Opinions and use of neoadjuvant therapy for resectable, borderline resectable, and locally advanced pancreatic cancer: international survey and case-vignette study

The treatment of pancreatic ductal adenocarcinoma (PDAC) has seen large changes during recent years. While surgery was only an option for resectable and non-metastatic PDAC in the past [1], modern poly-chemotherapeutic regimens (e.g. FOLFIRINOX or gemcitabine/nab-paclitaxel) with increased efficacy have changed the attitude to the management of PDAC: multimodality regimens are increasingly used to improve survival after curative resection, and increase resectability in both metastatic and non-metastatic disease [2‐4]. In addition, the surgical understanding of a curative resection is changing: recent literature suggests that the long term outcome is markedly improved if the circumferential tumor free margins is at least 1 mm (R0 wide) [5].

Several expert panels and associations have established criteria for resectability of PDAC [6]. These definitions contain a subgroup of (borderline resectable) tumors, which may be considered resectable or unresectable based on the available surgical expertise. Most centers liberally offer venous resections despite an impaired long-term survival in deep venous infiltration (portal/superior mesenteric vein, pv/smv) and are highly restrictive with arterial resections in PDAC [7‐9] because the limited oncological outcome adds to the risk for surgical morbidity. Consequently, the assessment of resectability of PDAC and the management of this subgroup varies widely based on the (surgical) expertise and the oncological practice of a center, but this heterogeneity complicates the interpretation of the current literature [2].

In parallel, neoadjuvant therapy is gaining popularity, but again indications may vary. Current literature suggests that a significant proportion of locally unresectable PDAC (=locally advanced PDAC, LAPC) may undergo radical resection with favourable outcome after adequate treatment response [10]. Recently, a randomized phase II-trial has reported a longer recurrence-free survival and R0-resection rate after neoadjuvant chemoradiation therapy for borderline resectable PDAC [11]. Even in primarily resectable tumors, resection margins appear wider, the proportion of lymph node metastases smaller, and survival superior following neoadjuvant chemotherapy. Since the definition and preoperative prediction of an R0 resection is difficult, any treatment with the potential aim of curative resection is considered neoadjuvant resulting in significant heterogeneity in the literature [2].

The aim of this survey was to analyze the current understanding of experts in the field regarding treatment aims and attitudes in the (neoadjuvant) management of pancreatic cancer.

A web-based survey was distributed through the administrative offices of the European-African Hepato-Pancreato Biliary Association (E-AHPBA) and the pancreas group of the European Organization for Research and Treatment of Cancer (EORTC) to the respective members using Surveymonkey™. In addition to questions regarding the individual experience in the treatment of pancreatic cancer, the survey was designed to evaluate the attitude of European experts regarding the definition of local resectability and the indications for neoadjuvant therapy of PDAC of the pancreatic head.

Since participation to this European survey was anonymous and participation was voluntary and only offered to experts in the field, an approval of an ethical review board was not considered necessary.

This first pan-European survey performed by E-AHPBA and EORTC on neoadjuvant treatment in pancreatic cancer found that the definition of borderline resectability (and unresectability) varies among European experts. These different definitions often do not affect the clinical management of patients with PDAC. This study confirms a paradigm shift in the understanding and treatment of patients with locally advanced and oligo-metastatic cancer, but also a broad heterogeneity in the management of such patients. For resectable PDAC, primary resection followed by adjuvant chemotherapy was the preferred concept, and a neoadjuvant approach was only recommended by a few participants. About half of the respondents would initiate neoadjuvant therapy in borderline resectable and in LAPC. Finally, neoadjuvant therapy was only recommended by a minority of participants for metastatic PDAC.

Following most classifications, a resectable tumor has a maximal tumor contact to the pv/smv of less than 180° without any contact to the celiac or superior mesenteric artery. Borderline resectability includes tumors which have more than 180° contact to the pv/smv and up to 180° contact to the celiac/superior mesenteric arteries. Tumor infiltration of the celiac/superior mesenteric arteries beyond 180° defines unresectable disease [6]. Although differences in the assessment of resectability are generally attributed to the surgical expertise of centers in the literature, this survey demonstrates that such differences are more likely related to different perceptions of PDAC disease and different oncological attitudes considering the experience in this field of most participating surgeons: while some may primarily consider the technical success of surgery, others predominantly see the questionable oncological benefit of an R1 resection. Accordingly, resectability rates in series on neoadjuvant therapy depend on the inclusion policy in addition to the surgical expertise and the applied treatment.

The heterogeneity in treatment concepts for unresectable or oligo-metastatic cancer is even more pronounced and mainly reflects the controversial literature: a few recent publications suggest favorable long-term outcome of patients with unresectable or even oligo-metastatic disease after response to FOLFIRINOX [10, 12, 13]. Although these are retrospective studies, some centers have already adopted these concepts for selected patients. Thus, a large proportion of participants considered both scenarios as potentially curable in a multimodality concept.

In contrast to well established definitions of adjuvant and palliative treatment aims, a clear definition of treatment aims for a neoadjuvant treatment has not been established, yet. Accordingly, about half of the recipients considered the conversion from an unresectable to a resectable disease as aim of neoadjuvant therapy, while ¾ of them considered the increase in the R0-resection rate as aim of neoadjuvant therapy. Considering the aim of an adjuvant treatment (reduction of tumor recurrence) and the homology of the terms “adjuvant” and “neoadjuvant”, the primary aim of a neoadjuvant treatment should be to decrease the recurrence risk after a curative resection, which may be attributed to fewer lymph node metastases as well as the treatment of micrometastases and circulating tumor cells as indicated by most participants. Further beneficial effects may be larger resection margins, a lower R1 resection rate and smaller tumors. In locally unresectable tumors or whenever the risk of an R1 (wide) resection is high, the primary treatment aim of a preoperative therapy is to shrink the tumor and convert a (potentially) unresectable to a resectable disease. In this scenario, the treatment should be considered a down-sizing or conversion therapy [2]. If secondary resectability is not attempted (or unprobable), a palliative therapy is indicated. Similarly, neoadjuvant therapy for other tumors such as breast or upper GI cancer implies the preoperative therapy of technically resectable tumors with the aim of a reduction of tumor recurrence [14‐16]. Also for colorectal liver metastases, many authors differentiate a “conversion” and a “neoadjuvant” chemotherapy [17‐19].

Such differentiation in the terminology of treatment concepts would particularly increase the comparability of results in the borderline resectable group: a borderline resectable tumor (e.g. infiltration of the smv) according to current definitions could be treated by a neoadjuvant therapy with an expected high resectability rate – since it is most probably primarily resectable and would be resected upfront in many experienced centers. Since a deep infiltration may be associated with impaired outcome [9], a real neoadjuvant therapy could be given with the aim of an improvement of recurrence-free and overall survival in this scenario. The same idea is the basis of neoadjuvant therapy in resectable tumors [6]. A primarily unresectable disease, however, would undergo conversion therapy with an estimated lower secondary resection rate (depending on the tumor response rate). The same would be true for a borderline resectable tumor with tumor contact to the superior mesenteric artery – a scenario with a high risk of R1- or even un-resectability. Moreover, in the case of a resectable primary tumor with resectable liver metastases, the aim of a neoadjuvant therapy would be to decrease recurrence rate (e.g. new metastases) after the resection of both the primary tumor and metastases (e.g. case 4).

These two separate definitions of “neoadjuvant” treatment would not necessarily change the management of the patient, since the same treatment is often applied in both situations. However, they might help to better define inclusion criteria for clinical studies and the reading of the literature. If a high response rate is required for a locally unresectable tumor, the conversion therapy should be a maximally potent local treatment (including chemo-radiation), while a systemic treatment could be more suitable to prevent metastatic recurrence in some borderline resectable tumors (e.g. portal vein infiltration).

In conclusion, this first E-AHPBA/EORTC pan-European survey documents a shift in attitudes regarding the management of advanced PDAC among HPB surgeons and oncologists with fairly large support for neoadjuvant treatment. Moreover, this survey depicts different understandings of borderline resectability and neoadjuvant therapy which should trigger a discussion on an adaption of definitions in the setting of PDAC.