Increased risk of second malignancy in pancreatic intraductal papillary mucinous tumors: Review of the literature

Abstract

AIM: To analyze the available evidence about the risk of extrapancreatic malignancies and pancreatic ductal adenocarcinoma associated to pancreatic intraductal papillary mucinous tumors (IPMNs).

METHODS: A systematic search of literature was undertaken using MEDLINE, EMBASE, Cochrane and Web-of-Science libraries. No limitations for year of publication were considered; preference was given to English papers. All references in selected articles were further screened for additional publications. Both clinical series and Literature reviews were selected. For all eligible studies, a standard data extraction form was filled in and the following data were extracted: study design, number of patients, prevalence of pancreatic cancer and extrapancreatic malignancies in IPMN patients and control groups, if available.

RESULTS: A total of 805 abstracts were selected and read; 25 articles were considered pertinent and 17 were chosen for the present systematic review. Eleven monocentric series, 1 multicentric series, 1 case-control study, 1 population-based study and 3 case report were included. A total of 2881 patients were globally analyzed as study group, and the incidence of pancreatic cancer and/or extrapancreatic malignancies ranged from 5% to 52%, with a mean of 28.71%. When a control group was analyzed (6 papers), the same incidence was as low as 9.4%.

CONCLUSION: The available Literature is unanimous in claiming IPMNs to be strongly associated with pancreatic and extrapancreatic malignancies. The consequences in IPMNs management are herein discussed.

Core tip: This paper enter a clinical debate which has real relevance in the daily practice in the field of pancreatic cystic neoplasm, as a number of asymptomatic intraductal papillary mucinous tumors (IPMNs) are diagnosed every day by abdominal imaging, and the actual most common clinical attitude provides for a weak diagnostic pathway. However, in the light of an unanimous literature, this attitude would be a while dangerous, as up to 30%-40% of IPMN cases would develop a second malignancy in their life. The same should be underlined when discussing the follow-up protocols.

INTRODUCTION

Intraductal papillary mucinous tumors (IPMNs) belong to the group of cystic neoplasms of the pancreas, and are characterized by a communication with the Wirsung duct and by the absence of ovarian-type stroma[1]; the cystic expansion of the main or a branch pancreatic duct is due to a papillary growth of the epithelium, with rich mucin production. IPMNs were firstly described in 1982[2], and subsequently classified as adenoma, borderline or carcinoma (in situ or invasive) tumors[3,4]. In the following years, from a number of clinical series[5-9] based both upon surgical treatment[10-13] and radiological follow-up[14-17], clinical and pathological features characteristic of IPMNs clearly emerged: the neoplasm is symptomatic in many cases, recurrent acute pancreatitis without recognised biliary or alcoholic etiology being the most frequent expression. The diagnosis is based on CT and MRI imaging, which further distinguish 2 different variants, the main duct tumor (MDT) and the branch duct tumor (BDT), having different incidence of malignancy and prognosis. The natural history of IPMNs remains partially unclear. At diagnosis, 10%-20% of cases harbour invasive carcinoma, 10%-20% in situ carcinoma and the remaining 60%-80% intraductal or intracystic papillomatosis with simple dysplasia. In the published series, resectability was 90%-100%, while mortality and morbidity were similar to those related to pancreatic surgery for ductal adenocarcinoma; 5-year survival is 100% for adenomas and border-line tumours, 80%-90% for in situ carcinomas and 50%-70% for invasive carcinomas.

One of the most surprising and most interesting characteristic of IPMNs is to be frequently associated with other neoplasms, both pancreatic and extrapancreatic. Some initial reports of this feature date back at least 20 years ago[18], and were later confirmed by numerous studies. Though it is not yet known the underlying mechanism, however this figure substantially has never been questioned ever since; it has been reported that patients affected by IPMN would be more likely to die of other neoplasm than from IPMN itself[17]; of course this entails major practical implications, particularly with respect to the diagnostic phase and to the follow-up.

The present article discusses the literature focused on this subject, and commits the clinical questions, to date unsolved, resulting therefrom.

MATERIALS AND METHODS

Search methods

A systematic search of literature was performed using MEDLINE, EMBASE, Cochrane and Web-of-Science libraries. No limitations for year of publication were considered. Search terms were: IPMN, IPMN-associated cancers, extrapancreatic neoplasm. MeSH terms were: Adenocarcinoma, Mucinous/diagnosis, Adenocarcinoma, Mucinous/surgery, Adenocarcinoma, Mucinous/prognosis, Pancreas/pathology, Pancreatic Neoplasms/diagnosis, Pancreatic Neoplasms/prognosis, Pancreatic Neoplasms/surgery.

Preference has been given to English publications, but abstracts from other languages were also included. All references in selected articles were further screened for additional publications. Articles were retrieved according to the Preferred Items for Reporting of Systematic Reviews and Meta-Analyses guidelines (Figure 1).

Figure 1.

Articles retrieval strategy, according to the Preferred Items for Reporting of Systematic Reviews and Meta-Analyses guidelines. IPMN: Intraductal papillary mucinous tumor.

Study selection

Articles were selected if their abstract showed an association between IPMN and pancreatic adenocarcinoma and/or extrapancreatic neoplasms before, after or at the same time of IPMN diagnosis. Both clinical series and Literature reviews were selected. Papers reporting genetic characteristic or proteomic assessment in IPMN, studies of risk factors for malignancy, congress presentations and letters to the editor were excluded.

Data extraction

For all eligible studies, a standard data extraction form was filled in and the following data were extracted: study design, number of patients, prevalence of pancreatic cancer (PC) and extrapancreatic malignancies (EPM) in IPMN patients and control groups, if available. For statistical analysis, it was considered significant P < 0.05, when published.

RESULTS

The titles and abstracts of a total of 805 articles were screened separately by 2 authors (Gian Luca Baiocchi, Sarah Molfino) for eligibility. Out of these, 780 articles were excluded, either for having abstract unavailable (n = 30), or for reporting risk factors for malignancy (n = 90), or being irrelevant to our topic (n = 355), or because they described other features such as genetic proteomic assessment of IPMN (n = 203); excluded congress presentations and letters to the editor were 102. This resulted in 25 articles, 8 of which were further excluded after full text examination, either for being narrative review (n = 1), or comments to other papers (n = 4), or only available in abstract form (n = 3).

The remaining 17 articles were fully analyzed and data were extracted as summarized in Table 1[19-35].

Table 1.

Studies reporting pancreatic cancer and extrapancreatic malignancies incidence in patients with intraductal papillary mucinous tumors

Ref.	Year	Study design	n (IPMN)	Prevalence EPM	Colon Ca	Gastric Ca	NET	Group 23	n (group 2)	Prevalence EPM - group 2 (%)	P vaule
Lubezky et al[19]	2012	Retrospective	82	20%	31%	NR	NR	Pancreatic ductal adenocarcinoma	150	6	0.002
Riall et al[20]	2007	Population-based	992	10%	3%	0.1%	NR	Other neoplasm1	18655	10	NS
Yoon et al[21]	2008	Retrospective	210	33.8%	7%	14%	NR	Non-IPMN cystic pancreatic neoplasm	175	12	< 0.001
Kamisawa et al[22]	2005	Retrospective	79	35%	9%	15%	NR
Ishida et al[23]	2008	Retrospective	61	24.6%	8%	10%	NR
Ishida et al[24]	2013	Case report	1	100%	0	0	100
Goh et al[25]	2006	Case report	3	100%	0	0	100
Eguchi et al[26]	2006	Retrospective	69	42%	12%	6%	NR	Other neoplasm2	301	12	< 0.001
Choi et al[27]	2006	Retrospective	61	39%	7%	13%	NR	Other pancreatic neoplasm	38	8	< 0.001
Calculli et al[28]	2010	Retrospective	142	14.1%			NR
Reid-Lombardo et al[29]	2010	Retrospective	471	52%	4%	0	NR
Oh et al[30]	2009	Retrospective	37	38%	8%	8%	NR
Baumgaertner et al[31]	2008	Case-control	178	17%	10%	NR	NR	Aged and gender-matched control	356	8.4	0.003
Tanno et al[32]	1999	Retrospective	42	48%	12%	10%	NR
Larghi et al[33]	2013	Multicentric cohort	390	23.6%	12.4%	NR	NR
Tewari et al[34]	2013	Case report	3	100%	0	0	100%
Uehara et al[35]	2008	Retrospective	60	54	NR	NR	NR

¹Data from SEER Tumor Registry;

²Data from Osaka Cancer Registry;

³Control-group, if present;

⁴Pancreatic adenocarcinoma. NR: Not reported; IPMN: Intraductal papillary mucinous tumor.

A total of 2881 patients were globally analyzed as study group, and the mean incidence of PC/EPM was 28.71%. When a control group was analyzed (6 papers), the same incidence was as low as 9.4%.

The majority of studies are retrospective (n = 11), while a case-control study and a multicentric-cohort were also included. Prevalence of PC/EPM in patients with IPMN was 14.1%-52%, with the exception of Uehara et al[35] (2008), reporting a prevalence of PC of only 5% (but this paper does not consider the extrapancreatic malignancies). Another paper reporting a somewhat lower incidence of EPM was published in 2007 by Riall and Coll.; in this study data of more than 18500 patients with pancreatic cancer were collected from the SEER Tumor Registry, and in this group the incidence of EPM was compared between patients with pancreatic adenocarcinoma and patients with IPMN. Thus, only malignant or invasive IPMNs were included.

Six studies reported a control group analysis; in all of them, a significantly higher prevalence of PC/EPM in IPMN patients than in control-group was reported; furthermore, in all the studies the incidence of EPM exceeded the expected rate of the same malignancies in the general population. When reported, in the majority of cases, the extrapancreatic neoplasm associated to IPMN patients were colorectal cancers (3%-12%).

DISCUSSION

Pancreatic IPMNs are observed with increasing frequency in asymptomatic patients[36], and still represent in the majority of cases a clinical dilemma. Identifyng the small subgroup pf malignant cases is of upmost importance: from one side, resecting all the IPMNs cases would represent an over-treatment in more than 75% of patients, but on the other side unrecognising malignancy would significantly worsen the prognosis of those cases. Thus, while in presence of clear malignant features the surgical indication is supported by good survival results, nevertheless a number of controversies remain for IPMNs at low risk of malignancy, relative to the diagnostic protocol to be adopted and to the surgical indication. Finally, it remains still not clear what kind of follow-up is indicated in both patients with resected IPMN and patients not submitted to surgery, as the likelihood of developing malignant recurrence and of progress into malignancy, respectively, is very low from the available series.

Comprehensive imaging, including ultrasonography (US) and contrast-enhanced US, CT scan, magnetic resonance cholangiopancreatography (MRCP) and endoscopic ultrasound, is highly sensitive and very low specific (down to about 20%[37]) in recognising malignancy, when the Sendai criteria, evidence-based treatment guidelines, published by a consortium of the International Association of Pancreatology in 2006 and 2012[38,39], are strictly applied. Endo-US fine-needle cytology significantly increases the specificity, but it is fairly invasive. ¹⁸FDG-PET was also investigated in a few published series, whit a reported accuracy rate of 85%-95%[40-44], and a specificity rate up to 100%[37]. However, Sendai guidelines do not take into account 18FDG-PET, which is besides very expensive and finally poorly employed in the majority of Centres.

When evaluating a rationale strategy for IPMNs staging and follow-up, an important and unique feature should be taken into consideration: it appears from the Literature that patients harbouring IPMNs are at higher risk of developing a second malignancy, both before, concurrent or after IPMN diagnosis[17,18].

An extensive literature search does not reveal a clear aetiopathogenetic explanation. The merely assumption that patients, once diagnosed to harbor IPMN, are subjected to instrumental examinations which therefore detect with greater frequency other malignancies, is not confirmed either by the fact that most of the associated tumors were diagnosed before the IPMN, either because the examination currently regarded as the gold standard for IPMNs, namely the MRCP, is an organ-specific and much less panoramic test compared to other investigations (including CT, which is less often used for the great number of necessary examinations and then the elevated radiation dose). The only speculative indications come from the analysis of risk factors for PC/EPC in IPMNs: female gender, white race, and above all advanced age [as early as 55-64 years, but with a peak to 75-84 years; for this age group the odds ratio for EPM, compared to the general population, is as high as 4.0 (95%CI: 2.4-6.7)]. Very few studies sought to analyze the molecular characteristics of IPMNs cases with PC/EPC; among them, Lee and Coll. found only correlation with the expression of the MUC2 gene and not of other genes, including p53[45]. As in other fields of oncology, the thorough study of patients with IPMN and associated malignancies would provide an interesting model for the study of carcinogenesis. Unfortunately, currently available data does not allow the precise identification of the subset of IPMN patients at greater risk of developing PC/EPM. Even in patients with small BDT, high incidences have been reported.

The most significant claim coming from the analysis of this data, to some extent surprising and still partially unexplained, concerns the clinical pathway. Indeed, the concept that at least one third of patients with IPMN develop in the course of their live PC/EPM cannot be overlooked. The first implication is related to the initial management of patients with newly diagnosed IPMN. There is currently a strong tendency to consider the BDT- IPMNs, without clinical and radiological signs which lead them in groups “high risk stigmata” and “worrisome features” according to Sendai guidelines, a disease of limited clinical relevance, for which no further investigation are warranted[46]. Although such a trend appears to be fully justified regarding the IPMN itself, what is described in the present paper leads to a reflection about the opportunity to submit these patients to instrumental examinations whit total-body value.

The second consideration concerns the IPMNs at high risk of malignancy (“high risk stigmata” according to Sendai): in these cases, since surgery is frequently performed, it might be smart to subject the patients to an even more comprehensive diagnostic preoperative work-up, including for example a colonoscopy and gastroscopy, in sight of a possible combined intervention.

The third consideration refers to the follow-up protocols: despite the low risk of malignant degeneration of resected IPMNs, and of not resected IPMNs without Sendai criteria, it appears at least careless to abandon these patients and end the follow-up, even though it is clearly expensive.

Of course, the points made above have consequences on the most widely used imaging modality in IPMNs: in addition to the MRCP, other panoramic examinations such as thorax-abdomen contrast medium enhanced CT and 18FDG-PET become relevant.

In conclusion, the literature appear unanimous in reporting that patients with IPMNs have a high risk, in the order of at least 30%, to develop PC and EPM; hence the opportunity to deepen, at the time of the first finding, and in any case preoperatively, the search for such tumors; and to maintain an high index of suspicion, and a constant follow-up, in all IPMNs patients, with total body examinations.

COMMENTS

Background

Intraductal papillary mucinous tumors (IPMNs) belong to the group of cystic neoplasms of the pancreas, and are characterized by a communication with the Wirsung duct and by the absence of ovarian-type stroma.

Research frontiers

One of the most surprising and most interesting characteristic of IPMNs is to be frequently associated with other neoplasms, both pancreatic and extrapancreatic. Some initial reports of this feature date back at least 20 years ago[18], and were later confirmed by numerous studies.

Innovations and breakthroughs

Comprehensive imaging, including ultrasonography (US) and contrast-enhanced US, computed tomography (CT) scan, magnetic resonance cholangiopancreatography (MRCP) and endoscopic ultrasound, is highly sensitive and very low specific (down to about 20%) in recognising malignancy, when the Sendai criteria, evidence-based treatment guidelines, published by a consortium of the International Association of Pancreatology in 2006 and 2012, are strictly applied.

Applications

The points made above have consequences on the most widely used imaging modality in IPMNs: in addition to the MRCP, other panoramic examinations such as thorax-abdomen contrast medium enhanced CT and ¹⁸FDG-PET become relevant.

Peer-review

This in an interesting review article presented by Baiocchi et al which focuses on intraductal papillary mucinous neoplasmas of the pancreas and their association with pancreatic and extrapancreatic malignancies.