|Home | About | Journals | Submit | Contact Us | Français|
Progress in the diagnosis and management plan for this pancreatic neoplasm
A large prospective study by Salvia et al1 published in this issue of Gut shows that a follow‐up protocol for branch duct intraductal papillary mucinous neoplasms (IPMNs) appears feasible and safe (see page 1086). IPMNs of the pancreas were originally described in a number of case reports and short series of patients in the early 1990s,2,3 and mucinous cystic neoplasms of the pancreas had been reported in the 1980s.4,5,6 In 1996 the World Health Organisation (WHO) revised the criteria for the pathological diagnosis of IPMN,7 thus allowing the differentiation of IPMN from other mucinous/cystic neoplasms of the pancreas. This meant that a number of “different” neoplasms could now come under the single diagnostic umbrella of IPMN (table 11).
The apparent incidence of IPMN has increased dramatically over the past 10–15 years and now represents up to 10–20% of the pancreatic resection workload of specialist units.8 This is due in part to improved imaging techniques, greater recognition of this clinicopathological entity and a larger number of asymptomatic patients undergoing cross‐sectional imaging.9
Histologically, IPMNs progress along a pathway from adenoma to borderline IPMN with dysplasia to IPMN with carcinoma in situ and eventually to invasive carcinoma. It is not always possible to differentiate high‐grade and low‐grade lesions on imaging alone, and the time to progression is not known. The association of IPMNs (particularly the gastric type variant) with the precursor lesions of pancreatic ductal adenocarcinoma known as pancreatic intra‐epithelial neoplasms (PanINs) or PanIN‐like lesions may provide further information concerning their development.10,11 Molecular analysis may provide biomarkers of disease progression in the near future with further follow‐up of patients with characterised lesions. The expression of different mucins has been associated with prognosis in previous studies.12 Various genetic and epigenetic alterations are associated with IPMNs (table 22).). As with pancreatic ductal adenocarcinoma, point mutations of the K‐ras oncogene occur in up to 80% of IPMN lesions, but in contrast to pancreatic ductal adenocarcinoma mutations of the tumour suppressor genes p53, p16 and SMAD4 occur less frequently or not at all (0–15%). IPMNs exhibit aberrant methylation at a variety of loci and there are certain upregulated genes demonstrated by microarray analyses.13,14,15 These alterations tend to occur at higher frequency in IPMNs with an invasive component.
Certain inherited and familial syndromes have an association with IPMN, including Peutz‐Jeghers syndrome and adenomatous familial polyposis.16,17,18,19 IPMN is also associated with other extrapancreatic malignancies such as colorectal cancer.20
Because of its variable malignant potential, the diagnostic and therapeutic challenge presented by IPMN is a singular one. The main questions are: who requires surgery and who does not and who can be followed up safely and, if so, what should the follow‐up involve? In this issue of Gut, Salvia et al1 attempt to answer this important question or at least provide further evidence to support critical clinical decisions. This paper reports on the follow‐up of 109 patients with branch duct IPMN (BD‐IPMN). Their diagnostic methods reflect the current standards for IPMN, as described in the following paragraphs, including clinical symptoms and imaging techniques such as magnetic resonance cholangiopancreatography (MRCP).
Most IPMNs occur in the head of the pancreas (>60%) and represent around 1% of pancreatic neoplasms.21 Patients may present with abdominal pain, jaundice, weight loss, acute pancreatitis or endocrine and/or exocrine insufficiency, or patients may be asymptomatic with an incidental diagnosis of IPMN. The diagnosis of IPMN relies on imaging to reveal typical features such as irregular dilation of the main pancreatic duct, a gaping ampulla (with copious mucin production), cystic lesion(s) communicating with the main pancreatic duct or side branches, or filling defects in the pancreatic duct. Multi‐detector CT scanning, MRCP, endoscopic retrograde cholangiopancreatography and endoluminal ultrasound (EUS) are all useful in the diagnosis of IPMN.22,23 Analysis of cyst fluid for mucin, carcinoembryonic antigen and estimation of serum CA19.9 may also add further diagnostic value.24,25 The study by Salvia et al1 secured a diagnosis of BD‐IPMN using transabdominal ultrasound and MRCP. Criteria for a suspicion of malignancy were cyst size >3.5 cm, the presence of nodules and/or a thick wall, serum levels of CA19.9 >25 U/l, recent onset of diabetes mellitus or worsening of diabetes mellitus and/or the presence of clinical symptoms.
Once the diagnosis of IPMN is made, surgery is the main therapeutic option because of the risk of malignant transformation but, until recently, it has not been possible to separate lesions with relatively different risks of progression. This is important as these patients are often in their sixth or seventh decade and the risks of surgery are not inconsiderable.26 At the present time patients with a main duct (MD)‐IPMN should be offered surgery if they are fit. The type of resection will depend on the distribution of the lesions. Most lesions occur in the head and will entail a pylorus‐preserving or classic Kausch‐Whipple resection, whereas lesions of the tail will require a left pancreatectomy. At the time of surgery, frozen sections must be sent from the resection margin. This will dictate whether further pancreatic resection and possible total pancreatectomy is required. Patients with invasive IPMN who have positive resection margins have disease recurrence rates of over 50% and poor long‐term survival.27 Thus, all patients must be made fully aware that a planned partial pancreatic resection may need to be converted to a total pancreatectomy depending on the result of the intraoperative frozen section. Patients who have undergone resection for invasive IPMN have 5 year survival rates of 13–78%. The survival for those with non‐invasive IPMN at the time of resection can be up to 90% or 100% (table 33).7,21,26,28,29,30,31,32 The role of adjuvant treatment following resection for invasive IPMN has not been established.
Prognostic factors which are significant following resection for IPMN include lymph node involvement, the presence of invasive IPMN, neural and vascular invasion, resection margin status (for invasive IPMN) and whether there is MD‐IPMN or BD‐IPMN.7,21,26,28,29,30,31,32
Using the above mentioned diagnostic criteria, the 109 patients with BD‐IPMN in the study by Salvia et al1 were divided into two groups, one of 20 patients thought to be at high risk for malignancy and therefore suitable for resection, and 89 patients thought to be low risk. Of the 20 patients who had a resection, 2 (10%) had invasive carcinoma and eventually developed hepatic metastases and died 20 and 40 months after resection. It is interesting that none of the asymptomatic patients who underwent surgery had invasive carcinoma or carcinoma in situ. The overall follow‐up for this arm was a median of 44.5 months.
Eighty‐nine patients were followed up with either EUS or MRCP at 6 monthly intervals. These patients were asymptomatic and, interestingly, approximately two‐thirds had multifocal disease. After a median follow up of 32 months, five patients underwent resection because their lesions increased in size (none had invasive carcinoma), two died of myocardial infarction at 60 and 50 months after the first observation, and the rest are still under follow‐up and remain asymptomatic.
Because surgical intervention may be associated with a high morbidity rate26 and IPMNs have a variable malignant potential, there has been considerable interest in the possibility of a safe follow‐up protocol for patients with non‐invasive IPMN. Patients with BD‐IPMN exhibit lower rates of invasive disease and also better survival than those with MD‐IPMN, even with limited resections.28,33 It would seem reasonable that follow‐up protocols for the present should focus on these patients. The International Association of Pancreatology (IAP) guidelines34 recommend yearly follow‐up for BD‐IPMN lesions <10 mm in size, 6–12 monthly for those 10–20 mm in size and 3–6 monthly for lesions >20 mm.
There are very few prospective series of follow‐up protocols that have long‐term results on a reasonable number of patients, which is why the paper by Salvia et al1 is important and provides good evidence that adopting a protocol based on the IAP guidelines can work in a carefully diagnosed and monitored group of patients. The strengths of this paper are that it is a prospective study with a large number of patients, the accuracy of the diagnosis was impressive (confirmed by a 100% correct diagnosis of IPMN in those patients who were resected) and complete follow‐up data were available. The main drawbacks are that adopting this protocol may lead to overtreatment as only 10% of patients (in the resection group) had invasive carcinoma (the presence of symptoms appears to be extremely important in this group of patients) and the median follow‐up in the asymptomatic group is only 32 months. A much longer follow‐up period will be required before stronger recommendations can be made, so further prospective studies with substantial follow‐up times are still needed.
Nevertheless, at the present time, surgery should be recommended for patients with MD‐IPMN. Those patients with BD‐IPMN who are symptomatic or exhibit features which fulfil the criteria for a high suspicion of invasive IPMN (cyst size >3.5 cm, nodules and/or a thick wall, raised serum CA19.9 level) should also be offered surgery. The remaining patients with BD‐IPMN may be followed up at 6–12 monthly intervals using non‐invasive imaging such as MRCP, EUS or CT and serum CA19.9 estimation.
The study by Salvia et al1 relied on radiological and endoscopic procedures and tumour marker analysis. Future studies should be designed to include molecular markers to add an extra component to the protocol. Several studies have assessed molecular markers. In one study, methylated ppENK was detected in four (44%) of nine samples of pancreatic juice from patients with high‐grade (carcinoma in situ) but not in pancreatic juice samples from two patients with low‐grade (borderline) IPMNs.35 The expression of two or more of the four proteins S100A4, mesothelin, claudin 4 and CXCR4 was commonly observed in IPMNs with an invasive carcinoma (16 [73%] of 22 invasive IPMNs) but was not seen in any of 16 IPMNs without an invasive phenotype.12 Nakamura et al reported that prognosis of IPMN was associated with mucin production.11 The subset of IPMNs expressing gastric type mucin (MUC5AC) had the most favourable prognosis, and the subset of IPMNs expressing intestinal type secretory mucin (MUC2) had a better prognosis than that of IPMNs expressing membrane‐bound type mucin (MUC1).11
In a relatively short period of time the diagnosis and management plan of this pancreatic neoplasm has progressed at an encouraging rate. Advances in diagnosis and surgical approaches have improved the outlook for these patients, and in the near future it is anticipated that the contribution of the underlying molecular and genetic mechanisms of this disease will further impact on clinical management.
The authors are grateful to Cancer Research UK for programme and project funding for pancreatic cancer research.
Competing interests: None.