The influence of neural invasion on survival and tumor recurrence in pancreatic ductal adenocarcinoma – A systematic review and meta-analysis

Surgical Oncology 26 (2017) 105e115

Contents lists available at ScienceDirect

Surgical Oncology journal homepage: www.elsevier.com/locate/suronc

Review

The influence of neural invasion on survival and tumor recurrence in pancreatic ductal adenocarcinoma e A systematic review and metaanalysis Stephan Schorn a, Ihsan Ekin Demir a, Bernhard Haller b, Florian Scheufele a, Carmen Mota Reyes a, Elke Tieftrunk a, Mine Sargut a, Ruediger Goess a, Helmut Friess a, Güralp Onur Ceyhan a, * a b

Department of Surgery, Klinikum Rechts der Isar, Technical University of Munich, Germany €t München, Munich, Germany Institute of Medical Statistics and Epidemiology, Klinikum Rechts der Isar, Technische Universita

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 November 2016 Received in revised form 28 January 2017 Accepted 31 January 2017

Objectives: To assess the impact of neural invasion/NI on overall survival/OS and tumor recurrence in pancreatic ductal adenocarcinoma/PDAC. Summary background data: NI is a histopathological hallmark of PDAC. Although some studies suggested an important role for NI on OS, disease-free/DFS and progression-free survival/PFS in PDAC, there is still no consensus on the actual role of NI on survival and local recurrence in PDAC. Methods: Pubmed, Cochrane library, Ovid and Google Scholar were screened for the terms “pancreatic ductal adenocarcinoma”, “pancreatic cancer”, “survival”, “tumor recurrence” and “perineural invasion”. The PreferredeReporting-Items-for-Systematic-review-and-Meta-Analysis/PRISMA-guidelines were used for systematic review and meta-analysis. Articles meeting predefined criteria were critically analysed on relevance, and meta-analyses were performed by pooling univariate and multivariate hazard ratios/HR. Results: A total number of 25 studies on the influence of NI on tumor recurrence, and 121 studies analysing the influence of NI on survival were identified by systematic review. The HR of the univariate (HR 1.88; 95%-CI 1.71e2.07; p < 0.00001) and multivariate meta-analysis (HR 1.68; 95%-CI 1.47e1.92; p < 0.00001) showed a major impact of NI on OS. Likewise, NI was associated with decreased DFS (HR 2.53; 95%-CI: 1.67e3.83; p ¼ 0.0001) and PFS (HR 2.41; 95%-CI: 1.73e3.37: p < 0.00001) multivariate meta-analysis. Conclusions: Although the power of this study is limited by missing pathological procedures to assess the true incidence of NI, NI appears to be an independent prognostic factor for OS, DFS and PFS in PDAC. Therefore, NI should be increasingly considered in patient stratification and in the development of novel therapeutic algorithms. © 2017 Published by Elsevier Ltd.

Keywords: Pancreatic cancer Neural invasion Survival Overall survival Recurrence-free survival Disease-free-survival Systematic review Meta-analysis

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.1. Study design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.2. Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.2.1. Histological characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.2.2. Survival time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.2.3. Patient characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

* Corresponding author. Department of Surgery, Klinikum Rechts der Isar, Technical University of Munich, Ismaninger Str. 22, D-81675 München, Germany. E-mail address: [email protected] (G.O. Ceyhan). http://dx.doi.org/10.1016/j.suronc.2017.01.007 0960-7404/© 2017 Published by Elsevier Ltd.

106

3.

4. 5.

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

2.2.4. Selection bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 2.2.5. Survival data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 2.3. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 2.4. Investigation of publication bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 3.1. Search results and characteristics of the included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 3.2. The impact of neural invasion on overall survival . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 3.3. The impact of neural invasion on tumor recurrence in PDAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 3.4. Analysis of publication bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Author contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

1. Introduction

conclusive statement on the impact of Pn in PDAC.

Pancreatic ductal adenocarcinoma (PDAC) is associated with a dismal 5-year survival rate <5% [1]. Despite intensive research, the median survival time of patients with PDAC has remained nearly constant along the last decades [1]. In curative therapy, surgery followed by adjuvant therapy remains the standard treatment for patients with resectable, non-metastatic PDAC [2]. Nevertheless, rates of failure, i.e. tumor recurrence, after even radical resections are high. Here, several randomized controlled trials including patients treated with curative intention demonstrated failure rates of tumor therapy exceeding 70% [2,3]. Moreover, autopsy results of patients, who were resected for PDAC, showed metastasis and local recurrence rates of 75%e88% [4]. So far, margin status [5], tumor stage and grade [6], lymph node metastasis [7] and invasion into adjacent organs and structures [8], could be identified as important prognostic factors in patients with PDAC. Neural invasion (Pn) is a pathohistological hallmark of PDAC [9]. Although it is similarly encountered in other solid tumors like colorectal cancer, prostate cancer and gastric cancer, the reported incidence of Pn in PDAC reaches up to 100%, surpassing any known solid tumor [10]. Furthermore, the severity of Pn was reported to be strongest in PDAC when compared to all other gastrointestinal malignancies [10]. This observation implies that Pn is a very characteristic, omnipresent and ominous feature of PDAC [10]. Intractable pain is one of the dominant symptoms of patients with PDAC [11,12]. Here, several studies could show that the severity of Pn is closely associated with the abdominal pain symptoms of PDAC patients [11]. Due to the ubiquitous neuropathic changes in PDAC including Pn, intrapancreatic neural hypertrophy and increased neural density, the pain sensation in PDAC was recognized as neuropathic [11]. Moreover, this neuropathic pain was found to be directly associated with impaired survival in patients with PDAC and identified as a novel prognostic factor in pancreatic cancer [11,12]. Until now, some studies demonstrated that Pn may impact survival and tumor recurrence in PDAC. However, a clear consensus about the exact impact and importance of Pn on overall (OS), disease-free (DFS) and progression-free survival (PFS) has not yet been defined. Whereas some studies could show a strong correlation between Pn and OS [13,14], there were also some studies showing no significant correlation between OS and Pn [15] or even improved OS for patients with Pn [16,17]. Therefore, in the current study, we performed a systematic review of the literature for the influence of Pn on survival, prognosis and local recurrence in PDAC. Furthermore, we conducted a metaanalysis of studies correlating Pn with OS, PFS and/or DFS to reach a

2. Methods 2.1. Study design To perform this systematic review and meta-analysis, we adhered to the Preferred Reporting Items for Systematic review and Meta-Analysis (PRISMA) [18] guidelines. As recommended by the PRISMA guidelines, this systematic review and meta-analysis was registrated on the international prospective register of systematic reviews (CRD42016038775). The following databases were systematically screened for literature: Pubmed, Cochrane library, Ovid and Google Scholar. The search strategy included the following items: “pancreatic cancer”, “pancreatic ductal adenocarcinoma”, “neural invasion”, “perineural invasion”, “survival” and “recurrence”. All studies of patients with PDAC that analysed the influence of Pn on survival and/or tumor recurrence were eligible for inclusion. 2.2. Data extraction After removing duplicates, topics and abstracts were independently screened by three reviewers (SS, IED and FS) for possible inclusion. Articles, which did not show any reference to PDAC or survival, were excluded from further analysis. Studies containing numeric data about the influence of Pn in PDAC on survival and recurrence were included. Studies published in other languages other than English were also excluded. Full-text articles were analysed for the influence of Pn on OS, DFS and PFS and for data match, and the following inclusion criteria were considered in the systematic review and meta-analysis. 2.2.1. Histological characteristics Only patients with proven histology of pancreatic ductal adenocarcinoma were included in the study. Publications dealing with other pancreatic tumors were excluded. Moreover, due to a different tumor biology and often a less aggressive clinical course, studies that included solely IPMN-based cancers were excluded from further analysis. Studies that compared the survival intervals or rates between patients without histological signs for neural invasion (Pn0) and patients with neural invasion (Pn1) were included in the meta-analysis. 2.2.2. Survival time Here, OS was defined as the time between resection and death. DFS was defined as the time interval between resection and tumor

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

107

recurrence. PFS was defined as time between operation and tumor relapse or death. Only publications that provided concrete information about DFS/PFS definitions were included in the metaanalysis. Meta-analysis was performed by pooling univariate and multivariate hazard ratios (HR) considering their associated confidence-intervals (CI) or their exact p-values of OS data. In the same way meta-analysis of DFS and PFS were performed by pooling HR from multivariate studies including either the 95%-CI or the exact p-value. 2.2.3. Patient characteristics Studies, which contained patients who all received neoadjuvant therapy or total pancreatectomy were excluded from the study. 2.2.4. Selection bias To avoid any selection bias caused by inclusion of same patients, only the latest study of each study group were included. 2.2.5. Survival data Since the handbook of the Cochrane library strongly recommit to use HR for any time-to-event-analysis [19] we decided to exclude any other studies which provided other information but HR to assess the true impact of Pn on OS, DFS or PFS. 2.3. Statistical analysis Statistical analysis was performed using Review Manager Software (Review Manager/RevMan, Version 5.3, Copenhagen, The Nordic Cochrane Centre, The Cochrane Collaboration, 2012). In the meta-analysis, heterogeneity between the included studies was quantified using the inconsistency statistic (I2). If I2 was less than 50%, the Generic inverse variance method for fixed effects was performed to pool HR, whereas I2 was more than 50%, the DerSimonian method for random effect was used to pool HR [20], providing an estimate for an average HR. Results of meta-analysis are expressed as pooled HR (95%-CI). Pooled HRs >1 implied an overall negative effect of Pn on OS, DFS and PFS in patients with PDAC. A two-sided p-value was calculated and a level of significance of a ¼ 0.05 was used. The results of the systematic review were expressed as the median of median survival times and its range from minimal to maximal median survival time. 2.4. Investigation of publication bias Using RevMan 5.3, a Forest blot was performed of each metaanalysis. Publication bias could be excluded if each point is evenly distributed and lying within the 95%-CI. In the univariate analysis, the 95% is indicated by the virtual triangle. 3. Results 3.1. Search results and characteristics of the included studies The initial database research yielded a total of 1300 clinical studies (Fig. 1). After removing duplicates and excluding studies without any reference to survival time or tumor recurrence and Pn, 146 studies could be identified, which met inclusion criteria for the systematic review. However, after exclusion of 118 studies which didn't contain any HR or did not provide any further data like a 95%-CI or an exact pvalue 23 studies [13e15,21e40] meeting all inclusion criteria for meta-analysis containing 3538 patients were included in the univariate survival meta-analysis. Among these 2535 patients 71.7% (median over studies: 64.2%, range: 43.2e91.9%) - showed Pn, whereas 1,003, i.e. 28.3% (Median: 35.8%, Range: 8.1e56.8%)

Fig. 1. Screening and identification process of the Systematic Review and MetaAnalysis: Pubmed, Cochrane Library, Ovid and Google Scholar were screened for the terms “Pancreatic Cancer”, “Pancreatic Ductal Adenocarcinoma”, “Survival”, “Recurrence”, “Neural Invasion” and “Perineural invasion”. According PRISMA-Guidelines, only HR were pooled for the meta-analyses. Here, 23 studies out of a total of 1300 identified studies were included in the meta-analysis of univariate OS, and 36 studies were eligible for the meta-analysis of multivariate OS, and 4 studies qualified for the multivariate meta-analysis of DFS and 3 studies for meta-analysis of PFS.

showed no tumor invasion into intrapancreatic nerves. Furthermore, 36 clinical studies [5,7,13e17,22,24,26e28,31e33,36,37,40e58] including 6885 patients qualified to be included in the multivariate meta-analysis. Here, 5392 out of 6885 patients (78.3%, Median: 76.2%, Range: 47.8e91.9% %) showed Pn. For the multivariate meta-analysis, 4 studies with a total number of 478 patients for DFS [16,59e61] and 3 studies with a total amount of 206 for PFS [26,27,37] met the inclusion criteria. However, from the 206 patients in the meta-analysis of PFS, only 125 (60.7%, Median: 59.9%, Range: 65.3%e54.5%) were reported to exhibit Pn, whereas 365 out of 478 patients in the DFS analysis (76.4%, Median: 63.5%, Range: 86.9%e50.9%) were Pn-positive. 3.2. The impact of neural invasion on overall survival To investigate the influence of Pn on survival, the median survival times of patients with PDAC were correlated to the presence

108

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

or absence of Pn (Table 1). Here, patients with Pn showed a notably decreased median survival time of 16 months (Median: 15.4 months, range 7.0e29.9 months) compared to 29.2 months (Median: 24.4 months; range: 8.0e56.1 months) among patients without Pn. Since median survival times are not eligible for meta-analyses [62], univariate and multivariate hazard ratios were extracted and pooled for final meta-analysis. Here, 23 studies [13e15,21e40,44,63e66] could be identified matching the inclusion criteria for the univariate meta-analysis of survival time. Importantly all of these 23 studies showed a HR higher than 1, underlining the negative impact of Pn on patients' survival. The meta-analysis showed a clear unfavourable effect of Pn on survival of PDAC patients (HR 1.88; 95%-CI 1.71e2.07; p < 0.00001, Fig. 2). Heterogeneity was absent in the univariate OS meta-analysis (I2 ¼ 0%; Chi2 ¼ 20.08; p ¼ 0.64) (Fig. 2). To further determine the prognostic impact of Pn on OS, a multivariate analysis was performed. For this, a total of 36 multivariate HRs [5,7,13e17,22,24,26e28,31e33,36,37,40e58] were extracted to investigate the association of Pn on OS. Here Pn could again be identified as an independent dismal prognostic factor on patient survival in PDAC (HR 1.68; 95%-CI 1.47e1.92; p < 0.00001; Fig. 3). Each multivariate HR which was included in our metaanalysis of multivariate OS Pn was adjusted for several histopathological features which showed any impact on OS in univariate OS analysis. However, Fig. 5 summarises the most recent clinical and histopathological factors, for which the single multivariate HRs used in the meta-analysis, were adjusted (Fig. 5).

with a clear definition of DFS/PFS were included. After screening the medical literature, a total of 4 studies containing HRs of DFS time [16,59e61] and a total of 3 HRs of PFS [26,27,37] matching all inclusion criteria could be pooled in the meta-analysis. Corresponding to its effect on OS, Pn was identified as a dismal prognostic factor which doubled the risk of local recurrence in PDAC when regarding DFS (HR 2.53; 95%-CI: 1.67e3.83; p ¼ 0.0001, Fig. 4a) and PFS (HR 2.41; 95%-CI: 1.73e3.37: p < 0.00001; Fig. 4b]. Heterogeneity was only 16% in the meta-analysis of DFS (I2 ¼ 16%; Chi2 ¼ 3.56; p ¼ 0.31, Fig. 4a), whereas heterogeneity was absent in the meta-analysis of PFS (I2 ¼ 0%; Chi2 ¼ 2.60; p ¼ 0.63, Fig. 4b). Like in the meta-analysis of the impact of Pn on OS, the most common used histopathological factors for which Pn was adjusted in the different multivariate OS analysis of the DFS and PFS metaanalyses were shown in Fig. 5 (Fig. 5). 3.4. Analysis of publication bias To exclude any publications bias in the univariate and multivariate meta-analysis, a Funnel plot of the meta-analyses of univariate and multivariate OS was created. Because The Cochrane Collaboration advise against performing Funnel plot of metaanalysis including less than 10 studies [67] no Funnel plot were created of meta-analysis of DFS and PFS. Importantly, all HRs of univariate meta-analysis (Fig. 6) were within the pseudo 95%-CI, suggesting the absence of any publication bias in each metaanalysis. Furthermore, the homogeneous distribution of the HRs of meta-analysis of multivariate OS HRs also underlines the absence of publication bias (Fig. 6).

3.3. The impact of neural invasion on tumor recurrence in PDAC 4. Discussion To determine the potential impact of Pn on tumor recurrence in PDAC, DFS as the time between operation and tumor recurrence and PFS as the time between operation and death of any cause or tumor progression were analysed. To avoid any bias, only studies

The present study underlines the unfavourable impact of perineural invasion/Pn on the clinical outcome in PDAC, and identifies Pn as a key independent prognostic factor for survival and tumor

Table 1 The influence of Pn on median survival. Author

Median Survival (months) Pn-positive/ Pn-negative

p-Value

Author

Median Survival (months) Pn-positive/Pn-negative

p-Value

Andren-Sandberg et al. 2011 [77] Aprile et al. 2012 [79] Bachellier et al. 2012 [64] Biankin et al. 2002 [81] € ttinger et al. 1990 [82] Bo Bouvet et al. 2000 [84] Cameron et al. 1991 [86]

22 7 17 9 13.1 12 12.4

36 10 8 11.2 18.4 29 43

p < 0,05 p 0.12 p 0.063 p 0.0140 ns e ns

Lenz et al. 2011 [78] Meyer et al. 2000 [80] Murata et al. [46] Ozaki et al. 1999 [7] Perini et al. 2008 [83] Perini et al. 2013 [85] Poultsides et al. 2010 [87]

p 0.760 e e p < 0.001 p 0.026 p 0.05 e

Chang et al. 2009 [5] Chatterjee et al. 2012 [69] Chu et al. 2010 [21] Gebhardt et al. 2000 [90] Geer et al. 1992 [92] Goh et al. 2008 [93] Griffanti-Bartoli et al. 1994 [95] Hatzaras et al. 2011 [42] Hong et al. 2012 [98] Jamieson et al. 2010 [99]

16.2 28.5 15 13.2 15 15 13 17.7 15.6 16.7

24.4 56.1 20 18 18 24 15 47.9 20.8 18.2

p 0.002 p 0.0002 p 0.13 e p 0.99 p 0.295 ns p < 0.00001 p 0.49 p 0.82

Sahin et al. 2012 [48] Salem et al. 2015 [88] Schnelldorfer et al. 2008 [89] Shimada et al. 2006 [91] Shimada et al. 2011 [72] Shin et al. 2014 [94] Skalicy et al. 2006 [96] Sperti et al. 1996 [97] Sperti et al. 1997 [3] Takahashi et al. 1997 [70]

Jamieson et al. 2012 [100] Jamieson et al. 2012 [22] Jin et al. 2012 [14] Kazanjian et al. 2008 [24] Lee et al. 2012 [59] Lee et al. 2014 [54]

16.2 16.1 11.5 19.5 16.5 15.4

27.5 18.2 18.5 43.1 38.1 38.1

p 0.003 p 0.033 p 0.414 e p 0.083 p 0.041

Takai et al. 2002 [71] van Roest et al. 2008 [52] Zeng et al. 2014 [101]

16.7 13.2 17.3 15 23.7 29.9 PDAC: 18 IPMN: 25 29.4 19.6 15 Grade 2/3: 20 Grade 2/3: 7 15.63 11.2 14 7 grade 2: 11.2 grade 3: 8.8 Grade 2/3: 10,63 13.1 13.12

22.1 18 25 38.9 39.6 49.2 25 92 51.7 22.7 41 Grade 0/1: 44 Grade 0/1: 37 34.7 11 24 10 grade 0: 78.3 grade 1: 29 Grade 0/1: 10,75 36 21.2

p < 0.001 p 0.24 p 0.06 e e e p 0.11 e p 0.37 p 0.002 p 0.7374 p < 0.0001 p 0.006

In all eligible studies, patients with Pn showed a shorter median survival time compared to patients without Pn. Here, the median survival time of patients with Pn was 16 months (range 7e29.9 months) whereas patients without nerve invasion showed a median survival time of 29.19 months (range: 8e56.1 months). ns ¼ not significant.

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

109

Fig. 2. The univariate Meta-analyses of OS demonstrate a strong unfavourable effect of Pn on OS: To evaluate the effect of Pn, univariate HRs were pooled in the meta-analysis. Here, the pooled HR was 1.88 (95%-CI 1.71 - 2.07; p < 0.00001) indicating Pn as a histopathological factor that is associated with decreased survival among patients with PDAC. Heterogeneity was absent in this meta-analysis.

progression in PDAC. Although the majority of the eligible studies were retrospective, the results of the different meta-analyses in the present study are in line with the current literature [10,68,69] that points out Pn as a prognostic relevant, unfavourable feature of PDAC. Moreover, retrospective studies harbour the risk for selection bias which mostly occurs when systematic differences in baseline characteristics between interventional and control group affects outcome of the study [19]. Importantly, due to inclusion of observational studies no essential difference in association between study outcome (OS, DFS and PFS) and Pn could be assumed. Therefore, the risk for publication bias should be minimised within this study. Moreover, it should be mentioned that a major limitation of this study is the high variance in the frequency of Pn between studies, ranging between 60 and 100% in different series [11]. This strong variance is undoubtedly attributable to the lack of standardized slicing and analysis techniques in the pathological processing of PDAC specimens. In our meta-analysis, we detected Pn-rates varying between 43.2% and 100% [10]. One should consider that the site of Pn may be intra- or extrapancreatic neural plexus, and depending on the site of Pn and the analysed tissue, one could classify Pn as intra-pancreatic intratumoral, intra-pancreatic extratumoral or extra-pancreatic retroperitoneal Pn. For this purpose, we screened our database of the systematic review for the incidence of intrapancreatic and extrapancreatic neural invasion. Here, only four studies provided sufficient information about the incidence of intrapancreatic Pn [7,70e72]. Here, the incidence of intrapancreatic Pn ranged from 76.2% to 97.8% [7,70]. Moreover, three studies reported the incidence of extrapancreatic Pn [70e72]. Here, the incidence of extrapancreatic Pn varied from 52.2% to 75.8% [70,71]. Thus, it is plausible that this variation of the incidence of intrapancreatic and extrapancreatic Pn may also contribute to the variation of the reported incidence of Pn in the included studies. Nevertheless, we assume that representative

tumor slides were available in all studies that were included in our meta-analysis. Indeed, even in our series, although Pn was detectable in every patient, its severity (e.g. frequency of tumor penetration) was variable among the patients. Therefore, it is presumable that primary PDAC tumors with a low severity of Pn may have been falsely classified as free of Pn in the included studies of this meta-analysis. Therefore, despite this limitation caused by “false-Pn-negative cases”, a role for Pn seems to be still underlined by the results of the current meta-analysis. The clinical impact of neural invasion in tumor specimens has not been sufficiently analysed since the UICC's inclusion of Pn into the 2010 TMN classification [73]. Particularly, a consensus about the clinical consequences of the presence of Pn has not been reached yet. Beside, in a recent study, we systematically analysed the tumor specimens from nearly 2000 patients with 10 different GImalignancies for the incidence and severity of Pn by means of our novel Pn severity score [10,11]. Strikingly, patients suffering from PDAC showed the highest frequency and the highest severity of Pn [10] among all gastrointestinal tumor entities. Here, after systematic revaluation, the prevalence of Pn in patients with PDAC was 100%, which implies the omnipresence of Pn in PDAC [10]. Importantly, increasing severity of Pn was associated with dramatically diminished OS and could also be identified as an independent factor for overall survival in PDAC [10]. Indeed, our meta-analysis verifies the observations from our previous study by demonstrating Pn as an unfavourable prognostic factor of OS, PFS and DFS in PDAC. The reasons behind the prognostic significance of Pn should be subject to further future investigations. In this context, a metaanalysis of clinical trials published by Butturini et al., in 2008 [74] demonstrated that the resection margin alone (R0 vs. R1) did not affect survival in subgroups of patient who were treated either with adjuvant chemotherapy or with chemoradiotherapy [74]. The 2and 5-year survival rates of R0-resected patients were 33% and 16%

110

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

Fig. 3. The multivariate Meta-analyses of OS demonstrate Pn as an independent, unfavourable factor on OS: In the multivariable meta-analysis, HRs from multivariable regression models were pooled to investigate Pn as an independent factor on patient survival. Importantly, the meta-analysis of 36 studies also identified Pn as an independent, unfavourable factor of OS with a pooled HR of 1.68 (95%-CI 1.47e1.92; p < 0.00001). Because heterogeneity (I2 ¼ 61%) was strongly present in this meta-analysis, the DerSimonian method for random effect models was used to perform the meta-analysis.

and not different from the 29% and 15% among R1-resected patients. Accordingly, the pooled HR also did also not show any clinically relevant differences (HR 1.10; 95%-CI 0.94e1.29; p ¼ 0.24). One can thus conclude that even after curative treatment of patients with “resectable” pancreatic cancer, the commonly used classification of residual tumor cells in the resection margin may not be sufficient to predict their prognosis. One possible explanation for this phenomenon is that tumor cells that have found early access into retropancreatic nerves may have “escaped” from resection and may be blamed for the high frequency of therapy failure. Hence, Pn may account for the potentially low relevance of R status on OS. Moreover, a peculiar feature of Pn is that it can even be encountered in nerves within normal pancreatic tissue regions that are far from the tumor [70]. Takahashi et al. could show that this intra-pancreatic, extra-tumoral perineural invasion could be observed in nearly over 50% of investigated PDAC specimens [70]. In that study, tumor tissue with higher severity of Pn also exhibited a more frequent (85.7%) infiltration of tumor cells into the extrapancreatic, retroperitoneal nerve plexus (pl) [70]. Importantly, in the absence of intrapancreatic Pn, no invasion of the extrapancreatic nerve plexus was encountered [70]. Therefore,

regarding the observations of Ozaki et al. [7] and Takahashi et al. [60], it is presumable that Pn contributes to the occurrence of local recurrence, lymph node metastases and to tumor spread into extrapancreatic nerve plexus. Hence, our findings together with early extrapancreatic Pn and the presence of PN in healthyappearing pancreatic regions may explain why Pn decreases OS, DFS and PFS and why the resection margin (R0) on its own does not allow a strong enough conclusion on the prognosis of completely resected PDAC. In addition to a hidden role in “false R” status, it is possible that Pn is mechanistically linked to the generation of lymph node metastases. In their study, Kayahara et al. analysed consecutive sections of surgical resection PDAC specimen and firstly described a continuous tumor cell growth along nerves and the perineural sheath into adjacent lymph nodes [75]. These observations suggest that Pn may have a direct influence on lymph node involvement of PDAC and may be the underlying determinant of PDAC spread into lymph nodes. Correspondingly, Ozaki et al. [7] and Takahashi et al. [60] scanned PDAC cohorts and found that PDAC patients without any nodal involvement and without Pn showed a strong improvement in survival compared to PDAC patients with nodal or neural involvement.

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

111

Fig. 4. The unfavourable impact of Pn on DFS and PFS: A. After screening studies that included the HRs of DFS, 4 studies were included in the meta-analysis of DFS, which yielded a HR of 2.53 (95%-CI 1.67e3.83; p ¼ 0.0001), indicating Pn as an independent predictor of tumor recurrence in PDAC. B. Similar to DFS, the meta-analysis of PFS. Heterogeneity was absent in PFS meta-analysis (I2 ¼ 0%), and only low grade of heterogeneity (I2 ¼ 16%) was present in the meta-analysis of DFS. Here, the majority of studies had to be excluded due to lack of an exact definition of DFS or PFS by the authors.

Fig. 5. Adjusted factors of the meta-analysis: A. These factors represent the most common used histopathological factors of the different multivariate survival analysis which could be pooled in the multivariate meta-analysis of OS. Interestingly other prognostic histopathological factors like resection margin, tumor grading, vascular involvement and adjuvant therapy were included in these multivariate OS analyses. B. Beside Pn, also other well-known factors of early tumor recurrence like nodal status, grading, TMN-stage and adjuvant therapy were considered as adjusted factors in the meta-analysis of multivariate DFS. C. Well-known factors of early tumor recurrence like nodal status, grading, TMNstage and adjuvant therapy were considered as adjusted factors in the meta-analysis of multivariate PFS.

112

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

Fig. 6. Analysis of publication bias in the meta-analyses: To assess possible publication bias, Funnel plot of each meta-analysis containing more than 10 studies was performed. A. Here, the majority of HRs of the meta-analyses of univariate OS were within the pseudo-95%-CI-lines, which indicates absence of publication bias in the present study. B. In addition to the fixed effect model, the symmetrical distribution of the multivariate HRs also indicates the absence of publication bias in the present study.

Recently, Chatterjee et al. analysed a cohort of 212 PDAC patients who underwent resection after neoadjuvant radiochemotherapy and compared the histopathological findings of this cohort to 60 patients who received no preoperative treatment [69]. Interestingly, patients with neoadjuvant therapy showed a noticeable decrease in the frequency of Pn compared to patients who had no neoadjuvant treatment (58% in the treated group vs. 80% in the untreated group). Moreover, in this study, the presence of Pn

positively correlated with the presence of lymph node metastasis and unfavourably affected OS and time to tumor progression. According to the observations of Chatterjee et al., Ferrone and colleagues analysed the influence of neoadjuvant FOLFIRINOXtreatment of 40 patients with PDAC and compared the results with patients who had primary tumor resection. Importantly, patients with neoadjuvant chemotherapy showed a distinct decrease in perineural invasion (95.4% vs. 72.5%) and noticeably less number

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115

of tumor-positive lymph nodes [68]. Furthermore, a recent study of Roland et al. also investigated the influence of neoadjuvant treatment on histopathological features of PDAC but also on OS and DFS. Strikingly, patients after neoadjuvant treatment showed a less frequency of neural invasion (70% vs. 88%), a decreased tumorpositive lymph node status (50% vs. 57%), a decreased invasion of tumor cells into lymphovascular invasion (31% vs. 38%), an increase of their 2- (59.7% vs. 47.6%) and 5-year survival rate (26.1 vs. 18.9%) and a decrease of their 2-year-(23.3% vs. 42.6%) and 5-year-localrecurrence rates (29.7 vs. 45.9%) [76]. It is possible that the increase of OS and DFS may be the consequence of the effect of neoadjuvant therapy on unfavourable prognostic factors like tumor positive lymph nodes or the presence of Pn. This observation draws attention to the potential benefit from neoadjuvant treatment with regard to OS after resection for PDAC. Similarly, it is conceivable that adjustment of adjuvant treatment regimens according to the Pn severity may contribute to prolonged OS, DFS and PFS. 5. Conclusion In summary, this meta-analysis demonstrated a clearly unfavourable impact of Pn on OS, DFS and PFS among PDAC patients. Furthermore, we could establish Pn as a strong independent prognostic factor for diminished overall survival and early tumor recurrence in PDAC. This prominent influence of Pn suggests that PDAC patients may benefit from adaptation of their adjuvant therapy regimens according to their severity of Pn. Hence, randomized controlled trials including these stratifications may considerably modify the current clinical management of PDAC. Author contributions SS, IED and FS performed the database search and article extraction. SS, IED, CMR, FS, ET, MS and RG screened the abstracts for relevance. SS and BH performed the statistical analyses. HF and GOC supervised the study. All authors critically read the manuscript. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] R. Siegel, D. Naishadham, A. Jemal, Cancer statistics, CA a cancer J. Clin. 2012 (62) (2012) 10e29. [2] H. Oettle, S. Post, P. Neuhaus, K. Gellert, J. Langrehr, K. Ridwelski, et al., Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial, JAMA 297 (2007) 267e277. [3] C. Sperti, C. Pasquali, A. Piccoli, S. Pedrazzoli, Recurrence after resection for ductal adenocarcinoma of the pancreas, World J. Surg. 21 (1997) 195e200. [4] S. Hishinuma, Y. Ogata, M. Tomikawa, I. Ozawa, K. Hirabayashi, S. Igarashi, Patterns of recurrence after curative resection of pancreatic cancer, based on autopsy findings. Journal of gastrointestinal surgery, Off. J. Soc. Surg. Alimentary Tract 10 (2006) 511e518. [5] D.K. Chang, A.L. Johns, N.D. Merrett, A.J. Gill, E.K. Colvin, C.J. Scarlett, et al., Margin clearance and outcome in resected pancreatic cancer. Journal of clinical oncology, Off. J. Am. Soc. Clin. Oncol. 27 (2009) 2855e2862. [6] J. Luttges, S. Schemm, I. Vogel, J. Hedderich, B. Kremer, G. Kloppel, The grade of pancreatic ductal carcinoma is an independent prognostic factor and is superior to the immunohistochemical assessment of proliferation, J. Pathol. 191 (2000) 154e161. [7] H. Ozaki, T. Hiraoka, R. Mizumoto, S. Matsuno, Y. Matsumoto, T. Nakayama, et al., The prognostic significance of lymph node metastasis and intrapancreatic perineural invasion in pancreatic cancer after curative resection, Surg. today 29 (1999) 16e22. [8] A. Hamidian Jahromi, G.B. Zibari, E. Jafarimehr, Q. Chu, G.P. Wellman, R. Shi, et al., Peripancreatic soft tissue involvement: independent outcome predictor in patients with resected pancreatic adenocarcinoma, Int. Surg. 99 (2014)

113

62e70. [9] D.E. Bockman, M. Buchler, H.G. Beger, Interaction of pancreatic ductal carcinoma with nerves leads to nerve damage, Gastroenterology 107 (1994) 219e230. [10] F. Liebl, I.E. Demir, K. Mayer, T. Schuster, J.G. D'Haese, K. Becker, et al., The impact of neural invasion severity in gastrointestinal malignancies: a clinicopathological study, Ann. Surg. 260 (2014) 900e907 discussion 7e8. [11] G.O. Ceyhan, F. Bergmann, M. Kadihasanoglu, B. Altintas, I.E. Demir, U. Hinz, et al., Pancreatic neuropathy and neuropathic painea comprehensive pathomorphological study of 546 cases, Gastroenterology 136 (2009), 177e86 e1. [12] I.E. Demir, H. Friess, G.O. Ceyhan, Nerve-cancer interactions in the stromal biology of pancreatic cancer, Front. Physiol. 3 (2012) 97. [13] S.G. Farid, G.A. Falk, D. Joyce, S. Chalikonda, R.M. Walsh, A.M. Smith, G. Morris-Stiff, Prognostic value of the lymph node ratio after resection of periampullary carcinomas, HPB Off. J. Int. Hepato Pancreato Biliary Assoc. 16 (6) (2014 Jun) 582e591. [14] Q. Ben, M. Xu, Y. Jiang, Y. Yuan, K. Wang, J. Fang, et al., Clinical profiles and long-term outcomes of patients with pancreatic ductal adenocarcinoma and diabetes mellitus, Diabetes/Metabolism Res. Rev. 28 (2012) 169e176. [15] T.M. Pawlik, A.L. Gleisner, J.L. Cameron, J.M. Winter, L. Assumpcao, K.D. Lillemoe, et al., Prognostic relevance of lymph node ratio following pancreaticoduodenectomy for pancreatic cancer, Surgery 141 (2007) 610e618. [16] G. Sergeant, N. Ectors, S. Fieuws, R. Aerts, B. Topal, Prognostic relevance of extracapsular lymph node involvement in pancreatic ductal adenocarcinoma, Ann. Surg. Oncol. 16 (2009) 3070e3079. [17] J. Helm, B.A. Centeno, D. Coppola, M. Melis, M. Lloyd, J.Y. Park, et al., Histologic characteristics enhance predictive value of American Joint Committee on Cancer staging in resectable pancreas cancer, Cancer 115 (2009) 4080e4089. [18] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, P. Group, Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Int. J. Surg. 8 (2010) 336e341. [19] J. Higgins, S. Green, Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011], The Cochrane Collaboration, 2011. [20] R. DerSimonian, N. Laird, Meta-analysis in clinical trials, Control. Clin. trials 7 (1986) 177e188. [21] C.K. Chu, A.E. Mazo, M. Goodman, V. Egnatashvili, J.M. Sarmiento, C.A. Staley, et al., Preoperative diabetes mellitus and long-term survival after resection of pancreatic adenocarcinoma, Ann. Surg. Oncol. 17 (2010) 502e513. [22] N.B. Jamieson, M. Mohamed, K.A. Oien, A.K. Foulis, E.J. Dickson, C.W. Imrie, et al., The relationship between tumor inflammatory cell infiltrate and outcome in patients with pancreatic ductal adenocarcinoma, Ann. Surg. Oncol. 19 (2012) 3581e3590. [23] H. Jiang, C. He, S. Geng, H. Sheng, X. Shen, X. Zhang, et al., RhoT1 and Smad4 are correlated with lymph node metastasis and overall survival in pancreatic cancer, PloS One 7 (2012) e42234. [24] K.K. Kazanjian, O.J. Hines, J.P. Duffy, D.Y. Yoon, G. Cortina, H.A. Reber, Improved survival following pancreaticoduodenectomy to treat adenocarcinoma of the pancreas: the influence of operative blood loss, Archiv. Surg. 143 (2008) 1166e1171. [25] J. Kim, H.A. Reber, O.J. Hines, K.K. Kazanjian, A. Tran, X. Ye, et al., The clinical significance of MAGEA3 expression in pancreatic cancer, Int. J. cancer J. Int. du cancer 118 (2006) 2269e2275. [26] R. Kim, A. Tan, K.K. Lai, J. Jiang, Y. Wang, L.A. Rybicki, et al., Prognostic roles of human equilibrative transporter 1 (hENT-1) and ribonucleoside reductase subunit M1 (RRM1) in resected pancreatic cancer, Cancer 117 (2011) 3126e3134. [27] H. Xie, W. Jiang, S.Y. Xiao, X. Liu, High expression of survivin is prognostic of shorter survival but not predictive of adjuvant gemcitabine benefit in patients with resected pancreatic adenocarcinoma, J. Histochem. Cytochem.: Off. J. Histochem. Soc. 61 (2013) 148e155. [28] E. Buc, A. Couvelard, F. Kwiatkowski, S. Dokmak, P. Ruszniewski, P. Hammel, et al., Adenocarcinoma of the pancreas: does prognosis depend on mode of lymph node invasion? Eur. J. Surg. Oncol. 40 (2014) 1578e1585. [29] L. Duan, X.Q. Hu, D.Y. Feng, S.Y. Lei, G.H. Hu, GPC-1 may serve as a predictor of perineural invasion and a prognosticator of survival in pancreatic cancer, Asian J. Surg. 36 (2013) 7e12. [30] B.E. Kadera, D.B. Sunjaya, W.H. Isacoff, L. Li, O.J. Hines, J.S. Tomlinson, et al., Locally advanced pancreatic cancer: association between prolonged preoperative treatment and lymph-node negativity and overall survival, JAMA Surg. 149 (2014) 145e153. [31] D.D. Li, Z.Q. Fu, Q. Lin, Y. Zhou, Q.B. Zhou, Z.H. Li, et al., Linc00675 is a novel marker of short survival and recurrence in patients with pancreatic ductal adenocarcinoma, World J. Gastroenterol. 21 (2015) 9348e9357. [32] S. Qu, X. Yang, W. Song, W. Sun, X. Li, J. Wang, et al., Downregulation of lncRNA-ATB correlates with clinical progression and unfavorable prognosis in pancreatic cancer, Tumour Biol. 37 (3) (2016 Mar) 3933e3938. [33] P.R. De Reuver, A. Mittal, M. Neale, A.J. Gill, J.S. Samra, Extended pancreatoduodenectomy as defined by the International Study Group for Pancreatic Surgery is associated with worse survival but not with increased morbidity, Surgery 158 (2015) 183e190. [34] M. Sugimoto, S. Mitsunaga, K. Yoshikawa, Y. Kato, N. Gotohda, S. Takahashi, et al., Prognostic impact of M2 macrophages at neural invasion in patients

114

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115 with invasive ductal carcinoma of the pancreas, Eur. J. Cancer 50 (2014) 1900e1908. D. Tang, J. Zhang, Z. Yuan, J. Gao, S. Wang, N. Ye, et al., Pancreatic satellite cells derived galectin-1 increase the progression and less survival of pancreatic ductal adenocarcinoma, PloS One 9 (2014) e90476. F. Wang, A.J. Gill, M. Neale, V. Puttaswamy, S. Gananadha, N. Pavlakis, et al., Adverse tumor biology associated with mesenterico-portal vein resection influences survival in patients with pancreatic ductal adenocarcinoma, Ann. Surg. Oncol. 21 (2014) 1937e1947. L.M. Wang, M.A. Silva, Z. D'Costa, R. Bockelmann, Z. Soonawalla, S. Liu, et al., The prognostic role of desmoplastic stroma in pancreatic ductal adenocarcinoma, Oncotarget 7 (4) (2016 Jan 26) 4183e4194. J. Wong, W.J. Fulp, J.R. Strosberg, L.K. Kvols, B.A. Centeno, P.J. Hodul, Predictors of lymph node metastases and impact on survival in resected pancreatic neuroendocrine tumors: a single-center experience, Am. J. Surg. 208 (2014) 775e780. H. Zhao, Q. Wang, X. Wang, H. Zhu, S. Zhang, W. Wang, et al., Correlation between RAB27B and p53 expression and overall survival in pancreatic cancer, Pancreas 45 (2016) 204e210. H.L. Ye, D.D. Li, Q. Lin, Y. Zhou, Q.B. Zhou, B. Zeng, et al., Low RASSF6 expression in pancreatic ductal adenocarcinoma is associated with poor survival, World J. Gastroenterol. 21 (2015) 6621e6630. S. Crippa, S. Partelli, G. Zamboni, G. Barugola, P. Capelli, M. Inama, et al., Poorly differentiated resectable pancreatic cancer: is upfront resection worthwhile? Surgery 152 (2012) S112eS119. I. Hatzaras, N. George, P. Muscarella, W.S. Melvin, E.C. Ellison, M. Bloomston, Predictors of survival in periampullary cancers following pancreaticoduodenectomy, Ann. Surg. Oncol. 17 (2010) 991e997. K. Kato, S. Yamada, H. Sugimoto, N. Kanazumi, S. Nomoto, S. Takeda, et al., Prognostic factors for survival after extended pancreatectomy for pancreatic head cancer: influence of resection margin status on survival, Pancreas 38 (2009) 605e612. J. Kim, H.A. Reber, S.M. Dry, D. Elashoff, S.L. Chen, N. Umetani, et al., Unfavourable prognosis associated with K-ras gene mutation in pancreatic cancer surgical margins, Gut 55 (2006) 1598e1605. S.K. Maithel, I. Coban, P.J. Kneuertz, D.A. Kooby, B.F. El-Rayes, J.S. Kauh, et al., Differential expression of ERCC1 in pancreas adenocarcinoma: high tumor expression is associated with earlier recurrence and shortened survival after resection, Ann. Surg. Oncol. 18 (2011) 2699e2705. Y. Murata, T. Hamada, M. Kishiwada, I. Ohsawa, S. Mizuno, M. Usui, et al., Human equilibrative nucleoside transporter 1 expression is a strong independent prognostic factor in UICC T3-T4 pancreatic cancer patients treated with preoperative gemcitabine-based chemoradiotherapy, J. hepato-biliarypancreatic Sci. 19 (2012) 413e425. K. Nagai, R. Doi, F. Katagiri, T. Ito, A. Kida, M. Koizumi, et al., Prognostic value of metastin expression in human pancreatic cancer. Journal of experimental & clinical cancer research, CR 28 (2009) 9. I.H. Sahin, M.A. Shama, M. Tanaka, J.L. Abbruzzese, S.A. Curley, M. Hassan, et al., Association of diabetes and perineural invasion in pancreatic cancer, Cancer Med. 1 (2012) 357e362. D.F. Schaeffer, D.R. Owen, H.J. Lim, A.K. Buczkowski, S.W. Chung, C.H. Scudamore, et al., Insulin-like growth factor 2 mRNA binding protein 3 (IGF2BP3) overexpression in pancreatic ductal adenocarcinoma correlates with poor survival, BMC cancer 10 (2010) 59. S.C. Schiffman, C.K. Chu, J. Park, M. Russell, S. Keilin, D.A. Kooby, et al., Is prior cholecystectomy associated with decreased survival in patients with resectable pancreatic adenocarcinoma following pancreaticoduodenectomy? Am. J. Surg. 201 (2011) 519e524. O. Turrini, F. Paye, P. Bachellier, A. Sauvanet, A. Sa Cunha, Y.P. Le Treut, et al., Pancreatectomy for adenocarcinoma in elderly patients: postoperative outcomes and long term results: a study of the French surgical association, Eur. J. Surg. Oncol. 39 (2013) 171e178. M.H. van Roest, A.S. Gouw, P.M. Peeters, R.J. Porte, M.J. Slooff, V. Fidler, et al., Results of pancreaticoduodenectomy in patients with periampullary adenocarcinoma: perineural growth more important prognostic factor than tumor localization, Ann. Surg. 248 (2008) 97e103. A.M. Nagrial, D.K. Chang, N.Q. Nguyen, A.L. Johns, L.A. Chantrill, J.L. Humphris, et al., Adjuvant chemotherapy in elderly patients with pancreatic cancer, Br. J. Cancer 110 (2014) 313e319. S.H. Lee, H. Kim, J.H. Hwang, E. Shin, H.S. Lee, D.W. Hwang, et al., CD24 and S100A4 expression in resectable pancreatic cancers with earlier disease recurrence and poor survival, Pancreas 43 (2014) 380e388. N.A. Lockney, M. Zhang, Y. Lu, S.C. Sopha, M.K. Washington, N. Merchant, et al., Pyruvate Kinase Muscle Isoenzyme 2 (PKM2) expression is associated with overall survival in pancreatic ductal adenocarcinoma, J. Gastrointest. Cancer 46 (2015) 390e398. K. Urbonas, A. Gulbinas, G. Smailyte, D. Pranys, A. Jakstaite, J. Pundzius, et al., Factors influencing survival after pancreatoduodenectomy for ductal adenocarcinoma depend on patients' age, Dig. Surg. 32 (2015) 60e67. S. Yamada, T. Fujii, Y. Shimoyama, M. Kanda, G. Nakayama, H. Sugimoto, et al., SMAD4 expression predicts local spread and treatment failure in resected pancreatic cancer, Pancreas 44 (2015) 660e664. M. Zhang, R. Fang, Y. Mou, R. Chen, X. Xu, R. Zhang, et al., LDP vs ODP for pancreatic adenocarcinoma: a case matched study from a single-institution, BMC Gastroenterol. 15 (2015) 182.

[59] S.H. Lee, H. Kim, J.H. Hwang, H.S. Lee, J.Y. Cho, Y.S. Yoon, et al., Breast cancer resistance protein expression is associated with early recurrence and decreased survival in resectable pancreatic cancer patients, Pathol. Int. 62 (2012) 167e175. [60] H. Takahashi, H. Ohigashi, O. Ishikawa, K. Gotoh, T. Yamada, S. Nagata, et al., Perineural invasion and lymph node involvement as indicators of surgical outcome and pattern of recurrence in the setting of preoperative gemcitabine-based chemoradiation therapy for resectable pancreatic cancer, Ann. Surg. 255 (2012) 95e102. [61] T. Fouquet, A. Germain, L. Brunaud, L. Bresler, A. Ayav, Is perineural invasion more accurate than other factors to predict early recurrence after pancreatoduodenectomy for pancreatic head adenocarcinoma? World J. Surg. 38 (2014) 2132e2137. [62] S. Michiels, P. Piedbois, S. Burdett, N. Syz, L. Stewart, J.P. Pignon, Metaanalysis when only the median survival times are known: a comparison with individual patient data results, Int. J. Technol. Assess. health care 21 (2005) 119e125. [63] Q.W. Ben, J.C. Wang, J. Liu, Y. Zhu, F. Yuan, W.Y. Yao, et al., Positive expression of L1-CAM is associated with perineural invasion and poor outcome in pancreatic ductal adenocarcinoma, Ann. Surg. Oncol. 17 (2010) 2213e2221. [64] P. Bachellier, E. Rosso, I. Lucescu, E. Oussoultzoglou, J. Tracey, P. Pessaux, et al., Is the need for an arterial resection a contraindication to pancreatic resection for locally advanced pancreatic adenocarcinoma? A case-matched controlled study, J. Surg. Oncol. 103 (2011) 75e84. [65] N.B. Jamieson, P. Glen, D.C. McMillan, C.J. McKay, A.K. Foulis, R. Carter, et al., Systemic inflammatory response predicts outcome in patients undergoing resection for ductal adenocarcinoma head of pancreas, Br. J. Cancer 92 (2005) 21e23. [66] H. Xie, J. Lin, D.G. Thomas, W. Jiang, X. Liu, Ribonucleotide reductase M2 does not predict survival in patients with resectable pancreatic adenocarcinoma, Int. J. Clin. Exp. Pathol. 5 (2012) 347e355. [67] H. JPT, S G, Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011], The Cochrane Collaboration, 2011. [68] C.R. Ferrone, G. Marchegiani, T.S. Hong, D.P. Ryan, V. Deshpande, E.I. McDonnell, et al., Radiological and surgical implications of neoadjuvant treatment with FOLFIRINOX for locally advanced and borderline resectable pancreatic cancer, Ann. Surg. 261 (2015) 12e17. [69] D. Chatterjee, M.H. Katz, A. Rashid, H. Wang, A.C. Iuga, G.R. Varadhachary, et al., Perineural and intraneural invasion in posttherapy pancreaticoduodenectomy specimens predicts poor prognosis in patients with pancreatic ductal adenocarcinoma, Am. J. Surg. Pathol. 36 (2012) 409e417. [70] T. Takahashi, H. Ishikura, T. Motohara, S. Okushiba, M. Dohke, H. Katoh, Perineural invasion by ductal adenocarcinoma of the pancreas, J. Surg. Oncol. 65 (1997) 164e170. [71] S. Takai, S. Satoi, H. Toyokawa, H. Yanagimoto, N. Sugimoto, K. Tsuji, et al., Clinicopathologic evaluation after resection for ductal adenocarcinoma of the pancreas: a retrospective, single-institution experience, Pancreas 26 (2003) 243e249. [72] K. Shimada, S. Nara, M. Esaki, Y. Sakamoto, T. Kosuge, N. Hiraoka, Intrapancreatic nerve invasion as a predictor for recurrence after pancreaticoduodenectomy in patients with invasive ductal carcinoma of the pancreas, Pancreas 40 (2011) 464e468. [73] L.T.N.M. Sobin, Classification of Malignant Tumours, seventh ed., WileyBlackwell, Oxford, 2009. [74] G. Butturini, D.D. Stocken, M.N. Wente, H. Jeekel, J.H. Klinkenbijl, K.E. Bakkevold, et al., Influence of resection margins and treatment on survival in patients with pancreatic cancer: meta-analysis of randomized controlled trials, Archiv. Surg. 143 (2008) 75e83 discussion. [75] M. Kayahara, H. Nakagawara, H. Kitagawa, T. Ohta, The nature of neural invasion by pancreatic cancer, Pancreas 35 (2007) 218e223. [76] C.L. Roland, A.D. Yang, M.H. Katz, D. Chatterjee, H. Wang, H. Lin, et al., Neoadjuvant therapy is associated with a reduced lymph node ratio in patients with potentially resectable pancreatic cancer, Ann. Surg. Oncol. 22 (2015) 1168e1175. [77] A. Andren-Sandberg, Prognostic factors in pancreatic cancer, North Am. J. Med. Sci. 4 (2012) 9e12. [78] J. Lenz, P. Karasek, J. Jarkovsky, K. Muckova, P. Dite, Z. Kala, et al., Clinicopathological correlations of nestin expression in surgically resectable pancreatic cancer including an analysis of perineural invasion. Journal of gastrointestinal and liver diseases, JGLD 20 (2011) 389e396. [79] G. Aprile, C. Avellini, M. Reni, M. Mazzer, L. Foltran, D. Rossi, et al., Biglycan expression and clinical outcome in patients with pancreatic adenocarcinoma, Tumour Biol. 34 (2013) 131e137. [80] W. Meyer, C. Jurowich, M. Reichel, B. Steinhauser, P.H. Wunsch, C. Gebhardt, Pathomorphological and histological prognostic factors in curatively resected ductal adenocarcinoma of the pancreas, Surg. today 30 (2000) 582e587. [81] A.V. Biankin, DPC4/Smad4 expression and outcome in pancreatic ductal adenocarcinoma, J. Clin. Oncol. 20 (2002) 4531e4542. [82] T. Bottger, J. Zech, W. Weber, K. Sorger, T. Junginger, Relevant factors in the prognosis of ductal pancreatic carcinoma, Acta Chir. Scand. 156 (1990) 781e788. [83] M.V. Perini, A.L. Montagnini, J. Jukemura, S. Penteado, E.E. Abdo, R. Patzina, et al., Clinical and pathologic prognostic factors for curative resection for pancreatic cancer. HPB, Off. J. Int. Hepato Pancreato Biliary Assoc. 10 (2008) 356e362.

S. Schorn et al. / Surgical Oncology 26 (2017) 105e115 [84] M. Bouvet, R.A. Gamagami, E.A. Gilpin, O. Romeo, A. Sasson, D.W. Easter, et al., Factors influencing survival after resection for periampullary neoplasms, Am. J. Surg. 180 (2000) 13e17. [85] M.V. Perini, A.L. Montagnini, R. Coudry, R. Patzina, S. Penteado, E.E. Abdo, et al., Prognostic significance of epidermal growth factor receptor overexpression in pancreas cancer and nodal metastasis, ANZ J. Surg. 85 (2015) 174e178. [86] J.L. Cameron, D.W. Crist, J.V. Sitzmann, R.H. Hruban, J.K. Boitnott, A.J. Seidler, et al., Factors influencing survival after pancreaticoduodenectomy for pancreatic cancer, Am. J. Surg. 161 (1991) 120e124 discussion 4e5. [87] G.A. Poultsides, S. Reddy, J.L. Cameron, R.H. Hruban, T.M. Pawlik, N. Ahuja, et al., Histopathologic basis for the favorable survival after resection of intraductal papillary mucinous neoplasm-associated invasive adenocarcinoma of the pancreas, Ann. Surg. 251 (2010) 470e476. [88] A.I. Salem, M. Alfi, E. Winslow, C.S. Cho, S.M. Weber, Has survival following pancreaticoduodenectomy for pancreas adenocarcinoma improved over time? J. Surg. Oncol. 112 (2015) 643e649. [89] T. Schnelldorfer, M.G. Sarr, D.M. Nagorney, L. Zhang, T.C. Smyrk, R. Qin, et al., Experience with 208 resections for intraductal papillary mucinous neoplasm of the pancreas, Archiv. Surg. 143 (2008) 639e646 discussion 46. [90] C. Gebhardt, W. Meyer, M. Reichel, P.H. Wunsch, Prognostic factors in the operative treatment of ductal pancreatic carcinoma, Langenbeck's Archiv. Surg./Deutsche Gesellschaft fur Chir. 385 (2000) 14e20. [91] K. Shimada, Y. Sakamoto, T. Sano, T. Kosuge, Prognostic factors after distal pancreatectomy with extended lymphadenectomy for invasive pancreatic adenocarcinoma of the body and tail, Surgery 139 (2006) 288e295. [92] R.J. Geer, M.F. Brennan, Prognostic indicators for survival after resection of pancreatic adenocarcinoma, Am. J. Surg. 165 (1993) 68e72 discussion-3. [93] B.K. Goh, Y.M. Tan, P.C. Cheow, Y.F. Chung, P.K. Chow, W.K. Wong, et al., Outcome of distal pancreatectomy for pancreatic adenocarcinoma, Dig. Surg. 25 (2008) 32e38. [94] S.H. Shin, S.C. Kim, S.M. Hong, K.B. Song, J.H. Lee, K.M. Park, et al., Can

[95]

[96]

[97] [98]

[99]

[100]

[101]

115

statistically determined prognostic factors predict the long-term survival of patients with pancreatic ductal adenocarcinoma following surgical resection?: Clinicopathological analysis of 82 long-term survivors, Pancreas 43 (2014) 571e577. F. Griffanti-Bartoli, G.B. Arnone, P. Ceppa, G. Ravera, S. Carrabetta, D. Civalleri, Malignant tumors in the head of the pancreas and the periampullary region. Diagnostic and prognostic aspects, Anticancer Res. 14 (1994) 657e666. D.A. Skalicky, J.G. Kench, D. Segara, M.J. Coleman, R.L. Sutherland, S.M. Henshall, et al., Cyclin E expression and outcome in pancreatic ductal adenocarcinoma. Cancer epidemiology, biomarkers & prevention, a Publ. Am. Assoc. Cancer Res. cosponsored by Am. Soc. Prev. Oncol. 15 (2006) 1941e1947. C. Sperti, C. Pasquali, A. Piccoli, S. Pedrazzoli, Survival after resection for ductal adenocarcinoma of the pancreas, Br. J. Surg. 83 (1996) 625e631. S.M. Hong, M. Goggins, C.L. Wolfgang, R.D. Schulick, B.H. Edil, J.L. Cameron, et al., Vascular invasion in infiltrating ductal adenocarcinoma of the pancreas can mimic pancreatic intraepithelial neoplasia: a histopathologic study of 209 cases, Am. J. Surg. Pathol. 36 (2012) 235e241. N.B. Jamieson, A.K. Foulis, K.A. Oien, J.J. Going, P. Glen, E.J. Dickson, et al., Positive mobilization margins alone do not influence survival following pancreatico-duodenectomy for pancreatic ductal adenocarcinoma, Ann. Surg. 251 (2010) 1003e1010. N.B. Jamieson, D.C. Morran, J.P. Morton, A. Ali, E.J. Dickson, C.R. Carter, et al., MicroRNA molecular profiles associated with diagnosis, clinicopathologic criteria, and overall survival in patients with resectable pancreatic ductal adenocarcinoma. Clinical cancer research, Off. J. Am. Assoc. Cancer Res. 18 (2012) 534e545. L. Zeng, Y. Guo, J. Liang, S. Chen, P. Peng, Q. Zhang, et al., Perineural invasion and TAMs in pancreatic ductal adenocarcinomas: review of the original pathology reports using immunohistochemical enhancement and relationships with clinicopathological features, J. Cancer 5 (2014) 754e760.