Six hundred eighty-six patients were diagnosed with extrahepatic bile duct cancer at Samsung Medical Center between February 1995 and June 2011. At first 138 patients who had overlapping midportion or perihilar bile duct cancer and 14 patients whose cell type were not ductal adenocarcinoma of distal bile duct were excluded. To avoid confusion between middle and distal bile duct, the authors defined distal bile duct cancer that originating below the superior margin of pancreas head. 14 who were diagnosed other primary cancer earlier and four patients who were diagnosed another malignancy during the follow-up period were excluded. Inclusion criteria were pylorus-preserving pancreaticoduodenectomy (PPPD) and Whipple procedure for distal bile duct cancer, which were performed as curative intent and were confirmed as ductal adenocarcinoma by specialized pathologists in our hospital. One patient with distant metastasis operated with palliative intent was excluded consequently. Finally, 241 patients were enrolled.

Data collection was achieved by reviewing electric medical records retrospectively. Follow-up day began on the operation day and ended on the last treatment day. We analyzed age, sex, preoperative carbohydrate antigen 19-9 (CA 19-9) level, preoperative serum total bilirubin, preoperative biliary drainage, types of biliary drainage, transfusion, tumor size, TNM staging and differentiation of cancer cell for recognized risk factors affecting survival postoperatively.

Preoperative biliary drainage such as percutaneous transhepatic biliary drainage (PTBD), endoscopic nasobiliary drainage (ENBD), and percutaneous transhepatic biliary drainage (PTBD) were performed for the patients with hyperbilirubinemia caused by biliary obstruction which were identified with preoperative imaging tools.

Adjuvant therapy was applied as concurrent chemoradiation manner. 5-Fluorouracil or gemcitabine or capecitabine were used as main chemotherapeutic agent and 44 Gy in 22 fractions were used for radiation. In palliative therapy, three ways such as chemotherapy only, radiation only and concurrent chemoradiation were carried out. The majority of them underwent chemotherapy with GEMOX (gemcitabine plus oxaliplatin) regimen or GP (gemcitabine plus cisplatin) regimen with or without radiation.

Resection margin was classified as R0, R1 and R2 whether tumor exists or not observed by macroscopically or microscopically. Macroscopic and microscopic tumor negative resection margin is considered as R0 resection. Macroscopic tumor negative and microscopic tumor positive resection margin is regarded as R1 resection. Macroscopic tumor positive resection margin is termed R2 resection. Transfusion was performed when patients' hemoglobin level was under 8.5 g/dL or hypoxia induced by bleeding or unstable vital sign caused by sustained hemorrhage. We defined complication and mortality as which appeared within a month after the surgery. Postoperative pancreatic fistula was graded as A, B, or C according to an international study group on pancreatic fistula definition [14]. TNM staging of distal bile duct cancer based on American Joint Committee on Cancer (AJCC) 7th edition.

Overall survival and survival of each variable were calculated with Kaplan-Meier formula. Univariate analysis of risk factors affecting survival was calculated with log-rank test and multivariate analysis was calculated with Cox-regression model. We defined significant data as a P-value < 0.05 and used IBM SPSS ver. 18.0 (IBM Co., Armonk, NY, USA) as a tool of statistical analysis.

Average age of the 241 patients was 62.4 years (range, 31 to 88 years). There were 166 males (68.9%) and 75 females (31.1%). Table 1 shows that the patient's chief complaint at the time of their first visit to our clinic. Jaundice was the most frequent symptom and was evident in 135 patients (56.0%), followed by abdominal pain (26.1%) and fever (5.4%). Preoperative biliary drainage was performed in 219 cases (90.9%). There were 115 cases of PTBD (52.5%), 80 cases of ENBD (36.5%), and 24 cases of ERBD (11.0%).

PPPD were performed in 171 patients and Whipple procedure were performed in 70 patients. 73 patients (30.3%) had lymph node metastasis. 6 patients were revealed to be R1 resection. Cancer cell was identified at the proximal margin of bile duct in five cases and at the soft tissue around superior mesenteric artery (SMA) in one case. R2 resection cases were four. One of them had portal vein invasion, one had SMA invasion, one had portal vein and SMA invasion and the last one showed proximal margin of bile duct invasion. Forty-one patients (17.5%) were pathologically confirmed as well differentiated form of ductal adenocarcinoma, 119 patients (50.9%) were moderate and 74 patients (31.6%) were poorly differentiated form. The last 7 patients had no pathological reports about tumor cell differentiation. According to AJCC 7th edition, stage IA were 11 cases (4.6%), stage IB were 42 cases (17.4%), stage IIA were 111 cases (46.1%), stage IIB were 62 cases (25.7%), stage III were 15 cases (6.2%), and stage IV was only one case but excluded as mentioned above.

Complications within a month from the operation date were recorded (Table 2). One hundred eight patients (44.8%) experienced various types of complications. The most frequent complication was intra-abdominal fluid collection (40 cases, 37.0%), which needed intervention or prolonged conservative treatment, followed by bleeding (15 cases, 13.9%). Three mortalities occurred as a consequence of postoperative bleeding and one mortality was caused by liver failure after left hepatic artery sacrifice during surgery.

The average follow-up period of 237 patients, during which four immediate postoperative mortality cases were excluded, was 31.3 months (range, 2 to 177 months). One hundred twelve patients (47.3%) experienced recurrence during the follow-up and one hundred twenty five patients (52.7%) had no follow-up recurrence. Among the 112 recurrent patients, local recurrence were 53 (47.3%), intrahepatic recurrence were 33 (29.5%), and systemic recurrence were 26 (23.2%). Eventually, 86 patients died because of the recurrence.

Adjuvant therapy was performed in 6 patients with stage IIA, 14 patients with stage IIB, and 3 patients with stage III. Despite of the adjuvant therapy, 13 patients recurred and died. Palliative therapy was conducted in 44 recurrent patients and 2 patients with R2 resection. Twenty-two patients, including one R2 resection case, showed partial response and survived during the follow-up. However 24 patients died after or during the palliative therapy. Only one patient was lost to follow-up.

Four postoperative mortality cases were also excluded for survival analysis. The 3-, 5- and 10-year survival was 55.3%, 48.3% and 33.7%, respectively, after curative resection of distal bile duct cancer in our hospital (Fig. 1). Median survival duration was 73.0 months (range, 30.8 to 115.1 months). Univariate analysis revealed preoperative CA 19-9 >35 U/mL, tumor size >2 cm, presence of biliary drainage, complications, advanced stage and tumor cell differentiation and positive resection margin as poor prognostic factors (Table 3). In multivariate analysis, only CA 19-9, resection margin, T stage and N stage were significant factors (Table 4). Among them, positive resection margin was the strongest factor (hazard ratio, 5.078) indicating a negative influence on survival.

Survival difference between Whipple procedure and PPPD was presently evident (P = 0.006) (Fig. 2), despite the fact that their proportion of tumor staging was not significantly different (P = 0.659) (Fig. 3). For that reason, to rule out a cohort effect, we divided patients into four groups by period and by operation type. Patients who underwent Whipple procedure before 2003 were classified as group A and those who underwent the procedure after 2002 constituted group B. Patients underwent PPPD before 2003 were classified as group C and after 2002 were group D. When comparing groups A and B, it is obvious that the use of Whipple procedure had declined from 51 to 18 cases. On the other hand, PPPD had increased from 30 to 138 cases. There were no survival differences between groups A and C (P = 0.955). Groups B and D also showed no survival differences (P = 0.150). However, survival of group D patients was significantly better than group C patients (P = 0.001) (Fig. 4).