Intraductal papillary-mucinous neoplasm (IPMN) of the pancreas is a relatively new tumour entity that has unique clinicopathological features.1^–7 They are defined as an intraductal papillary growth of neoplastic columnar cells producing mucin that arise in the main pancreatic duct or its major branches. These neoplasms are characterised by variable biological behaviour and are classified as adenoma, borderline and carcinoma, the last category including both in situ and invasive varieties.8 Although current thinking is that IPMNs have the potential to be malignant,9 it is not known whether all IPMNs have malignant potential or what the time course of progression may be.

IPMNs can be subdivided into “main duct” and “branch duct” types depending on the location of the lesion in the main pancreatic duct or the side branch, respectively.8 10 There are significant differences in the prevalence of cancer between main duct IPMNs and branch duct IPMNs, ranging from 57% to 92% and 6% to 46%, respectively.9 11^–15 The largest published series on patients with main duct IPMNs showed that approximately 70% are symptomatic, 60% had carcinoma and 46% had invasive carcinoma.9 In contrast to main duct IPMNs, most branch duct IPMNs are asymptomatic and it is believed that the risk of malignancy is low in cystic lesions <30 mm in size.14 15 Therefore, the classification of IPMNs has substantial value in determining the management strategy for this neoplasm.

The majority of IPMNs are branch duct IPMNs, which are being diagnosed with increasing frequency. Despite frequent reporting, management of branch duct IPMNs has been controversial. The treatment controversy is based on the fact that IPMNs contain a wide range of histological findings from apparently benign to overtly malignant, and accurate preoperative prediction of malignancy is not possible by imaging modalities. Recently, it has been reported that one of the acceptable predictive signs of malignancy for branch duct IPMNs is the presence of mural nodules or protruding lesions in dilated branch ducts.15^–20 Therefore, branch duct IPMNs without mural nodules seem to be less aggressive compared with those with mural nodules or main duct IPMNs, and may possibly be followed-up conservatively. However, there is no clear consensus on how best to manage branch duct IPMNs without mural nodules. Understanding the natural history of branch duct IPMNs without mural nodules would allow us to improve management strategies for this neoplasm. In this study, the natural history of branch duct IPMNs without mural nodules during long-term follow-up was studied prospectively.

METHODS

Patients

Eighty-two consecutive patients with branch duct IPMNs attending the medical gastroenterological units of our institution: Department of Gastroenterology, Asahikawa Medical College, Asahikawa, between 1990 and the end point of this study (December 2006), who fulfilled the diagnotic criteria of branch duct IPMN cited below, were included in the study.

Diagnosis of branch duct IPMNs was based on the pathognomonic findings of endoscopic retrograde pancreatography (ERP)21—that is, communication between the cyst and the main pancreatic duct (MPD), and the presence of mucin, as indicated by filling defects in the pancreatic duct—and was confirmed by ERP in all cases at diagnosis of branch duct IPMNs. To exclude mucinous cystic tumours of the pancreas or pseudocysts, characteristic findings of imaging studies, a grape-like multilocular cystic lesion communicating with the MPD was confirmed by CT scan, magnetic resonance cholangiopancreatography (MRCP) and endoscopic ultrasonography (EUS) in all cases. The maximum diameter of the dilated branch duct and MPD was measured by EUS. Absence of mural nodules in dilated branch ducts was diagnosed by EUS in combination with MRCP and CT. To exclude pancreatic intraepithelial neoplasia, the size of IPMNs had to be >1 cm.9

Follow-up

Eighty-two branch duct IPMNs without mural nodules were selected for follow-up without surgical resection as outpatients. All followed-up patients gave consent for observation instead of surgical procedures. Imaging modalities including EUS in combination with CT scan and/or MRCP were performed once or twice a year. Serial changes of the maximum cystic diameter and the appearance of mural nodules were studied during the observation periods, ranging from 14 to 148 months (median, 61 months). Dilated branch ducts showing changes >10 mm in diameter were defined as enlarged or reduced.

Statistical analysis

Statistical significance was analysed by Student t and Fisher exact χ² tests. Differences were considered significant when the p value was <0.05.

RESULTS

Patient characteristics

The characteristics of 82 branch duct IPMN patients attending the medical gastroenterological unit of the Asahikawa Medical College are summarised in table 1. They consisted of 57 men and 25 women, with a median age of 68 years (range, 40–84 years). Only one case (1.2%) had recurrent pancreatitis; the remaining patients were asymptomatic. The median size of the dilated branch duct was 20 mm (range, 11–45 mm) at diagnosis of branch duct IPMNs. The location of the lesion was 49 (59.8%) in the head and 33 (40.2%) in the body to tail of the pancreas. The median maximum diameter of the MPD was 3 mm (range, 3–12 mm). The median follow-up period was 61 months (range, 14–148 months). Ten of 82 (12.2%) branch duct IPMNs had an initial dilated branch duct diameter of ⩾30 mm, with a median diameter of 30 mm (range, 30–45 mm) (table 2). Fifty-four (65.9%) of those with branch duct IPMNs were older than 65 years in age, with a median age of 73 years (range, 65–84 years). There is no significant association between age and the initial size of the lesions in branch duct IPMN patients (p = 0.314).

Clinical outcome of branch duct IPMNs without mural nodules

Table 3 shows the clinical outcome of 82 branch duct IPMNs without mural nodules during long-term follow-up. Nine (11.0%) of 82 patients exhibited obvious progression of cystic dilatation after a median follow-up of 59 months (range, 14–113 months). There were no patients with a reduction in size of dilated branch ducts. Four patients (4.9%) showed the appearance of mural nodules in dilated branch ducts after a median follow-up of 105 months (range, 45–148 months). None of the remaining 69 patients (84.1%) showed any changes in dilated branch ducts after a median follow-up of 57 months (range, 14–138 months).

The characteristics of nine branch duct IPMNs with increased size of dilated branch ducts are shown in table 4. The nine cases consisted of one woman and eight men, with a median age of 74 years (range, 52–81 years) at diagnosis of branch duct IPMNs. Four (44.4%) of these nine patients had a MPD diameter of ⩾7 mm. Cystic enlargement was observed between 14 and 113 months (median, 59 months) after the diagnosis of branch duct IPMNs. Although one patient had recurrent pancreatitis (patient 9), the remainder were asymptomatic. Of nine patients with cystic enlargement, six continued with regular follow-up examinations. Three cases underwent surgical resection and the IPMNs were pathologically diagnosed as adenoma in two and borderline in one.

The characteristics of the four patients with branch duct IPMNs with the appearance of mural nodules in dilated branch ducts are shown in table 5. They consisted of four men, with a median age of 62 years (range, 43–74 years) at diagnosis of branch duct IPMNs. All patients were asymptomatic. One patient had an MPD diameter of 12 mm (patient 1). Mural nodules appeared between 45 and 148 months (median, 105 months) after diagnosis of branch duct IPMNs. Median maximum size of the mural nodules was 6.5 mm (range, 6–8 mm). All of the patients with newly developed mural nodules underwent surgical resection. At histological analysis, three cases were classified as adenoma and one as carcinoma in situ.

Table 6 shows a comparison of various clinical factors between patients with no cystic changes, with increased size of dilated branch ducts and with the appearance of mural nodules. The latest size of dilated branch ducts with the appearance of mural nodules was significantly larger than those without cystic changes, but not statistically different from those with cystic enlargement. The MPD diameter in the patients without cystic changes was significantly narrower than those with cystic enlargement or with the appearance of mural nodules. The time of follow-up in patients with appearance of mural nodules was significantly longer than those without cystic changes, but not statistically different from those with cystic enlargement. There were no significant differences in age, gender, location or initial size of the dilated branch duct between patients with cystic size changes and patients with the appearance of mural nodules.

DISCUSSION

Branch duct IPMN has been increasingly recognised and is now being diagnosed with increasing frequency. Many of these patients are elderly and asymptomatic. The majority of these lesions are believed to be premalignant, and the only curative treatment is surgical resection.22 Despite the more frequent reporting of branch duct IPMN, the natural history of this disease is not well understood, and a number of controversies regarding its treatment persist. The majority of branch duct IPMNs appear to have no apparent mural nodules; however, little precise information is available on the natural history of branch duct IPMNs without mural nodules.

Although current thinking is that all IPMNs have the potential to be malignant,9 the frequency and rate of progression to carcinoma in situ and invasive carcinoma are not known and may be different for branch duct IPMNs with and without mural nodules. It is clear that branch duct IPMN has a more favourable prognosis than standard ductal pancreatic adenocarcinoma23; however, the factors that determine outcome in these patients are not well understood. This is the first report demonstrating the evidence that the absence of mural nodules may be an important factor in determining a conservative management strategy for this tumour, as most branch duct IPMNs without mural nodules remained unchanged during long-term follow-up. Our data emphasise that branch duct IPMNs without mural nodules can be followed-up with careful examination to detect newly developed mural nodules in the dilated branch duct. The advantages of our study include a prospective study design, a near complete follow-up with a median of 61 months, and a much larger sample size of branch duct IPMNs than in previous studies.24^–29

In this study, 69 (84.1%) of 82 branch duct IPMN patients without mural nodules showed no changes in the dilated branch duct during follow-up lasting 14–138 months (median, 57 months) after diagnosis of branch duct IPMNs, whereas progression of branch duct IPMNs was found in 13 patients (15.9%) during follow-up lasting 14–148 months (median, 67 months). Of 13 patients exhibiting progression of IPMNs, nine patients (69.2%) had cystic enlargement alone. Three cases underwent surgical resection and were pathologically diagnosed as adenoma in two and borderline in one. Of four patients with newly developed mural nodules who underwent surgical resection, histological analysis showed that three cases were classified as adenoma and one as carcinoma in situ. Importantly, our study demonstrated that there was no invasive carcinoma in surgically resected specimens of branch duct IPMNs without mural nodules that had an increase in dilated branch duct size or the appearance of mural nodules. These findings suggest that close observation with careful attention to the appearance of mural nodules or cystic enlargement will obviate unnecessary surgery. However, the presence of mural nodules should be considered to be an indication for surgery in order to avoid the wrong treatment, because mural nodules have been reported to be indicative of malignancy in branch duct IPMNs,15^–20 and the natural history of branch duct IPMNs with mural nodules is not understood.

In the past, several authors have noted that a cyst diameter of ⩾30 mm suggests malignancy in branch duct IPMNs.30 31 In this study, the 82 branch duct IPMN patients without mural nodules included 10 patients (12.2%) with a cystic size of ⩾30 mm in diameter. Among them, two patients (20%) showed cystic enlargement alone and one (10%) had the appearance of a mural nodule, at 64.9 and 60.3 months on average after diagnosis of branch duct IPMNs, respectively. At pathological analysis, two cases were classified as adenoma and one as borderline. The remaining seven patients (70%) remained unchanged during the follow-up lasting 21–50 months (median, 30 months). In our experience, therefore, cystic lesions of ⩾30 mm in diameter did not correlate well with cystic enlargement or the appearance of mural nodules. A similar observation was made by Chari et al.32 In their study, 73 non-invasive IPMNs had an average diameter of 52 mm, whereas 40 invasive IPMNs had an average diameter of 66 mm. Thus, these data suggest that the size of a branch duct IPMN alone cannot be used to determine the treatment strategy.

Several authors reported that a dilated MPD of ⩾7 mm may predict malignancy of branch duct IPMNs in studies based on pathological analysis of surgically resected pancreas.24^–27 In this study, the mean maximum diameter of the MPD in branch duct IPMNs without cystic size changes was significantly narrower than those with the appearance of mural nodules or enlargement of the dilated branch duct. Eleven (13.4%) of 82 patients had a dilated MPD of ⩾7 mm, and six of these (54.5%) showed progression of branch duct IPMN during follow-up lasting 14–148 months (median, 73 months). Furthermore, all three cases which had a MPD diameter of ⩾10 mm showed cystic enlargement in two and the appearance of mural nodules in one. Thus, our follow-up data support the previous reports24^–27 and suggest that a dilated MPD of ⩾7 mm may also predict the progression of branch duct IPMNs without mural nodules.

In this study, 54 (65.9%) branch duct IPMN patients were 65 years or older. Among them, eight patients (14.8%) showed cystic enlargement in six and the appearance of mural nodules in two. Of 28 patients <64 years old, five patients (17.9%) showed cystic enlargement in two and appearance of mural nodules in three. The mean age of those with branch duct IPMNs without cyst size changes at diagnosis of IPMNs was not significantly less than those with the appearance of mural nodules or enlargement of the dilated branch duct. In addition, the mean follow-up period of branch duct IPMNs without cyst size changes was not shorter than those with the appearance of mural nodules or enlargement of the dilated branch duct. Thus, these data suggest that age may not be a predictive factor for progression of branch duct IPMNs without mural nodules. Quality of life must certainly be taken into account in electing to perform surgery for branch duct IPMNs in elderly patients, especially with very old patients or those with impairments, because the operation involves a high level of morbidity.

In conclusion, the current study demonstrated the long-term follow-up results of branch duct IPMN patients without mural nodules. The absence of mural nodules supports a conservative management policy without surgery in branch duct IPMNs, although follow-up with careful examination is required to detect the appearance of mural nodules in dilated branch ducts. It is recommended that careful attention should be paid to the presence or absence of mural nodules in order to determine the treatment strategy for branch duct IPMNs.