Pancreatic Cysts

A 48-year-old, healthy male applied for $1M of life insurance. In 2013 he had undergone ultrasound examination for intermittent vague abdominal pain. This showed a 1.5 cm cyst in the uncinate process of the pancreas. Computed tomography of the abdomen confirmed a cyst measuring 0.9 × 0.9 × 1.5 cm communicating with the distal pancreatic duct. The radiologic impression was a branch duct intra-pancreatic mucinous neoplasm (IPMN). An endoscopic ultrasound was non-diagnostic. The abdominal pains resolved spontaneously. Observation was recommended. Over the subsequent 2 years, annual magnetic resonance imaging (MR) and endoscopic ultrasonography (EUS) demonstrated a minimal increase in size without any worrisome features. Continued observation with alternating 6-monthly abdominal MR and EUS was recommended.

Medical directors and underwriters encounter pancreatic cysts in 2 contexts: 1) a pancreatic cyst that has been detected by imaging for which there is no tissue diagnosis; 2) a pancreatic cyst for which a precise tissue diagnosis is available. The former is particularly challenging as the natural history and current management strategies for pancreatic cysts is quite complex.

DESCRIPTION OF PANCREATIC CYSTS

Pancreatic cysts are common. It is estimated that 2%-3% of abdominal computed tomography (CT) and 15% of abdominal magnetic resonance (MR) studies will detect a pancreatic cyst.¹ As the sensitivity of abdominal imaging continues to improve and as the number of imaging studies continues to climb, the absolute number of incidentally discovered cysts will increase.

There are 3 types of pancreatic cyst: inflammatory cysts (previously known as ‘pseudocysts’), non-neoplastic cysts and pancreatic cystic neoplasms (PCN). (See Figure 1.)

View Figure

• Download as Powerpoint
• Download Figure

Inflammatory cysts, as the name suggests, are complications of acute pancreatitis. As they develop in the context of acute pancreatitis their diagnosis is usually straightforward. Their mortality risk is that of the underlying process.

Non-neoplastic cysts are rare. They are usually asymptomatic and are most often diagnosed following surgical resection of lesions that were felt to be PCNs. A number of different histological types are recognized; all are benign.

Pancreatic cystic neoplasms (PCN) are relevant to risk selection on 2 counts: they constitute more than 50% of all pancreatic cysts and, in contrast to the other 2 cyst types, are capable of malignant transformation. That said, the vast majority are benign at time of diagnosis and will follow a benign course. The challenge, as omnipresent in the clinical as in the insurance world, is to identify those PCNs that are malignant or will undergo malignant transformation with time. The purpose of this article is to provide pointers that can help meet this challenge while also summarizing the ongoing debate about their optimal management.

Pancreatic Cystic Neoplasms

There are 4 types of PCN: serous cystadenoma, solid pseudopapillary neoplasm, mucinous cystic neoplasm and intraductal papillary mucinous neoplasm (IPMN). The latter 2 are mucin-producing PCNs. The epidemiologic and imaging characteristics of PCNs are summarized in Table 1. IPMNs are the most common (about 40%), followed by mucinous cystic neoplasms, then serous cystadenomas and lastly solid pseudopapillary neoplasms. However, these frequencies are derived from operated cohorts; as neither serous cystadenomas nor IPMNs are routinely resected it is likely that they comprise a higher percentage of the total.

Table 1. Characteristics of Pancreatic Cystic Neoplasms

View Fullscreen

Most PCNs are detected in the 5^th to 7^th decades; solid pseudopapillary neoplasms are the exception, presenting at much earlier ages. Men and women are equally involved; the exception here is mucinous neoplasms, which affect women exclusively. Each PCN type has a different propensity to become malignant: serous cystadenomas have minimal malignant potential whereas mucinous cystadenomas, solid pseudopapillary neoplasms and IPMNs have moderate malignant potential (see Table 1).

IPMN MALIGNANCY RISK

Main-duct IPMNs have a 70% risk of present or future malignancy.^2-8 In contrast, branch-duct IPMNs are usually benign with conversion to malignancy a relatively rare event. This justifies the ‘watch and see’ approach to branch-duct IPMNs that many centers have adopted. However, in one study the 5- and 10-year actuarial rates for pancreatic cancer were 2% and 20% respectively.⁹ Size is the most important determinant of malignancy for branch-duct IPMNs. Malignancy is rare in tumors <3 cm in size.¹⁰ The presence of symptoms, older age and/or solid components on imaging are additional determining factors. The presence of mural nodules and an increase in cyst size during follow-up imaging are further risk factors for malignancy.¹¹

DIAGNOSIS AND MANAGEMENT

Abdominal MR with cholangiopancreatography (MR/MRCP) or pancreatic CT are the preferred diagnostic techniques to sort out the different cyst types. If imaging suggests a serous cystadenoma no further follow up is required. If the findings are diagnostic or highly suggestive of mucinous cystic neoplasm, solid pseudopapillary neoplasm, main duct or mixed main-duct/branch-duct IPMN, resection is undertaken. Surgery is also indicated for any PCN causing symptoms, regardless of type.

If imaging is inconclusive further management will be dictated by 3 key imaging characteristics: 1) cyst >3 cm in diameter, 2) the presence of a solid component, and 3) dilatation of the main pancreatic duct.

A 2015 American Gastroenterological Association (AGA) systematic review of pancreatic cysts recommended that cysts with 2 or more of these adverse imaging characteristics be biopsied. Cysts that have fewer than 2 adverse imaging characteristics do not require further investigation and can be re-imaged in 1 year and then biennially until year 5. At that point imaging should be discontinued. Should any interval study show 2 or more adverse imaging features the cyst should be biopsied.¹²

This recommendation, whose main purpose is to reduce the number of unnecessary biopsies, has come under fire from some quarters.^13,14 Critics argue that it is not supported by data and suggest that any cyst >1 cm or the presence of any 1 of the 3 unfavorable imaging characteristics is sufficient to warrant biopsy.

This was demonstrated in a 2008 study¹⁷ where 4 out of 57 patients with a BD-IPMN <3 cm without a solid component had invasive carcinomas. A further study¹⁸ suggested that cyst size did not discriminate benign from malignant lesions.

In a similar vein, the critics argue that it is unwise to abandon follow-up after 5 years of imaging stability, as cysts may still become malignant at this stage.

Interestingly, both the AGA and its critics acknowledge that there is scant evidence to support either management path. The AGA points out that most information comes from referral centers and referral bias is skewing the data in favor of intervention; the critics counter-argue that possible over-intervention is the price to pay to prevent pancreatic cancer, a nearly-always fatal disease. Furthermore, they point out, clinical practice over the past 5 years attests to the success of the more conservative approach. While debate continues it is likely that some physicians, in spite of the AGA recommendations, will continue to recommend biopsy for cysts <3 cm and with fewer than 2 adverse imaging features.

Definitive pathologic diagnosis requires tissue acquisition via endoscopy-guided fine needle aspiration (EUS-FNA). This technique has a sensitivity of approximately 60% and a specificity of 90% to diagnose malignancy. Sensitivity of EUS-FNA is impaired by the difficulty of acquiring adequate cellular content to permit cytological analysis; in this instance tumor and molecular markers may provide supportive evidence. Proteomic profiling of cyst fluid may also play a role.

Carcinoembryonic antigen (CEA) is present in higher quantities in mucinous cysts. A cut value of 192 ng/ml confirms the diagnosis of a mucinous, (either a mucinous cystadenoma or IPMN) rather than a serous cyst. However, an elevated CEA is not reliably associated with malignant transformation, although there is growing consensus that a value >2500 ng/ml confirms a malignant cyst. An elevated CEA in combination with high levels of 2 other tumor antigens, CA19-9 and CA72-4, is an even more reliable marker of malignancy.

Molecular analysis of cyst fluid is of increasing interest. It has 2 purposes: 1) to differentiate between mucinous and non-mucinous cysts, and 2) to distinguish benign from malignant cysts. Analysis includes the detection of KRAS mutations, allelic loss and DNA concentration. These markers have excellent specificity but average to poor sensitivity for differentiating mucinous from non-mucinous lesions. When combined with cystic fluid CEA the sensitivity improves markedly. The utility of molecular markers to distinguish between benign and malignant cysts has not yet been established. A recent report suggested that a specific combination of molecular markers could accurately predict lesions requiring surgery.¹⁵

Proteomic profiling of cyst fluid allows different mucins to be identified. Each of the 4 different histologic types of IPMN produces identifiable combinations of mucin and can permit histologic identification with a high degree of accuracy. Proteomic profiling may eventually prove to be more accurate than either cytology or CEA for identifying lesions with malignant potential.¹⁶

In the future it is likely that various combinations of tumor and molecular markers will permit more accurate diagnoses in instances where tissue sampling is inadequate. In the meantime, the decision to resect a pancreatic cyst must often be individualized, based on the imaging characteristics, cyst fluid analysis and clinical judgement.

UNDERWRITING

For pancreatic cysts that have been resected, the pathologic diagnosis will guide the underwriting decision. Serous cystadenomas, mucinous cystic neoplasms, solid pseudopapillary neoplasms and benign IPMNs, once resected, have an excellent prognosis. Cysts that demonstrate any degree of dysplasia (these are almost exclusively IPMNs) will be approached more cautiously; many will be insurable after an appropriate delay and with careful follow-up. Pancreatic cysts that have not been diagnosed and are being followed with periodic imaging provide the more difficult dilemma. Available clinical data suggest that most of these small pancreatic cysts have limited malignant potential. This is particularly the case for cysts that are <3 cm, have no solid component, no pancreatic duct dilatation and do not develop any of these characteristics over a 1-3 year follow-up period.

BACK TO OUR CASE

The clinical diagnosis is a branch-duct IPMN. The treatment plan calls for an annual MR alternating with EUS, without any clearly stated temporal end-point. The ACG would not support such intensive surveillance. Rather, they would recommend against endoscopic ultrasound/biopsy and suggest an MRI in 1 year, and then 3 and 5 years later. This recommendation would be justified by the small cyst size, the lack of duct dilatation, and the absence of any complex cyst features. These differing approaches reflect the wide variety of clinical practices that are currently in vogue and underscore the prevalent concern of preventing pancreatic cancer. However, there is now consensus that many small pancreatic cysts do not present a future cancer risk. Many of these will be insurable. Nevertheless, one might argue that in cases similar to the one presented a substandard rating could adequately address the modest risk of future malignant transformation.

CONCLUSION

• 1. Pancreatic cysts are more common than before, largely as a result of widespread abdominal imaging. Although the vast majority are benign, they pose an underwriting challenge as some may be malignant at time of presentation; others may be benign but progress to carcinoma.

• 2. Amongst all pancreatic cysts, only pancreatic cystic neoplasms (PCN) progress to carcinoma.

• 3. Intraductal papillary mucinous neoplasms (IPMN) are the most common type of PCN.

• 4. There is controversy about the correct management of incidentally discovered IPMN; this reflects the lack of high-quality scientific evidence to support any particular management approach.

• 5. For IPNMs with favorable characteristics, observation rather than surgery has become the accepted practice.