Home Current issue Ahead of print Search About us Editorial board Archives Submit article Instructions Subscribe Contacts Login 

 Table of Contents  
Year : 2018  |  Volume : 37  |  Issue : 3  |  Page : 368-374

Nontraditional criteria for prediction of patient survival following pancreaticoduodenectomy for malignancy

1 HBP Surgery, Faculty of Medicine, Minia University, Minya, Egypt
2 General Surgery and Laparoscopy Department, Minia University Hospital, Minya, Egypt
3 General Surgery Department, Faculty of Medicine, Misr University for Science and Technology, Giza, Egypt
4 General Surgery Department, Minia University Hospital, Minya, Egypt
5 Pathology Department, Minia University, Minya, Egypt

Date of Submission05-Mar-2018
Date of Acceptance15-Mar-2018
Date of Web Publication17-Aug-2018

Correspondence Address:
Mohamed M.T Zaazou
General Surgery Department, Faculty of Medicine, Misr University for Science and Technology, Giza
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/ejs.ejs_36_18

Rights and Permissions

Introduction This study analyzed the histological features modifying the outcome after pancreaticoduodenectomy operation for periampullary tumors.
Patients and methods This study is a prospective cohort. A total of 35 cases of pancreaticoduodenectomy operations were performed from March 2011 to February 2013. Of the procedures, 23 cases were diagnosed as pancreatic carcinoma, and the rest were ampullary carcinomas (7), cholangiocarcinomas (3), and duodenal carcinomas (2). Statistical analysis was completed by using log-rank and Cox regression multivariate analyses.
Results The 5-year survival rate was 29% for all patients who went through pancreaticoduodenectomy. For periampullary carcinomas other than pancreatic carcinoma, the 1, 3, and 5-year survival rates were 100, 66.7, and 58.3%, respectively. The 1, 3, and 5-year survival rates for pancreatic carcinoma were 42.1, 10.5, and 10.5%, respectively (P=0.01). In the multivariate analysis, the existence of both perineural and lymphovascular invasions were the only independent factors influencing outcome. The 5-year survival rate was 88.9% in patients negative for both factors and 0% in patients positive for both (P=0.02).
Conclusion Cases with both perineural and lymphovascular invasions on histopathological analysis have poor 5-year survival outcomes after Whipple’s procedure for pancreatic and periampullary malignant tumors.

Keywords: histopathological lymphovascular invasion, histopathological perineural infiltration, pancreatic tumors, pancreaticoduodenectomy, Whipple operation

How to cite this article:
Mohammed MM, El-Heiny AE, Zaazou MM, Mohamed AE, Toni ND, Khalil ZI. Nontraditional criteria for prediction of patient survival following pancreaticoduodenectomy for malignancy. Egypt J Surg 2018;37:368-74

How to cite this URL:
Mohammed MM, El-Heiny AE, Zaazou MM, Mohamed AE, Toni ND, Khalil ZI. Nontraditional criteria for prediction of patient survival following pancreaticoduodenectomy for malignancy. Egypt J Surg [serial online] 2018 [cited 2020 Apr 4];37:368-74. Available from: http://www.ejs.eg.net/text.asp?2018/37/3/368/239120

  Introduction Top

Long-term survival of patients with cancer in the head of pancreas remains insufficient compared with other abdominal malignancies despite improvements in perioperative and hospital mortality. Traditional factors that affect long-term patient survival include tumor type, tumor size, lymph nodal status, resectability, patient age, and associated diseases. Perioperative mortality had decreased from 30% in the early 1980s [1] to up to 3% in the past decade [2], whereas the 5-year patient survival rates after curative resection remain poor. Most published large series demonstrate postresection survival rates of only 4–17% for pancreatic carcinoma [3],[4]. The picture in periampullary carcinoma is better, where 5-year survival after resection is 60% [5].

The traditional criteria for determination of patient survival after pancreaticoduodenectomy operation include tumor type, size, margin, differentiation, and positive lymphatic infiltration. Two parameters that have not been fully studied are the lymphovascular and perineural infiltration. Groningen in his study reported that perineural involvement is a reliable factor in the prediction of patients’ survival after pancreatic resection with high significance [6]. Two studies have reported that patients’ survival decreases in pancreatic neuroendocrine tumors, when lymphovascular infiltration is present [7],[8]. It is proposed that perineural infiltration may have a role in local failure because of tumor growth along nerves that innervate the pancreas and eventually form the periarterial nerves [9]. In a similar way, the lymphovascular infiltration is responsible for regional or distant metastasis in lymph nodes or other organs.

This study aimed to analyze the histopathological factors affecting long-term survival of patients after pancreaticoduodenectomy focusing on invasion of perineural and lymphovascular components as predictors of 5-years survival.

  Patients and methods Top

Study design

This cohort is a prospective study. The work has been described in line with the STROCSS criteria [10].

Study participants and setting

All patients at Hepatobiliary Surgery Unit of Minia University Hospital and Surgical Department of Misr University for Science and Technology who underwent pancreaticoduodenectomy in the period between March 2011 and February 2013 were included in the study.

Data collection

The general data, including sex, age, clinical data, preoperative laboratory or imaging investigation, preoperative interventional and preoperative biopsy, any neoadjuvant chemotherapy or radiotherapy, of included cases were reviewed. Operative and postoperative data including postoperative morbidity based on Clavien–Dindo classification [11] were recorded.

The gross specimens were examined by a pathologist who was blind to any prior knowledge about the clinical data. Histopathological data were entered into a prospective database maintained since 2011 by the departments of surgery. The histopathological factors analyzed were tumor type, tumor size, tumor grade, regional lymph node status, lymph node capsular invasion, resection margin, and the presence of perineural and lymphovascular invasion. If there was incomplete information about the required histopathological parameters, the histopathological slides (n=4 patients) were re-evaluated.

There were two groups of tumors according to the analysis of gross specimens: pancreatic and periampullary carcinoma. The latter group includes ampullary carcinoma, duodenal carcinoma, and cholangiocarcinoma. Tumors were classified according to size as less than 2 cm and more than 2 cm. Tumor differentiation was classified as well, moderate, or poor. The resection-free margins were analyzed as positive up to 2 mm or negative more than 2 mm. Typically, specimens were inked and then fixed before sectioning for margin analysis. Nodal status was categorized into negative or positive according to the presence or absence of malignant infiltration.

If malignant cells were identified inside the lymphovascular channels, lymphovascular invasion was considered as positive ([Figure 1]a). Perineural infiltration was considered to be present if tumor cells were identified within the perineural space and/or nerve fibers ([Figure 1]b). For statistical analysis purposes, hematoxylin and eosin stained slides in which lymphovascular invasion were not detected on histopathological examination while lymph nodes examination was positive were considered as negative for lymphovascular invasion.
Figure 1 Hematoxylin and eosin sections showing (a) invasion of the tumor in lymphovascular channel(white arrow) away from the main tumor and (b) perineural infiltration (black arrow) by tumor along nerve fibers in the pancreas.

Click here to view


Follow-up visits were obtained by patient interview or personal contact with the attending physician once every 3 months during the first year, re-examined once every 6 months during the second and third years, and re-examined once a year later. Patients were examined during the follow-up visits by routine laboratory investigations, tumor markers, chest radiography, abdominal ultrasound, and computed tomography/MRI. The follow-up deadline was June 2017.

Ethical approval

The protocol of the study was discussed and approved regarding ethics of research in general surgical department. The study had been approved by the ethical committee for human studies in our institution. The study has a registry number (research registry 3064). Full written, informed consent was signed from all participants.

Data analysis

Univariate analysis was applied using the log-rank test with differences in patients’ survival as the outcome measured. Continuous variable, such as tumor size, was converted into categorical variables. Variables that were significant on univariate analysis were then subjected to multivariate analysis using Cox’s multiregression analysis. All statistical analyses were carried out using SPSS, version 23 (SPSS Inc., Chicago, Illinois, USA). Overall survival at 1, 3, and 5 years was determined and analyzed by the Kaplan–Meier method, and differences in survival were compared using the log-rank test.

  Results Top

A total of 35 pancreaticoduodenectomy cases were performed in our institutions and were included in the study over the 2-years period. Of the 35 patients, 33 went through a pylorus-preserving pancreaticoduodenectomy, whereas two patients underwent a total pancreatectomy. In five patients, parts of the portal vein or superior mesenteric vein were removed with the specimen, and the venous structure was reconstructed. No patient underwent resection with arterial involvement during this time period ([Table 1]).
Table 1 Malignancies treated by pancreaticoduodenectomy (n=31)

Click here to view

The median patient age was 58 years (range: 25–65 years). There were 23 men among the patients. The time required for the operative procedure ranged from 5.0 to 10.0 h, with a median of 7.0 h. Estimated blood or fluid loss ranged from 100 to 3000 ml, with a median of 1000 ml. Requirements of blood transfusion ranged from 0 to 6.0 U, with a median of 2.0 U ([Table 2]).
Table 2 Preoperative and operative data of patients

Click here to view

Postoperative morbidity was based on Clavien–Dindo classification: grade 4 was seen in 11.4% of cases, grade 3 in 11.4%, grade 2 in 31.4%, and grade 1 occurring in 37.1%. The most common complications were intra-abdominal abscess; pancreatic fistula and biliary fistula were less frequently seen. There were four (11.4%) in-hospital/30-day deaths, being Clavien–Dindo grade 5.

There was no difference in 5-year survival in term of sex ([Table 3]). The overall 5-year actuarial survival for all patients with malignant tumors was 29% ([Figure 2]).
Table 3 Five-year survival rates by tumor characteristics (n=31)

Click here to view
Figure 2 Cumulative survival in patients with pancreatic adenocarcinoma, and patients with periampullary malignancies. At 5-year survival, the periampullary group had achieved a significantly better 5-year actuarial survival of 58.3%.

Click here to view

Tumor characteristics and patient survival

Most tumors were pancreatic in origin 65.7 versus 20% ampullary, 5.7% duodenum, and 8.6% distal common bile duct. Similar results were obtained in a cohort of patients from the SEER cancer registry who went through pancreaticoduodenectomy between 1993 and 2003: 62.5% pancreatic, 18.9% ampullary, 7% duodenal, and 11.6% distal bile duct [12]. Patients who went through resection for pancreatic adenocarcinoma had poorer survival (10.5% at 5 years) than patients with periampullary malignancies (58.3% at 5 years; P=0.01) ([Figure 3]).
Figure 3 Cumulative patient survival in the presence of perineural infiltration. A significantly poorer survival outcome was noted in the presence of perineural infiltration.

Click here to view

The patient survival was significantly better in smaller and well-differentiated tumors (P<0.0001). The survival was reduced when lymph nodes were positive (P<0.0001). Well-differentiated tumors were related significantly to better survival compared with moderately and poorly differentiated tumors (P<0.0001) ([Table 3]). Long-term survival was poor with positive margin up to 2 mm.

Perineural infiltration and lymphovascular invasion

The absence of any of these factors, malignant perineural infiltration or lymphovascular invasion, was related with high significance to better survival ([Figure 3] and [Figure 4]).
Figure 4 Cumulative patient survival in the presence of tumor invasion in lymphovascular channels. A worse survival outcome was marked in the presence of lymphovascular invasion.

Click here to view

When all the significant histopathological variables on univariate analysis were assessed using a multivariate regression model, perineural infiltration (P<0.0001) and lymphovascular invasion (P<0.0001) were independently significant factors for prediction of long-term survival.

Prognostic modeling using perineural infiltration and lymphovascular invasion

As these two factors were independently significant in multivariate analysis, the effects of combining them in terms of predicting long-term survival were analyzed. The 31 patients with malignancies were divided into four categories consisting of those who were positive for both factors, those who were positive for either one of the factors, and those who were negative for both. [Figure 5] illustrates that significantly better survival result was obtained in patients who were negative for both parameters rather than other groups (P=0.02). The 5-year actuarial survival is tabulated in [Table 4].
Figure 5 Cumulative patient survival in the presence or absence of perineural infiltration and lymphovascular invasion after resection. Significant improvement was noted if both features are absent on microscopic examination.

Click here to view
Table 4 Five-year survival rates after resection of pancreatic malignancies in patients positive and negative for perineural infiltration and lymphovascular invasion (n=31)

Click here to view

  Discussion Top

During the 1980s and 1990s, the morbidity and mortality for Whipple procedure were so high, that many thought the operative procedure should be abandoned for tumors of head of pancreas. During the 2000s, however, results improved, and now many centers are reporting hospital mortality rates of less than 5% [13]. Many factors are definitely responsible. Better intensive care and nutritional support in caring for postoperative complications have almost certainly contributed. During the 1990s, general surgeons specialized so that now a significantly high proportion of pancreaticoduodenectomies are being operated by hepatobiliary surgeons experienced in pancreatic surgery. This specialization has resulted in less time, with intraoperative less blood loss and with better results.

There were four hospital deaths among the 35 patients in our series who underwent a pancreaticoduodenectomy for malignancy of the pancreas. When the 31 hospital survivors are compared with the four patients who died after their surgery, they are demographically similar. The differences were that those patients who died after their procedure had more advanced stage of tumor with positive margin and poor degree of differentiation, lost more blood, required more transfusions, and had a longer duration of operative procedure than those who survived.

There was marked decrease in hospital morbidity and mortality in addition to the improvement in long-term survival of these patients. Most of this improvement is not owing to the drop-in hospital mortality but owing to other factors. Some have attributed the recent improvements to the better advanced radiological diagnosis that allows surgeons to operate on patients with pancreatic carcinoma at earlier stages. However, most patients with cancer of head of pancreas still do not seek for medical advice until the development of jaundice. Moreover, there is similar outcome when tumor size and stage are taken into account in patients going through pancreaticoduodenectomy for pancreatic carcinoma today compared with those patients treated a decade or more ago [14].

Five-year survival rates following Whipple procedure for pancreatic malignancy remain low (<20%) even in large volume institutions. In this study, the actuarial survival was 29% for the 5-year follow-up. This is comparable with the recently published MD Anderson series [14]. The 5-year survival for patients who went through pancreaticoduodenectomy for periampullary malignancies other than pancreatic adenocarcinoma was 58.3%. This result is also similar to previously published results for. periampullary malignancies, including a recently published series from Birmingham which reported an actuarial 5-year survival of 60% for ampullary carcinoma following resection [5].

The patient outcome is poor when the pancreatic resection margin is positive, that is, the surgical margin up to 2 mm [15]. The European Study Group for Pancreatic Cancer defines an R0 resection as having a margin of 1 mm. In our series, a surgical margin of less than 2 mm was considered negative and was associated with better patient survival. A recent study reported a margin of 1.5 mm to be associated with better outcome [16]. If this is borne out by other investigators, a re-evaluation of the R0 margin will be required. The TNM classification and its modifications have been the standard prognostic parameters used for most malignancies, including those of pancreatic origin [17].

Multivariate analysis in this cohort identified that perineural and lymphovascular invasions were independent significant factors for long-term survival, and they are not included as a part of the TNM system. Perineural infiltration as a significant prognostic factor after pancreatic head resection has been reported previously and has been proven to be related to local failure by various authors [13]. As the tumor grows along nerves in the pancreas, it infiltrates distally to follow an arterial channels, reducing the chances of complete microscopic clearance [6],[9].

Lymphovascular infiltration would provide a channel for metastasis as indicated by poorer outcome despite aggressive vein resection, as reported in more patients [9]. Lymphovascular invasion has been noted as a poor prognostic factor following resection of pancreatic neuroendocrine tumors [7]. In cases with no lymphovascular or perineural infiltration, the 5-year survival was 82.4% in our patients who had pancreatic malignancy (n=31). It is apparent that these two parameters, which are easily identified on hematoxylin and eosin sections, are highly suggestive pathological prognostic criteria for prediction of patient survival following pancreatic resection for malignancies.

Currently, perineural and lymphovascular invasions are not included among the traditional prognostic factors after resection of pancreatic or other periampullary cancers. These results would strongly recommend the inclusion of these parameters in the postoperative staging system as standard criteria. We propose that a new staging system specific for periampullary tumors is required for both research and clinical purposes.

  Conclusion Top

The presence of perineural infiltration and lymphovascular invasion on histopathology is highly significant in predicting 5-year outcomes after Whipple’s procedure for periampullary and pancreatic malignancies.

Strengths of the study

  1. This is a prospective cohort study which is one of the strong studies in research.
  2. We operated on cases with pancreatic cancer, which is an invasive cancer in a hidden area in the body.
  3. We focused on perineural infiltration and lymphovascular invasion as predictors of long-term survival.

Limitation of the study

The sample size needed to be increased, but the cases analyzed were the only patients who attended our institutes during the study duration. More studies with a bigger sample size needed to be performed.


The authors thank all patients who participated in the study.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

  References Top

Chen JW, Bhandari M, Astill DS, Wilson TG, Kow L, Brooke‐Smith M et al. Predicting patient survival after pancreaticoduodenectomy for malignancy: histopathological criteria based on perineural infiltration and lymphovascular invasion. HPB (Oxford) 2010; 12:101–108.  Back to cited text no. 1
Cameron JL, Riall TS, Coleman J, Belcher KA. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006; 244:10–15.  Back to cited text no. 2
Cleary SP, Gryfe R, Guindi M, Greig P, Smith L, Mackenzie R et al. Prognostic factors in resected pancreatic adenocarcinoma: analysis of actual 5-year survivors. J Am Coll Surg 2004; 198:722–731.  Back to cited text no. 3
Howard TJ, Krug JE, Yu J, Zyromski NJ, Schmidt CM, Jacobson LE et al. A margin-negative R0 resection accomplished with minimal postoperative complications is the surgeon’s contribution to long-term survival in pancreatic cancer. J Gastrointest Surg 2006; 10:1338–1346.  Back to cited text no. 4
Morris-Stiff G, Alabraba E, Tan Y-M, Shapey I, Bhati C, Tanniere P et al. Assessment of survival advantage in ampullary carcinoma in relation to tumour biology and morphology. Eur J Surg Oncol 2009; 35:746–750.  Back to cited text no. 5
Van Roest MH, Gouw AS, Peeters PM, Porte RJ, Slooff MJ, Fidler V et al. Results of pancreaticoduodenectomy in patients with periampullary adenocarcinoma: perineural growth more important prognostic factor than tumor localization. Ann Surg 2008; 248:97–103.  Back to cited text no. 6
Kazanjian KK, Reber HA, Hines OJ. Resection of pancreatic neuroendocrine tumors: results of 70 cases. Arch Surg 2006; 141:765–770.  Back to cited text no. 7
Nikfarjam M, Warshaw AL, Axelrod L, Deshpande V, Thayer SP, Ferrone CR et al. Improved contemporary surgical management of insulinomas: a 25-year experience at the Massachusetts General Hospital. Ann Surg 2008; 247:165.  Back to cited text no. 8
Christians K, Evans DB. Pancreaticoduodenectomy and vascular resection: persistent controversy and current recommendations. Ann Surg Oncol 2009; 16:789–791.  Back to cited text no. 9
Agha RA, Borrelli MR, Vella-Baldacchino M, Thavayogan R, Orgill DP, Pagano D et al. The STROCSS statement: strengthening the reporting of cohort studies in surgery. Int J Surg 2017; 46:198–202.  Back to cited text no. 10
Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240:205.  Back to cited text no. 11
Gutierrez JC, Franceschi D, Koniaris LG. How many lymph nodes properly stage a periampullary malignancy? J Gastrointest Surg 2008; 12:77–85.  Back to cited text no. 12
Zhao X, Dong J, Huang X, Zhang W, Jiang K. Prognostic factors for survival of patients with ampullary carcinoma after local resection. ANZ J Surg 2015; 85:567–571.  Back to cited text no. 13
Katz MH, Wang H, Fleming JB, Sun CC, Hwang RF, Wolff RA et al. Long-term survival after multidisciplinary management of resected pancreatic adenocarcinoma. Ann Surg Oncol 2009; 16:836.  Back to cited text no. 14
Bilimoria KY, Talamonti MS, Sener SF, Bilimoria MM, Stewart AK, Winchester DP et al. Effect of hospital volume on margin status after pancreaticoduodenectomy for cancer. J Am Coll Surg 2008; 207:510–519.  Back to cited text no. 15
Chang DK, Johns AL, Merrett ND, Gill AJ, Colvin EK, Scarlett CJ et al. Margin clearance and outcome in resected pancreatic cancer. J Clin Oncol 2009; 27:2855–2862.  Back to cited text no. 16
Bilimoria KY, Bentrem DJ, Ko CY, Ritchey J, Stewart AK, Winchester DP et al. Validation of the 6th edition AJCC pancreatic cancer staging system. Cancer 2007; 110:738–744.  Back to cited text no. 17


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]

  [Table 1], [Table 2], [Table 3], [Table 4]


Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
Patients and methods
Article Figures
Article Tables

 Article Access Statistics
    PDF Downloaded49    
    Comments [Add]    

Recommend this journal