Multivisceral resection for adenocarcinoma of the pancreatic body and tail—a retrospective single-center analysis

Background Adenocarcinoma of the pancreatic body and tail is associated with a dismal prognosis. As patients frequently present themselves with locally advanced tumors, extended surgery including multivisceral resection is often necessary in order to achieve tumor-free resection margins. The aim of this study was to identify prognostic factors for postoperative morbidity and mortality and to evaluate the influence of multivisceral resections on patient outcome. Methods This is a retrospective analysis of 94 patients undergoing resection of adenocarcinoma located in the pancreatic body and/or tail between April 1995 and December 2016 at our institution. Uni- and multivariable Cox regression analysis was conducted to identify independent prognostic factors for postoperative survival. Results Multivisceral resections, including partial resections of the liver, the large and small intestines, the stomach, the left kidney and adrenal gland, and major vessels, were carried out in 47 patients (50.0%). The median postoperative follow-up time was 12.90 (0.16–220.92) months. Median Kaplan-Meier survival after resection was 12.78 months with 1-, 3-, and 5-year survival rates of 53.2%, 15.8%, and 9.0%. Multivariable Cox regression identified coeliac trunk resection (p = 0.027), portal vein resection (p = 0.010), intraoperative blood transfusions (p = 0.005), and lymph node ratio in percentage (p = 0.001) as independent risk factors for survival. Although postoperative complications requiring surgical revision were observed more frequently after multivisceral resections (14.9 versus 2.1%; p = 0.029), postoperative survival was not significantly inferior when compared to patients undergoing standard distal or subtotal pancreatectomy (12.35 versus 13.87 months; p = 0.377). Conclusions Our data indicates that multivisceral resection in cases of locally advanced pancreatic carcinoma of the body and/or tail is justified, as it is not associated with increased mortality and can even facilitate long-term survival, albeit with an increase in postoperative morbidity. Simultaneous resections of major vessels, however, should be considered carefully, as they are associated with inferior survival.


Background
Pancreatic adenocarcinoma is a fatal malignant disease with 5-year survival rates below 10% and an increasing incidence worldwide [1]. As typical symptoms often occur at a late stage, 80 to 90% of the patients present themselves with unresectable tumors and dismal prognosis despite efforts of improving non-surgical therapies including chemotherapy regimens [2,3].
In case of resection, severable risk factors impeding long-term survival have been identified in the past, including lymph node metastases, advanced tumor stage, positive resection margins, and distant disease, among others [4,5].
Although the influence of the tumor localization has been a matter of great debate, data from large patient registries have demonstrated that lesions of the pancreatic body and tail, accounting for around 20 to 25% of the pancreatic adenocarcinomas, are associated with inferior survival, most likely as a result of a delayed diagnosis and a more aggressive tumor biology [6][7][8][9]. Distal pancreatectomy is the standard procedure for these tumors; however, advanced tumor stages often require additional resection of extrapancreatic tissue, including large vessels. Data on the effect of such resections is scarce. Recently, Malinka et al. and Panzeri et al. each published results of single-center studies, indicating that multivisceral resections are justifiable in selected patients [10,11]. However, especially the role of vascular resections remains to be elusive. Recent recommendations, for example, discourage arterial resections of the coeliac axis without prior neoadjuvant chemotherapy, whereas venous resections are generally regarded as feasible and safe. Nonetheless, reports are contradictory throughout the available literature, not least because data of patients with pancreatic head and body/tail lesions are often pooled [12,13].
The aim of our study was to evaluate the influence of different types of multivisceral resections, among other potential prognostic factors, on postoperative morbidity and mortality in patients with adenocarcinoma of the pancreatic body and/or tail.

Study design and patient cohort
This is a retrospective, single-center analysis of 94 patients with adenocarcinoma of the pancreatic body and/ or tail undergoing surgery between April 1995 and December 2016 at the Department of General, Visceral and Transplant Surgery, Hannover Medical School, in Hannover, Germany.

Inclusion and exclusion criteria
Included were all patients with intraoperatively and histologically confirmed ductal adenocarcinoma of the pancreatic body and/or tail undergoing resection. No exclusion criteria were defined.

Definition of variables
Multivisceral resections were defined as distal or subtotal pancreatectomy (including lymphadenectomy of the stations 10, 11, and 18 and splenectomy) with additional resection of contiguous or distant organs, including large vessels, also referred to as "extended distal pancreatectomy" and "distal pancreatectomy with resection of non-contiguous organs" according to the International Study Group for Pancreatic Surgery [14,15].
For classification of medical co-morbidities and preoperative fitness prior surgery, the American Society of Anesthesiologists (ASA) Physical Status Classification System was applied [16].
Preoperative anemia was defined according to the World Health Organization as hemoglobin concentrations lower than 12.0 g/dl in female patients and lower than 13.0 g/dl in male patients [17].
Surgical complications were defined as postoperative complications requiring surgical revision or intervention.
Postoperative pancreatic fistulas were defined according to the latest definition of the International Study Group for Pancreatic Surgery [18]. Postoperative pancreatic fistulas grade B and C were defined as clinically relevant.
For the classification of the respective tumor stage, the current AJCC/UICC 8 th edition was applied [19].
The lymph node ratio was analyzed as continuous variable in percent for regression analysis. For visualization of postoperative survival, the lymph node ratio was analyzed as binary variable (cut-off ≥ 20%).
Follow-up time was defined as time between surgery and last contact or death.
Survival times are reported as Kaplan-Meier median estimates, unless stated otherwise.

Study endpoints
The primary endpoint was postoperative survival after pancreatic resection. Secondary endpoints were clinically relevant postoperative pancreatic fistulas and postoperative complications in general.

Statistical analysis
The influence of nominal and ordinal variables on binary endpoints was analyzed with chi-squared test and Fisher's exact test where appropriate.
Continuous endpoints, such as mean and median values, were compared with the Student's t test in case of normal distribution or the Mann-Whitney U test.
Risk factors for postoperative survival were analyzed with Cox regression analysis. Independent risk factors were identified by purposeful selection of variables with a rate of missing values of < 10% and p values in univariable Cox regression of < 0.300 and consecutive stepwise forward selection. Kaplan-Meier analyses including Logrank tests were performed where appropriate.
The identification of risk factors for clinically relevant postoperative pancreatic fistulas was achieved by univariable and multivariable binary logistic regression analysis, as described above.
Statistical significance was set at a p value of < 0.050 and is shown in bold (tables) or marked with an asterisk (figures).
The collected data was implemented and analyzed using SPSS statistical software (version 26; SPSS Inc.; IBM Corporation, Armonk, NY USA), and respective figures were created with GraphPad Prism (version 8.4.0 for Windows, GraphPad Software, La Jolla, CA USA).

Study cohort and preoperative course
Females (51.1%) and males (48.9%) were equally distributed among the cohort of 94 patients. The median age was 65 (41-84) years.
Elevated tumor markers (CA19-9 and CEA) were observed in 42 (44.7%) and 19 (20.2%) patients, respectively, with a high rate of missing values, especially in the early observation period.
A summary of the biometrical and preoperative data is provided in Table 1.

Pancreatic surgery
Distal pancreatectomy was performed in 85 (90.4%) patients. Subtotal pancreatectomy was performed in 9 (9.6%) patients and was neither associated with increased postoperative morbidity (i.e., complications with or without the need of surgical revision; p ≥ 0.050) nor inferior survival (14.36 versus 12.42 months; p = 0.869).
Simultaneous splenectomy was performed in all but one patient, who underwent splenectomy after polytrauma prior to the pancreatic resection.
Due to the invasion of extrapancreatic tissue, 47 (50.0%) patients underwent multivisceral resections, including partial resections of the liver (12 patients), the large and small intestine (12 and 8 patients, respectively), the stomach (14 patients), the left kidney (5 patients), the left adrenal gland (13 patients), the portal vein (15 patients), and the coeliac trunk (5 patients). Of note, in 21 (44.7%) of these patients, preoperative imaging led to the suspected diagnosis of local invasion of neighboring tissue or distant metastases prior surgery, whereas in 24 (51.1%) patients, extrapancreatic disease was diagnosed intraoperatively (missing information in 2 (4.3%) patients). Supplementary Table 1 (Additional file 1) gives an overview of the performed resections. Patients undergoing multivisceral resection were of similar mean age (64.02 versus 66.04 years; p = 0.296), but displayed a significantly lower rate of ASA scores > 2 (16.  Table 2).
Of note, none of the surgical procedures was carried out minimally invasive. Table 1 provides further information on surgical details.
Patients stayed in hospital for a median of 15 (5-109) days.  Six (6.4%) patients died in the postoperative course. Three patients died due to severe bleeding (two patients) and gastric perforation with consecutive sepsis (one patient) as a result of pancreatic fistula. One patient died after severe bleeding caused by a postoperative intraabdominal abscess, and two further patients died after pulmonary embolism with consecutive multiple organ failure.
The median estimated survival was 12.78 months with 1-, 3-, and 5-year survival rates of 53.2%, 15.8%, and 9.0%. Eighty-eight (93.6%) patients were deceased at the time of analysis. Table 2 summarizes selected variables regarding the postoperative course and outcome of the study cohort.

Risk factors for clinically relevant pancreatic fistulas
Univariable regression analysis identified simultaneous partial colectomy as risk factor for the incidence of clinically relevant pancreatic fistulas (OR = 4.000; CI 95% = 1.138-14.063; p = 0.031); however, none of the analyzed variables was an independent risk factor in multivariable analysis (Supplementary Table 2, Additional file 1).

Risk factors for postoperative survival
The results of the univariable Cox regression analysis are displayed in Table 3

Discussion
Patients suffering from pancreatic adenocarcinoma continue to have an extremely poor prognosis. In cases  In-hospital mortality 6 (6. of unresectable disease, median survival of around 7 months has been reported, and although surgical resection is regarded as the only chance for long-term survival, 5-year survival rates after resection of up to around 20% are still unsatisfying [1,[20][21][22]. Studies investigating prognostic factors for postoperative survival often include patients irrespective of the exact tumor localization, although abundant data suggests that lesions of the pancreatic head differ not only in terms of local tumor extent, invasion of adjacent tissue, and necessary surgical strategy to achieve negative resection margins, but also in tumor biology [21][22][23][24][25][26].
In order to optimize preoperative patient selection, we have evaluated the effects of multivisceral resections, among other variables, on the postoperative outcome in patients undergoing distal or subtotal pancreatectomy for ductal adenocarcinoma. In summary, multivisceral resections in general were associated with increased morbidity and the risk of reoperation, but not with increased shortor long-term mortality. Although there are currently no   meta-analyses on the matter, a large multi-center study by Paye et al. and previously published single-center reports support our observations [10,11,[27][28][29][30]. Of note, most of these publications included resections for different types of pancreatic tumors, whereas the current study focused explicitly on histologically confirmed pancreatic ductal adenocarcinoma.
The rate of arterial and venous resections in our patient cohort was comparable to previous publications; however, reports on the effects of vascular resections in case of distal pancreatectomy are conflicting [10,27]. In general, the need for arterial resections to achieve tumor-free resection margins in case of pancreatic adenocarcinoma is seen critical or even as a contraindication for surgery, as it is associated with increased morbidity and mortality [13]. In the current study, resection of the coeliac trunk was associated with a considerable increase in postoperative mortality with two of five patients dying in the postoperative course and a resulting median survival of only 3.52 months. Nonetheless, a recent systematic review and meta-analysis by Gong et al. revealed that distal pancreatectomy with en bloc celiac axis resection can result in favorable survival, despite increased postoperative morbidity, and improved quality of life in selected patients, especially when compared to palliative treatment [31].
As a result of refined surgical techniques, simultaneous resections of the portal and/or the superior mesenteric vein are generally regarded as safe and can facilitate longterm survival [13,32]. Data on the effect of such venous resections in case of distal pancreatectomy is scarce, and recommendations result mostly from reports on pancreatic head carcinoma [33][34][35][36]. Interestingly, recent metaanalyses as well as large single center studies have shown rather discouraging results [35][36][37]. In the current study, simultaneous resection of the portal vein resulted in a median survival of only 7.56 months and only one patient survived more than 5 years. Previous publications did not identify resections of the mesenteric-portal axis as risk factors for survival in case of distal pancreatectomy; however, missing details of the extent of venous resection or mainly minor resections (such as wedge or tangential resections) limit the significance and comparability of such observations [11,27]. In our study, all but one patient underwent segmental resections of the portal vein due to macroscopically suspected invasion. Such resections, unlike tangential resections, were recently identified as an independent risk factor for postoperative survival after resection of pancreatic adenocarcinoma by Serenari et al. [38].
Further prognostic factors identified in the current study were intraoperative blood transfusions, which were demonstrated to result in adverse oncologic effects such as shorter disease-free survival in the past, and a higher lymph node ratio [39]. The latter has been confirmed by several authors including a systematic review for pancreatic adenocarcinoma in general. Thus, apart from the classical nodal status, the lymph node ratio should be an inherent part when planning and optimizing postoperative adjuvant strategies [23,[40][41][42][43][44][45].
The incidence of clinically relevant pancreatic fistulas was comparable to reports of the past [46]. Although we observed a trend towards more fistulas in patients after multivisceral resection (30.4%) and especially after simultaneous partial colectomy (50.0%), multivariable analysis did not identify extended surgery as an independent risk factor. This is consistent with reports from previous single-center studies [30,47]. Recently, Ecker et al. analyzed 2026 consecutive distal pancreatectomies and identified risk factors such as young age, obesity, non-malignant histology, or concomitant splenectomy, among others; however, the authors concluded that despite the considerable size of the study, postoperative pancreatic fistulas cannot be reliably predicted, since individual surgeon performance, applied techniques, and patient-specific pancreatic texture limit the significance of clinical applicability of mentioned observations [46].
Despite long follow-up durations-only one surviving patient was followed up less than 5 years-the retrospective nature of the study and the limited number of patients hamper definitive conclusions. Further limitations of our study are missing information on disease recurrence and systemic therapy, especially as the latter is of increasing importance in a more and more multimodal approach for pancreatic cancer [48,49]. Although reliable data on this matter is still scarce, a review of resections in patients with oligometastatic pancreatic disease underlines the importance of standardizing neoadjuvant and adjuvant chemo(radio)therapy strategies and defines a response towards neoadjuvant systemic treatment as important patient selection criterion along with an adequate performance status as well as resectability of the primary tumor and the metastases [50]. Larger patient cohorts, meta-analyses, and prospective trials are urgently required to confirm our observations in this lethal disease.

Conclusions
Our data indicates that multivisceral resection in case of advanced left-sided pancreatic adenocarcinoma is justified, as it is not associated with increased mortality and can even facilitate long-term survival, albeit with an increase in postoperative morbidity. However, vascular resections are associated with a dismal prognosis and should be performed after thorough consideration and only in selected patients.