AIM: To identify the influence of the surgery type and prognostic factors in middle and distal bile duct cancers.

METHODS: Between August 1990 and June 2011, data regarding the clinicopathological factors of 194 patients with surgical and pathological confirmation were collected. A total of 133 patients underwent resections (R0, R1, R2; n = 102, 24, 7), whereas 61 patients underwent nonresectional surgery. Either pancreaticoduodenectomy (PD) or bile duct resection (BDR) was selected according to the sites of tumors and co-morbidities of the patients after confirming resection margin by the frozen histology in all cases. Univariate and multivariate analyses of clinicopathologic factors were performed, utilizing the Kaplan-Meyer method and Cox hazard regression analysis.

RESULTS: The overall 5-year survival rate for the 133 patients who underwent resection (R0, R1, and R2) was 41.2%, whereas no patients survived longer than 3 years among the 61 patient who underwent nonresectional surgeries. The 5-year survival rate of the patients who underwent a PD (n = 90) was higher than the rate of those who underwent BDR (n = 43), although the difference was not statistically significant (46.6% vs 30.0% P = 0.105). However, PD had a higher rate of R0 resection than BDR (90.0% vs 48.8%, P < 0.0001). If R0 resection was achieved, PD and BDR showed similar survival rates (49.4% vs 46.5% P = 0.762). The 5-year survival rates of R0 and R1 resections were not significantly different (49.0% vs 21.0% P = 0.132), but R2 resections had lower survival (0%, P = 0.0001). Although positive lymph node, presence of perineural invasion, presence of lymphovascular invasion (LVI), 7^th AJCC-UICC tumor node metastasis (TNM) stage, and involvement of resection margin were significant prognostic factors in univariate analysis, multivariate analysis identified only TNM stage and LVI as independent prognostic factors.

CONCLUSION: PD had a greater likelihood of curative resection and R1 resection might have some positive impact. The TNM stage and LVI were independent prognostic factors.

Cholangiocarcinoma is a rare malignant tumor of the biliary tree that can arise anywhere from the intrahepatic to the extrahepatic bile duct just proximal to the duodenal ampulla, excluding the gallbladder. Extrahepatic bile duct cancer, defined as the presence of malignant tumors arising at the biliary tree distal to second-order branches, accounts for 80% to 90% of all cholangiocarcinomas. Extrahepatic bile duct cancer can be further divided into hilar or middle/distal bile duct cancers. Among these extrahepatic bile duct cancers, middle and distal bile duct cancers comprise approximately 20% to 30%[1-4].

The site of an extrahepatic bile duct cancer has clinical importance because it affects the selection of the appropriate type of surgical resection and the outcomes after surgery. Compared to hilar bile duct cancer, which requires concomitant bile duct and liver resection, the surgical resection for a middle and distal bile duct cancer requires either segmental bile duct resection or pancreaticoduodenectomy with lymph node dissection. The type of surgery selected depends on the possibility of achieving tumor-free resection margins, and the prognostic influence of the type of resection on long-term prognosis has been controversial[5,6].

In some patients, the superficial spreading nature of the disease, as well as comorbid diseases, can make it hard to achieve R0 resection, inevitably resulting in R1 resection. Because such cases have been anecdotal, the effect of R1 resection on long-term prognoses must be investigated.

Factors such as tumor node metastasis (TNM) stage, perineural invasion, and lymphovascular invasion have been reported to affect survival, but their influences on prognosis have not always been universal. Tumors with higher T stages tend to run a greater risk of distant metastasis and poorer prognosis, but in some cases, tumors are confined to the bile duct without lymph node metastasis, showing early recurrence in distant areas, such as the liver, lungs, or multiple bones. The unpredictable prognoses of such cases might be explained by bile duct cancer progressing not only by directly invading into the depths of the bile duct wall but also into the perineural, vascular, and lymphatic spaces[7]. Nagahashi et al[8] recently insisted that the presence of lymphovascular invasion resulted in poor prognosis, comparing a pT1 tumor invading the fibromuscular layer to a pT1 tumor confined to the mucosa. The role of such factors as perineural invasion (PNI) and low viscosity index (LVI) must be investigated.

In this study, we analyzed the clinicopathological data of 133 patients who underwent surgical resection for middle and distal bile duct cancers, to identify the influence of the type of surgery selected and of clinicopathological factors on the long-term prognoses of the patients.

Middle and distal bile duct cancers are rare malignancies with poor prognoses, and only complete surgical resection of bile duct cancer offers a chance for long-term survival. The prognosis of middle and distal bile duct cancers remains poor, even with radical resection, due to the high incidence of recurrence. Furthermore, it is difficult to determine the optimal extent of resection and to identify the prognostic factors. Although several prognostic factors, such as surgical radicality, nodal status, depth of invasion, differentiation, perineural invasion, and lymphovascular invasion, have been reported for middle and distal extrahepatic bile duct cancers[9-13], the prognostic values of these factors have not been consistent. Among the aforementioned factors, nodal metastasis, differentiation, and R0 resection have been widely recognized to be associated with long-term survival[3,14-17].

In this study, the overall survival was longer in patients with resection, compared to patients without resection. However, the prognosis of patients with R2 resection was much poorer compared that of patients with R0 or R1 resection. As shown in the results, patients with R0 and R1 resections showed 5-year survival rates of 49.0% and 21.0%, respectively, whereas none of the patients with nonresection or R2 resection survived longer than 3 years (medians of 10 mo and 12 mo , respectively).

Many studies have reported that R0 resection is necessary for long-term survival[5,17-19]. However, it is sometimes very difficult to achieve a tumor-free margin of the bile duct due to superficial microscopic spreading and multifocal tumors. In such cases, R1 resection is inevitable, and some authors have reported that patients with R1 resections have survived longer than expected[3,5,6,20].

In our study, R1 resection was performed in 24 patients, including eight patients with R1 PD and 16 patients with R1 BDR. Positive uppermost bile duct margins accounted for the 20 patients with R1 resections (8 PD and 12 BDR), and positive lowermost bile duct margins accounted for the four patients with R1 BDR resections, for whom the addition of further PD was precluded by the presence of co-morbidities. The 5-year survival rates after R0-resection were higher than those after R1-resection, but the difference did not reach statistical significance (49.0% vs 21.0%; P = 0.132) (Lymphovascular. 2). Thus, it can be assumed that R1 resection, if inevitable, should be recognized as prolonging survival, according to our data, which reflected that no patients with R2 resections or nonresections survived longer than 36 mo The definition of RI resection in this study was a microscopically positive margin of the resected bile duct. The major drawbacks of this study were that the pathological description was not made by a single pathologist and that the invasiveness of the positive margin was not sub-classified into carcinoma in situ or invasive carcinoma. These means of analysis will be further investigated in the near future.

According to the primary site and the extent of the tumor, different surgical procedures can be applied. PD is most commonly performed for tumors because bile duct resection is sometimes not sufficient to achieve adequate surgical margins, compared to PD[5,6]. In our study as well, the microscopic involvement of margins was more frequent after bile duct resection, compared to PD [37.2% (16/43) vs 8.9% (8/90); P < 0.0001]. The 5-year survival rates of patients with PD and BDR were 46.6% and 30.0% (P = 0.105), respectively; this difference, although not statistically significant, can be attributed to a higher rate of positive resection margins in patients with bile duct resection. The lack of statistical significance in the difference between the two groups might have been due to the limited number of patients. However, if R0 resection had been achieved, the 5-year survival rates after bile duct resection would have been similar after PD (46.5% vs 49.4%; P = 0.762) (Figure 3A). Thus, it is possible that PD offers a greater likelihood of complete resection and, consequently, better survival. Therefore, PD is recommended unless patient co-morbidities preclude its implementation. BDR is an option only if the co-morbidities of patients preclude PD or if an R0 resection seems to be achieved for tumors limited to the area surrounding the cystic duct.

The frequency of lymph node metastasis has been reported to range from 23.8% to 68%, and lymph node involvement has been determined to be an important predictor of survival in patients with middle and distal bile duct cancers[3,6,16,19,21,22]. In our study, the frequency of nodal metastasis was 25%, and patients with nodal metastasis had worse survival than patients without nodal metastasis in univariate analysis (5-year survival rate 17.0% vs 51.3%; P = 0.033)

Perineural invasion is one of the pathways through which local infiltrations spread and metastasize. Although perineural invasion has recently been accepted as a prognostic factor in a number of different malignancies, its clinical significance for mid-distal extrahepatic cholangiocarcinoma remains unclear. According to Bhuiya et al[7], perineural invasion had a profound impact on the survival of patients with extrahepatic bile duct cancer. The 5-year survival rate for patients with perineural invasion was 32%, compared to 67% for patients without invasion. A distinct and significant correlation was potentially reported between the depth of tumor invasion and perineural invasion. In our study as well, the frequency of perineural invasion was 41.4%, and the rates of perineural invasion for T1, T2 and T3 tumors were 18.4%, 60% and 46.2%, respectively. Thus, a significant association with depth of invasion was shown (P = 0.017). The survival of patients with perineural invasion was worse than that of patients without perineural invasion in univariate analysis (5-year survival rate 17.5% vs 55.8%; P = 0.022).

Lymphovascular channel invasion is one of many different manners by which tumors spread. The clinical significance of LVI was first described as far back as 1967, when studies reported higher recurrence and poorer survival in cervical cancer patients with lymphovascular invasion[23]. Nevertheless, few reports have been published regarding the significance of lymphovascular invasion, and its clinical significance has not been established in middle and distal bile duct cancers[8,24]. Our study showed that the presence of lymphovascular invasion was an independent prognostic factor and unfavorably influenced long-term survival. Lymphovascular invasion was present in 19.5% of patients, and the overall 5-year survival rates of these patients with lymphovascular invasion were poor, compared to the rates of patients without lymphovascular invasion (6.7% vs 52.3%, P < 0.0001).

Additionally, lymphovascular invasion was significantly associated with the depth of invasion (T classification) (P = 0.013). The frequency of lymphovascular invasion, according to the depth of the invasion (T1, T2 and T3), was 5.3%, 20% and 26.2%, respectively. Although pathologic T1 (pT1) tumors generally have favorable prognoses after resection, the presence of lymphovascular invasion, even in pT1 tumors, affected survival unfavorably, compared to pT1 tumors without lymphovascular invasion (5-year survival rate, 0% vs 57.4%; P < 0.0001). Our results were similar to those from a study published by Nagahashi et al[8], who reported poorer prognoses in patients with invasion into the fibromuscular layer, compared to those with mucosal layer invasion. These authors insisted that the cause of these adverse effects was related to lymphovascular invasion. Contrary to our expectations, the present study did not find that lymphovascular invasion was strongly associated with nodal metastasis (P = 0.071). This lack of association has also been reported in other studies examining nodal status and lymphovascular invasion in breast cancer[25,26]. A direct correlation between the presence of lymphovascular invasion and nodal metastasis might not always be apparent. The presence of lymphovascular invasion in patients with nodal metastasis did not affect survival, but in patients without nodal metastasis, the presence of lymphovascular invasion affected 5-year survival (0% vs 60.4%; P < 0.0001). This result suggests that lymphovascular invasion has a profound impact on the survival of patients with lymph node-negative middle and distal extrahepatic bile duct cancers.

Pancreaticoduodenectomy had higher rate of R0 resection than bile duct resection, although the long-term survival between the two groups was not significantly different. BDR can be an applied when only R0 resection is possible or if the co-morbidity of the patient precludes PD. R1 resection, if inevitable, can be performed because it offers better survival than R2 or no resection. The TNM stage and LVI were identified as independent factors influencing the survival of patients with middle and distal bile duct cancers.