SPNs commonly occur in young females in their 2nd and 3rd decades of life [24]. SPNs are usually large in size at the time of diagnosis with a mean long-axis tumor diameter of 6.8 cm. Despite its larger size, there is usually no biliary or pancreatic duct obstruction [25].

SPNs are characterized as a well-encapsulated thick-walled mass with varying solid, cystic, necrotic, and hemorrhagic components giving it a heterogeneous appearance. Hemorrhagic components of the lesion demonstrate higher than water attenuation on unenhanced CT and appear hyperintense on T1-weighted imaging. Solid components of the lesion show little enhancement on post-contrast imaging [26–28]. Peripheral/rim or septal calcifications may also be seen on CT scans, especially in larger lesions [29]. On MRI, the cystic components are T2 hyperintense while the characteristic thick capsule is T2 hypointense and demonstrates progressive delayed enhancement on post-contrast images (Figure 1) [30]. These features can help differentiate SPN from other PCL, such as SCA and MCN.

On EUS, SPN may appear solid or may have mixed solid-cystic appearance. Though EUS may not add much to cross-sectional imaging in terms of morphological characterization of SPN, it is vital in obtaining histological diagnosis via EUS-guided FNA [31]. The most common findings seen on histopathology are an increase in cellularity, with neoplastic cells developing loose papillary clusters with fibrovascular central cores. Foamy macrophages, multinucleated giant cells, and hemorrhagic remains are also seen in the background [32, 33].

The recommended treatment for all SPN is surgical resection [34, 35]. Although they are low-grade neoplasms and generally follow an indolent course, they have a low but significant risk of metastatic spread in 5–15% mainly involving the liver and/or peritoneum [36, 37]. Therefore, upfront surgical treatment is generally recommended, even for asymptomatic lesions.

SCAs are commonly seen in females in their 5th to 6th decade of life and characteristically have microcystic morphology with a central area of calcification [38]. The classic “honeycomb” pattern consists of numerous sub-centimeter micro-cysts that cannot be differentiated on imaging. The oligocystic pattern exhibits cysts of > 2 cm and occasionally overlaps with MCN and IPMN in morphological features [39]. Hence, cyst fluid analysis further aids in differentiating between SCA and IPMN/MCN [40].

On MRI, SCAs appear as a cluster of small cysts grouped together within the pancreas (Figure 2A). Signal characteristics are as follows: Cystic foci demonstrate hyperintensity on T2 (Figure 2B), hypointensity on T1 (Figure 2C) and the central fibrous scar is hypointense on T2, T1 C+ (Gd): on delayed contrast enhanced imaging, the fibrous septa between the cysts appear enhanced [41].

On EUS, the SCA commonly appears as a cystic lesion with multiple anechoic cystic spaces with septations and an occasional central hyperechoic calcification [42]. The key findings on EUS-FNA of SCA include cytologically bland, non-mucinous epithelium that tests positive for melan-A, inhibin, CDK19, and negative for chromogranin and synaptophysin [42].

The management of SCAs is primarily driven by the presence of symptoms as they have no malignant potential. If the diagnosis is definitive and the lesion is asymptomatic, observation with consideration for serial imaging can be pursued. Surgery is warranted if malignancy cannot be excluded or if the lesion is symptomatic, regardless of size. Currently, offering surgical resection to all SCAs > 4 cm regardless of the presence of symptoms is anecdotal [43, 44].

MCNs are predominantly diagnosed in women in their 4th–6th decade of life and are often located in the body or tail of the pancreas. All MCNs are associated with a significant risk of advanced neoplasia with a rate of high-grade dysplasia (HGD) (6–13%) and invasive carcinoma (IC) (4–12%) [45–49].

On CT, MCNs of the pancreas may appear as unilocular macrocysts that do not communicate with the pancreatic duct (Figure 3A). They are occasionally seen with peripheral eggshell calcification and mostly have heterogeneous cyst contents. The presence of thick, enhancing septations or mural nodules increases the likelihood of malignant transformation [50, 51].

Typical MRI features of MCN include a well-defined, unilocular, or oligolocular cyst. On T1-weighted images, MCNs may exhibit variable signal intensity, depending on the proteinaceous content of the cyst fluid (Figure 3B). On T2-weighted images, MCNs are hyperintense. Unlike IPMNs, the lesions do not show communication with pancreatic duct. MRI is superior to CT in identifying internal thick septations, mural nodularity, and any enhancing components. After gadolinium injection, enhancement of thick cyst wall is seen best on delayed phases of imaging owing to its fibrous nature [52].

IPMNs occur in both sexes, with a slight predominance in older males in their 5th to 6th decade of life. They are considered precursor lesion that can potentially progress to pancreatic ductal adenocarcinoma (PDAC). A distinctive feature that helps differentiate IPMN from mucinous cystadenoma/cystadenocarcinoma is that these tumors communicate with the MPD or its branches, and thus are classified into three types based on ductal involvement: MD-IPMN, BD-IPMN, and mixed-type IPMN [2, 53].

CT shows IPMNs as single or multiple cystic hypodense lesions. MD-IPMNs directly involve the MPD associated mucin hypersecretion resulting in ductal dilation over 5 mm often with progressive atrophy of overlying pancreatic parenchyma (Figure 4A and 4B). BD-IPMNs may be unilocular, multicystic, or may have a tubular morphology (Figure 5A). Mixed-type IPMNs have features of both the BD-IPMNs and MD-IPMNs (Figure 5B) [18]. On MRI, IPMNs are hyperintense on T2-weighted imaging and demonstrate low signal on T1-weighted imaging due to its mucinous content. Communication with pancreatic duct is commonly noticeable. Enhancing mural nodules and pancreatic ductal dilation are better identified on MRI when compared to CT imaging [54].

MD-IPMNs and mixed-typed IPMN demonstrate malignant transformation in approximately 45% of resected samples [6, 8]. In comparison with BD-IPMN, which exhibits a rate of advanced neoplasia in around 10–20% of resected specimens [6, 8]. Guidelines aimed at identifying worrisome and high-risk features for BD-IPMN have been set fourth, including the Sendai criteria followed by the Fukuoka guidelines which have been revised in 2017 and most recently the Kyoto guidelines [9, 55]. A recent study hypothesized that dilation of the uncinate duct is a radiographic indicator of high-risk disease, specifically IPMN associated with HGD or IC [56, 57]. IC may develop at a site distant from the primary location of the IPMN at a later stage, supporting the concept of field defect and dual carcinogenesis. This emphasizes the importance of radiological evaluation of the entire pancreas and further surveillance of the residual pancreas following resection of the cystic lesion [58].

Although CT and MRI are able to detect PCL at high frequencies, these modalities have limitations when it comes to distinguishing the different types based on morphologic features alone [5]. While EUS has been shown to have a superior image resolution of the pancreas, studies have shown that morphologic features alone remain suboptimal in delineating the different subtypes or accurately predicting the risk of advanced neoplasia.

The effectiveness of EUS is highlighted when combined with FNA of pancreatic cyst fluid. EUS-guided-FNA for cyst fluid analysis including fluid cytology and different biochemical markers help bridge the void of inaccurate diagnosis as well as providing a reliable tool to accurately predict the risk of advanced neoplasia among the different PCL types [59–61].

Fluid cytology is a valuable diagnostic tool that helps differentiate benign versus malignant cystic lesions albeit the diagnostic sensitivity of fluid cytology obtained by EUS-FNA is low (approximately 28.7%). This is due inadequate sample retrieval, several overlapping pathologic features and false negative results. It does, however, have a relatively high specificity for confirming malignancy when other radiologic or clinical findings are present. According to the World Health Organization (WHO) definition, the cytologic grade demonstrates that the absolute risk of HGD/IC of “suspicious” and “positive” results are 91–100% and 100%, respectively [62–65].

Indications for pursuing EUS-guided FNA are typically driven by the presence of the aforementioned worrisome features to confirm the diagnosis, especially if surgery is being contemplated. Further, ambiguity on imaging would also mandate cytologic and fluid evaluation for diagnostic confirmation [8, 12, 66]. The approach to aspirate and analyze the cyst fluid is variable and is based on several factors including cyst location, size, and accessibility. EUS-guided aspiration is typically the preferred approach due to its high accuracy, yield and low risks. In the cases where EUS-FNA is not feasible, alternative approaches such as percutaneous aspiration or even surgical aspiration may be required. Percutaneous drainage is pursued in the setting of prior surgical bypass or altered anatomy [67]. On the other hand, surgical drainage is mostly reserved for drainage procedures in the setting of chronic pancreatitis and walled off pancreatic necromass [68].

Even though EUS-FNA is generally considered a safe procedure, however for cystic lesions there is some degree of risk involved with a complication rate 13.6% [69]. Factors that may increase the risk of adverse events include cyst location, size, wall thickness, operator experience, vascular involvement, needle size and protease inhibitor use [70]. Some of the more common complications acute pancreatitis (0.44%), bleeding (0.13%), and infection (0.05%) [69].

Despite advancements in imaging, endoscopic, cytopathologic and biomarker analysis, identifying precisely which PCL harbors malignancy or is at risk of progression remains a challenge. Given that most precancerous lesions have a low risk of malignant transformation, molecular and genetic profiling including next generation sequencing (NGS) of cell-free DNA or mi-RNA further aids in proposing treatment and surveillance that can guide clinical decision making [5]. For instance, mutations in the mitogen-activated protein kinase genes such as KRAS and GNAS are specific for mucinous cysts, whereas mutations in TP53, SMAD4 are linked with more advanced lesions [87]. Alterations in KRAS, GNAS, and/or BRAF are present in 89% mucinous cysts. Among mucinous cysts with advanced neoplasia, mutations in TP53, SMAD4, and the mammalian target of rapamycin (mTOR) genes were identified in 86% of cases [87].

KRAS is a proto-oncogene that plays a crucial role in various cellular pathways. Several studies have identified links between pancreatic neoplasms and KRAS mutation. In a study by Nikiforova et al. [88], 618 pancreatic cyst specimens were obtained and evaluated for molecular mutations, of which 38% KRAS mutations. However, KRAS mutations had an overall 100% specificity and a sensitivity of 89% for mucinous differentiation [88]. Thus, the utility of probing for KRAS mutations have been shown to be a valuable marker to be able to differentiate between mucinous cysts such as IPMN and MCN, and non-mucinous cysts [89].

GNAS is a gene that encodes for a subunit of the G-protein that produced cAMP. While KRAS and GNAS are classically evaluated in tandem, they do play different roles in tumor development. When mutated it can lead to abnormal and overactive cells growth [90]. In a study by Wu et al. [91], 66% of IPMN cyst fluid showed a mutation in GNAS, while 96% of cysts exhibited either a KRAS or a GNAS mutation. In another investigation, researchers analyzed EUS-FNA samples of in both IPMN and non-IPMN cysts. GNAS mutation was detected in 47.2% of IPMN samples [92]. In another study by Paniccia et al. [87], GNAS mutations were present in 30% of cystic fluid samples, and occurred with either KRAS (78%), or BRAF (10%). As such, GNAS mutation are highly specific to IPMN and coupling it with a KRAS mutation or elevated CEA improves diagnostic accuracy as they are not found in MCNs.

RNF43 is a tumor suppressor gene, which encodes a protein that is involved in several biological processes such as cell division and differentiation. RNF43 mutations are associated with different types of malignancies such as PDAC [93]. However, RNF43 has been shown to be a potential marker to predict malignant transformation. In a study by Sakihama et al. [94], 106 MCN with varying grades of dysplasia, high grade dysplasia and IC had a higher frequency of RNF43 mutations in comparison to lower grade lesions [94]. Another study by Zhou et al. [95], used mice that had a RNF43 knockout and an activated KRAS, like in IPMN, these mice showed increased cystic papillary lesions as well as PDAC [95]. When these mice were treated with a growth inhibitor that selectively inhibited mutant RNF43 cell, increased survival as well as decreased malignant transformation were observed in these mice. As such loss of RNF43 is an important prognostic tool to monitor PDAC development in precancerous pancreatic cysts [95].

Proteomic analysis is starting to play an equally crucial role in profiling lesions. Through proteomic examination, Chen et al. [109] found nine differently expressed proteins in mucinous compared with non-mucinous cystic fluids. Four of these identify mucinous lesions with a sensitivity of 81%, specificity of 90%, and accuracy of 86% [109]. Other studies have looked at the various mucin proteins found in PCL such as thymosin-beta-4, ubiquitin, and VEGF-A. High levels of VEGF-A (over 5,000 pg/mL) as being significantly associated with benign serous cystic neoplasms [110].

A variety of other molecular markers in pancreatic cyst fluid have also been examined and found to hold prognostic value. Hao et al. [111] found that higher concentrations of cytokines, including interleukins (IL)-1b, 5, and 8, are associated with cysts containing HGD or malignancy [111]. IL-1b alone was found to be an independent predictor of high-risk vs low risk cysts, with a positive predictive value of 71%. Prostaglandin E2 (PGE2) was also found to be elevated in pancreatic cancer tissue and hypothesized to differentiate IPMNs and MCNs with the former showing higher levels [112]. Finally, telomere fusion status has been assessed as a prognostic marker of malignancy, with evidence of shortened chromosomal telomeres in dysplastic IPMNs [111].

Molecular analysis of the malignant potential of pancreatic cysts is not limited to pancreatic cyst fluid. Several authors have suggested using pancreatic fluid or plasma as a source of molecular biomarkers. The use of pancreatic juice collected from the duodenum avoids the risk of direct sampling of the pancreas and could be more representative of lesions throughout the pancreas, as opposed to a single cyst [5]. Similarly, plasma analysis presents an avenue for less invasive bio-specimen collection and could be used for the surveillance of at-risk patients. Levink et al. [113], compared mutation detection rate in plasma versus pancreatic fluid. They hypothesized that due to the proximity that pancreatic fluid analysis would show higher mutation detection rates. Despite finding higher concentrations of cell free DNA, there was no difference in mutation dection rate, supporting the theory that plasma analysis could be used for genomic analysis [113]. In another study by Visser et al. [114], they showed that the presence of DNA mutation in pancreatic juice—particularly of key oncogenic drivers such as TP53, SMAD4, or CDKN2A, as well as hypermethylation of NPTX2—had a near 100% specificity for the presence of HGD [114]. Similarly, Yang et al. [115] analyzed 25 different protein biomarkers in plasma extracellular vesicles and found they had high discriminatory power in detecting atypical SCA that usually missed on typical imaging findings.

With multiple systematic reviews and meta-analyses confirming the prognostic power of molecular analyses of cyst fluid (as well as pancreatic juice and plasma), future challenges will largely center around integrating these new biomarkers into existing diagnostic and surveillance frameworks. Different sequencing strategies, gene coverage, and thresholds for mutation calling can impact the consistency of results among institutions [116]. Techniques for identifying circulating tumor cells vary and target different characteristics of tumor cells [5]. It will be essential to integrate these analyses with currently available imaging modalities.