Cytology and Core Biopsy for Cystic Lesions of the Pancreas: A Comparison of Three Cytological Grading Systems

Abstract

Introduction:

Cytological appearances may be insufficient to establish the diagnosis of pancreatic cystic lesions, especially for premalignant neoplasms. This study aimed to evaluate three commonly used cytological grading systems [C1–C5, Papanicolaou (Pap), and World Health Organization (WHO) classification systems] in assessing pancreatic cystic lesions.

Materials and Methods:

A total of 210 pancreatic cytology specimens were classified based on the aforementioned grading systems, 127 of which had supporting histology reported across 6 years, at a single-tertiary referral center.

Results:

We excluded 26 cases because of inadequate cytology. The most common cystic lesions were intraductal papillary mucinous neoplasms (IPMNs, n = 71) and pseudocysts (n = 55). Among IPMN, 27 were moderate/high-grade, and 44 were low-grade. There were 12 mucinous cystic neoplasms, 75 benign cysts (pseudocysts, serous cystadenomas, lymphoepithelial cysts, and others), and 15 malignant cases, with 11 cysts being of uncertain etiology. There were 21 high-risk and potentially malignant (WHO grade V) cases in comparison with 42 cases using the C grading (C4 and C5). However, according to Pap grading, 87 had varying risks of malignancy (Category IVB). According to the WHO classification, 63 cases were classified as low-risk pancreaticobiliary neoplasms (WHO grade IV). The remaining cases 85 were benign to very low-risk malignant potential and, therefore, were less likely to be considered for surgical intervention. Diagnostic concordance was higher between cytology and core biopsies in low-grade IPMN.

Conclusion:

The WHO system provides better risk stratification for neoplasms, optimizing surgical management. However, the C1–C5 system does not recognize cystic lesions with malignant potential.

INTRODUCTION

Cystic lesions of the pancreas are a heterogeneous group comprising both benign and malignant entities.[1] These lesions are incidental findings in 2%–20% of all examinations, with the likelihood of malignancy increasing in high-risk subjects, such as those with a familial history of pancreatic cancer.[2,3] Mucinous cystic lesions of the pancreas such as intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs) are particularly important as these are considered precursors of pancreatic cancer. Thus, early detection and management are paramount.[4] Fine needle aspiration (FNA) cytology and core biopsies taken under endoscopic ultrasound guidance are reliable, minimally invasive, and cost-effective diagnostic tools. The tissue obtained from core biopsies is essential for not only diagnostic purposes but also molecular profiling, which is now being increasingly utilized as part of targeted therapy.[5] However, cytomorphological features alone may not be sufficient for an accurate diagnosis, and clinicopathological correlation in combination with imaging studies are essential considerations in a multidisciplinary setting.[6,7,8,9,10]

This study aimed to evaluate the use of three commonly used cytological grading systems [C1–C5 system,[11] the Papanicolaou (Pap) system,[12] and the World Health Organization (WHO) classification system[13]] for the assessment of cystic lesions of the pancreas and compare them with histology where available.

MATERIALS AND METHODS

We investigated the cytological database held in the Department of Cellular Pathology between 2017 and 2023. We used the systematized nomenclature of medicine (SNOMED) classification system to select pancreatic cytology, core biopsies, and resections (distal, partial, and Whipple’s specimens) performed at our center or referred from external sites for a second opinion [Figure 1].

Figure 1.

Flow diagram with the spectrum of cystic lesions of the pancreas identified on cytology

The FNA material was collected after centrifugation into cytology fixative—3% acetic acid in 70% industrial methylated spirit, spread over glass slides, and stained with the Pap stain and those including core biopsy obtained via 22-gauge needle were stained with hematoxylin and eosin (H&E). Cell blocks for FNA needle core aspirates were prepared from the pellet of centrifuged cell suspension by adding plasma and thrombin to enmesh the cellular material in a clot. Pancreatic resection specimens were dissected and embedded according to standard protocols, and the blocks were stained with H&E. Immunohistochemistry was performed as necessary for accurate diagnosis.

We stratified the diagnoses according to the “C1–C5” grading system,[11] cytological grading of Pap grading system devised by The Papanicolaou Society of Cytopathology Guidelines,[12] and the WHO international grading system[13] [Table 1]. The “C1–C5” grading was initially introduced for breast FNA,[14] but this was adopted for grading biliary brushings and pancreatic cytology specimens.[11,15] We aimed to identify the differences between these systems concerning clinical outcome and management. With regard to mucinous cystic lesions, we compared the diagnostic concordance (levels of diagnostic agreement including grading of dysplasia) of histological specimens, especially core biopsies, with the FNA cytology findings.

Table 1.

The three grading systems used for reporting pancreatic cytology

“C1–C5” grading	Papanicolaou Society of Cytopathology	World Health Organization (WHO)
C1—Insufficient	I—Non-diagnostic	I—Insufficient/Inadequate/Non-diagnostic
C2—Negative for malignancy	II—Negative for malignancy	II—Benign/ negative for malignancy ▪Serous cystadenoma
C3—Atypical	III—Atypical	III—Atypical
	IVA—Neoplasm—benign ▪Serous cystadenoma	IV—Pancreaticobiliary neoplasm Low-risk/grade ▪IPMN/MCN (low/moderate-grade)
	IVB—Neoplasm—other ▪IPMN/MCN (low/moderate grade) ▪IPMN/MCN (high-grade) ▪PanNET ▪SPN	V—Pancreaticobiliary neoplasm High-risk/grade ▪IPMN/MCN (high-grade)
C4—Suspicious for malignancy	5—Suspicious for malignancy	VI—Suspicious for malignancy
C5—Malignant	6—Malignant	VII—Malignant ▪PanNET ▪SPN

IPMN, intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm; PanNET, pancreatic neuroendocrine tumor; SPN, solid pseudopapillary neoplasm

RESULTS

We found a total of 210 patients with cystic lesions of the pancreas, 26 being excluded because of inadequate cytology [Figure 1]. There were 94 (51.1%) male and 90 (48.9%) female patients. The most common diagnosis was IPMN (38.5%, n = 71), followed by benign pseudocyst (29.9%, n = 55). Thirty out of 55 cases (54.5%) of pseudocysts were accompanied by inflammatory cells, whereas the rest of 25 (45.5%) did not show an inflammatory component. MCNs were diagnosed in 12 (6.5%) cytological specimens. Among the cases reviewed by histology (n = 127), 48 (26%) included resected specimens, while the remaining were examined via core biopsy (n = 29, 22.8%) or through a cell block of FNA biopsy (n = 50, 39.3%) [Table 2]. The morphological features of the cystic lesions of the pancreas are illustrated in [Figures 2,3,4,5,6,7,8,9,10]. Figure 11 illustrates the cytological grading of cystic lesions of the pancreas by the “C1–C5” grading system,[11,15] Pap,[12] and WHO grading systems.[13] Table 3 shows the levels of concordance of core biopsy and resection specimens with the FNA sampled material for mucinous cystic lesions. Around 84.2% of low-grade IPMN were concordant on subsequent histology with a 100% agreement rate on core biopsies alone. In contrast, 46% of high-grade cases were diagnostically concordant on histology.

Table 2.

Spectrum of cystic neoplasms of the pancreas on histology including resected and core biopsy specimens

Diagnosis	Number	%
IPMN—Low-grade	44	34.6
IPMN—Moderate-grade	6	4.7
IPMN—High-grade	21	16.5
Mucinous cystic neoplasm	12	9.4
Lymphoepithelial cyst	6	4.7
Serous cystadenoma	7	5.5
Other benign cysts	7	5.5
Cysts of uncertain etiology	11	8.6
Neuroendocrine tumor (with cystic change)	3	2.3
Pseudopapillary neoplasm	1	0.7
Adenocarcinoma (with cystic change or associated with IPMN)	9	7
Total	127
Histology*	127
Core biopsies	29
Resections	48
Percentage of combined core biopsies and resections	60.6%

IPMN, intraductal papillary mucinous neoplasm.

^*Histology data include surgical resections and cases sampled via core biopsy including cell blocks from fine needle aspirated material

Figure 2.

(A and B) FNA pancreas (×200) (A) The PAP-stained slide is sparsely cellular and shows dispersed lymphohistiocytic collections. (B) The cell clot (H&E) in the right field shows similar collections. There are no epithelial cells present and no evidence of malignancy

Figure 3.

(A and B) FNA tail of the pancreas (×200). Lymphoepithelial cyst (A) Pap stain: Cellular sample containing neutrophils admixed with debris and keratinized squames. (B) Cellblock H&E: Clusters of benign squamous cells, lymphocytes, neutrophils, and keratinous debris. (C–F) Pancreatic cyst with occasional squamous cells, FNA Pap (×400) (C). H&E (×200) (D) shows a lymphoepithelial cyst with sebaceous glands. On immunohistochemistry, the squamous epithelial lining is focally positive for CK7 (E) and strongly positive for CK14 (×200) (F). (G and H) Distal pancreatectomy specimen. Benign squamous cyst. (G) H&E stain (×400): Sections from the wall of a pancreatic cyst showing epithelial lining with features of squamous differentiation. (H) on immunohistochemistry (×400), there is strong nuclear positivity for p63 in the epithelium

Figure 4.

(A) This H&E clot (×100) from an FNA specimen shows a fragment of fibrous tissue lined by flattened cuboidal cells. Immunohistochemistry is positive for calponin (B) and inhibin (C) × 100

Figure 5.

(A) H&E clot (×100) shows cyst debris with mixed inflammatory cells, set in a serous-like background. There is collapsed cuboidal epithelium due to crushing artifact however, on immunohistochemistry, this lining stains strongly positive for calponin (B) and AE1/3 (C) × 100

Figure 6.

(A and B) The H&E clot (×100) shows stringy mucoid material. (C) Occasional mucinous epithelium with mild nuclear atypia (Grade C3) is seen (H&E ×200). (D) A special stain for Alcian blue (×200) highlights the mucoid material

Figure 7.

(A) H&E stain (×200) reveals a cystic lesion lined by low-grade epithelium which is positive for CAM5.2 (B) on immunohistochemistry. The ovarian-like stroma is positive for ER (C), PR (D), and CD10 (E) ×200. Ki67 is low ~1% (F) ×400

Figure 8.

(A and B) Pap preparations (×200) show mucin material in the background with neoplastic epithelium. (C) The H&E clot (×200) shows fragments of low-grade neoplastic epithelium

Figure 9.

(A and B). Pap stains (×400). This is a cytology preparation containing a few crowded clusters of epithelial cells with a high nuclear–cytoplasmic ratio. (C) H&E clot preparation (×400) shows a fragment of high-grade intraductal papillary mucinous neoplasm

Figure 10.

(A) Pap stain (×400) and (B) H&E Clot (×200) contain bland epithelial cells arranged singly, as well as in groups. Cells have moderate amounts of cytoplasm and prominent nucleoli. The background preparation contains acellular debris, a few lymphocytes, neutrophils, and macrophages. (C) Core biopsy, H&E stained (×40). Fragments of tumor arranged in cords with intervening fibrous tissue showing focal lumen formation. (D) Immunohistochemistry for CD10 shows strong membranous staining (×100). (E) Immunohistochemistry for beta-catenin shows strong nuclear staining (×40)

Figure 11.

Cytological grading of cystic lesions of the pancreas

Table 3.

Concordance of cytology with core biopsy and resection for mucinous cystic lesions of the pancreas

Cytological diagnosis	Total number of cases	Concordant core biopsies	Concordant resections	Combined concordance
IPMN—Low-grade	44	9/9	7/10	16/19
IPMN—Moderate-grade	6	1/2	0/0	1/2
IPMN—High-grade	21	4/5	2/8	6/13
Mucinous cystic neoplasm	12	0/0	1/2	1/2

IPMN, intraductal papillary mucinous neoplasm

Mucinous cystic lesions of the pancreas

Most IPMNs (n = 44) displayed low-grade dysplasia and were categorized as C3, PAP IVB, and WHO IV. Cases with moderate dysplasia (n = 6) were included under C4, PAP IVB, and WHO IV, whereas IPMN with high-grade dysplasia (n = 21) were categorized as C4, PAP IVB, and WHO V.

The majority of MCNs (n = 7) were clustered in the C3, PAP IVB, and WHO IV grading groups. One case showed high-grade nuclear dysplasia (C4, PAP IVB, and WHO V), whereas four cases showed no epithelial atypia with thick mucin material or ovarian-like stroma in the background (C2, PAP IVB, and WHO IV) [Figure 10].

In this study 87 cases (41.4%) were classified as Group IVB on Pap; this included 21 cases of high-grade IPMN and 12 cases of MCN, which together account for approximately 41% of cases that may require surgical intervention. According to the WHO grading system, the low-risk category (Group IV) included 62 cases (IPMN and MCN of low- and moderate-grade), whereas the high-risk category (Group V) accounted for 21 examples of high-grade IPMN and one case of high-grade MCN.

Benign cystic lesions of the pancreas

There were 85 cases (40%) that were benign or had very low-risk malignant potential that did not require surgical intervention (PAP grades II, III, and IVA and WHO grades II and III).

All pancreatic pseudocysts (n = 55, 29.9%) and lymphoepithelial cysts (n = 6, 3.2%) were classified as C2, PAP II, and WHO II. Cystic lesions of uncertain origin (n = 7, 3.8%) but with benign characteristics on cytology were included in the same group. Serous cystadenomas (n = 7, 3.8%) were classified separately as C2, PAP IVA, and WHO II [Figure 10].

Malignant cystic lesions of the pancreas

All cases of adenocarcinoma with cystic changes or with associated IPMN (n = 11) were classified as C5, PAP VI, and WHO VII. Our cohort also included three cases of neuroendocrine tumors with cystic changes (graded C5, Pap IVB, and WHO VII) and one case of pseudopapillary neoplasm with cystic change (C3, Pap IVB, and WHO VII) [Figure 10].

Cysts of uncertain nature with scanty atypical cells (n = 11, 5.9%) were categorized as C3, PAP III, and WHO III.

DISCUSSION

Our study highlights the nuances in classifying cystic lesions of the pancreas using three different grading systems. The Pap system accounts for lesions of malignant potential through the incorporation of Group IVB (neoplastic—other),[12] whereas the WHO classification takes this further by separating entities of low malignant potential (Group IV), for example, low- and moderate-grade IPMN/MCN, from those of high malignant potential (Group V) for example, high-grade IPMN/MCN.[13] As the “C1–C5” grading system does not take into consideration malignant potential and the associated clinical management, its value in the assessment of these lesions appears limited. Concordance on histology, particularly for core biopsies, was higher in cases of low-grade cystic neoplasms when compared with high-grade lesions.

Mucinous cystic lesions of the pancreas

The management of both IPMN and MCN typically involves surveillance or surgical resection. This approach is considerably more aggressive than most benign lesions, and this demonstrates the importance of distinguishing these lesions from other benign pancreatic cystic lesions.[3] The absolute indication for surgery is more strongly associated with Group V (WHO); however, this option could not be excluded for the remainder of IPMN and MCN cases in category IV (WHO). For example, in the absence of high-grade cytological atypia, the management of MCN tends to be more conservative, but the final decision depends on other clinical and radiological findings and patient choice. Therefore, the WHO grading system performs better in selecting high-risk cases for surgery with PAP allowing some flexibility in the management of low-risk cases.

Benign cystic lesions of the pancreas

Pancreatic pseudocysts are common non-neoplastic entities that arise as a consequence of acute or chronic pancreatitis. In our cohort, we found evidence of pancreatitis in 54.5% of cases. In the absence of clinical symptoms, pseudocysts are followed up, otherwise, drainage or surgical intervention may be considered.[16] These lesions along with lymphoepithelial cysts and other benign cysts were categorized similarly across all grading systems (C2, Pap grade II, and WHO grade II).

Serous cystadenomas are benign cystic lesions that can occur anywhere in the pancreas and very rarely display features of malignancy.[17] Serous cystadenomas were shifted into WHO category II (negative for malignancy) from PAP category IVA (benign-neoplastic). We consider this a positive change since these lesions have no risk of malignancy and management is generally conservative. Treatment is non-surgical, except for rapidly progressive lesions, giant tumors with obstructive symptoms, or when malignancy cannot be excluded.[16,17]

Otherwise, all grading systems performed similarly in categorizing benign-neoplastic and non-neoplastic cystic lesions.

Malignant cystic lesions of the pancreas

Obtaining tissue for confirmation of a malignant cystic process suspected clinically or on imaging studies is essential for initiation of neoadjuvant therapies and enrolment into clinical trials.[18] Entities included in the malignant categories include ductal carcinoma with cystic changes and IPMN with invasion. Surgical resection remains the mainstay of treatment for these cases (C5, PAP grade VI, and WHO grade VII).

Neuroendocrine tumors (PanNETs) may display cystic changes, and three examples were included in our study. WHO classification allows for reduced ambiguity when compared with the “C1–C5” system and Pap grading systems with a reduced proportion of cases being classified as suspicious of malignancy instead of malignant.[15]

In our study, one solid pseudopapillary neoplasm with cystic change was included. This is a low-grade malignant pancreatic tumor that often shows mixed cystic and solid appearance on imaging studies. The recurrence/metastatic rates are relatively low (~15%) and these lesions do not behave as aggressively as the ones categorized as “malignant”.[19,20] In the “C1–C5” grading system,[11] this lesion is categorized as C3, whereas in PAP is IVB and in the WHO international system is categorized as VII. The WHO classification system appears to better reflect the risk of malignancy of these neoplasms, including the indication for surgical intervention.

Strengths and limitations

Our study provides novel data on cystic lesions of the pancreas using known (C1–C5[11] and PAP)[12] and recently described (WHO)[13] grading systems in pancreaticobiliary cytology. The data show that each system classifies these lesions independently, particularly concerning malignant potential and clinical management including surgery. This directly addresses the practical implications for clinicians managing pancreatic cystic lesions.

It is important to note the limitation of cytology in the assessment of pancreatic lesions. We corroborated our cytological findings with histopathology in 127 cases. In the absence of a histological diagnosis, cytological findings were discussed at multidisciplinary team meetings and corroborated with clinical presentation, radiological findings, and fluid biochemical markers such as carcinoembryonic antigen and amylase.

In this study, 12.3% of samples were inadequate; however, a thick film of mucoid material without cellular component is considered significant and categorized under Group IVB on PAP[12] (premalignant neoplasms and low-grade malignant neoplasms) and Group IV on WHO[13] (low-risk pancreaticobiliary neoplasms). Furthermore, within a cystic lesion, atypical cells may be sparse in comparison with cyst contents so the sample may not be representative of the cystic tumor and may affect grading. Therefore, the role of grading dysplasia (especially moderate/intermediate) in FNA cytology within the management algorithms, especially in cases of mucinous cystic lesions, is not well established and remains open to prospective research.[21] The appearances of MCN on cytology can also vary considerably, ranging from acellular aspirates with mucin material to samples demonstrating only mild cellular atypia, a further reason why cytological assessment may be less appropriate for these lesions.

FNA sampling in itself for benign cystic lesions of the pancreas carries the risk of bleeding and fluid aspiration is often challenging in cases of small cysts with a risk of pancreatitis of around 3%.[22,23]

Limitations of FNA cytology in malignant cystic lesions include both sampling and interpretation errors. In cases of pancreatic ductal adenocarcinoma, desmoplastic stroma may be overrepresented, and this could hinder the morphological assessment of the tumor cells. Similarly, well-differentiated cases may resemble normal pancreatic tissue which could alter cytological interpretation.[18]

Due to the aforementioned limiting factors, reaching a precise diagnosis may require additional biochemical [carcinoembryonic antigen, Kirsten rat sarcoma viral oncogene homolog (KRAS), and guanine nucleotide-binding protein, alpha-stimulating activity polypeptide (GNAS)] and ancillary tests along with radiological correlation in a multidisciplinary setting.

Future work

With regards to molecular pathology testing, future studies on assessing important driver gene mutations (either for diagnostic or therapeutic aid) in pancreatic neoplasms including malignancies should be considered.[24] Furthermore, large-scale studies are recommended to revalidate the utility of these three grading systems in pancreatic cytology.

CONCLUSION

The “C1–C5” grading system does not recognize lesions with malignant potential and so appears to be of limited utility in the assessment of pancreatic cystic lesions. The Pap system accounts for lesions of malignant potential through the incorporation of Group IVB (neoplastic—other). However, the WHO classification supersedes the Pap system by separating entities of low malignant potential from those of high malignant potential (Group V). This allows for diminished ambiguity and, therefore, provides optimal flexibility for managing patients with cystic lesions of the pancreas, which has become increasingly more conservative, especially in the realm of low-grade IPMNs and MCNs. While this study demonstrates that cytology provides a useful minimally invasive diagnostic tool that helps in surgical decision-making, it also highlights the limitations of cytological interpretation and sampling variability in FNA.

Data availability statement

Data will be available for sharing from a.zaitoun@nhs.net upon reasonable request.

Author contributions

Study design: AB, LA, DNL, and AMZ. Data collection: AB, MH, and AMZ. Data interpretation: all authors. Writing of manuscript: AB, DNL, and AMZ. Review of the manuscript: all authors. Critical review: DNL and AMZ.

Conflicts of interest

There are no conflicts of interest.

Abbreviations

• IPMN intraductal papillary mucinous neoplasm

• MCN mucinous cystic neoplasm

• SNOMED systematized nomenclature of medicine

• SPN solid pseudopapillary neoplasm

• PanNET pancreatic neuroendocrine tumor

• Pap Papanicolaou

• H&E hematoxylin and eosin

• WHO World Health Organization

• FNA fine needle aspiration

• GNAS guanine nucleotide-binding protein, alpha-stimulating activity polypeptide

• KRAS Kirsten rat sarcoma viral oncogene homolog

Funding Statement

This work was supported by the National Institute for Health Research Nottingham Biomedical Research Centre (grant number NIHR203310) and a charitable grant from the Legacy of Norton and Anne Collier. The funders had no role in the design or conduct of the work, or in the decision to publish. This paper presents independent research. The views expressed are those of the authors and not necessarily those of the funders, NHS or the Department of Health.