Introduction

Hepatolithiasis, intrahepatic duct stones, is frequent in Asian countries and also it is one of the major risk factors for cholangiocarcinoma [1, 2]. The overall incidence of cholangiocarcinoma in association with hepatolithiasis has been reported to be 4–12 % [3–5]. Intrahepatic stone makes chronic recurrent cholangitis and biliary carcinogenesis is caused by prolonged inflammation of the biliary epithelium secondary to bile duct stones, bile stasis, and bacterial infections [6–8].

Stone removal using percutaneous transhepatic choledochoscopic lithotomy (PTCSL) can resolve the cholangitis due to stone [9] and also surgical treatment such as hepatic resection can eliminate the inflammatous hepatic segment including stones [10]. Despite the various efforts to remove the stones and advances in diagnostic modality, early diagnosis of cholangiocarcinoma is still challenging in these patients [11].There have been insufficient data regarding the long-term outcome of hepatectomy as the treatment modality of hepatolithiasis. Furthermore, the literature contains few data for occurrence of cholangiocarcinoma after treatment of hepatolithiasis.

This study was aimed to evaluate the incidence of cholangiocarcinoma and clinical characteristics after treatment in hepatolithiasis patients.

Material and methods

Patients

We retrospectively reviewed that the patients were treated for hepatolithiasis from January 2004 to December 2013, at Korea University Anam and Guro Hospital, Korea University College of Medicine, Seoul, Korea.

Among 257 patients included in this study, 236 patients were eligible for analysis and total 21 patients were excluded. Exclusion criteria were follow-up period less than 9 months, preoperative diagnosis of cholangiocarcinoma, occurrence of cholangiocarcinoma within 1 year after treatment.

All patients were received adequate procedures including hepatic resection for stone removal. Hepatic resection was preferred in the patients with atrophy of involved liver segments or lobe on initial CT. PTCSL and endoscopic retrograde cholangioscopy were offered in all indicated patients during follow-up period. As medical treatment, broad spectrum antibiotics were used when cholangitis is suspected.

Completeness of stone clearance was defined when there was no intrahepatic duct stone during whole follow-up period. Either remnant or new stone was regarded as residual stone. Recurrence of stone was defined to be detected on previously treated bile duct. The definition of development of cholangiocarcinoma was occurrence of tumor in hepatic lobe with stone.

Patients were followed up 1 year after treatment and then whenever cholangitis was recurred. Abdominal CT or MRI was taken before treatment and 1 year after. In case of hepatic resection, additional exam was taken within 1 year after operation.

We compared the two groups during the same study period in regard to the patients’ characteristics, complete stone clearance, and occurrence of cholangiocarcinoma as the long-term outcomes.

Statistical analysis

The t test and Mann–Whitney U test were used for continuous variables. Categorical variables were tested by the Person χ 2 test. The cumulative rates of cholangiocarcinoma were analyzed by the Kaplan–Meier method and compared using the log-rank test. Cox proportional hazards regression model (variables: age, gender, CA19-9, location of stone, bile duct stenosis, liver atrophy, residual stone, stone recurrence, and liver resection) was used for risk development of cholangiocarcinoma. p values less than 0.05 were considered statistically significant.

Results

Patient characteristics

Ninety-five patients underwent hepatic resection (resection group) and 141 patients were treated without hepatic resection (non-resection group). Mean follow-up duration was 41 ± 41 months.

We compared the two groups with regard to the clinical characteristics. The patient’s characteristics are described in Table 1. Mean age was significantly younger in hepatic resection group than the non-resection group. There was no significant difference in level of CA 19-9 and symptoms between the two groups.

Table 1 Patient characteristics of the hepatic resection and non-resection groups
Full size table

Intrahepatic duct stone location

The most common site of the intrahepatic duct stone was left hepatic lobe. Stone distribution in liver was 54.7 % (129/236) in the left lobe, 23.7 % (56/236) in the right lobe, and 21.6 % (51/236) in both lobes. Stone location in non-resection group and resection group was 44.7 % (63/141) and 69.5 % (66/95)  % in the left lobe, 31.9 % (45/141) and 11.6 % (11/95) in the right lobe, 23.4 % (33/141) and 18.9 % (18/95) in both lobes, respectively. Left intrahepatic duct stone was more frequent in resection group (p < 0.001).

Treatment for intrahepatic duct stone

In 141 patients of non-resection group, PTCSL was done in 62 patients, CBD exploration with choledochoscopic examination in 9 patients, and intraoperative choledochoscopic examination in 4 patients. Bilio-enteric anastomosis was done in 4 cases (choledochojejunostomy). Thirty-three patients had stones in both hepatic lobes at initially, and new intrahepatic duct stones were found in 50 patients during follow-up. These 83 patients (59 %) were considered to have residual stones.

In hepatic resection group, left hemihepatectomy or lobectomy was frequently used in 87.4 % (83/95) and right lobectomy or segmentectomy was done 12.6 % (12/95). Choledochojejunostomy was done in 3 patients. Peri-operative PTCS was performed in 10 patients of 18 patients with stones in both hepatic lobes. Of 18 patients with both hepatic lobe stones, only 8 patients had residual stones constantly after resection. One patient who failed to remove due to stricture underwent PTCSL after operation. In 10 patients, new stone was detected in other hepatic segment after surgery. Nine patients with both hepatic lobes stones had recurrence after PTCSL. Total 28 patients (29 %) were considered to have residual stone in resection group. There was no peri-operative death.

Complete stone clearance was possible in 41 % (58/141) of non-resection group and 71 % (67/95) of resection group. Residual stone rate was significantly low in resection group. (p < 0.001) There was no significant difference in bilio-enteric anastomosis of treatment (p = 0.887).

Incidence of cholangiocarcinoma during follow-up

Among the 236 patients with hepatolithisis, intrahepatic cholangiocarcinoma occurred in total 16 patients (6.8 %) (Fig. 1).The clinical features are summarized in Table 2. The median tumor occurrence time was 28 months (13–111 months). Among 16 patients with cholangiocarcinoma, 6 patients (6.3 %) developed in hepatic resection group and 10 patients (7.1 %) developed in non-resection group. There was no significant difference in the cholangiocarcinoma incidence between the two groups (p = 0.254) (Fig. 2).

Fig. 1
figure 1

Cumulative incidence of cholangiocarcinoma during follow-up

Full size image
Table 2 Clinical characteristics of cholangiocarcinoma patients
Full size table
Fig. 2
figure 2

Cholangiocarcinoma proportion of hepatic resection group and non-resection group

Full size image

Cholangiocarcinoma according to completeness of stone clearance

When analyzed according to completeness of stone clearance regardless of treatment modality, cholangiocarcinoma occurred in four patients among the 125 patients with complete stone clearance (3.2 %). On the other hand, 12 patients developed cholangiocarcinoma among the 111 patients with incomplete stone clearance (10.8 %). Cholangiocarcinoma incidence was high in patients with residual stone, but there is no difference in survival between two groups by Kaplan–Meier log-rank test (p = 0.263) (Fig. 3).

Fig. 3
figure 3

a Kaplan–Meier one minus survival curves for cholangiocarcinoma. b One minus survival curves according to residual stone

Full size image

In hepatic resection group, cholangiocarcinoma occurrence was significantly low in patients with complete stone clearance (3 %, 2/67) compared with incomplete stone clearance (14.2 %, 4/28) (p = 0.042). However, cholangiocarcinoma developed even after complete stone clearance in two patients, one in caudate lobe 14 months after left hemihepatectomy and the other in medial segment after left hemihepatectomy, 28 months later.

Also in non-resection group, cholangiocarcinoma occurred in 3.4 % (2/58) of patients with complete stone clearance and 9.6 % (8/83) of patients with incomplete stone clearance, there is no difference in survival between two groups (p = 0.510).

Risk factors associated with cholangiocarcinoma

In cox regression analysis, residual stone and hepatic resection did not show the relation with incidence of cholangiocarcinoma. Only CA 19-9 was found to be significant risk factor for survival (Table 3). However, none of the factors was found to be significant in multivariate analysis.

Table 3 Univariate analysis of risk factors for cholangiocarcinoma
Full size table

Discussion

Although there have been an effort to manage stones effectively including hepatic resection, hepatolithiasis is associated with a recurrent cholangitis, biliary stricture, and stone formation. Furthermore, the biliary carcinogenesis is caused by prolonged inflammation through recurrent hepatolithiasis and cholangitis. Epidemiological, pathological, and genetic studies have showed the relationship between hepatolithiasis and cholangiocarcinoma [6–8]. The presence of cholangiocarcinoma-associated hepatolithiasis is the main factor compromising long-term survival in patients with hepatolithiasis [10] but unfortunately early diagnosis of concomitant cholangiocarcinoma is difficult [11]. Hepatic resection is frequently considered for the definitive treatment of hepatolithiasis, because it can remove stones and stenotic bile duct simultaneously, reduce the risk of recurrent stones and development of cholangiocarcinoma [12–15]. Nevertheless, subsequent late development of cholangiocarcinoma may appear months to years after the initial hepatectomy [10, 16, 17].

To the best of our knowledge, few studies on the comparing of treatment modality for hepatolithiasis reported that surgical treatment showed better long-term outcome [9, 18]. But it is not clear whether hepatic resection could reduce the occurrence of cholangiocarcinoma.

We analyzed the data of patients who were treated for hepatolithiasis. Long-term outcomes especially occurrence of intrahepatic cholangiocarcinoma were compared according to treatment modality. During the 41 months of mean follow-up period, total 16 patients (6.8 %) developed cholangiocarcinoma. No significant difference was proven in hepatic resection group and non-hepatic resection group (6.3 vs. 7.1 %, p = 0.526). Furthermore, there was no difference in clinical manifestation and laboratory data to predict the development of malignancy between the two groups. Hepatic resection did not show any advantage for cholangiocarcinoma in this study, especially if incomplete clearance of stones was performed. Although it was generally agreed that hepatic resection could remove not only stones and ductal stricture but also premalignant pathologic change and concomitant cholangiocarcinoma; however, it did not showed any difference in survival according to subsequent cholangiocarcinoma in this study. Furthermore, cholangiocarcinoma developed in two patients with complete stone clearance. In these cases, cancer was developed in near hepatic lobe from resected hepatic segment. This may be caused that any premalignant change such as biliary intraepithelial neoplasia exists in nearby bile duct. Long-term inflammation derived from neighboring segment stones may promote liver carcinogenesis and aggressive hepatic resection including neighboring segments may prevent cholangiocarcinoma.

Incidence of cholangiocarcinoma was higher in patients with residual stone than patients without residual stone (10.8 vs. 3.2 %, p = 0.054). This finding was similar in hepatic resection group and non-resection group (14.2 vs. 3 %, 9.6 vs. 3.4 %, respectively) but significant difference was shown only in hepatic resection group (p = 0.042). This suggests that complete stone removal is important in the prevention of cholangiocarcinoma. Previous studies on long-term outcomes after treatment reported that residual stone and bile duct stricture are the risk factors for complication [9].

But the treatment strategy for complete stone removal is not simple especially in patients with both hepatic lobes stones. Although it looks like high proportion of the patients with residual stone in resection group, purely new stone detection was in 10 patients (10.5 %) because 8 patients had stones in both hepatic lobes. In our cases, 4 patients with residual stones after hepatic resection had all both hepatic lobe stones and cholangiocarcinoma developed in right hepatic lobe. This fact would decrease the positive effect of hepatic resection in survival. This may cause the different results compared with other reports. We think that the incidence of cholangiocarcinoma would be low in patients with complete stone clearance and it yields the importance of complete stone removal in hepatolithiasis, which reduce the incidence of cholangiocarcinoma and will increase the survival.

Limitation of this study is the different follow-up period between two groups. And 41 months would be short interval to clarify the survival difference and the risk factor of cholangiocarcinoma. Second, this study was limited by retrospective nature.

In summary, hepatic resection for hepatolithiasis is considered to have a limited value in preventing the risk of cholangiocarcinoma. Neither did for survival after treatment. Hepatic resection should be offered to selected patients and careful follow-up for cholangiocarcinoma is needed even after hepatic resection. A combination of different treatment modalities is necessary to decrease the residual stone and improve the outcome of the patients with hepatolithiasis.