor itself, excluding the rupture factor. Using these nine covariables in the regression analysis, a propensity score was calculated for each patient. The discrimination and calibration abilities of the propensity score models were assessed with the C-statistic and the HosmerLemeshow statistic. The model was then used to obtain a one-to-two match for the ruptured HCC with staged hepatectomy group and the non-ruptured HCC with hepatectomy group. Finally, each patient with ruptured HCC and a staged hepatectomy was matched to one patient with nonruptured HCC and a hepatectomy, and the identical propensity score. After the propensity score-matched sample was formed, we assessed baseline variable balance between the two propensity-matched cohort groups. The effect of HCC rupture on the outcomes and its statistical significance were estimated using appropriate statistical methods for matched data. Continuous variables were compared with the paired t-test or the Wilcoxon signed rank test, as appropriate, and categorical variables were compared with the McNemar’s or marginal homogeneity test, as appropriate. Data are presented as patient numbers (percentages) or as median value (range). Categorical variables were analyzed with the chi-square test or Fisher’s exact test. Continuous variable were analyzed with the Mann–Whitney U-test. Survival rates were calculated by the Kaplan–Meier method, and differences between the groups were assessed with the log rank test. Overall survival was calculated as the time from hepatectomy until death from any cause, or until the observation period conclusion. Disease-free survival was the time from hepatectomy until recurrent disease detection, or until the observation period conclusion without recurrence. To identify prognostic factors for OS in the staged hepatectomy group and the TACE-alone with resectable tumor group, Cox-regression analysis was performed. Variables found to be significant prognostic factors by univariate analysis were considered in the multivariate analysis. Statistical analyses were performed with SAS (version 9.2; SAS Institute Inc, Cary, NC). P-values \ 0.05 were considered statistically significant. This study was approved by the Institutional Review Board of Yonsei
University College of Medicine (Seoul, Republic of Korea).
Results Patient Characteristics Patients with ruptured HCC were managed according to the flow chart presented in Fig. 1. Of the 119 patients diagnosed with ruptured HCC, 53 patients had newly diagnosed ruptured HCCs. Among these 53 newly diagnosed ruptured HCC patients, 3 patients had concomitant extrahepatic metastasis and 6 patients had Child-Pugh grade [ B disease; these 9 patients were managed conservatively. The remaining 44 patients with Child-Pugh grade A were initially treated with TAE for hemostasis. Shock was present in 23/53 patients (52.2 %) on admission, and 3 patients underwent emergency laparotomy due to hemostasis failure after TAE (bleeder ligation in 2, tumor enucleation in 1 patient). Among the three emergency laparotomy patients, one died of liver failure 1 month postoperatively, and one died of tumor rebleeding at postoperative day 3. After TAE, 18 patients underwent staged hepatectomy, and 23 underwent TACE alone. Among those 23 patients, 10 patients were seen on retrospective data review to have had resectable tumors. The reasons for TACE-alone treatment in the 10 patients with resectable tumor were: patient refusal in 8 and comorbid disease in 2. The median duration between TAE and staged hepatectomy was 23 days (range: 9–375 days). Major hepatectomy was conducted in 7 patients (38.9 %): right hepatectomy in 3, left hepatectomy in 3, and central bisectionectomy in 1. Minor hepatectomy was performed in 11 patients (61.1 %): lateral sectionectomy was performed in 2 and segmentectomy performed in 9. The median follow-up duration after staged hepatectomy was 68 months (range: 4–160 months). Staged Hepatectomy Complications Eight staged hepatectomy patients (44 %) experienced complications. Grade I complications developed in 6 patients and were managed conservatively: pleural effusion developed in 3, wound problems in 2, and ileus in 1. Grade II complications developed in two patients: fluid accumulation at the resected liver surface in 1 and acute renal failure in 1. The patient with fluid collection developed fever at postoperative day 6 and recovered after antibiotic optimization. The renal failure patient received intermittent diuretic therapy and recovered normal renal filtration rates.
123
World J Surg Fig. 1 Patient selection schematic
group showed significantly higher OS than the TACEalone group (P \ 0.001). Prognostic Factors for OS in the Staged Hepatectomy Group and the TACE-alone with Resectable Tumor Group
Fig. 2 Comparison of overall survival between the staged hepatectomy group and the transarterial chemoembolization (TACE)-alone with resectable tumor group
There were no complications above type IIIa; nor were there any postoperative deaths. Comparison of OS between the Staged Hepatectomy Group and the TACE-alone with Resectable Tumor Group The 1-, 3-, and 5-year OS rates were, respectively, 40.0 %, 0 %, and 0 % in TACE-alone with resectable tumor patients, and 88.5 %, 64.4 %, and 55.2 % in the staged hepatectomy patients (Fig. 2). The staged hepatectomy
123
The patient characteristics at HCC rupture diagnosis in both groups are summarized in Table 1. Patients in the TACEalone group had significantly higher total bilirubin levels and international normalized ratios than patients in the staged hepatectomy group. In addition, tumors were significantly larger in the TACE-alone group than in the staged hepatectomy group. On univariate analysis, tumor size and type of treatment were found to be significantly associated with OS (Table 2). However, in multivariate analysis, type of treatment was the only independent prognostic factor for OS. Comparison of Long-term Surgical Outcomes between Non-ruptured HCC Patients and Ruptured HCC Patients As noted earlier, propensity score-matching considered nine variables (Table 3). The two groups showed significant differences in tumor size and staging, as well as microvascular and gross vascular invasion, before matching. These differences between groups were eliminated after matching. Before matching, 1-, 3-, and 5-year OS rates were, respectively, 90.3 %, 68.7 %, and 57.2 % in non-ruptured HCC patients and 88.5 %, 64.4 %, and 55.2 % in staged hepatectomy patients. The 1-, 3-, and 5-year DFS rates were,
World J Surg Table 1 Patient characteristics at diagnosis of tumor rupture in the staged hepatectomy group versus the TACE-alone with resectable tumor group Variables
Staged hepatectomy (n = 18)
Age, years Male gender
53.0 (40–78) a
HBsAg positivitya Shock on presentationa
TACE-alone (n = 10) 66.5 (36–79)
P value 0.072
15 (83.3 %)
9 (90.0 %)
16 (88.9 %)
8 (80.0 %)
0.520
5 (50 %)
0.778
8 (44.4 %)
0.629
Hemoglobin, g/dL
11.4 (8.0–16.0)
10.0 (7.6–11.2)
0.002
Creatinine, mg/dL
0.9 (0.5–2.7)
1.0 (0.5–1.7)
0.796
AST, IU/L ALT, IU/L Platelets (/mm3) Total bilirubin, mg/dL INR ICG R15 (%)
49.5 (15–9) 45.5 (5–1) 170,000 (53,000–467,000)
82.0 (1–6) 55.5 (7–1) 161,500 (76,000–249,000)
0.121 0.245 0.796
0.8 (0.3–2.2)
1.2 (0.5–2.4)
0.041
1.07 (0.86–1.47)
1.27 (1.04–1.77)
0.014 0.159
7.0 (4.0–31.5)
9.0 (5.5–31.2)
Alphafetoprotein, IU/mL
106.8 (1.1–83,000)
285.0 (1.8–83,000)
0.356
PIVKA-II, mAU/mL
806.5 (13–20,498)
479.0 (40–2,000)
0.900
Tumor size, cm Tumor number
6.1 (3.0–15.0) 1 (1–3)
8.0 (6.0–13.0) 1 (1–3)
0.045 0.944
Values in parentheses are median with range unless indicated otherwise a
Values are number with percentages
HCC Hepatocellular carcinoma, TACE Transarterial chemoembolization, HBsAg Hepatitis B virus surface antigen, AST Aspartate aminotransferase, ALT Alanine aminotransferase, INR International normalized ratio, ICG R15 Indocyanine green retention rate at 15 min, AFP Alphafetoprotein, PIVKA-II Protein induced by vitamin K absence or antagonist-II
respectively, 64.3 %, 42.8 %, and 36.5 % in non-ruptured HCC patients and 54.9 %, 39.2 %, and 7.8 % in staged hepatectomy patients. The DFS rates were significantly higher in the non-ruptured HCC group than in the staged hepatectomy group (P = 0.019), although OS rates were not significantly different between the two groups (Fig. 3a). After matching, 1-, 3-, and 5-year OS rates were, respectively, 89.0 %, 55.1 %, and 46.5 % in the matched control group and 88.5 %, 64.4 %, and 55.2 % in the staged hepatectomy group. The 1-, 3-, and 5-year DFS rates were, respectively, 55.8 %, 35.5 %, and 35.5 % in the non-ruptured HCC group and 54.9 %, 39.2 %, and 7.8 % in the staged hepatectomy group. No significant intergroup differences in OS and DFS rates were observed between the two matched groups (Fig. 3b). In the staged hepatectomy group, HCC recurrence developed in 14/18 patients (77.8 %), and the peritoneal recurrence rate was 14.3 % (Table 4). In the matched control group, HCC recurrence developed in 20/37 patients (54.1 %), and the peritoneal recurrence rate was 10.0 %. There was no significant difference in recurrence pattern between the two matched groups.
Discussion Comparing surgical outcomes between spontaneously ruptured HCC and non-ruptured HCC is difficult because
of different patient characteristics at diagnosis. Most patients with spontaneously ruptured HCC had higher tumor stage and poorer liver function than non-ruptured HCC patients. Thus, hepatectomy in spontaneously ruptured HCC show lower survival outcomes than hepatectomy in non-ruptured HCC. Liu et al [10] demonstrated that hepatectomy in ruptured HCC shows lower OS and DFS rates than hepatectomy in non-ruptured HCC. Yeh et al [5] reported no significant difference in OS rates after hepatectomy between ruptured HCC and non-ruptured HCC, although non-ruptured HCC show significantly better DFS rates. However, previous studies did not adjust for patient characteristics that affect prognosis after surgery, and required recalibration of patient characteristics between ruptured HCC and non-ruptured HCC groups. Mizuno et al [9] compared survival rates of patients with ruptured and non-ruptured HCC based on the same background factors, and they found no significant differences between groups in OS and DFS. However, this may not be an accurate comparison between the groups because ruptured HCC patients were compared with non-ruptured HCC patients with stage IVa disease (AJCC TNM classification T4N0/1M0). Although tumor rupture is considered T4 stage, most resectable ruptured HCCs are small or mediumsized solitary tumors. Likewise, when the factor of tumor rupture is excluded, there were no T4 tumors in ruptured HCCs treated with staged hepatectomy in our study.
123
World J Surg Table 2 Cox regression analysis of prognostic factor for overall survival in the staged hepatectomy group and the TACE-alone with resectable tumor group (n = 28)
Variable
Category
Overall survival Hazard ratio
95 % Confidence interval
P value
1.035
0.996–1.076
0.076
Univariate analysis Age (1-year increase) Gender
Female vs male
1.762
0.400–7.762
0.454
HBsAg positivity
No vs yes
1.524
0.345–6.727
0.578
Shock on presentation
No vs yes
0.732
0.258–2.075
0.557
Hemoglobin B 10 g/dL
No vs yes
1.944
0.701–1.944
0.202
Creatinine [ 1.2 mg/dL Serum platelet B 100,000 mm3
No vs yes No vs yes
1.100 0.995
0.245–4.950 0.324–3.059
0.901 0.993
Total bilirubin [ 1.1 mg/dL
No vs yes
1.835
0.721–4.667
0.203
INR [ 1.4
No vs yes
1.680
0.462–6.104
0.431
ICG R15 [ 10 %
No vs yes
AFP (continuous), IU/mL
2.183
0.645–7.382
0.209
1.000
0.998–1.002
0.906
PIVKA-II (continuous), mAU/mL
1.000
0.976–1.024
0.981
Tumor size (continuous), cm
1.177
1.012–1.368
0.034
Tumor number
Solitary vs multiple
1.279
0.493–3.316
0.613
Type of treatment
Staged hepatectomy vs TACE-alone
8.635
2.172–34.336
0.002
1.169 7.876
0.983–1.390 1.939–31.996
0.077 0.004
Multivariate analysis Tumor size (continuous), cm Type of treatment
Staged hepatectomy vs TACE-alone
Table 3 Baseline variables in the non-ruptured HCC with hepatectomy group versus the ruptured HCC with staged hepatectomy group Variable
Category
Before matching Non-ruptured HCC (n = 326)
a
After matching Staged hepatectomy (n = 18)
P value
Non-ruptured HCC (n = 37)
Staged hepatectomy (n = 18)
P value
53 (22–72)
53 (40–78)
0.647
52 (32–71)
53 (40–78)
0.831
254 (77.9 %)
15 (83.3 %)
0.588
28 (75.7 %)
15 (83.3 %)
0.731
HBV
259 (79.4 %)
16 (88.9 %)
0.543
31 (83.8 %)
16 (88.9 %)
1.000
HCV
13 (4.0 %)
0 (0 %)
0 (0 %)
0 (0 %)
No hepatitis
54 (16.6 %)
2 (11.1 %)
6 (16.2 %)
2 (11.1 %)
Tumor size, cma
4.5 (0.9–18.0)
6.1 (3.0–15.0)
0.047
6.5 (0.9–18.0)
6.1 (3.0–15.0)
0.905
Multiple tumors
108 (33.1 %)
6 (33.3 %)
0.986
20 (54.1 %)
6 (33.3 %)
0.149
Cirrhosis
176 (54 %)
11 (61.1 %)
0.543
21 (56.8 %)
11 (61.1 %)
0.759 0.505
Age, years
Male gender Etiology
Gross vascular invasion
44 (13.5 %)
8 (44.4 %)
0.000
13 (35.1 %)
8 (44.4 %)
Microvascular invasion
294 (90.2 %)
12 (66.7 %)
0.002
26 (70.7 %)
12 (66.7 %)
0.786
II III
273 (83.7 %) 47 (14.4 %)
4 (22.2 %) 14 (77.8 %)
0.000
9 (24.3 %) 28 (75.7 %)
4 (22.2 %) 14 (77.8 %)
0.403
IV
6 (1.8 %)
Tumor stage
0 (0 %)
Values in parentheses are percentages unless indicated otherwise a
Values are median (range)
HCC Hepatocellular carcinoma, HBV Hepatitis B virus, HCV Hepatitis C virus
123
0 (0 %)
0 (0 %)
World J Surg Fig. 3 Comparison of diseasefree survival (DFS) and overall survival (OS) between staged hepatectomy in the ruptured hepatocellular carcinoma (HCC) group and hepatectomy in the non-ruptured HCC group
Table 4 Recurrence pattern in the staged hepatectomy group versus the matched control group Recurrence pattern
Group
P value
Staged hepatectomy (n = 14)
Matched controls (n = 20)
Intrahepatic
7 (50.0 %)
10 (50.0 %)
Lung
2 (14.3 %)
6 (30.0 %)
Peritoneal
2 (14.3 %)
2 (10.0 %)
Lymph node
3 (21.4 %)
2 (10.0 %)
0.632
Therefore it seems optimal that ruptured HCCs were compared to non-ruptured HCCs with similar tumor characteristics excluding the rupture factor. In this regard, our study redefined tumor characteristics by excluding the factor of tumor rupture, and the patients were restricted to those characterized as Child-Pugh class A. We performed propensity score-matching to overcome differences in patient characteristics that affect postoperative prognosis between the ruptured HCC and non-ruptured HCC groups. After matching, ruptured HCC showed no significant differences in OS and DFS than non-ruptured HCC. Thus, ruptured HCC patients may have long-term survival
comparable to those with non-ruptured HCC, with similar tumor characteristics and liver functional status. There are many reports of peritoneal recurrence after spontaneous HCC rupture [19–22], and some reports insist that spontaneous HCC rupture evokes peritoneal metastasis [3]. However, it remains debatable whether spontaneous rupture of HCC increases peritoneal recurrence rates and decreases positive long-term surgical outcomes. Lin et al [23] reported that most patients with peritoneal metastasis after hepatectomy exhibit no prior evidence of overt HCC rupture. Mizuno et al [9] demonstrated that the recurrence pattern after hepatectomy is similar between ruptured HCC and non-ruptured HCC. Our results also demonstrate no difference in recurrence pattern between ruptured HCC patients and matched non-ruptured HCC controls. Peritoneal lavage is important in reducing peritoneal recurrence rates; however, tumor control by staged hepatectomy may play a larger role. Cheng et al [11] assessed the AJCC 7 staging system validity for resectable HCC prognosis, and tumor rupture does not influence survival after hepatectomy in stage III patients. Kwak et al [24] reported that peritoneal metastasis is not an independent prognostic factor of HCC, and maintenance of
123
World J Surg
favorable hepatic function and controlling intrahepatic HCC are more important prognostic factors. These findings are consistent with our observations that patients in the TACE-alone with resectable tumor group had significantly lower OS rates than those in the staged hepatectomy group. The TACE-alone group had Child-Pugh grade A liver function, and the tumor size was not a crucial factor for resectability. Nonetheless, poor OS rates in the TACE-alone group might be attributed in part to lower baseline liver function and a higher proportion of larger tumors. However, in multivariate analysis of prognostic factors for OS, type of treatment was identified as the only independent prognostic factor. This finding suggests that surgical resection may be the most important factor for long-term survival in patients with ruptured HCC. Consequently, in patients with preserved liver function and resectable tumors, staged hepatectomy should be considered as a potential approach for achieving tumor control and long-term survival. Recently, Hai et al [4] advocated emergency hepatectomy for achieving both hemostasis and definitive treatment in a single operation; mortality rates were 5.8 %. However, morbidity rates were not described and lower mortality rates resulted because emergency hepatectomy was only performed in patients with peripherally located tumors. Ruptured HCC patients have acutely damaged livers and abnormal coagulation profiles [10, 25, 26]. Therefore, emergency hepatectomy may diminish residual hepatic function, leading to postoperative bleeding and liver failure. Lai et al [27] reported that emergency hepatectomy in spontaneously ruptured HCC shows a 71.4 % mortality rate, and hepatic artery ligation has 76.6 % mortality. Vergara et al [28] reported that mortality and morbidity rates after emergency hepatectomy in spontaneously ruptured HCC were 16.5 and 50 %, respectively. In our study, one patient died of postoperative liver failure and one patient died of postoperative bleeding among the three emergency laparotomy patients. Thus, emergency surgery for ruptured HCC may be associated with high morbidity and mortality. We observed no complications above type IIIa after staged hepatectomy. Full preoperative clinical evaluations indicated major hepatectomy for seven patients (38.9 %). Thus, staged hepatectomy is facilitated by full preoperative clinical evaluation, thereby allowing safe and curative treatment in ruptured HCC versus emergency hepatectomy. This study has several limitations. It was performed in single center with a small sample size. In addition, it was designed retrospectively, so selection bias is possible. Although propensity score-matching reduced potential selection bias, the possibility of data loss still exists in the matching process [17, 29].
123
In summary, spontaneously ruptured HCC with staged hepatectomy shows comparable long-term survival and recurrence pattern versus non-ruptured HCC, with similar tumor characteristics and liver functional status. Spontaneous HCC rupture may not increase peritoneal recurrence rates and decrease long-term survival after hepatectomy. Additionally, spontaneously ruptured HCC may need staging reappraisal according to its own tumor characteristics and excluding the rupture factor, when patients have resectable tumors and preserved liver function.
Conflict of interest No author has any conflict of interest to declare, financial or otherwise. Funding
None
References 1. Lai EC, Lau WY (2006) Spontaneous rupture of hepatocellular carcinoma: a systematic review. Arch Surg 141:191–198 2. Edge SB, Byrd DR, Compton CC et al (2009) AJCC cancer staging manual, 7th edn. Springer, New York 3. Sonoda T, Kanematsu T, Takenaka K et al (1989) Ruptured hepatocellular carcinoma evokes risk of implanted metastases. J Surg Oncol 41:183–186 4. Hai L, Yong-Hong P, Yong F et al (2005) One-stage liver resection for spontaneous rupture of hepatocellular carcinoma. World J Surg 29:1316–1318. doi:10.1007/s11805-007-0322-2 5. Yeh CN, Lee WC, Jeng LB et al (2002) Spontaneous tumour rupture and prognosis in patients with hepatocellular carcinoma. Br J Surg 89:1125–1129 6. Leung CS, Tang CN, Fung KH et al (2002) A retrospective review of transcatheter hepatic arterial embolisation for ruptured hepatocellular carcinoma. J R Coll Surg Edinb 47:685–688 7. Kim JY, Lee JS, Oh DH et al (2012) Transcatheter arterial chemoembolization confers survival benefit in patients with a spontaneously ruptured hepatocellular carcinoma. Eur J Gastroenterol Hepatol 24:640–645 8. Buczkowski AK, Kim PT, Ho SG et al (2006) Multidisciplinary management of ruptured hepatocellular carcinoma. J Gastrointest Surg 10:379–386 9. Mizuno S, Yamagiwa K, Ogawa T et al (2004) Are the results of surgical treatment of hepatocellular carcinoma poor if the tumor has spontaneously ruptured? Scand J Gastroenterol 39:567–570 10. Liu CL, Fan ST, Lo CM et al (2001) Management of spontaneous rupture of hepatocellular carcinoma: single-center experience. J Clin Oncol 19:3725–3732 11. Cheng CH, Lee CF, Wu TH et al (2011) Evaluation of the new AJCC staging system for resectable hepatocellular carcinoma. World J Surg Oncol 9:114 12. Yoshida H, Onda M, Tajiri T et al (1999) Treatment of spontaneous ruptured hepatocellular carcinoma. Hepatogastroenterology 46:2451–2453 13. Shimada R, Imamura H, Makuuchi M et al (1998) Staged hepatectomy after emergency transcatheter arterial embolization for ruptured hepatocellular carcinoma. Surgery 124:526–535 14. Zhu LX, Wang GS, Fan ST (1996) Spontaneous rupture of hepatocellular carcinoma. Br J Surg 83:602–607
World J Surg 15. Choi GH, Park JY, Hwang HK et al (2011) Predictive factors for long-term survival in patients with clinically significant portal hypertension following resection of hepatocellular carcinoma. Liver Int 31:485–493 16. Clavien PA, Barkun J, de Oliveira ML et al (2009) The ClavienDindo classification of surgical complications: five-year experience. Ann Surg 250:187–196 17. D’Agostino RB Jr (1998) Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 17:2265–2281 18. Braitman LE, Rosenbaum PR (2002) Rare outcomes, common treatments: analytic strategies using propensity scores. Ann Intern Med 137:693–695 19. Yunoki Y, Takeuchi H, Makino Y et al (1999) Intraperitoneal seeding of ruptured hepatocellular carcinoma: case report. Abdom Imaging 24:398–400 20. Okano J, Shiota G, Horie Y et al (1996) Rupture of metastatic nodule on the peritoneal surface secondary to hepatocellular carcinoma. Intern Med 35:783–784 21. Eriguchi N, Aoyagi S, Okuda K et al (2000) Successful surgical treatment for implanted intraperitoneal metastases of hepatocellular carcinoma. J Hepatobiliary Pancreat Surg 7:520–523 22. Lin CC, Chen CH, Tsang YM et al (2006) Diffuse intraperitoneal metastasis after spontaneous rupture of hepatocellular carcinoma. J Formos Med Assoc 105:577–582
23. Lin CC, Liang HP, Lee HS et al (2009) Clinical manifestations and survival of hepatocellular carcinoma patients with peritoneal metastasis. J Gastroenterol Hepatol 24:815–820 24. Kwak MS, Lee JH, Yoon JH et al (2012) Risk factors, clinical features, and prognosis of the hepatocellular carcinoma with peritoneal metastasis. Dig Dis Sci 57:813–819 25. Miyamoto M, Sudo T, Kuyama T (1991) Spontaneous rupture of hepatocellular carcinoma: a review of 172 Japanese cases. Am J Gastroenterol 86:67–71 26. Hsueh KC, Fan HL, Chen TW et al (2012) Management of spontaneously ruptured hepatocellular carcinoma and hemoperitoneum manifested as acute abdomen in the emergency room. World J Surg 36:2670–2676. doi:10.1007/s00268-012-1734-6 27. Lai EC, Wu KM, Choi TK et al (1989) Spontaneous ruptured hepatocellular carcinoma. an appraisal of surgical treatment. Ann Surg 210:24–28 28. Vergara V, Muratore A, Bouzari H et al (2000) Spontaneous rupture of hepatocelluar carcinoma: surgical resection and longterm survival. Eur J Surg Oncol 26:770–772 29. Austin PC, Grootendorst P, Anderson GM (2007) A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: a Monte Carlo study. Stat Med 26:734–753
123
