Prognostic Factors for Survival of Patients with Ampullary Carcinoma After Local Resection

UPPER GI

ANZJSurg.com

Prognostic factors for survival of patients with ampullary carcinoma after local resection

Xiangqian Zhao, Jiahong Dong, Xiaoqiang Huang, Wenzhi Zhang and Kai Jiang Hospital and Institute of Hepatobiliary Surgery, Chinese PLA General Hospital, Beijing, China

Key words Abstract ampullary cancer, local resection, pathology, prognosis. Background: Local resection (LR) is a potentially effective alternative to Correspondence pancreaticoduodenectomy for treatment of ampullary cancer, but the prognostic Professor Jiahong Dong, Hospital and Institute of factors remain undefined. The purpose of this study was to identify the prognostic Hepatobiliary Surgery, Chinese PLA General Hospital, factors for ampullary cancer patients who had undergone LR. 28 Fuxing Road, Beijing 100853, China. Methods: We retrospectively reviewed the clinical, pathological data and surgical Email: [email protected] approach of 34 ampullary cancer patients who had undergone LR during 1996–2009 X. Zhao MD; J. Dong MD, PhD; X. Huang MD; at People’s Liberation Army General Hospital. Prognostic factors for survival and W. Zhang MD; K. Jiang MD. recurrence were analysed. Results: The 1-, 3- and 5-year survival rates of the patients were 97.1, 69.5 and 53.7%, Accepted for publication 3 March 2014. respectively. The gender, age, preoperative bilirubin levels, CA19-9 levels and preoperative biopsy did not correlate with the survival rates. The survival rates of patient doi: 10.1111/ans.12600 with T1 and T2 tumours were superior to that of patients with T3 tumours (P = 0.000). Tumour size, surgical margin status and the extent of differentiation had no effect on survival rates (P = 0.464, P = 0.601 and P = 0.121, respectively). The survival rate of patients who had extraduodenal LR (12 cases) was superior to that of patients who had transduodenal LR (22 cases) (P = 0.026). Tumour recurrence occurred in 14 (41.2%) patients. Tumour infiltration (P = 0.014) correlated with the recurrence. Conclusion: The degree of tumour infiltration is the pathological factor that most affects the survival of ampullary cancer patients who undergo LR. Extraduodenal LR is a promising surgical procedure, the efficacy of which is superior to that of transduodenal LR. The depth of tumour invasion correlated with the recurrence.

Introduction Transduodenal LR is the conventional surgical approach for ampullary cancer. Its main drawback is that the resection ﬁeld is too Ampullary cancer is a rare cancer located in the duodenal papilla, small to permit complete resection of tumours at the distal end of the the mucosa around the papilla, the mucosa within the ampulla, the common bile duct, and extended resection is greatly limited. pancreatic duct opening, or the mucosa between the common bile However, the extraduodenal LR that we recently reported can expose duct and the duodenal wall. Ampullary cancer has a better prognosis the entire common bile duct, permit estimation of the resection than pancreatic cancer or cholangiocarcinoma,1 and surgical resec- length of the common bile duct according to the site of tumours and tion is associated with a satisfactory prognosis. The current surgical offer a better chance for complete removal of the tumours.5 In treatments for ampullary cancer are radical pancreatoduodenectomy addition, the extraduodenal approach permits excision of lymph (PD) and local resection (LR). Although PD is still the preferred nodes behind the pancreas and duodenum concurrently, if necessary. surgical procedure, it has a high rate of surgical trauma, post- However, predictive factors for survival of ampullary cancer operative mortality and complications;2 LR has a lower rate of these patients after operation vary in different reports. These factors outcomes. Nikfarjam et al.3 reported that the complication rate of include preoperative jaundice status,6 surgical margin status,7 the PD was as high as 71%, whereas the complication rate of LR was presence or absence of lymph node metastases,7–9 tumour size,10 20%. Feng et al.4 reported a complication rate of 34.8% and a blood vessel invasion,9 nerve invasion,8 tumour differentiation8 and mortality rate of 6.5% with PD compared with corresponding rates the degree of tumour inﬁltration.7,9–11 Additionally, the report about of 6.5 and 0% with LR. prognostic factors for LR is few. In view of this, we retrospectively

© 2014 The Authors ANZ Journal of Surgery published by Wiley Publishing Asia Pty Ltd on behalf of Royal Australasian College of Surgeons ANZ J Surg 85 (2015) 567–571 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 568 Zhao et al. analysed the clinical, pathological data and surgical approach of to the site entering the duodenum to find the tumour. A further ampullary cancer patients who had undergone LR at the Chinese circumferential incision along the duodenal papilla was made to People’s Liberation Army General Hospital from February 1996 to locate the main pancreatic duct at the connection between the duo- February 2009. We assessed the efficacy of LR and explored the denum and pancreas. The pancreatic duct was severed, the lower part prognostic factors for long-term survival and recurrence after the of the common bile duct and the papilla were excised. The pancre- operation. atic and common bile ducts were anastomosed to the duodenum.5

Pathology and follow-up Methods The general data (including surgical approach, gender, age, preop- Patients erative bilirubin levels, CA19-9 levels and preoperative biopsy) of Between 1996 and 2009, LR was performed on 35 ampullary carci- all 35 patients were reviewed. The gross specimens, the original noma patients at our hospital. pathology reports and the specimens were examined by two experienced pathologists who were blind to any prior knowledge regarding Inclusion criteria the clinical data. The histological type, diameter, degree of differen- All patients reviewed had histologically confirmed duodenal papilla tiation, depth of invasion, lymph node metastases, vascular invasion or ampullary adenocarcinoma. and surgical margin of the tumours were evaluated and recorded. Follow-up results were obtained for 34 of the 35 patients who had Exclusion criteria undergone LR, with a 97.1% follow-up rate. One patient was lost to Patients who had duodenal cancer, distal cholangiocarcinoma, pan- follow-up. Follow-up visits were performed once every 3 months creatic cancer, involvement of these cancers in the ampulla or duo- during the first year, re-examined once every 6 months during the denal papilla, and tumours of ambiguous origin. second and third years, and re-examined once a year later. Items checked during the follow-up visits included routine laboratory tests, Surgical indications and contraindications tumour markers, chest roentgenogram, abdominal ultrasound and computed tomography/magnetic resonance imaging. The follow-up LR was considered for patients who met one of the following con- deadline was 1 March 2011, and the follow-up duration ranged from ditions: (i) preoperative diagnosis of benign tumour or malignant 10 to 152 months, with a median duration of 36 months. transformation of benign tumour; negative biopsy of the base after local tumour resection; (ii) tumour diameter ≤3 cm; no infiltration outside the papilla of Vater; biopsy of the base after local tumour Statistical analysis resection was negative; (iii) well-differentiated cancer; histological All data were presented as figures and percentages and analysed by confirmation intraoperatively; negative biopsy of the base after local non-parametric tests, using SPSS 16.0 statistical software (SPSS, tumour resection; (iv) the patient was old or had poor general con- Inc., Chicago, IL, USA). Survival of ampullary carcinoma patients dition or other associated diseases that prevented PD but likely could was calculated using the Kaplan–Meier method, log-rank test and tolerate LR; (v) patients who refused PD but his or her physical Cox regression analysis. Logistic regression analysis was also per- condition allowed LR. formed to evaluate the prognostic parameters for recurrence. A value Patients with any of the following conditions were not considered of P < 0.05 was considered statistically significant. suitable for LR: (i) poor general condition, high risk of operative complications, unable to tolerate laparotomy; (ii) the presence of metastases to parenchymal organs or lymph nodes; (iii) duodenal Results obstruction; (iv) ulcerated bleeding tumour; (v) tumour invasion into the pancreas; (vi) positive basilar biopsy after LR; (vii) tumour Patient characteristics recurrence after LR. Patients’ clinical data are presented in Table 1. Of the 34 patients, 16 were men and 18 women, aged 30–75 years. There were 22 CA19- Surgery 9-negative patients and 12 CA19-9 positive. Preoperative jaundice Before 2000, LR was performed via the transduodenal approach; was present in 22 patients. Twenty-two patients were operated via after 2000, the patients were randomized to be operated via either the transduodenal approach and 12 via the extraduodenal approach. the transduodenal or the extraduodenal approach. Duodenoscopic biopsy of the papilla of Vater was performed before The transduodenal LR method has been described.12 For surgery in 22 patients. After LR, one patient received chemotherapy, extraduodenal LR, a Kocher manoeuvre was performed to mobilize another one received radiotherapy and the others received neither the duodenum. The head of the pancreas was turned over, the radiotherapy nor chemotherapy. postero-superior pancreaticoduodenal artery was cut off, the second There were 14 cases of recurrent cancer during the follow-up part of the duodenum was turned to the medial side, and the attach- period, including local recurrence in nine patients, liver metastases ments of the pancreas and duodenum were freed. A probe was placed in four and abdominal wall metastases in one. Post-operative into the common bile duct via the cystic duct opening or the complications occurred in six patients: four cases of pancreatic common bile duct opening in order to identify the distal end of the leakage, one case of gastrointestinal bleeding and one case of wound common bile duct. The common bile duct was opened 2.5 cm prior dehiscence.

Table 1 Clinical and pathologic characteristics Further analysis indicated that there was no significant difference between T1 and T2 groups (P = 0.169), but the survival rates of T1 Age (years) Median (range) 62 (30–75) and T2 groups were significantly superior to those of T3 group (P = Gender (Male : Female) 16:18 (47.1%:52.9%) = CA19-9 (U/L) Median (range) 34.3 (7–3160) 0.000 and P 0.000). Positive : Negative 12:22 Tumour recurrence occurred in 14 (41.2%) patients. Logistic TB (μmol/L) Median (range) 47.2 (9–757) regression analysis showed that tumour infiltration (P = 0.014) cor- Positive : Negative 22:12 Approach Transduodenal 22 related with recurrence (Table 3). No significant differences were Extraduodenal 12 found between T1 and T2 patients in recurrence (P = 0.096). The Duodenoscopic biopsy Yes 22 recurrence was significantly higher in the T3 group than the T1 No 12 = Diameter (cm) Median (range) 2 (0.8–3.5) group (P 0.016). ≤219 >215 Differentiation High 21 (61.8%) Discussion Moderate 9 (26.5%) Low 4 (11.8%) Ampullary tumour is a relatively rare gastrointestinal tumour that Infiltration T1 12 (35.3%) appears in the duodenal papilla, the mucosa around papilla, and at T2 17 (50.0%) T3 5 (14.7%) the biliary and pancreatic duct openings. Tumours in this location Margins Positive 3 (8.9%) early block the common bile duct and cause jaundice, which is easily Negative 31 (91.1%) detected. Most ampullary tumours are relatively small when they Recurrence Yes 14 (41.2%) No 20 (58.8%) become symptomatic, so LR is an advisable choice for their treatment. Although LR has played an important role in ampullary cancer treatment, prognostic factors for LR are still not clearly defined. In this retrospective analysis of 34 ampullary cancer patients treated with LR, we found that the degree of tumour infiltration was the Results of pathology examinations pathological factor that most affected post-operative survival. In Table 1 illustrates features of the 34 resected tumours. The tumour addition, extraduodenal LR was superior to the conventional diameters ranged from 0.8 to 3.5 cm (median, 2.0 cm). There were transduodenal LR approach. 12 cases of T1, 17 cases of T2 and five cases of T3 stage tumours; T1 It has been reported that the 5-year survival rates of ampullary and T2 cases accounted for 85.3% of all cases. Twenty-one tumours cancer patients who underwent PD were 43.3–61.0%4,6,13,14 and the (61.8%) were well differentiated, nine were moderately differenti- 5-year survival rates of LR patients were 48.0–51.4%;4,15,16 the ated and four were poorly differentiated. Three resected cancers had results suggested comparable efficacy of LR and PD. Consistent positive surgical margins. Local lymph node dissection was per- with these reports, the 5-year survival rate of our ampullary carci- formed in five patients who underwent extraduodenal LR; metastasis noma patients who underwent LR was 53.7%. Therefore, LR is still was found in one case. There was no adjacent vascular involvement a choice of ampullary carcinoma treatment if a patient’s condition is case. suitable for the operation. The commonly used transduodenal LR12 procedure for ampullary Survival and recurrence analysis carcinoma suffers from the drawback that exposure of the common The impact of clinicopathological factors on patients’ survival rates bile duct is difficult to attain and the resection range is restricted. are shown in Table 2. Seventeen patients died during the follow-up Thus, if the tumour has infiltrated along the mucosa of bile duct and period. The 1-, 3-, and 5-year survival rates were 97.1, 69.5 and the pancreatic duct, complete resection of the tumour is difficult to 53.7%, respectively. Survival rates did not differ among different achieve, and if the scope of the resection is extended, re-anastomosis gender or age groups (P = 0.174 and P = 0.935, respectively). of the ducts is also difficult. In 2000, Huang et al. reported the Survival rates were not statistically different between CA19-9- resection of ampullary tumours via an extraduodenal approach.17 negative and CA19-9-positive patients, or between patients with This approach permits exposure of the entire common bile duct, different levels of CA19-9 (P = 0.245 and P = 0.566, respectively). determination of resection length of the common bile duct according The survival rates of patients with or without preoperative jaundice, to the site of tumours, and resection of pancreaticoduodenal lymph or of patients with different bilirubin levels, also were not signifi- nodes if necessary. Thus, this procedure can result in a nearly com- cantly different (P = 0.143 and P = 0.516, respectively). Preoperative plete excision of the cancer. Our results indicate that among patients biopsy had no impact on survival rates (P = 0.659). Most notably, with similar clinical features, the survival rates of those who had extraduodenal LR led to a better survival rate than did the received extraduodenal LR were superior to the rates of those who transduodenal LR approach (P = 0.026) (Table 2 and Fig. 1). had received conventional LR. Surgical approach, therefore, is an Tumour size, surgical margin status and the extent of differentia- important determinant of survival. Although extraduodenal LR can tion had no effect on survival rates (P = 0.464, P = 0.601 and P = be performed in patients who are suitable for LR, this procedure is 0.121, respectively), but the depth of tumour invasion correlated relatively complex and demanding even for experienced surgeons. with survival (P = 0.000) (Table 2 and Fig. 2). Because only five It has been reported that TNM, pT, pN stage, tumour size and patients underwent regional lymph node dissection, we did not resection margin status affect patient survival.4 In contrast, our data perform survival analysis regarding lymph nodes and tumour stages. showed that the survival rates of patients were independent of the

Table 2 Impact of clinicopathological factors on survival rates

Outcome Cases 3-year survival 5-year survival P-value

Gender Male 16 10 (59.2%) 8 (40.6%) 0.174 Female 18 14 (77.8%) 12 (64.2%) Age (years) ≤62 18 12 (66.2%) 10 (51.5%) 0.935 >62 16 12 (73.1%) 10 (54.8%) CA19-9 (U/L) Positive 12 9 (72.9%) 6 (33.3%) 0.245 Negative 22 15 (67.9%) 14 (62.2%) ≤100 25 16 (63.5%) 15 (58.6%) 0.566 >100 9 8 (88.9%) 5 (37.0%) TB (μmol/L) Positive 22 14 (61.7%) 11 (43.6%) 0.143 Negative 12 10 (83.3%) 9 (71.4%) ≤120 26 19 (72.3%) 16 (56.3%) 0.516 >120 8 5 (57.1%) 4 (42.9%) Approach Transduodenal 22 11 (62.9%) 6 (38.1%) 0.026 Extraduodenal 12 9 (82.5%) 6 (82.5%) Duodenoscopic biopsy Yes 22 17 (76.3%) 14 (57.6%) 0.644 No 12 8 (50%) 6 (37.5%) Diameter (cm) ≤2 19 11 (57.9%) 11 (57.9%) 0.464 >2 15 13 (84.8%) 9 (43.1%) Differentiation High 21 16 (75.9%) 13 (57.8%) 0.121 Moderate 9 6 (66.7%) 5 (55.6%) Low 4 1 (0%) 1 (0%) Inﬁltration T1 12 11 (90.0%) 9 (67.5%) 0.000 T2 17 13 (76.5%) 11 (60.8%) T3 5 0 (0%) 0 (0%) Margins Positive 3 2 (50.0%) 2 (50.0%) 0.601 Negative 31 21 (66.7%) 18 (53.9%)

Wilcoxon rank sum test and chi-square/Fisher’s exact test were used for data analysis.

Fig. 1. The survival curves of extraduodenal and transduodenal groups. Fig. 2. Survival curves of patients with different degree of tumour infiltration. patients’ gender, age, preoperative bilirubin levels, CA19-9 levels rates between T1 and T2 patients. These findings indicate that the and preoperative biopsy. We also found that tumour size, differen- prognosis of ampullary cancer will be poor if the cancer has invaded tiation and surgical margin status had no relationship to survival into the pancreas. rates, a result that differs from results previously reported.7,8,10 Con- There are few studies in the literature that have focused on the sistent with previous studies,7,9–11 though, we found that the degree recurrence of ampulla cancer after surgical resection. Lindell et al.18 of tumour invasion and survival were closely correlated. Moreover, showed that 22% of patients after PD and 80% of patients after LR also consistent with previous reports,4,10 we found that the 5-year developed a recurrence. Feng et al.4 reported that recurrence was survival rates of T1, T2 and T3 stage patients were 67.5, 60.8 and diagnosed in 23.3% of patients after PD and 48.0% of patients after 0%, respectively, and there was no significant difference in survival LR. The risk factors of recurrence include TNM stage, pT stage and

Table 3 Clinical features and tumour histopathological characteristics of recurrent group and disease-free group patients

Outcome Recurrent group Disease-free group P-value

Age (years) median (range) 63 (30–75) 60 (36–73) 0.540 Gender (male : female) 7:7 9:11 0.777 CA19-9 (U/L) median (range) 35.8 (11–3160) 28.5 (7–1750) 0.336 TB (mol/L) median (range) 45.2 (10–757) 47.2 (9–218) 0.726 Preoperative biopsy Yes 8 14 0.447 No 6 6 Diameter (cm) median (range) 2 (0.8–3.0) 2.0 (1.0–3.5) 0.684 ≤2 8 (57.1%) 11 (55%) >2 6 (42.9%) 9 (45%) Differentiation Well 10 (71.4%) 11 (55%) 0.209 Moderately 4 (28.6%) 5 (25%) Poorly 0 (0%) 4 (20%) Infiltration T1 2 (14.3%) 10 (50%) 0.014 T2 8 (57.1%) 9 (45%) T3 4 (28.6%) 1 (5%) Margins Positive 1 (7.1%) 2 (10%) 0.776 Negative 13 (92.9%) 18 (90%) pN stage;4 tumour differentiation;6 and lymph node ratio, 6. Choi SB, Kim WB, Song TJ, Suh SO, Kim YC, Choi SY. Surgical lymphovascular.19 In our study, tumour recurrence occurred in 14 outcomes and prognostic factors for ampulla of Vater cancer. Scand. J. (41.2%) patients. Logistic regression analysis showed that tumour Surg. 2011; 100: 92–8. infiltration (P = 0.014) was correlated with recurrence. No signifi- 7. Sessa F, Furlan D, Zampatti C, Carnevali I, Franzi F, Capella C. Prog- cant differences were found between T1 and T2 patients in recur- nostic factors for ampullary adenocarcinomas: tumor stage, tumor his- tology, tumor location, immunohistochemistry and microsatellite rence (P = 0.096). The recurrence was significantly higher in the T3 instability. Virchows Arch. 2007; 451: 649–57. group than the T1 group (P = 0.016). These findings indicated that 8. Lee JH, Lee KG, Ha TK et al. Pattern analysis of lymph node metastasis pancreatic invasion increased risk of recurrence for ampulla cancer and the prognostic importance of number of metastatic nodes in after LR. ampullary adenocarcinoma. Am. Surg. 2011; 77: 322–9. Because only five patients underwent local lymph node dissection, 9. Sakata E, Shirai Y, Yokoyama N, Wakai T, Sakata J, Hatakeyama K. metastasis was found in one case. Therefore, we did not perform Clinical significance of lymph node micrometastasis in ampullary car- survival analysis in accordance with lymph nodes or metastasis cinoma. World J. Surg. 2006; 30: 985–91. staging. Because ampullary cancer is usually found at early stage, the 10. Yoon YS, Kim SW, Park SJ et al. Clinicopathologic analysis of early tumours rarely affect the portal vein or superior mesenteric vein. In ampullary cancers with a focus on the feasibility of ampullectomy. Ann. this study, no vascular invasion was observed in any patient, so we did Surg. 2005; 242: 92–100. et al not take blood vessel invasion into account for analysis of survival. 11. Morini S, Perrone G, Borzomati D . Carcinoma of the ampulla of Vater: morphological and immunophenotypical classification predicts In summary, we found that the degree of tumour infiltration is a overall survival. Pancreas 2013; 42: 60–6. pathological factor that affects the post-operative outcome of ampullary 12. Distler M. Grützmann R. Transduodenal resection of ampullary tumors. cancer patients who have undergone LR. Also, extraduodenal LR is Zentralbl. Chir. 2012; 137: 319–21. associated with a higher survival rate than that associated with conven- 13. Qiao QL, Zhao YG, Ye ML et al. Carcinoma of the ampulla of Vater: tional LR. The depth of tumour invasion correlated with the recurrence of factors influencing long-term survival of 127 patients with resection. ampulla cancer after LR. However, because our study is a single-centre World J. Surg. 2007; 31: 137–43. retrospective study, our conclusions need to be validated by additional, 14. Lazaryan A, Kalmadi S, Almhanna K, Pelley R, Kim R. Predictors of preferably prospective, studies. clinical outcomes of resected ampullary adenocarcinoma: a single- institution experience. Eur. J. Surg. Oncol. 2011; 37: 791–7. References 15. Liu N, Liang H, Li Q et al. Determinants of long-term survival in 38 1. Hatzaras I, George N, Muscarella P, Melvin WS, Ellison EC, Bloomston patients with carcinoma of ampulla of Vater treated by local resection. M. Predictors of survival in periampullary cancers following Zhonghua Zhong Liu Za Zhi. 2005; 27: 629–31 (in Chinese). pancreaticoduodenectomy. Ann. Surg. Oncol. 2010; 17: 991–7. 16. Meneghetti AT, Safadi B, Stewart L, Way LW. Local resection of 2. Balachandran P, Sikora SS, Kapoor S et al. Long-term survival and ampullary tumors. J. Gastrointest. Surg. 2005; 9: 1300–6. recurrence patterns in ampullary cancer. Pancreas 2006; 32: 390–5. 17. Huang X, Cai S, Huang Z. Improvement in the local resection of benign 3. Nikfarjam M, Muralidharan V, McLean C, Christophi C. Local resection tumors in the distal bile duct. Chin. J. Surg. 2000; 38: 328 (in Chinese). of ampullary adenocarcinomas of the duodenum. ANZ J. Surg. 2001; 71: 18. Lindell G, Borch K, Tingstedt B et al. Management of cancer of the 529–33. ampulla of Vater: does local resection play a role? Dig. Surg. 2003; 20: 4. Feng J, Zhou X, Mao W. Prognostic analysis of carcinoma of the 511–5. ampulla of Vater: pancreaticoduodenectomy versus local resection. 19. Roland CL, Katz MH, Gonzalez GM, Enell EL, Ihse I. A high positive Hippokratia. 2012; 16: 23–8. lymph node ratio is associated with distant recurrence after surgical 5. Zhao XQ, Huang XQ, Zhang WZ, Liu Z. Comparison between two types resection of ampullary carcinoma. J. Gastrointest. Surg. 2012; 16: 2056– of local resection in the treatment of ampullary cancer. ANZ J. Surg. 63. 2014; 84: 255–9.