Prognostic factors for long-term survival after pancreaticoduodenectomy for periampullary adenocarcinoma. A retrospective cohort study

Background Periampullary adenocarcinoma (PAAC) had a poor prognosis, and pancreaticoduodenectomy (PD) remains the only potentially curative treatment. The study aimed to identify the impact of different clinicopathological factors on long-term survival following PD for PAAC. Patients and methods This study is a retrospective cohort study for the patients who underwent PD for pathologically proven PAAC from January 2010 to January 2019. Statistical analysis was done using Cox regression multivariate analyses for independent risk factors for survival. Result There were 137 patients with PAAC who underwent PD, 79 patients (57.7%) underwent pylorus-preserving PD. Pancreatico-jejunostomy was done in 108 patients (78.8%). The primary analysis showed that risk factors for poor long-term survival include patients with co-morbidities like hypertension or ischemic heart disease, Carbohydrate Antigen 19-9 > 400U/ml, tumor size > 3 cm, poor tumor differentiation, positive lymph nodes invasion, lymphovascular invasion, and Perineural invasion. Multivariate analysis demonstrated that large tumor size > 3 cm (HR: 0.177, 95%CI: 0.084–0.374, P = 0.002), poorly differentiated tumor (HR: 0.059, 95%CI: 0.020–0.0174, P = 0.016), and perineural invasion in the pathological study (HR: 0.101, 95%CI: 0.046–0.224, P = 0.006) were independent risk factors for poor 5-years survival. The prognosis was better in ampullary adenocarcinoma (5-year survival was 42.1%) than pancreatic adenocarcinoma (5-year survival was 24.3%). The 1, 3, 5 and 7-year overall survival rates were 84.5%, 57.4%, 35.9% and 20.1% respectively. Conclusion It seems from the current study that Tumor size > 3 cm, poor tumor differentiation, and Perineural invasion were independent predictors of poor survival in patients with PAAC.


Introduction
Periampullary adenocarcinoma (PAAC) including adenocarcinoma (AC) of pancreatic head, the distal common bile duct (CBD), the second portion of the duodenum, and the ampulla of Vater, it accounts for approximately 0.2% of all gastrointestinal tract tumors. In recent years, the occurrence of periampullary tumors has an increasing trend although is relatively uncommon neoplasm [1][2][3].
Pancreaticoduodenectomy (PD) is the treatment of choice for PAAC, however, only 10-15% are resectable at the time of diagnosis. Patient survival after radical resection of periampullary tumors greatly varies, the different biology of the tumor origin could result to some degree into the difference of prognosis [4][5][6].
Pancreaticoduodenectomy surgery is associated with high morbidity and mortality, therefore it is important to determine which patient can receive benefits from surgery to avoid unnecessary intervention and to facilitate treatment planning of neoadjuvant and adjuvant treatments [12,13]. This study aimed to investigate the prognostic factors for long-term survival in resectable PAAC.

Patients and methods
We conducted a retrospective study to patients who underwent PD for PAAC between January 2010 to January 2019 at the department of Hepato-pancreato-biliary surgery, National Liver Institute, Menoufia University, Egypt. Data were retrieved from the prospectively collected pancreatic database and patients' medical files, after local Institutional Review Board approval. The research goes with the standards of the Declaration of Helsinki and ethical guidelines and was registered in the clinical trial no ChiCTR2000034782. The study was written in line with the Strengthening the Reporting of Cohort Studies in Surgery (STROCSS) criteria [14]. Patients with confirmed PAAC in the pathological study of the specimen after surgery were included in our study. Other pathological types of lesions after PD were excluded from the study. Data on preoperative, Intraoperative, and postoperative care were collected and analyzed.

Preoperative evaluation
Magnetic resonance image (MRI) or multi-detector abdominal computed tomography (CT) with three-dimensional reconstructions are used to evaluate the periampullary tumors and its relation to vascular structures. Endoscopic ultrasound was done for cases with suspicious diagnosis and for determining the relation of the mass with the surrounding vessels. Preoperative endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous trans-hepatic drainage (PTD) was done in case of cholangitis or delayed surgery.

Surgical procedure and pathological evaluation
Laparotomy was done by bilateral subcostal or midline incision. Patients underwent classical Whipple's operation or pylorus-preserving pancreaticoduodenectomy (PPPD). Pancreatic reconstruction was done by either pancreatico-jejunostomy (PJ) or pancreatico-gastrostomy (PG). Wedge or segmental resection of the portal vein or superior mesenteric vein was performed if a pancreatic head mass was inseparable from the vein. The histopathological features of the specimens were analyzed according to; tumor origin, size, grade, resection margin, lymph node (LN) invasion, perineural and lymphovascular invasion. According to Royal College of Pathologists' guidelines on reporting histological outcomes after major pancreatic resections [15], perineural infiltration was considered positive if tumor cells were identified within the perineural space and/or nerve fibers whereas lymphovascular invasion was defined as the presence of tumor within an endothelial lined or lymphatic space (Fig. 1).

Postoperative follow-up
Follow-up has been arranged in the outpatient clinic or through personal contact, every three months in the first year, in the second and third years every six months then yearly later on. The follow-up was from the date of surgery until July 2020 with a median period of follow up 39 months. Long-term survival was considered ≥ 5-year of survival. Postoperative complications were classified according to the Dindo-Clavien grading system [16]. Postoperative pancreatic fistula (POPF), post pancreatectomy hemorrhage (PPH), and delayed gastric emptying (DGE) were defined and graded according to the International Study Group for Pancreatic Surgery (ISGPS) [17][18][19].

Statistical analysis
Statistical analysis was done using SPSS 23 (SPSS Inc., Chicago, IL). Fisher's exact or Chi-square X 2 test was used for categorical variables comparison. For comparing 2 or more continuous variables, the Mann-Whitney U test or Kruskal-Wallis test was used respectively. Overall survival (OS) rates in different groups were done by using the Kaplan-Meier method, while the survival rate comparison was applied by the Log-rank test. Cox's regression model was appraised for the multivariate analysis in case of any significant variable in the univariate analysis. Pvalue was considered to be statistically significant if less than 0.05.

Demographic and preoperative data of the patients
During this study, 137 patients underwent PD for PAAC. Patient demographics and characteristics are shown in Table 1. Of these patients 84 (61.3%) were male and the mean age was 56.8 years. The main complaint was jaundice in 112 patients (81.8%). Preoperative biliary drainage was done in 58 patients (42.3%); by ERCP in 42 patients and PTD in 16 patients.

Operative, pathological, postoperative data and complications
Seventy-nine patients (57.7%) underwent PPPD and 65.7% patients had PDAC in the pathological study. According to the type of pancreatic reconstruction; PJ was done in 108 patients (78.8%), mainly by duct to mucosa in 68 patients, whereas PG was done in 29 patients (21.2%). Seventeen patients (12.4%) underwent vascular reconstruction; PV or SMV reconstruction was done by lateral venoraphy in 10 patients and end to end primary repair using 6/0 proline in 7 patients. Postoperative pancreatic fistula was found in 26 patients (19%) and mainly grade A POPF. Other operative pathological and postoperative data were shown in Tables 2 and 3.

Discussion
Pancreaticoduodenectomy operation remains the standard curative approach for periampullary tumors. Despite several refinements in the surgical technique with the improvement of postoperative mortality, the long-term prognosis still disappointing with a 5-years survival rate rarely to exceed 20% in some centers. These results raised the enthusiasm to search for the main factors that can improve the prognosis of periampullary tumors with the optimal resection [1,4,6,12].
In the PAAC the long-term survival rate varies in a wide range related to the different anatomical locations in the periampullary region. El Nakeeb et al., in their study, showed that 5-year survival was 20.6% in PAAC with a median survival of 34 months. The worst prognosis was reported in pancreatic head AC with 5%-20% 5-year survival, and a better prognosis was in ampullary and duodenal AC with 5-year survival 30%-65% [20]. Also, Zakaria et al., demonstrated that 5-year survival rate in patients with PDAC was 23.4% [21].
Serum CA19-9 has manifested as a clinically valuable biomarker of pancreatic cancer, and it has proved that higher serum CA19-9 level preoperatively can predict poorer survival of pancreatic cancer after resection [27]. However, there were few studies on the prognostic value of CA19-9 in periampullary cancer. Gao et al. have suggested that periampullary cancer patients with preoperative serum CA19-9 > 35 U/ml are prone to have a poorer survival [28]. Also, El Nakeeb et al., showed that preoperative serum CA19-9 > 37 U/ml was associated with a poor survival rate [20]. In the present study the elevated CA 19-9 (>400U/ml) had a statistical significance risk for poor survival in univariate analysis, similar to the previous studies.

Table 1
Demographic and preoperative data of the patients.

Variables
Patients with PD (n = 137) According to patients with IHD, a previous investigation reported that there was an association between IHD and mortality after PD that did not remain significant in the multivariate model [29], as seen in our study.
Tumor size is a well-established predictor of survival. In general tumor size <3 cm has a better prognosis. In some studies it was only significant in univariate analysis [30]. Other studies, reported that the size of the tumor was independent predictors of survival [22]. In our study, both univariate and multivariate analysis demonstrated that tumor size > 3 cm was significantly independent risk factor for poor survival.
Venous reconstruction can be done if there is an invasion of portomesentric access to achieve R0 resection with accepted postoperative morbidity and mortality. Some studies showed that the resection margin was an independent risk factor for survival, and R0 achieved significantly better OS [6,7,20,31]. In contrast, the meta-analysis study by Butturini et al., found that resection margin was not a significant prognostic factor for survival [32] as seen in our study, it may be due to the difference in the pathological definitions and findings of the resection margin.
Also, tumor differentiation has been reported to be associated with the progression of PAAC. Most studies in the multivariate analysis reported that poor tumor differentiation was a poor prognostic factor for survival [26,33,34], similarly, to what is seen in our series.
Other histopathological characters like lymph node metastasis and lymphovascular invasion should be regarded as an independent predictor of survival and may have therapeutic and prognostic implications for patients [11,33,34]. The poor OS reported in the study by Al-Jumayli et al., was likely due to the high rate of tumor invasion and extension [35]. As the tumor grows along nerves in the pancreas, it infiltrates distally to follow an arterial channel, reducing the chances of complete microscopic clearance [36]. Zhao et al., reported in their study that the perineural infiltration was a significant prognostic factor after pancreatic head resection and has been proven to be related to local failure [37]. The perineural invasion appeared to be the most significantly associated with 1-year mortality [10,12,38]. In our study, the perineural invasion was significantly independent risk factors for poor OS.
Panaro et al., concluded that PDAC is considered a systemic disease, and microvascular invasion is a major prognostic factor after PD as it can PD (pancreaticoduodenectomy), PPPD (pylorus preserving pancreaticoduodenectomy), CBD (common bile duct), SD (standard deviation). LN (lymph nodes). lead to distant metastasis, but unfortunately, we cannot predict microvascular invasion in the preoperative image, so it raised the question about the significance of neoadjuvant therapy for all resectable pancreatic cancer, that needs further studies [12]. The limitations of this study are its retrospective nature and singlecenter experience that is liable for statistical bias, PD surgery was done by different surgeons but they almost have equal experience, the biological behavior of the different PAAC that may have also racial variations may affect the result between centers and needs further study and there is no complete data about the postoperative adjuvant therapy.
In conclusion: It seems from the current study that the predictors of poor long-term survival in patients with PAAC were patients with comorbidities like HTN or IHD, CA19-9 > 400U/ml, tumor size > 3 cm, poor tumor differentiation, LNs invasion, lymphovascular invasion, and Perineural invasion. However, after multivariate analysis tumor size > 3 cm, poor tumor differentiation, and Perineural invasion were independent risk factors of poor long-term survival. Patients with ampullary AC had better mean survival than patients with pancreatic AC.

Provenance and peer review
Not commissioned, externally peer reviewed.

Statement of ethics
The research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The patients have given their written informed consent on admission and pre-operatively to use their prospective database and files for research work. The study protocol was approved by the National Liver Institute committee and review board.

Consent
The work has been approved by the National Liver Institute ethical committees, in which the study was performed and the patients gave informed consent to use their retrospectively collected data from files for study and research work.

Financial support
No.