IPMNs have an even greater chance of being malignant than MCNs (35). About 40 % of patients at the time of diagnosis of main-duct IPMN already have an established invasive malignancy [36, 37]. Accurate classification of IPMNs is of special clinical importance, because the risk of malignancy depends on the changes of the pancreatic ducts and its branches [38]. The risk of invasive malignancy is relatively high (≥40 %) in both main-duct and mixed-duct types.

Currently, risk factors for the presence of malignancy within IPMN include [9, 36, 39]:

Histologic changes (atypia, dysplasia, or frank carcinoma in situ or invasive cancer) can be present concurrently in discontinuous areas throughout the pancreas, thereby raising the question of whether IPMN represents a generalized global disorder of the epithelium of the pancreatic duct or rather a localized, field defect. True multicentricity of main-duct IPMN is not common (<10 %), but in branch-duct IPMNs, multicentricity has been recognized much more frequently [35]. Multicentricity is of clinical importance for the surgeon performing pancreatectomy (see below).

Because of the overt or latent malignant potential of IPMN, operative resection is the therapy of choice in most patients with main-duct and mixed-duct IPMN, provided they are operative candidates [3, 9, 10, 14, 16, 35, 36, 38, 39, 41]. Operative therapy is more controversial with branch-duct disease alone. Analytically, a formal oncologic resection is indicated for [35, 38]:

Recently, a conservative management (“watchful waiting”) has been proposed for selected patients with IPMNs. This approach can be considered in selected subgroups, including high-risk surgical patients if other predictors of malignancy are not present (e.g., for main-duct IPMNs, when the main pancreatic duct is smaller than 10 mm and there are no visible mural nodules). This conservative approach has been accepted with more support in selected patients with branch-duct IPMNs, given the much lesser incidence of invasive cancer in branch-duct disease [35, 36, 45, 46]. This group of patients includes asymptomatic patients with branch-duct IPMNs with cystic lesions <3 cm, no mural nodules or main-duct dilation, and no cytologic findings suspicious or positive for malignancy [35, 42, 47–49]; this approach is based on the low incidence of invasive malignancy (~2 %) in these patients, which approximates the mortality risk of a major pancreatic resection (i.e., pancreatoduodenectomy) [48]. Close follow-up using periodic cross-sectional imaging is required in these cases to detect suspicious findings suggesting a reevaluation of the situation and reassessment of the role of operative resection. The timing of surveillance is determined by cyst size [38]. Obviously, the “watchful waiting” approach is applicable only if the patient can be kept under close supervision. Tables 6.1 and 6.2 summarize the recommendations of the International Association of Pancreatology regarding optimal management of IPMNs and MCNs (Sendai 2006 and the more recent [2012] consensus guidelines) [35, 50].

The aim of operative resection is to remove all the adenomatous or malignant ductal epithelium if possible and if reasonable to minimize the probability of recurrence in the pancreatic remnant. The basic and as yet not fully answered question is whether or not IPMN represents a local, clonal expansion of a site of neoplastic transformation, a localized field defect limited to a segment of the pancreas, a global abnormality in the ductal epithelium with the potential to affect all of the pancreatic ductal epithelium, or even possibly an anatomically local or global environmental stimulus, either exogenous or endogenous. If we know IPMN is a localized process, as with typical ductal cancer of the pancreas, then a focused resection of the involved anatomic region of the gland would be all that is necessary. In contrast, if IPMN is a global disorder of all the pancreatic ductal epithelium, probably then all the pancreatic duct epithelium is “at risk” of malignant transformation, and therefore, in selected individuals, a total pancreatectomy might be indicated [9]. Total pancreatectomy with its obligate apancreatic state has its own often serious problems (brittle diabetes, exocrine insufficiency) and trades one disease (IPMN) with another (the “apancreatic” state with its endocrine and exocrine deficiencies). In some patients, a total pancreatectomy may not be appropriate, especially in the elderly or the medically unsophisticated patient who will not be able to manage the endocrine and exocrine insufficiency. After total pancreatectomy, episodic hypoglycemia can be a substantive problem.

The type of operation is determined based on the location of the IPMN and its subtype (i.e., the type of ductal distribution – main or mixed IPMN vs. branch-duct IPMN). For main-duct IPMN located in the pancreatic body/tail (10–25 % of patients) [51], distal pancreatectomy including splenectomy with frozen-section analysis of the proximal pancreatic margin is the procedure of choice [9, 39]. If the frozen section is negative for true adenomatous changes in the ductal epithelium (not reactive ductal hyperplasia), total pancreatectomy is not indicated in the absence of objective evidence that the proximal duct is involved. In contrast, when the margin is positive for invasive or noninvasive malignant IPMN, most surgeons would advocate a further “creeping” proximal pancreatic resection; if a tumor-free margin is not attainable after two further limited resections, most surgeons would proceed with total pancreatectomy, provided the patient is an appropriate candidate to manage the “disease” of the apancreatic state [10, 14, 35, 39]; obviously this discussion would have occurred preoperatively between patient and surgeon.

When the entire pancreatic duct is diffusely dilated, the assumption is that the disease is in the pancreatic head causing obstruction by growth of the neoplasm and/or by mucous production. Based on this assumption and provided no intraluminal or extraluminal solid mass is evident elsewhere in the duct outside the boundaries of a pancreatic head resection, a pancreatoduodenectomy is undertaken with intraoperative frozen-section analysis of the distal margin. A positive margin for adenomatous changes (again, not ductal hyperplasia) necessitates a further “creeping” distal resection, keeping in mind that IPMN may involve the pancreatic duct diffusely. If the frozen section remains positive after two attempts for further resection, total pancreatectomy should be entertained (in up to 10–20 % of patients) [35, 36, 39]. The concept of “prophylactic” total pancreatectomy is considered by most pancreatic surgeons as both unacceptably aggressive and unnecessary in most patients [52]. When evaluating the results of frozen section, it should be emphasized that the surgeon should keep in mind that even a negative margin does not assure the absence of neoplastic cells in the remaining pancreas. Unlike typical ductal carcinoma of the pancreas which is a contiguous clonal expansion and not a multicentric malignancy [53], IPMN can be a multifocal disease in up to 8–10 % of patients harboring main-duct disease with “skip” lesions, possibly indicating a generalized instability of the epithelium [35]. Intraoperative pancreatic ductoscopy to evaluate the pancreatic remnant has been tried with some success. While nodal metastases occur with invasive IPMN, at present, there is no evidence to justify an extended lymphadenectomy in the management of malignant IPMN.

In localized branch-duct IPMNs, a segmental but formal anatomic, oncologic pancreatectomy is the favored procedure, i.e., pancreatoduodenectomy (preferentially of the pylorus-preserving type) for neoplasms located in the pancreatic head/uncinate process, or distal pancreatectomy for body/tail lesions [39]. In multifocal branch-duct IPMN, therapeutic decisions are more difficult; ideally, these patients should be treated by total pancreatectomy, which eliminates all the foci of the disease, but the formidable and obligate long-term morbidity of total pancreatectomy must be considered seriously in this decision. A more conservative approach in this case would be an anatomic pancreatic resection removing the cystic lesions with worrisome characteristics (size >3 cm, mural nodules, abnormal cytology) and surveillance observation of the remnant gland/lesions for findings suggestive of malignancy in the remaining cystic lesions [35].

Solid pseudopapillary neoplasms (SPNs) are usually very-low-grade, malignant neoplasms with the potential for metastatic spread, and operative resection is, therefore, recommended in all patients [54]. More than 95 % of SPNs arise in women and usually before the age of 40 years; their appearance is rather characteristic especially in a young women. Cystic neuroendocrine neoplasms are considered premalignant or malignant lesions and should be treated operatively. Typically, however, accurate preoperative identification is not possible, even using sophisticated, modern, state-of-the-art imaging and molecular techniques, and thus operative resection usually establishes the diagnosis definitively [55].

Because of uncertainty of the diagnosis and concerns about malignancy, rare pancreatic neoplasms should be treated with a formal, anatomic pancreatic resection. Despite that theoretic concept that a more conservative approach could be acceptable in selected cases, a radical approach is preferred, especially in surgically fit patients to avoid the risk of undertreatment of a potentially curable disease [55, 56].