Pancreatic Cyst (Incidental)

• Incidental pancreatic cysts may be documented on CT, ultrasound or MRI performed for symptoms unrelated to the pancreas.

• Identification of increasing numbers of IPCLs in asymptomatic patients has become a clinical problem and is likely due to the expanding use of cross section imaging and the increased sensitivity of modern scanners, rather than an increase in incidence - supported by the apparent lack of correlation between increasing pancreatic cyst detection and pancreatic cancer incidence 1.

• The true prevalence of IPCLs is contentious, with a wide range (0.47%-44.7%) being reported in different series 2, but a study of patients undergoing MRI found a prevalence of 13.5%, increasing with age 3. However, a population-based study of >1000 participants 4 using MRI found a prevalence of about 49% (similar to autopsy studies) and a weighted incidence of cysts >2mm of 2.6%. Numbers and sizes of cysts increased with age.

• Only 6% of cysts were >1cm 4. The authors, in a mean follow-up period of 5.9 years, reported no participant dying of pancreatic disease and concluded that a restrictive follow-up approach to IPCLs can be supported. However, it should be noted that none of the detected cysts in this study was greater than 3cm.

• The development of sub-centimetre cysts may be part of the aging process.

• A number of IPCLs are potentially malignant (see later), although the overall risk of malignancy of IPCLs is low. However, the morbidity of pancreatic surgery is relatively high. Therefore, their increasing detection as asymptomatic incidental findings poses a diagnostic and management dilemma with regard to whether or not the patient should be reassured or subjected to a surveillance strategy or whether surgery is indicated. Pancreatic surgery carries a risk of morbidity (approximately 20%–40%) 5.

• The vast majority of IPCLs are small 3, and the majority of pancreatic cystic neoplasms are likely to be branch duct intraductal papillary mucinous neoplasms (BD-IPMNs) 6, 7, although firm pathologic proof that this is the case is lacking. As long as BD-IPMNs are limited to the branches of the pancreatic ductal system, the overall risk of these cysts harbouring high-grade dysplasia (HGD) or invasive cancer is low, although in one review HGD or cancer was present in 11-30% of resected BD-IPMN 8, compared to the mean frequency of 62% in patients undergoing resection with main pancreatic duct involvement.

• Guidelines regarding surveillance and management have been developed by various specialist professional bodies.

• These guidelines include:

  • The European Study Group on Cystic Tumours of the Pancreas 9
  • The American College of Gastroenterology (ACG) 10
  • Canadian Association of Radiologists 11
  • The American College of Radiology (ACR) 12
  • The IAP/Fukuoka guidelines (which deal only with Intraductal Papillary Mucinous Neoplasms (IPMN) 13. This was originally the Sendai guideline.)
  • The American Gastroenterological Association (AGA) 6

• Several of the above guidelines have been recently summarised and compared in useful reviews 14-17.

• In general, the European and IAP/Fukuoka guidelines provide advice on managing subtypes of PCLs – in the case of the IAP/Fukuoka guidelines the focus is on IPMNs- and the ACG and AGA guidelines tend to deal with management and surveillance in a broader way. The ACR guidelines provide advice on technique of CT and MRI and the reporting of findings 16.

• In one study 18 of 269 patients, the sensitivity and specificity of the AGA, the IAP/Fukuoka and the ACR guidelines for predicting advanced neoplasia were compared. The results were, respectively, 7% and 88% (AGA), 73% and 46% (IAP/Fukuoka) and 54% and 61% (ACR).

• However, importantly:

  • None of these guidelines have been evaluated or validated in a population-based setting.
  • The guidelines vary in scope, their foci, and their recommendations for resection and surveillance/follow-up 14-16.
  • They are based on retrospective studies and do not account or control for most tumour- or patient-specific factors 16.
  • When graded, the level of evidence is defined as low or very low quality and recommendations, for the most part, are conditional.

The remainder of this DIP pathway attempts to collate and incorporate components of these existing guidelines.

A simplified classification of pancreatic cysts is as follows: 19

1. Simple benign cysts:
   1. Epithelial
   2. Non-neoplastic mucinous cysts
   3. Retention cysts
   4. Congenital cysts
2. Inflammatory:

   Pseudocysts

3. Cystic neoplasms:
   • Mucinous

     Types of Mucinous cystic neoplasms:

     These comprise Intraductal Papillary Mucinous Neoplasms (main duct, branch duct and mixed) , and Mucinous Cystic Adenomas/carcinomas

     • Intraductal Papillary Mucinous Cystic Neoplasms (IPMNs) arising within pancreatic ducts. These tend to occur in the 5^th - 7^th decade and are of equal gender distribution. They may be solitary (most frequently in the head of the pancreas) or multifocal. They are the most common

       Neoplastic PCLs in clinical practice and may be of the following varieties.

       • Main duct (MD-IPMN), where there is involvement and dilatation of the main pancreatic duct alone. MD- IPMNs have the highest malignant potential (approximately 60%-80%) and are characterized by diffuse or segmental dilatation of the main pancreatic duct to > 5 mm from a cystic tumour producing mucus within the duct 13. Surgical resection is usually recommended for surgically fit patients with MPD > 10 mm, jaundice, or mural nodules 13.
       • Branch-duct (BD-IPMN) associated with a cyst but without involvement/dilatation of the main duct. BD-IPMNs arise within the branches of a nondilated main pancreatic duct with malignant risk ranging from 3% to 26% 5, 13. Many small uncharacterizable IPCLs may be BD-IPMNs.

       • Mixed type with one or more cysts and dilatation of the main duct. Mixed type IPMNs have features of both MD- IPMN and BD-IPMN with approximately 20% to 30% of BD-IPMN ultimately proven to be mixed type IPMN following surgical resection 20.

         There is a spectrum of dysplastic changes - benign to invasive cancer. IPMNs are macroscopic precursors of Pancreatic Ductal Adenocarcinoma, progressing from adenomas to invasive carcinomas (similar to the adenoma-carcinoma sequence of colon cancer) 2. MD-IPMN progress more frequently than BD-IPMN.

         There is an association between IPMN and diabetes mellitus (DM) - 10-45% of patients with IPMN have DM- as well as an increase in chronic pancreatitis.8

     • Mucinous Cystic Adenomas/carcinomas (MCNs), arising outside of the pancreatic ducts (Mucinous Cystic Adenomas -MCA) and not communicating with ducts. They have malignant potential; MCAs can progress to mucinous cystadenocarcinoma. Almost all (90-95%) occur in female patients. They are unilocular or septated/macrocystic. They typically arise in the body or tail of the pancreas.
     • Rarely cystic degeneration of Pancreatic Ductal Adenocarcinoma (PDAC)
   • Non-mucinous:
     • Serous Cystadenoma (SCA) - extremely low risk of malignancy. More common in females. They are typically microcystic (honeycomb) +/- central calcification in appearance on imaging. Variants include macrocystic/oligocystic, mixed micro/macrocystic and “solid”.
     • Solid Pseudopapillary Neoplasms (SPNs)- these are indolent low-grade malignant neoplasms occurring in young adults - 90% females- and are more commonly symptomatic than incidentally discovered. There is up to a 15% incidence of local invasion or metastatic disease 21. In the absence of metastases prognosis is excellent following resection 1, 9.
     • Cystic Pancreatic Neuroendocrine Tumours (PNETs) - benign or malignant
4. Other rarities:

   Lymphoepithelial cysts, congenital cysts

ACR guidelines recommend standardised reporting for cross-sectional imaging of pancreatic cystic lesions using the following criteria 12:

• Cyst morphology
• Location
• Size
• Multiplicity
• Possible communication with pancreatic duct
• High-risk stigmata or worrisome features
• Growth on follow-up imaging

There is limited ability for cross-sectional imaging to fully characterise IPCLs as the size of detectable lesions is decreasing 2. A correct pre-operative diagnosis varies in several series, from 22-86%. Cross-sectional imaging in combination with EUS +/- cyst aspiration +/- biopsy may yield characteristic features that allow a specific diagnosis. However, particularly in cysts <3cm, specific features may be absent 2. Surveillance programs, as recommended by the guidelines listed below, are necessary to detect the significant proportion of IPCLs that progress to malignancy.

Imaging modalities available for the investigation of IPCLs and included in the various guidelines (and in the current document) include:

• MDCT and MRI/MRCP have comparable, albeit somewhat disappointing 5 (see below), accuracy for characterising pancreatic cystic lesions and estimating their aggressiveness. However, MRI can improve reader confidence in characterizing cystic pancreatic lesions and more readily determine the relationship of the lesion with the pancreatic ducts 22-24 and is the preferred imaging modality. Secretin-enhanced MRCP improves the ability to demonstrate whether a cyst communicates with the pancreatic duct 25
• MRI/MRCP technique – MRI/MRCP
• Multidetector CT – CT Pancreas Protocol If MRI Unavailable
• Endoscopic ultrasound and adjunctive techniques such as Contrast-enhanced EUS, EUS-guided cyst fluid analysis and tissue acquisition, and Confocal Laser Endomicroscopy - EUS +/- FNA

Note regarding PET/CT

• A 2015 study of a relatively small number of patients found that the diagnostic accuracy for distinguishing malignant from benign Pancreatic Cystic Lesions was 94% for (18)F-FDG PET/CT, compared with 77% and 87% for MDCT (P < 0.05) and MR imaging, respectively 26.
• A further study reported the sensitivity of PET/CT in diagnosing malignancy was 85.7% to 100% with a reported accuracy of 88% to 95% 27.
• A network meta-analysis published in 2022 reported that 18 F-FDG PET is the best preoperative imaging modality for differentiating malignant from benign PCLs 28.

However, the ACG guidelines 10 (2018) state “There is insufficient evidence to support the use of PET-CT currently”.

The European Study Group guidelines (2018) 9, the American Gastroenterological Association guidelines (2015) 6, the Revisions of the International consensus Fukuoka guidelines (2017) 13, the ACR Appropriateness criteria (2017) 12 and the Canadian Association of Radiologists Incidental Findings Working Group (2022) 11 report do not mention PET/CT.

Dedicated MRI with MRCP has been recommended as the imaging procedure of choice to characterise a pancreatic cyst and for surveillance scanning, due to lack of ionising radiation (in contrast to CT), excellent interobserver agreement and superior contrast resolution to facilitate recognition of concerning/high risk cyst morphology and identification of ductal communication of the cyst (as in IPMN) 6, 9, 10, 13, 15.

In IPCLs, the aims of cross-sectional imaging (preferably by MRI/MRCP, or by CT if MRI is unavailable) are to:

• Determine the specific nature of the PCL, if possible, by identifying those lesions with characteristic features.
  • In a 2018 systematic review the accuracy of CT scan for reaching a specific diagnosis was 39% to 61.4%, whereas MRI showed a better accuracy in identifying a specific diagnosis of 50% to 86% 27
  • MRI (or CT) may be sufficient for diagnosis when pathognomonic features of certain cyst types are present (for example those of a serous cystadenoma) 29, 30
  • In one study, detection of debris within a cyst was reported as a 95% specific sign of pseudocyst 31
• Determine the degree of risk of malignancy, either current or potential.

  This is undertaken by:

  • Distinguishing mucinous (which have a risk of malignant change) from non-mucinous cysts (the latter of which, with some specific exceptions such as SPN and PNET, do not have malignant potential).
  • By detecting debris within the cyst, MRI is useful in differentiating pseudocysts from neoplastic cysts. This is a 95% specific sign 31.
  • There is a reported 88% accuracy, 91% sensitivity and 78% specificity in differentiating mucinous from non-mucinous cysts. The combination of MRI and EUS improves sensitivity to 100% 32.
• Defining concerning/high-risk features versus benign features that indicate the need for further investigation by EUS/EUS-FNA or for surgery.
  • In a 2021 review and meta-analysis, MRI and CT had similar sensitivity (76% and 72%, respectively) and specificity (80% and 74%, respectively) in distinguishing benign and malignant pancreatic cystic lesions 33.
  • In this same article, MRI was slightly more specific than EUS.
  • Ability to determine malignant from benign cystic lesions varies (accuracy 73.2-91%, sensitivity 50-94.3% and specificity 58.3-88.9%) and is lower in prospective studies limited by small sample size, but is improved by combination with EUS 32, 34-36.
  • A study 37 showed MRI and EUS to be equivalent in predicting malignancy (approximately 75% accuracy) and suggested that EUS be reserved for those patients likely to require FNA.
  • In the same 2018 review as above, 27, MRI was 56-87% accurate in distinguishing benign from malignant lesions, compared to CT (62-80%).
  • In the same 2018 systematic review, the use of magnetic resonance imaging was superior to CT scan in identifying septations, mural nodules, and ductal communication 27.
  • Other studies have shown that MRI is highly sensitive in determining morphological features such as septa (93.3-94.4%), mural nodules (58.3-87.5%), main pancreatic duct dilatation (85.7-92.9%) and superior to MDCT for smaller cysts 22, 34, 38.
  • High risk stigmata for IPMN includes enhancing mural nodule>5mm which is 87% specific 39, in keeping with the 2017 revised International Consensus Fukuoka guidelines 13
  • A 2016 study by Kang et al found MDCT and MRI/MRCP to be similar in performance in determining the malignant potential of IPMN. The study included branch duct and main duct IPMNs 40.
• Identifying whether the cyst communicates with the pancreatic duct system (i.e. diagnosing likely IPMN).
  • Importance is attached to differentiating BD-IPMN from other cystic lesions. Identifying communication with the main PD is the first step in this process 41.
  • MRI has superior ability to detect duct communication (93.1% accuracy, 91.4-100% sensitivity and 89.7% specificity) thus aiding in distinguishing BD-IPMNs from mucinous cystic neoplasms (MCNs). The latter have higher malignant potential 34, 37, 42.
  • MRI is highly accurate in preoperative characterisation of IPMNs: reportedly 93.2-99.5% accurate, 96.8% sensitive and 90.8% specific for differentiating IPMN from other cystic lesions 42.
  • A 1998 article reported MRCP to be superior to ERCP in depicting cystic dilated ductal branches 43
  • Secretin-enhanced MRCP improves the ability to demonstrate whether a cyst communicates with the pancreatic duct 25.
• Individualising the approach to each patient with an IPCL (towards reassurance, further investigation and/or surveillance or surgery) with regard to age, clinical status and comorbidities, risk of malignancy, and personal preferences.

Notes on MRI technique 12, 24

• It is recommended that MRI examination of IPCLs should routinely include MRCP.
• MRI/MRCP can be performed at 1.5T or 3T, usually after a period of fasting. An oral contrast agent is optional but is advisable - negative T2 oral contrast agent is preferred.
• Recommended sequences include: Axial T1 in-phase and opposed-phase gradient recalled echo (GRE) sequences, axial and coronal T2 single-shot fast spin echo breath-hold sequences, heavily T2-weighted 2D and/or 3D MRCP, and dynamic 3D fat-suppressed T1-weighted spoiled GRE axial sequences before and after administration of intravenous gadolinium contrast. Diffusion weighted imaging is also useful.
• MRCP sequences are obtained using heavily T2-weighted sequences performed in 2D and 3D volumetric acquisitions.
• Note that nomenclature and details of MRI sequences will differ among different scanner vendors.

Secretin-enhanced MRCP improves the ability to demonstrate whether a cyst communicates with the pancreatic duct 25.

Diffusion Weighted MRI (DWI)

• May help in distinguishing neoplastic cysts from simple cysts and pseudocysts 44 but results in distinguishing mucinous from non-mucinous cysts have been disappointing 45, 46.

Limitations of MRI

• Like all preoperative diagnostic techniques, MRI performs poorly in the specific diagnosis of lesions compared to resection histopathology (55.6-76.2%) 35, 36
• One study has shown only fair inter-observer agreement among radiologists in classifying pancreatic cysts by MRI 47.

CT technique 12, 24

Although most guidelines recommend MRI/MRCP as the preferred imaging, CT may be useful if MRI is contra-indicated or unavailable, but also in special circumstances where, for example, the identification of calcification is important (such as pseudocysts in chronic pancreatitis) 9.

• 16-slice or greater multiple detector CT scanners.
• Submillimeter slices with isotropic voxels with allowance for reformatted thicker slices (3-5 mm).
• A minimum dual phase (pancreatic parenchymal and portal venous) scan is recommended.
• However, a pre-contrast phase is useful to assess for subtle calcification and to detect the degree of any enhancement.
• A delayed post-contrast phase may detect small isodense solid masses. The acquisition of these extra phase scans must be balanced against the extra radiation dose, especially in younger patients.
• Pancreatic-phase (late arterial) images beginning about 40-50 seconds after start of injection of intravenous contrast medium. Timing will depend on scanner speed.
• Contrast (concentration 300-350 mgI/ml of iodine) injection rate of 4-5 mL/s.
• A second phase is recommended at approximately 70-80 seconds to evaluate the liver.

• The overall risk that a pancreatic cyst is malignant at the time of diagnosis is a maximum of 0.01%. If it is assumed that only those cysts >2 cm have malignant potential, then the risk that a cyst >2 cm is malignant is a maximum of 0.21%. 5
• Similarly, the transformation into mucin-producing adenocarcinoma in patients with pancreatic cysts is thought to be a very rare event 48.
• A recent publication suggested that most low-risk and indeterminate-risk lesions do not increase in size during a minimum follow-up period of 2 years 49. Another study showed that asymptomatic cysts of <1.5 cm in diameter can safely be followed-up by imaging and are expected to undergo little change during a median follow-up of 5 years 50.
• A population-based study using MRI in a mean follow-up period of 5.9 years reported no participant dying of pancreatic disease and concluded that a restrictive follow-up approach to incidental PCLs can be supported. However, importantly, none of the detected cysts in this study was greater than 3cm. 4
• Serous cystadenomas (SCAs) are benign neoplasms and, essentially, can be regarded as having negligible potential risk of malignancy. There are rare cases of aggressive SCAs reported 51.

Guidelines and risk assessment

• Existing guidelines from multiple sources attempt to aid risk stratification and, dependent on cyst features (size and other features), recommend further evaluation, follow-up (frequency and duration) 14 or resection.
• “Concerning” (or “Worrisome”) features, and “High-risk” stigmata are defined by the various guidelines to determine which lesions require further investigation and biopsy, or surgery, or more intense monitoring/surveillance. Indeterminate Cyst Without Worrisome/High risk Features for further details.
• Features identified of relevance in most guidelines include:
  • Cyst size
  • Solid components (including enhancement with contrast)
  • Mural nodules (including enhancement with contrast)
  • Thickened cyst walls
  • Calibre of the main pancreatic duct
  • Regional lymphadenopathy
  • Presence/development of symptoms such as jaundice
• Structured reporting templates for CT and MRI have been devised for reporting cystic (and solid) pancreatic lesions 52.
• The risk stratification in the AGA guidelines 6 depends on the presence / number of high risk stigmata, which have been narrowed down to three in number (size >3 cm, dilated pancreatic duct, solid component on MRI), with only those with ≥2 stigmata or those in whom management would change by further characterization, proceeding to Endoscopic Ultrasound (EUS).
• Since the estimated risk of any PCL being malignant at the time of diagnosis, according to the AGA Technical Review authors, is not greater than 1 in 10,000 5 (contentious, but very low prevalence), an increased odds ratio as indicated by the three criteria (above) still indicates malignancy to be a rare event in an incidental pancreatic cyst.
• However, guidelines differ as to whether at least two risk features must be present on cross-sectional imaging (as per the AGA guideline -see above) or one to lead to EUS-FNA.
• In particular, there is variation among the published guidelines regarding the follow-up intervals and duration of follow-up of incidental PCLs without “worrisome” or concerning features. In most guidelines, these timings will depend on cyst size. The recommendations from the different authorities have been summarised by Hasan et al 14 and by Navarro et al. 53. Given this variation and the lack of hard evidence for any one interval over another, the clinical context (e.g. age, co-morbidities, preferences) of the patient should be taken into consideration.
• Given the lack of clear, convincing data in the various guidelines, management of patients with pancreatic cysts should be based on expert guidelines but individualized, incorporating the clinical judgment of an experienced multidisciplinary team 5.

Endoscopic Ultrasonography (EUS)

Advantages

• EUS is an adjunct to other imaging modalities to aid:
  • Differentiating mucinous from non-mucinous PCL
  • Differentiating benign from malignant where CT and/or MRI are unclear
  • Detecting communication of a cyst with the pancreatic duct as well as mural nodules/solid components 54
  • Helping to identify features of pancreatic cysts that indicate consideration for surgical resection and is recommended if the PCL has either clinical or radiological features of concern
• EUS is the optimal method for targeted sampling of cystic fluid for cytology and fluid biochemical analysis, as well as sampling the cyst wall and other solid components (EUS-FNA) 55, 56
• Contrast-enhanced EUS is reportedly the best imaging tool to assess vascularity of mural nodules 57, 58 and solid components and septations of IPCLs and thus characterising the nodule(s) - for example by distinguishing mucin aggregations from solid nodules. The presence of solid components may indicate the need for FNA 9, 57-60. Lisotti et al., in a meta-analysis comprising over 500 patients, showed that CE-EUS has high sensitivity (88%) and specificity (79%) in detecting mural nodules within PCLs 58. Interobserver agreement is moderate to excellent 59, 61.

Limitations

• EUS alone (based on cyst morphology) is only modestly accurate (48-75%) in differentiating mucinous and non-mucinous lesions 32, 54, 62 and cannot reliably differentiate benign and malignant cystic lesions after a non-diagnostic cross-sectional modality.
• EUS (as for CT and MRI) is NOT consistently reliable in identifying the exact nature of pancreatic cysts.
• There is conflicting evidence regarding EUS accuracy in distinguishing benign from malignant lesions 9, 34, 63-65
• A 2014 study from Korea describes a scoring system based on EUS features that led to high specificity but only moderate sensitivity in predicting malignancy in BD-IPMNs 66
• EUS is subject to significant interobserver variation 67, 68

EUS-FNA

• In view of the only modest utility of EUS for depicting the morphology of PCLs over other cross-sectional imaging techniques, it has been suggested that EUS be reserved for those patients likely to require FNA i.e. EUS-FNA 37.
• EUS–FNA should only be performed when the results are expected to change clinical management.
• EUS-FNA is NOT indicated when the diagnosis is established by cross-sectional imaging or where surgery is already clearly indicated.
• The ACR Appropriateness criteria do NOT recommend EUS-FNA for cysts <2.5cm 69
• EUS-FNA is indicated when typical features of a specific diagnosis are absent but there are concerning morphological features such as size >3cm, associated solid component and/or mural nodule(s) – especially when enhancing - dilated MPD, rapid cyst growth or raised Ca 19-9. Guidelines differ as to whether at least two risk features must be present (as per the AGA guideline) to indicate EUS-FNA or one feature.
• The controversial AGA guideline increased the threshold for sending a patient to EUS-FNA as well as surgery from one to at least two risk features. While this may be expected to decrease the unnecessary resection of benign cysts, the impact on the negative predictive value has not been proven 70, 71.
• EUS-FNA is still a matter of debate 32, 72. However, a prospective study found that EUS-FNA had a management impact in 72% of patients with incidental pancreatic cysts 72
• However, the ACR White Paper on Incidental Pancreatic Cysts notes that some centres routinely perform EUS-FNA on all cysts >2.5cm whether or not concerning features are present 12.
• EUS-FNA may also be indicated in some circumstances when a non-mucinous cyst is suspected (other than when there are typical features of SCA).
• EUS-FNA has a low risk of complications (2%) 73. A 2011 systematic review reported a 0.98% overall complication rate 74
• Routine antibiotic cover for EUS-FNA is not required 75
• EUS-FNA for cytology is very specific (83-100%) but not very sensitive (27-48%). However, combining morphology, cytology and cyst fluid analysis provide greater accuracy 54, 62

Cytology should be combined with cyst fluid analysis. A combination of cyst cytology and cyst fluid analysis for the contents detailed below can increase cyst classification. For example cytology + CEA content increases cyst classification accuracy to 70% 16.

EUS-FNA for cytology and cyst fluid analysis

Aspirated fluid analysis should include assessment of:

• Cytology
  • Two large meta-analyses confirmed similar sensitivity (54% to 63%) and specificity (88% to 93%) for cytology in differentiating mucinous from non-mucinous cystic lesions 76, 77
  • The main limitation of cytology is that often insufficient material is obtained for diagnosis 78
  • An additional limitation is that cyst fluid obtained during EUS–FNA is often acellular.
  • Cytology should be combined with cyst fluid analysis for mucin, CEA and amylase levels and glucose levels, as well as DNA sequencing analysis where available (see below).
• Cytology in mucinous versus non-mucinous
  • A recent retrospective study found that the combination of CEA analysis, cytology and viscosity (i.e. mucin content) of pancreatic fluid obtained by EUS-FNA increases its utility in the diagnosis of mucinous pancreatic cystic lesions, with it being greater than 90% 79
• Cytology in malignant versus benign
  • Cytology is very specific (91%) but not very sensitive (65%) in detecting malignancy 80
  • However, combining morphology, cytology and cyst fluid analysis (see below) provide greater accuracy 2
• Mucin
  • The presence of mucin aids the diagnosis of MCN and IPMN.
  • This can be assessed macroscopically by simple assessment of fluid viscosity - the so-called string sign 81 - or microscopically by subjecting the fluid to mucin-stains (periodic acid-Schiff stain).
• Amylase
  • The presence in the cyst fluid of amylase >1000ng/ml is highly suggestive of a pancreatitis-related pseudocyst.
  • Concentration below 250ng/ml essentially excludes this diagnosis 2.
• Carcino-Embryonic Antigen (CEA)
  • The presence of CEA, especially at a level greater than 192ng/ml correlates strongly with MCN and IPMN (75% sensitivity; 84% specificity) 54.
  • A systematic review found a level of >800ng/ml was 98% specific and 48% sensitive for the diagnosis of a mucinous cyst 82.
  • CEA levels are not useful in predicting malignancy 5.
  • Conversely a level of CEA <5ng/ml makes a pseudocyst, neuroendocrine tumour or SCA highly likely (Sensitivity 50%, Specificity 95%) 82.
  • Combining the presence of multiple biomarkers in cyst fluid aids in differentiating benign vs malignant mucinous lesions 15
  • A minimum volume of cyst fluid of 0.5-1 ml is required for CEA analysis 83
• Glucose content
  • Intra-cystic glucose <50mg/dL suggests mucinous cyst or pseudocyst 83
  • If only a small volume of cyst fluid is available, a meta-analysis 84 suggests that glucose levels (<50mg/dl) are more accurate than CEA levels (>192ng/ml) in identifying mucinous (versus non-mucinous) cysts.
• Other
  • DNA-based tests. Molecular fluid analysis with DNA-based markers has shown promise in differentiation of mucinous and non-mucinous cysts, distinguishing MCNs and IPMNs and detecting grades of neoplasia 83.

  • Cyst wall biopsy

    • EUS-guided fine needle biopsy (EUS-FNB) by the technique of introduction of microforceps through the EUS-FNA needle to sample the cyst wall or solid components of the cyst may provide a superior yield compared to cytology 55, 85, 86 with the possibility of pathological examination of specimens.
    • A systematic review showed an improvement of diagnostic yield compared to cytology 87. However, the complication rate is significantly higher than EUS-FNA 85, 88-91, particularly in suspected IPMN, and therefore the risk of this procedure must be weighed against the use of other diagnostic techniques and the fact that for most patients a pancreatic cyst is benign.
    • The value of adding this technique to EUS-FNA and confocal laser endoscopy remains to be resolved 92

  • Confocal Laser Endomicroscopy

    • This technique utilises a needle probe during EUS-FNA or FNB to obtain images of the inner wall of the cyst. Three patterns can be defined 93.
    • The technique is associated with a significant learning curve but is highly accurate 93, 94.
    • However, there is a significant risk of adverse events, acute pancreatitis being the most common.

1. “Concerning” (or “worrisome”) features and “High-risk” stigmata

• “Concerning” or “worrisome” features of pancreatic cysts are used to determine which patients should proceed to EUS/FNA after CT or MRI or alternatively indicate the requirement for regular surveillance.

• “High-risk” stigmata are usually applied in the context of suspected IPMNs and should lead to consideration of surgery in patients of appropriate age and fitness.

• It is important to note that there are variations among the published society/international guidelines 16 as to what are “concerning” or “worrisome” morphological features of PCLs.

• However, there appears to be general agreement that cyst size/growth, the presence of mural nodules or solid components and dilatation of the main pancreatic duct (MPD) are criteria that are important in assigning risk of high-grade dysplasia (HGD) or malignancy.

• Worrisome features also include certain clinical (e.g. jaundice or acute pancreatitis attributed to the cyst; raised CA 19-9) 10

• Different international guidelines recommend EUS/FNA after 2 or more concerning features (AGA) or at least one concerning feature (ACG, ACR, European Study Group) 95

  • AGA 6: 2 or more of:

    • >3cm

      Dilated MPD >5mm

      Solid component

    • ACG 10: one or more of:

      Mucin-producing cyst >3cm

      Presence of a mural nodule or solid component

      Dilated MPD >5mm

      Focal dilatation of MPD

      Abrupt change in MPD calibre with distal pancreatic atrophy

    • The European Guidelines 9:

      4cm or mural nodule

  • There is little evidence that there is any difference in these approaches in the NPV for malignancy 95.

  The AGA technical review identified the following as the greatest risk factors for malignancy in incidental pancreatic cysts:

  • • solid component [odds ratio (OR) 7.7],
    • cyst size >3 cm (OR 3), and
    • dilated main pancreatic duct (OR 2.4) 5.
    • The studies on which the AGA recommendations were made used different definitions of MPD dilatation.

    • Studies have suggested that the degree of MPD dilation correlates with varying malignant potential with higher risk associated with duct >7-8 mm 95

  • MCNs <3cm in size with no concerning features (see below) have a low risk of high-grade dysplasia or malignancy 10 96, 97.

  2. “Worrisome” features in suspected BD-IPMN 13

  The features below (previously the Sendai criteria) derive from the International Association of Pancreatology Fukuoka Guidelines consensus group 13.

  • >3 cm

    • A study of 563 resected and radiologically diagnosed BD-IPMN noted that 18% of cysts >3 cm had high-grade dysplasia or invasive cancer, while no malignancy was detected in cysts <2 cm and no high-grade dysplasia was noted in lesions <1 cm 98

  • enhancing mural nodule <5 mm
  • thickened enhanced cyst walls
  • MPD size of 5-9 mm

    • Studies have suggested that the degree of main pancreatic duct dilation correlates with varying malignant potential with higher risk associated with duct >7-8 mm 95

  • Abrupt MPD calibre change and upstream atrophy
  • lymphadenopathy
  • an elevated serum level of carbohydrate antigen (CA)19-9
  • rapid rate of cyst growth >5 mm/2 year

  3. “High-risk” stigmata in suspected BD-IPMN 9, 13

  These high-risk stigmata are indications for resection in surgically fit patients.

  • Obstructive jaundice with lesion in pancreatic head
  • Enhancing mural nodule >5mm / solid component

    • A mural nodule has been shown to be an independent predictor of both invasive cancer and high-grade dysplasia (HGD) for all type of IPMNs (with the exception of HGD in BD-IPMNs), with size of the enhancing mural nodule having a significant effect on predicting advanced neoplasia 57

    • Contrast-enhanced ultrasound (if available) would appear to be the optimal method of examining for mural nodules 57

  • MPD >10mm
  • Positive cytology for high grade dysplasia or cancer on cyst aspiration

    Shin et al 99 devised a simple scoring system to predict the invasiveness /malignancy of IPMN using 5 factors (age >60, history of pancreatitis, mural nodules, elevated CA19-9, MPD >6mm). 3 or more of these features predicted malignancy with a sensitivity of 50.7% and specificity of 90%. The authors did not distinguish between BD-IPMN and MD-IPMN.

    4. Main duct IPMN

    • The “worrisome” features and “High risk stigmata” are essentially as for BD-IPMN.

    • Surgical resection is recommended for all surgically fit patients with MPD >10 mm, jaundice, or mural nodules 13

IPCL surveillance

• All pancreatic cyst guidelines agree that patients who are not surgically fit do not need ongoing surveillance
  • A study of 725 patients with IPMNs noted that patients with Charlson comorbidity index ≥ 7 were 11 times more likely to die from non-IPMN related causes within 3 years of cyst diagnosis 100.
  • Another study used the Adult Comorbidity Evaluation 27 scoring system in 793 patients with IPMNs followed for over a year with similar conclusions that patients with higher scores were more likely to die from non-IPMN related causes 101
• There is variation among the published guidelines regarding the follow-up intervals and duration of follow-up of incidental PCLs without “worrisome” or concerning features. In most guidelines, these timings will depend on cyst size.
• The recommendations from the different authorities have been summarised by Hasan et al, 2019 14 and Mohapatra (2023) 17. Given this variation and the lack of hard evidence for any one interval over another, the clinical context (e.g. age, co-morbidities, preferences) of the patient should be taken into consideration.
• The AGA guideline 6 published in 2015 suggested cessation of surveillance after 5 years of stable cyst appearance; this was met with much debate and was not adhered to due to worry of missing cancers.
• Most other guidelines, including the international consensus guideline have suggested lifelong surveillance of IPCL 17.
• However, it is not economically feasible to survey every IPCL for lifelong especially in the landscape of increasing detection of IPCL and aging society.
• It is estimated that a surveillance program would cost $3.6 million to detect one additional pancreatic cancer following the international guideline for lifelong surveillance versus ceasing surveillance after 5 years of stable cyst appearance 102
• Despite increase in detection of pancreatic cancer, surgical morbidity and mortality may not lead to better overall survival compared to more conservative surveillance program.
• There is increasing evidence showing cysts without worrisome/high risk features have a low risk of pancreatic cancer and that to consider cessation of surveillance after a certain duration should be considered.
• Pancreatic cancer risk in those with IPCL may remain elevated up to 15 years on long term large Japanese cohort study (3.5% at 5 years, 6.6% at 10 years and 15% at 15 years follow up) 103
• The overall risk of cancer does appear to reduce over time in those with stable cyst.
  • In a pooled meta-analysis, the incidence of pancreatic cancer decreased from 0.6 per year to 0.2 per year after 5 years of surveillance 104
• Size of the pancreatic cyst appears to be important risk factor but cysts less than 10mm in size have the lowest risk.
  • In a Japanese cohort, only pancreatic cysts with size larger than 10mm were associated with pancreatic cancer 105.
  • In another Japanese cohort of 450 patients 106 stable pancreatic cyst <20mm in size at 5 years without any worrisome features and connection to the MPD, there were no observed cases of pancreatic cancer over 10 years of follow up.
  • Similar findings were seen in a 577 US cohort of patients followed for median of 82 months, where pancreatic cancer risk was 0.1% vs 7.5% in those followed for more than 5 years with cyst size ≤1.5cm vs cyst size >1.5cm 107
  • In individuals with IPCL, the pancreatic cancer risk was 20 times compared to aged matched general population but the risk appears to reduce to 6 times if the cyst size is ~10mm and remains stable for over 5 years 108
• The risk of development of worrisome/high risk features in IPCL had median time of 4 years in a 540-patient long term Italian study 109
• This indicates that within 5 to 10 years of surveillance, majority of incidental pancreatic cyst would be able to be dichotomise into low risk and close surveillance/surgery.
• The only unpredictable risk is the presumed “field-change” in pancreatic parenchyma and concomitant PDAC development remote from the surveyed cyst which appears independent to the pancreatic cyst risk factor characteristics 103
• However, it is estimated that only 18% pancreatic cancers occur as an independent precursor to IPMN 110

Recommendations

• There is variation among the published guidelines regarding the follow-up intervals and duration of follow-up of incidental PCLs without “worrisome” or concerning features.
• The approach to each patient with a pancreatic cyst should be individualized based on age, clinical status and comorbidities, risk of malignancy, and patient preferences.
• Management should ideally be discussed by a multidisciplinary group of experts with full access to all of the patient’s clinical and laboratory data and all imaging.
• All pancreatic cyst guidelines agree that patients who are not surgically fit do not need ongoing surveillance.
• Ceasing surveillance of stable pancreatic cysts may result in infrequent cases of missed cancers but the risk stratification would minimise the case numbers whilst making the surveillance program economically feasible.
• Stable cysts <10mm should be surveyed for 5 years only.
• Stable cysts between 10mm to 20mm should be surveyed for 5 years only but in patients younger than 65, extending up to 10 years should be considered.
• Stable cyst between 20-30mm should be surveyed for 10 years only. However, in patients younger than 65, longer surveillance or MDT referral should be considered given that there is relative paucity in the evidence on the long-term natural history of the cyst at this size.
• Surveillance in patients above age 75 or those not fit for pancreatic surgery is not recommended.

Serous Cystadenoma (SCA) / Serous Cystic Neoplasm (SCN)

SCAs or SCNs are benign neoplasms with extremely rare examples of malignancy reported (0.1%) 111, 112.

Occasional aggressive tumours are associated with tumour size at diagnosis, and location in the pancreatic head 113.

However, the European Study Group on Pancreatic Tumours of the Pancreas reporting in 2018 9 states that cases reported as malignant do not fulfil the WHO criteria for SCN and that no deaths attributable to dissemination/malignant behaviour of serous cystic neoplasms have been reported.

Microscopically, SCNs appear as single layer of cuboidal or flattened cells lining the small cysts and have round nuclei and abundant clear cytoplasm.

The following clinical features are typical:

• Asymptomatic and discovered incidentally. Rarely, if large can produce symptoms by mass effect.
• Commoner in women (75%).
• Age 5th-7th decades, usually 50s.
• Most common in body and tail of pancreas 114
• Slow growth over time (2-6mm per year on average) 115
• Characteristically, SCAs are solitary, but patients with Von Hippel Lindau syndrome may have multiple pancreatic cysts and tumours as well as in other organs.

The typical morphological pattern of SCA on imaging is as follows: 116, 117

• Polycystic/multilocular/microcystic appearance (about 70% of cases) 51. Cystic locules vary from a few mm to about 2cm 29. Sometimes the multiple intralesional interfaces can lead to a pseudo-solid appearance on imaging - particularly on ultrasound. A honeycomb pattern is often described. This appearance is due to numerous internal tiny cysts mimicking a honeycomb but the lesion may appear well-marginated and show enhancement with contrast 117. Microcystic or honeycomb appearance if sufficiently typical can be assumed to be diagnostic of SCA.
• Lobulated (bosselated) external contour caused by the rounded protuberances of the large number of internal cysts.
• Stellate central scar (present in approximately 30%).
• Central calcification (about 20%).
• Location in body or tail of pancreas.
• A recent study found the combination of morphology and textural analysis on CT to be highly accurate 118.
• However, atypical appearances such as oligocystic/macrocystic variants are present in a relatively high proportion (approx. 60%) of cases 119. Unilocular SCNs are rare.
• Other atypical appearances include mixed microcystic/macrocystic and “solid” SCNs.
• Rarely, a peripheral enhancing wall or mural nodules may be seen 116.
• Macrocystic SCA may present a diagnostic dilemma, being difficult to distinguish morphologically from MCN or BD-IPMN. “Solid” SCA may be confused with SPN.
• Cyst fluid analysis in SCA shows very low CEA level and low amylase/lipase in fluid of low viscosity.

The various expert consensus guidelines are based on low quality evidence and differ in their recommendations regarding surveillance of typical-appearing SCA.

If a confident diagnosis of SCA can be made on the basis of a combination of typical features, as above, most guidelines suggest conservative management and no further imaging follow-up, unless the patient becomes symptomatic 83.

Unless the lesion is large and/or the patient develops symptoms due to mass effect, SCAs should be managed conservatively 9. Given the very low potential for malignancy, surgery is rarely indicated for SCN.

ACR guidelines 12 suggest that SCA >4cm in size should be considered for resection.

Follow-up recommendations for typical SCN

• The European Study Group 9 recommends that if there is a clear/certain diagnosis of SCN, asymptomatic patients should be followed up for a year and then radiological follow-up discontinued in favour of symptomatic follow-up.
• The ACG recommendation 10 that most asymptomatic patients with SCNs with classical features on imaging need no further follow up evaluation (conditional recommendation, low quality evidence).
• Both the ACR 12 and European committee 9 recommend follow-up only based on symptoms, although the European guidelines recommend initial follow-up for one year.

Further recommendation is that if the diagnosis is uncertain, follow-up should be as for branch-duct IPMN Indeterminate Cyst Without Worrisome/High risk Features.

However, oligo- or macro-cystic variants cannot reliably be distinguished on imaging alone from mucinous cystic neoplasms (which may progress to malignancy) and therefore require either further investigation, surveillance or surgery dependent on size and other features.

• All decisions regarding management of individual patients, especially where surgery is considered an option, should ideally be discussed by a multidisciplinary group of experts in the field who have full access to all of the patient’s clinical and laboratory data and all imaging.
• It is important to note that there are variations among the published society/international guidelines 16 as to what features should lead to consideration of surgery in patients of appropriate age and fitness. However, there appears to be general agreement that cyst size, the presence of mural nodules or solid components and dilatation of the main pancreatic duct are criteria that are important in assigning risk of high-grade dysplasia (HGD) or malignancy.

There appears to be a degree of agreement that the following are features that are indications for surgery in patients with mucinous cystic lesions of the pancreas:

• Symptom development (e.g. jaundice, acute pancreatitis, recent onset Diabetes Mellitus)
• Cyst size (? >3cm ? ≥4cm – varies among different guidelines)
• MPD ≥10mm, obstructed MPD or focal dilatation of MPD
• Rapid cyst growth (>5mm per year)
• Solid component / mural nodule >5mm especially when enhancing.
• Positive cytology for HGD or carcinoma on EUS-FNA
• Regional lymphadenopathy

Other indications for surgery:

• Solid pseudopapillary neoplasm 9
• Cystic Pancreatic Neuroendocrine neoplasm (>2cm 9 or symptomatic)
• Serous Cystadenoma that is >4cm 12 or development of symptoms or compression of adjacent structures

Cystic Neuroendocrine Tumours (Cystic NET)

Cystic Neuroendocrine Tumours (Cystic NET) 120

The pathological and imaging findings and classification of PNETs has been reviewed by Khanna et al in 2020 120, 121

• The majority of PNETS occur spontaneously but a small proportion are associated with syndromes such as Multiple Endocrine Neoplasia Type 1 (MEN 1), von Hippel-Lindau (VHL) disease, Tuberose Sclerosis, Neurofibromatosis Type 1 or Glucagon Cell Adenomatosis.
• Well-differentiated grade 1 and grade 2 (WHO classification) PNETS are typically small (<3 cm) well-circumscribed hypervascular masses with intense homogeneous enhancement in the arterial and portal venous phases.
• The degree of homogeneity and enhancement correlates with tumour grade.
• Biochemical investigation will identify which tumours are functioning:
  • Functioning lesions are usually treated with surgery, where clinically appropriate.
  • 68Gallium DOTATATE PET/CT with 68Ga-tetraazacyclododecane tetraacetic acid (DOTA)–somatostatin analog is useful to look for metastases.
• Non functioning lesions that are asymptomatic are generally treated as follows:
  • Surgery for >20mm
  • EUS/MRI surveillance at 6-12 months for lesions <20mm
  • If the change in size is >5 mm or the final diameter is >20 mm, the cyst should be resected.

Go to the pathway

Pancreatic Cancer (Staging)