Relevant prognostic factors influencing outcome of patients after surgical resection of distal cholangiocarcinoma

Cholangiocarcinoma is a malignant disease with poor prognosis and occurs with an incidence of 1.3 to 3.4 cases per 100,000 in Western countries [1].

The primary location of the carcinoma in reference to the biliary tract influences clinical manifestation and therapy. Therefore they are classified into intrahepatic, proximal extrahepatic and distal extrahepatic cholangiocarcinomas (DCC) [2].

DCC constitute up to 20–30% of all cholangiocarcinomas and are located between the cystic duct entry into the main hepatic duct and the Ampulla of Vater. Clinical symptoms comprise painless jaundice and/or weight loss [3].

Curative surgery is the only chance for long-term survival [3‐9]. Palliative chemotherapy can achieve median survival rates of 8.1 to 11.7 months [10].

Apart from negative lymph node status, a tumor-free resection margin has been reported to have crucial impact on disease free survival [4, 5, 11, 12]. Nevertheless, tumor stage-specific surgical strategies have not been clearly defined. Some centers perform segmental bile duct resections when the tumor is small and located in the middle segment of the extrahepatic bile duct. Major surgery including partial pancreaticoduodenectomy is necessary to achieve tumor-free resection margins in the majority of cases [6, 13]. Some patients present with advanced tumors with additional infiltration of central bile ducts on one side of the liver or local infiltration of the portal vein. For these patients extended pancreatic resection with en-bloc portal vein resection or additional resection of proximal extrahepatic bile ducts might be necessary to achieve tumor-free resection margins [8, 14, 15].

Most published data describe inhomogeneous patient groups with respect to the primary location of biliary tract carcinoma [3‐9]. The objective of this study was to evaluate the outcome of surgery in patients with DCC following resection in curative intent. In this context, the focus was put on the extent of surgical resection to clarify the role of organ preserving resection and extended surgery, including additional resection of the proximal extrahepatic bile ducts and/or partial resection of the portal vein.

Due to the aforementioned comparatively low incidence of DCC among the general population, especially from Western countries, reports on outcome after surgery, of large patient collectives, are scarce. This data from a single high-volume center was analysed to identify prognostic factors for unfavourable patient outcome.

Interestingly no published data investigated preoperative stenting in patients with DCC and its impact on survival yet. In multivariable analysis we identified preoperative stenting as strongest risk factor for survival. A significantly longer duration between first symptoms and surgery in these patients could provide a clinically plausible explanation for the inferior outcome (Table 2). A prospective trial aiming to elucidate the role of preoperative biliary stenting in patients with pancreatic head cancer showed a significantly higher rate of complications after preoperative stenting [20]. So far similar observations have not yet been reported for DCC. The current study did not find a significant influence of biliary stenting on the rate or severity of postoperative complications. Nonetheless, we are well aware that preoperative conditions, which were not analysed in the current study, like cholangitis or severe biliary obstruction which require prolonged biliary drainage could be an explanation for the inferior postoperative outcome [21].

Survival after surgical treatment in the current study was similar to several previous studies [22‐28].

Reported outcomes of patients with DCC after extended pancreatic surgery are extremely scarce. Solely Courtin-Tanguy et al. investigated extended surgery for treatment of DCC and found that combined organ resection was an independent risk factor for worse survival [29]. Interestingly, the present study showed that extended pancreatic resection was not significantly associated with inferior survival except in case of concomitant histological venous invasion as displayed by the identified interaction variable. Previous studies on risk factors for survival have not taken possible multiplicative factor interactions into account [28, 30]. The approach to multivariable Cox regression in this study deploys the previously published methodology for the detection of such factor interactions [18]. However, the displayed factor interaction is not suitable for intraoperative decision making regarding the extent of surgery due to the circumstance that histopathogical results verifying microscopic venous invasion are rarely available pre- or intraoperatively. Instead the prognostic value of venous invasion in patients undergoing extended surgery should be considered within the postoperative therapeutic strategy.

Partial portal vein resection due to macroscopic tumor invasion had no statistically significant influence on survival in Cox regression analysis (Table 2) in this study while several studies described poor survival for patients after portal vein resection, although only Miura et al. were able to show significant influence on survival in multivariable analysis [28, 31, 32]. This discrepancy could be due to the small number of cases in this study since a tendency towards worse survival in patients after partial portal vein resection has been observed. The benefit of portal vein resection therefore remains unclear.

Intraoperative frozen sections were crucial for flexible intraoperative decision making on the extent of resection, achieving tumor-free resection margins and partially led to an extension of the surgical strategy including partial portal vein, additional liver or perhilar bile duct resection. This is emphasized by the fact that the preoperative CT scans in the displayed cohort were not a reliant predictor of tumor extent. CT scans of 22 patients identified no measurable tumor mass although final histology revealed tumor sizes between T stage 1 and 4 in these patients. Even additional MR tomography did not provide significantly better results in preoperative detection of tumor extent. Therefore, in some cases endosonography might help for identification of tumor extent.

Finally patients with stricture of the distal bile duct suspicious of a malignant tumor should be explored surgically even if histology was negative, since 49.3% of preoperative histologies were false negative.

The impact of positive tumor infiltration at the surgical resection margin (R1/2 resection) on postoperative survival following curative resection of DCCs is widely discussed. We found no significant impact of R1/2 resection on overall survival in Cox regression analysis (Table 2). This finding is in line with some of the published series [15, 33‐35]. Other studies identified positive resection margins as an independent prognostic factor on overall survival [3, 28]. Interestingly the percentage of R1 or R2 resected patients differs widely between the aforementionend published series from 7.0 to 43.3% [3, 15, 28, 33‐37]. Our series shows a comparatively low rate of R1/R2 resections with 13.3%. A comparable retrospective study with 10.5% positive resections margins found no significant difference in survival for these patients in uni- and multivariable analysis [15]. Two published series found rates of 13.4 and 14.0% of R1/2 resected patients, respectively with significant influence on patient survival in uni- but not in multivariable analysis [33, 35]. In contrast series with higher rates of cancer positive resection margins identified R1/2 resections as an independent significant negative prognostic factor on survival in multivariable analysis [28, 38]. This could implicate a correlation of higher rates of positive resection margins with the identification of R1/2 resections as a significant negative prognostic factor in multivariable analysis in some series, which in turn would explain the conflicting results of our and these series [15, 33‐35]. As displayed above, further local or extended resection in case of initially positive intraoperative resection margins was performed on a regular basis especially in patients undergoing local excision of bile duct and did not impact recurrence rate or overall survival. Interestingly there is presently no further data on this matter in patients undergoing resection for distal cholangiocarcinoma. Until further clarification the primary goal of surgical resection of DCC should be to achieve R0 resection. We therefore recommend intraoperative frozen sections of the resection margins and when residual disease is detected, additional resection of the bile duct bifurcation or locally infiltrated structures should be performed if technically feasible.

The presence of tumor positive lymph nodes was identified as a significant prognostic factor in multivariable analysis (Table 2) as was reported by different meta-analysis and some multicenter studies [28, 30, 36, 38]. Apart from a German study by Petrova et al., all mentioned publications reported a significant impact of lymph node metastasis on survival in uni- and not in multivariable analysis [28]. The above mentioned German colleagues did not report on the impact of the number of metastatic lymph nodes on survival. This seems to be relevant as Kiriyama et al. demonstrated earlier that the total number of infiltrated lymph nodes influences patients outcome. They found a significant difference when comparing no lymph node metastasis with 1–3 and more than 4 metastases [39]. This correlates well with our findings, as we classified the number of lymph node metastasis according to the newest TNM 8th edition classification for DCC in N0 versus N1 (1–3 positive lymph nodes) versus N2 (≥ 4 positive lymph nodes) resulting in a negative prognostic factor for each grade with a hazard ratio of 2.18 in multivariable analysis.

Postoperative complications graded in points according to Clavien-Dindo had a significant impact on patient outcome (Table 2). The influence of postoperative morbidity on the outcome of patients with DCC was recently observed by Petrova et al. and Andrianello et al.. Aside from direct consequences leading to early death in hospital it is assumed that patients with complicated postoperative course experience delay in the start of adjuvant therapy options (i.e. chemotherapy) resulting in an inferior survival [28, 40]. However the efficiency of adjuvant chemotherapy in patients with cholangiocarcinoma, as currently investigated by the multicenter prospective ACTICCA trial, remains unclear [41].

Preoperative biliary stenting, positive lymph node staging, perineural invasion as well as extended surgery in patients with concomitant venous invasion and postoperative complications should be considered within the perioperative therapeutic strategy. It is important to state that it is difficult to draw definite conclusions from this retrospective study considering the small patient collective undergoing extended surgery, respectively. Therefore it would certainly be beneficial performing multi-center studies with prospective analysis to determine which surgical approach is justifiable regarding optimal patient outcome. In addition the effects of preoperative biliary drainage in patients with DCC and relevant cholestasis on survival should be thoroughly examined in respect of our presented data.