Background

During the last 40 years, pancreatic surgery has evolved as an important topic in visceral surgery and has developed from a dangerous and rarely performed niche discipline to a highly specialized field in centers worldwide [1]. From historic surgery-related mortality rates of up to 80% [2], improved surgical techniques as well as progress in perioperative care and especially complication management have led to dramatic improvements in both—postoperative as well as long-term—outcomes [3]. Today, despite morbidity rates of 30–50%, postoperative mortality has decreased to well below 5% in high-volume centers specialized in pancreatic surgery [1]. It is therefore generally accepted that centralization is an essential cornerstone to achieve best results in this complex field of surgery and—with regard to pancreatic cancer surgery—that multimodal approaches, especially adjuvant multiagent chemotherapy, must be additionally included to add oncological benefit to good surgical results [4]. This clinical development has been accompanied by an increasing body of scientific reporting on all aspects, including tumor stage and resection margin status definitions, preoperative evaluation of resectability as well as reporting of postoperative morbidity. In this context, it has become obvious during the last 15 years that standardized definitions are mandatory to allow any comparability of indications, procedures, and outcomes and therefore create more evidence. Various national and international groups and committees have consequently started to work on this topic and numerous definitions have been proposed, although not all of them have eventually gained acceptance in every day’s practice or the literature. However, a large progress has been made and has significantly improved standardization and scientific reporting and continuous efforts are ongoing for topics that are not clearly defined today. This review summarizes the development, current status and future perspective of standardization, and definitions in pancreatic surgery.

Definitions

Tumor staging

The definition of cancer tumor stages (TNM) is a long-standing and well-established procedure initiated and continuously maintained by the Union International Contre le Cancer (UICC) since 1943 which was adopted by the American Joint Committee on Cancer (AJCC) [5] and is presently used in its 8th edition [6]. Especially, the eighth modification of this system underlines the importance of both—clear and specific descriptions as well as the continuous update of such definitions under consideration of their clinical usefulness and prognostic validity. During the recent year, pancreatic ductal adenocarcinoma (PDAC) has been classified with two important differences compared to other solid tumors. On the one hand, the T4 stage was limited to an arterial tumor invasion alone (vs. invasion of adjacent organs for most other tumors) due to the general exclusion of resectability in this situation; on the other hand, the N stage has been classified in a two-stage system (N0 vs. N1) due to the consideration that the presence of lymph node metastases on any extent lead to a general poor prognosis and a further subclassification would not be reasonable [7]. In the new 8th edition, the T4 stage has been unchanged, while a new N classification has been introduced including N0 (no positive lymph nodes), N1 (1–3 positive lymph nodes), and N2 (4 or more positive lymph nodes) [6]. The latter change has been made due to numerous publications during the recent decade showing a distinct prognostic relevance of the number of positive lymph nodes as well as the lymph node ratio, which is calculated as the number of positive divided by the total number of examined lymph nodes [8,9,10]. The new classification has been validated in several institutions confirming its better prognostic discriminative power but also new modifications have been already proposed regarding TNM as well as clinical staging [11,12,13,14]. This ongoing process may be associated with the need for re-evaluation of current and past study results and may bear a potential for confusion [15] but above all reflects the need for clear and validated definitions that help in stratifying patients’ prognosis and therefore certainly have an important impact on scientific reporting.

IPMN definitions

A field of intense discussion is the definition and management of cystic pancreatic lesions, especially regarding intraductal papillary mucinous neoplasms (IPMN) of the pancreas. Although the entity of IPMN itself is clearly defined in terms of pathological characteristics and the extension to the branch ducts (BD), the main duct (MD), or the combination of both (mixed type), the evaluation of their malignant potential and consequently the recommendations for management with observation or surgical intervention remains highly controversial. Three major international guidelines are commonly used; the consensus of the International Association of Pancreatology (IAP, “Fukuoka guidelines”), the guidelines of the European Study Group of Cystic Tumors, and those of the American Gastroenterological Association (AGA) [16,17,18]. These three guidelines clearly show discrepancies not only in the terminology but also in the treatment recommendations and underline the dilemma of non-unified definitions that has considerable impact on reporting as well as on clinical decision-making. The major controversy arises in the description and evaluation of BD IPMN and their malignant potential. While the IAP guidelines describe “worrisome features” and “high-risk stigmata” in radiological morphology of these lesions, the European guidelines define “relative” and “absolute” indications for surgery; whereas, the AGA adopts “high-risk features” but also describes “concerning features” and “alarm symptoms” without clearly defining these in the guideline. Consequently, the terminologies are not comparable at all and studies switch between either guideline terminology which leads to a confusing variety of described management algorithms and finally allows to draw conclusion only with regard to the specific guideline that was applied in the respective study protocol. Even if the two classification in each—the IAP and the European—guidelines may be comparable to some extent a clear scientific discrimination is difficult in many publications which makes comparison of series on observation as well as resection of IPMN difficult. A special feature among others is the size of BD IPMN, which is an important topic in the IAP guidelines with cutoff margins of 1, 2, 3, and > 3 cm. These different sizes should be combined with additional worrisome features and high-risk stigmata to define further diagnostics and to decide on management [16]. As many studies have shown, that size itself is not a useful tool to define malignant potential, the European guidelines have abandoned size below 4 cm and have only defined BD IPMN > 4 cm to be an indication for surgery. All smaller IPMN should not be evaluated according to size but to other features indicating potential malignancy (solid components, tumor markers, symptoms) [17]. This reflects the ongoing uncertainty and underlines the different approaches towards this difficult entity of cystic lesions. Furthermore, the recommendation of resection for MD and mixed-type IPMN remain controversial between both guidelines. While the IAP remains to recommend resection from a MD diameter of > 10 mm, the European consensus has lowered the threshold to > 5 mm based on a number of recent studies showing a relevant risk of malignancy between 30 and 90% within the diameter range of 5–9.9 mm [19,19,21]. A harmonization of these guideline recommendations would therefore be desirable to overcome this gray zone in the management of IPMN.

Definition of resection status (R0/R1)

The pathological evaluation of local radicality after resection for PDAC is under intense debate since the historic definition of R0 (no tumor cells within the resection margin) and R1 (tumor cells found in the resection margin without macroscopic residual tumor in situ) has been challenged by Verbeke et al. in 2006 [22]. The concept of a different definition was based on a standardized pathological work-up due to the observation that in a non-standardized work-up, a large proportion of false-negative tumor margins were described and that with a defined axial slicing technique, many of the initial R0 findings became positive with a proximity of tumors cells of < 1 mm to the resection margin, besides additional direct invasions by single cells that were also missed in non-standardized specimen work-up [22]. The revised R-definition was adopted in Europe quite quickly and large patient cohort analyses from both—retrospective as well as prospective and randomized—studies, confirmed the prognostic significance of the classification [23,24,25]. As in other countries, especially the USA, the revised classification was not generally adopted; reports using the historic R1 definition remain difficult to compare with those using the revised definition.

As the revised definition led to a large proportion of R1 patient (including both R1 direct invasion and R1 < 1 mm), currently the terminology has been changed in a certain analogy to rectal cancer specimen description and is now defined including circumferential resection margin (CRM) as R0 (CRM−), R0 (CRM+), and R1, which equal to R0, R1 (< 1 mm), and R1 (direct invasion). This terminology has been adopted by the AJCC in the latest staging edition [6]. Consequently, a more unified reporting worldwide can be expected in the future although this item is still an object of intense debate.

ISGPF/ISGPS definitions

One of the earliest approaches towards a standardization was initiated by the International Study Group on Pancreatic Fistula (ISGPF) in 2005. The background of this effort was the observation that by 2004, authors used a large number of different definitions when reporting on postoperative pancreatic fistula as one of the main factors for morbidity after pancreatic surgery [26]. In total, there were 26 different definitions mentioned in the international literature which led to a relevant reporting bias, namely POPF incidences ranging between 1.5 and 30% for comparable surgical procedure in different institutions, which was not necessarily attributable to better surgical performance but to the specific POPF definition that different authors used [26]. As a result of this observation, an international panel of 37 pancreatic surgeons formed the ISGPF and introduced a standardized POPF definition with a grading system to reflect POPF severity [27]. This initial definition fulfilled the criteria of an easy clinical application, was therefore quickly accepted by most centers, and became the standard reporting tool for POPF in retrospective as well as prospective studies. With an average number of 187 citations per year, this publication has been cited more than 2.600 times (September 2018) which gives an impression of its relevance in the field of pancreatic surgery and the need for standardization [28]. The group then—renamed as International Study Group for Pancreatic Surgery (ISGPS)—extended its activity towards two other main topics of morbidity, delayed gastric emptying (DGE [29]; 902 citations, average 75/year) and postpancreatectomy hemorrhage (PPH [30]; 631 citations, average 53/year), both of which were successfully adopted and had also been reported with varying definitions before [28].

With specific regard to pancreatic cancer (PDAC) surgery, more standardized definitions and consensus publication by the ISGPS were published in 2014 which aim at standardized procedures and reporting. These cover the topics of lymphadenectomy [31], extended procedures [32], and borderline resectability [33] including the controversial field of neoadjuvant treatment vs. upfront resection in the respective patients. During preparation of these three recommendations, the current evidence was systematically collected, reviewed, and condensed under consideration of other already existing guidelines (i.e., the National Comprehensive Cancer Network (NCCN) guidelines). Finally, the expert panel of pancreatic surgeons voted on the statements of every topic and has approved the recommendations which are consequently given with a comment on the strength of evidence and agreement.

More recently, the ISGPS has approached a standardized definition and reporting of pancreatic anastomosis technique [34], postoperative chyle leakage [35], and nutritional status and its management in pancreatic surgery [36] (Table 1). The success of the initial definitions shows the strong need for standardization in the scientific literature and their implementation has led and will lead to a significant progress in more concise reporting of procedures and outcomes. An important aspect of the respective definitions and consensus statements is their plasticity. This is reflected in the updated POPF definition that was published by the ISGPS in 2016 [38]. This update was initiated as clinical reality between 2005 and 2016 showed that POPF management had substantially changed, especially with regard to percutaneous drainage in the management of these patients [39]. As percutaneous drainage was not explicitly addressed in the original paper, but became an important tool to prevent POPF-associated complications and avoid reoperation and the related mortality, this has been reflected in the updated POPF classification; furthermore, as the original POPF A was shown to be more or less clinically irrelevant, this grade of POPF has been abandoned and the respective situation has been “downgraded” to a biochemical leak without any clinical relevance [38]. Comparable to the original definition, the updated consensus has gained acceptance and is currently used as the standard in POPF reporting.

Table 1 Overview of the currently published ISGPS definitions and their scientific impact (web of science citation index [28])
Full size table

Definition of resectability

Definition of resectability of PDAC is one of the fields where numerous definitions by different groups or institutions are used worldwide and the requirement of a standardized use and reporting is still not fulfilled. Local resectability is defined as primary resectable PDAC, borderline resectable (BR-PDAC), or irresectable (locally advanced) PDAC. Resectable PDAC is—more or less uniformly—characterized by the absence of distant metastases and any vascular attachment (no distortion of SMV or PV and clearly preserved fat planes towards CA and AMS and therefore not an issue of debate [40]. Generally, a contrast-enhanced computed tomography (CE-CT) using a pancreas-specific protocol which should be the gold standard to determine local tumor extension, exclude liver metastases, and evaluate a possible vascular infiltration. The CE-CT should offer a visualization and differentiation of normal and tumorous pancreatic tissue in an arterial and venous phase including an optimal contrast imaging of the vascular structure in both phases as well as the liver parenchyma to exclude hepatic metastases.

Regarding BR PDAC, most definitions focus on anatomical issues defined in preoperative cross-sectional imaging. According to the criteria published by the ISGPS in 2014 [33], which are mainly based on the recommendations of the National Comprehensive Cancer Network [41], BR-PDAC comprises findings with a distortion/narrowing or occlusion of the respective veins but a technical possibility of reconstruction on the proximal and distal margin of the veins. With regard to the arterial structures, a semi-circumferential abutment (< 180°) of the SMA or an attachment at the hepatic artery (HA) without contact towards the CA is regarded as a borderline resectable finding. Finally, irresectable PDAC is defined as a more extended involvement of the SMA, CA, aorta or inferior vena cava, and a SMV/PV venous involvement without a possibility for surgical reconstruction of the venous tract due to the lack of a suitable luminal diameter of the feeding and/or draining vein. This situation is most likely to be found in tumor-associated portal cavernous transformation. Besides these definitions, two other classifications are in clinical use, namely the definition of the AHPBA/SSO/SSAT published in 2009 [42] and the M.D. Anderson criteria that were published in 2006 [43]. Both definitions are similar to the abovementioned in terms of resectable PDAC. Regarding arterial involvement, there are no relevant differences between all definitions either, however, with regard to PV involvement; the M.D. Anderson definition does not include contact or narrowing of the vein, but gives occlusion as the criterion for BR-PDAC. Furthermore, in this publication, irresectability is termed “locally advanced.” An additional aspect with regard to patient management is the recommendation of the AHPBA/SSO/SSAT consensus that diagnostic laparoscopy should be performed in resectable tumors of the pancreatic head > 3 cm, in all tumors of the body and tail and in patients who display CA 19-9 levels > 100 U/ml. This aspect is not mentioned in the other definition or consensus and is based on the observation that a considerable proportion of patients fulfilling these criteria show unexpected peritoneal or liver metastases despite local resectability.

The most recent definition of resectability has been provided by the International Association of Pancreatology (IAP) in 2017 and extends the consideration to define resectability beyond merely anatomical issues [44]. Besides the criterion of anatomy (A), which is mainly similar to the ISGPS consensus with a more detailed subclassification of venous tumor involvement, a biological (B) as well as conditional (C) aspect is introduced. Biological BR (BR-B) includes patients with high CA 19-9 levels (> 500 iU/l) and suspected or proven lymph node metastases (PET CT/biopsy) and conditional BR (BR-C) describes a poor performance status of the patient. Both considerations are based on previous studies that show poor oncological outcomes despite tumor resection in each of these subgroups. Consequently, patients with an anatomically resectable tumor may be shifted to biological or conditional BR PDAC and upfront surgery may be critically discussed in these situations, although the IAP consensus does not clearly recommend neoadjuvant therapy or general non-surgical treatment.

In summary, the definitions of resectability show one field of pancreatic surgery, where clear definitions are available, however, the number of different classifications and the—partially important differences between them—do not allow to expect a real standardization at the moment.

Conclusion

Important progress has been made in defining and standardizing pancreatic surgery during the last decade. This has led to a much more valid and comparable scientific reporting of procedures, complications, and outcomes. Currently, there are ongoing efforts to further define and standardize specific topics in pancreatic surgery within the ISGPS as well as other groups and associations. Besides new definitions that may address perioperative care or minimally invasive procedures, there is need for a unification of definitions that should be applied worldwide in the future. Especially three important topics—resection status definition, cyst management recommendations, and the definition of resectability (BR)—could be aimed at to create more standardization and facilitate management as well as reporting. This would be of specific importance as these three topics cover the most controversial fields in pancreatic surgery at the moment and a number of international data registries as well as prospective studies are dealing with them or are expected to start in the near future.