Diagnostic Utility of Cytological Examination in Neoplastic Lesions of Pancreas and Biliary Tract
Vijayalakshmi Gudivada, MD, FRCpath 1*, Siddaraju Raju, MD 1, Sree Rekha J, MD 1,
Debasis Gochhait, MD, DNB 1
*Correspondence to: Vijayalakshmi Gudivada. Specialist Clinical and anatomic pathologist, Premier Diagnostic Center, Prime Healthcare Group, Dubai, United Arab Emirates.
Copyright
© 2026 Vijayalakshmi Gudivada. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Received: 19 October 2026
Published: 01 November 2026
DOI: https://doi.org/10.5281/zenodo.23907994
Abstract
Cytology plays a crucial role in the diagnosis of neoplastic lesions within the pancreatobiliary system. This prospective diagnostic test accuracy study aimed to evaluate the effectiveness of pre- and intra-operative cytology in diagnosing these lesions. Over two years, cytology samples were obtained from 166 patients with clinicoradiological suspicions of pancreaticobiliary pathology, using histopathology as the gold standard. The study found a sensitivity of 75% and a positive predictive value (PPV) of 96% for fine-needle aspiration cytology (FNAC) in pancreatic lesions. In gallbladder malignancies, FNAC demonstrated a sensitivity of 100%. However, the sensitivity for ampullary and cholangiocarcinoma lesions was comparatively lower, at 40% and 50%, respectively. Overall, cytological examination yielded a diagnostic sensitivity of 52% and a PPV of 98%. Thus, FNAC emerges as a potent diagnostic tool for pancreatic and gallbladder lesions, while bile cytology is vital for inaccessible bile duct malignancies.
Key words: Cytology, FNAC, Pancreatico-biliary tract, sensitivity
Introduction
A variety of benign and malignant lesions involving the pancreaticobiliary tract, such as strictures and mass lesions, are commonly encountered in routine clinical practice. These lesions clinically manifest as jaundice, pruritus, abdominal pain, anorexia, weight loss, and steatorrhea, while biochemically with abnormal liver enzymes. Early detection of these lesions, especially cancers, is highly critical. Although detailed clinical examination and imaging modalities are highly effective in diagnosing these diseases, documentation of tissue diagnosis is crucial in the management of patients. In patients with inoperable tumors, a tissue diagnosis is essential to preclude unnecessary surgery and initiate chemo- and/ or radiotherapy. In patients who are surgical candidates, a preoperative tissue diagnosis assists in optimal planning of surgery. Depending on the nature and accessibility of pancreaticobiliary lesions, biopsy or a variety of cytological sampling techniques can be utilized. When biopsy is not possible, cytological assessment in combination with clinico-radiological examination remains the only and preferred modality of documenting tissue diagnosis.
An improved understanding of premalignant lesions and the advent of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) emphasize the importance of cytopathologists in the management of patients with strictures/ growth involving the pancreaticobiliary tract and solid or cystic lesions of the pancreas.[1]
Pancreaticobiliary malignancies are the fourth leading cause of death in developed countries. Between 2006 and 2010, the incidence of pancreatic cancer increased by 1.3 percent per year, and the death rate increased by 0.4 percent per year.[2] New-onset diabetes mellitus and acute onset of pancreatitis in elderly people should raise the suspicion of pancreatic malignancy and should be correlated with radiological findings.[3]
Aim of the Study
To assess the diagnostic power of pre- and intra-operative cytology in the diagnosis of neoplastic lesions of the pancreas and biliary tract.
Materials and Methods
The present study is a diagnostic test accuracy study, which was conducted in the department of Pathology, Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER), Puducherry, for a period of 17 months from September 2015 to January 2018, after obtaining the approval of the Institute Ethics Committee.
Patients with clinical or radiological suspicion/diagnosis of neoplastic and non-neoplastic lesions of the pancreas and biliary tract were included in the study.
Study Procedure
Cytological samples, such as pre-operative image-guided or intraoperative FNAC, bile cytology, and brushing cytology, were taken for neoplastic and non-neoplastic lesions of the pancreaticobiliary tract. A total of 166 samples were collected, of which 60 are non-neoplastic from the pancreas and 106 are neoplastic lesions from the pancreas, gall bladder, and biliary tract. There were no non-neoplastic lesions in the gall bladder and biliary tract. There was a male preponderance with an overall mean age of 52 years for the entire spectrum of pancreaticobiliary lesions. The chief clinical manifestations of patients with pancreatic lesions were abdominal pain with loss of appetite and weight (82%), and jaundice (46%); 36% patients gave a history of smoking and alcoholism. Eleven patients had a previous history of chronic pancreatitis.
Biochemical values were available for seven patients with pancreatic lesions, of whom four had increased CA-19.9 levels, while three patients had increased serum amylase levels. The clinical presentation of patients with gall bladder lesions was abdominal pain with loss of appetite and weight (70%) and jaundice (52%); 10% of patients had a history of smoking and alcoholism. The major clinical manifestations among the patients with biliary tract lesions were obstructive jaundice (97%) and abdominal pain with loss of appetite and weight (55%), while 15% had a history of smoking and alcoholism.
Cytological sampling was done according to the type and accessibility of the lesion. FNAC was done in mass lesions of the pancreas, proximal intrahepatic biliary tract, ampulla, and gall bladder. Bile fluid and brush cytology was taken in biliary strictures, gall bladder, and a few cases of pancreatic malignancies. There were 105 pancreatic lesions. Intraoperative FNAC was done in 60 non-neoplastic lesions, which included clinically and radiologically diagnosed cases of acute and chronic calcific pancreatitis, and the remaining 45 neoplastic lesions involving the pancreas are clinicoradiologically diagnosed/suspected benign, premalignant, and malignant lesions in which intraoperative FNAC was done in 18 patients, while 22 patients were subjected to preoperative FNA under USG; in one patient, FNA was done under CT guidance. Bile was collected through ERCP for cytological evaluation in four patients with pancreatic malignancy. Cytology samples were collected in 23 patients with gall bladder lesions; of these, 22 had thickening of the gall bladder wall, and one presented with a polypoidal lesion in the mucosa. USG-guided FNA was done in 14 patients, and bile cytology was collected in six patients with suspected gall bladder malignancy. USG-guided FNA was done in four patients with suspected hilar cholangiocarcinoma and two patients with suspected hepatocellular carcinoma (HCC). Intraoperative FNAC was done in eight patients with ampullary mass lesions. Bile fluid was evaluated in a total of 19 patients, of whom nine were suspected of having hilar cholangiocarcinoma, seven with ampullary lesions, and three with distal CBD strictures. Combined bile fluid and brush cytology was examined in five patients, of whom three had distal CBD strictures and the other two had hilar lesions.
Smears were reported according to the Papanicolaou Society of Cytopathology Guidelines into the following categories.[4]
Sixty non-neoplastic lesions were reported as Category-II of the Papanicolaou Society of Cytopathology Guidelines and were taken as ‘true negatives (TN)’. The remaining 45 neoplastic lesions, six were non-diagnostic with only blood elements (Category I). Thirteen cases were ‘Negative for malignancy’ (Category II), of which seven were cases of chronic pancreatitis. Three were granulomatous lesions with Langhans giant cells, one with proven Ziehl-Neelsen (ZN) stain, positive for acid-fast bacilli (AFB) (Figure1a,b), and reported as pancreatic tuberculosis. One case was interpreted as Category-III ‘Atypical’. There were three cases interpreted under the category IV, i.e., benign lesions of the pancreas; of these, two were solid pseudo papillary neoplasms (SPN) (Figure2e), and the third was a case of serous cystadenoma, which was diagnosed by scrape cytology. Four cases were reported under the category V, i.e., ‘Suspicious for malignancy’. Eighteen cases were ‘positive for malignancy’ (Category VI), of which 12 were moderately differentiated ductal adenocarcinomas, one was poorly differentiated carcinoma and two were mucin secreting adenocarcinomas; the remaining cases comprised, a case each of papillary adenocarcinoma, well differentiated carcinoma and low grade spindle cell lesion with possibility of gastrointestinal stromal tumor (GIST)/Malignant peripheral nerve sheath tumor (MPNST). Histopathological correlation was available in 19/45 cases. In the remaining 26 cases, clinicoradiological suspicion/diagnosis of ‘pancreatic neoplasm’ was taken as the gold standard. All these cases were considered as ‘true positive’ clinicoradiologically for data analysis. Of 19 histopathologically correlated cases, 11 were true positives and five ‘true negatives’.
There were a total of 23 gall bladder lesions. Eight were categorised as ‘Negative for malignancy', of which seven were from bile fluid cytology and one was an acute suppurative lesion on image-guided FNAC. FNAC from one polypoidal lesion was diagnosed as ‘benign epithelial lesion’ (Category IV) (Figure2g). Two were interpreted as Category V (Suspicious for malignancy). Twelve were reported as ‘Positive for malignancy’, of which 10 were diagnosed as adenocarcinoma on FNAC (Figure1d) and one as adenocarcinoma on bile fluid cytology, while the twelfth case was diagnosed as an adenosquamous carcinoma (Figure2f) on both FNAC and bile fluid cytology. As most gall bladder malignancies were unresectable, histopathological correlation was available only in four cases. Of these, one was a true positive and three were false negatives.
There were 38 cases of biliary tract lesions. Three of the USG-FNAs were non-diagnostic. Twenty-two were ‘Negative for malignancy’. No cases were reported under the ‘Atypical’ and ‘Benign’. Two bile fluid samples and one intraoperative FNA were categorised as ‘Suspicious for malignancy’. Twelve cases were reported as ‘Positive for malignancy’, of which three were diagnosed on image-guided FNAC; two of them were confirmed to be ‘intrahepatic cholangiocarcinoma’ on ICC. The third was diagnosed as ‘hilar cholangiocarcinoma’. Three cases were reported as adenocarcinoma on intraoperative FNAC from ampullary lesions (Figure1c). Bile fluid was positive for malignancy in three cases and brush cytology in one case (Figure2h). Histopathological correlation was available in 17 of the biliary tract lesions. Of these, six were ‘true positives”. There was a single true negative case, which was negative for malignancy both on bile cytology and histopathology. There were a total of 10 false negative cases, which included seven bile fluid and three FNA samples interpreted as ‘Negative for malignancy’ but histologically turned out to be adenocarcinoma in all ten cases.
The diagnostic utility of cytology in pancreaticobiliary lesions varied according to the site of the lesion and mode of sampling. For pancreatic lesions, the intraoperative FNAC showed sensitivity, specificity, and accuracy of 86.6%, 66.6% and 83.3% respectively. Pre-operative sampling by image-guided FNAC was found to be less sensitive (52.6%) but highly specific (100%) with a diagnostic accuracy of 60%. Bile cytology in pancreatic lesions had a 50% sensitivity in diagnosing malignant lesions. For gall bladder lesions, the sensitivity of USG-FNAC was 82%, whereas bile cytology had a very low sensitivity of 25%. As there were no non-neoplastic lesions to consider as true negatives, specificity and accuracy were not calculated. Pre and intraoperative FNAC in biliary tract lesions showed a sensitivity of 50%. As the sample size of brush cytology is small, combined bile and brush cytology sensitivity was calculated, which showed a sensitivity of 25% with a specificity of 100% and diagnostic accuracy of only 37%.
In pancreaticobiliary lesions with histopathology correlation, the sensitivity was 54.5%, while specificity was 85.7% with a PPV, NPV of 94.7%, 29% and 60% respectively. In pancreaticobiliary lesions with clinicoradiology correlation, the sensitivity was 53% and specificity was 100% respectively, with a PPV, NPV, and diagnostic accuracy of 100%, 66% and 75.3% respectively.
Discussion
The role of tissue diagnosis need not be overemphasized in patients with suspected malignancies, especially with inoperable ones. Although biopsy is a preferred method for histological diagnosis, it has its limitations not only due to the inaccessibility of pancreaticobiliary structures, but also due to serious complications associated with it.
Pancreas
The false negativity of pancreatic FNACs is more in image-guided FNACs than intraoperative FNACs, and it could be attributed to inadequate sampling through image guidance, the retroperitoneal location of the pancreas, and the low cell yield, which can be due to desmoplastic stroma of the tumor tissue and sometimes high vascularity.[5]
Clinicoradiologically, a non-neoplastic lesion like chronic calcific pancreatitis can mimic malignancy. Often, there can be overlap between chronic pancreatitis and well-differentiated pancreatic adenocarcinoma, leading to misinterpretation, resulting in a false positive or false negative diagnosis. Reactive epithelial changes in chronic pancreatitis, variable degree of differentiation in pancreatic adenocarcinoma, improper sampling of the lesion, drying artefact, a low cellular yield due to desmoplastic stroma, and the gastrointestinal epithelial contaminant are some of the reasons for misinterpretation.[6] One of our false positive cases interpreted as ‘well differentiated pancreatic adenocarcinoma’ is of relevance in this context. Interestingly, we encountered three rare cases of granulomatous lesions of the pancreas on FNAC. All these cases had been clinicoradiologically interpreted as malignancy. One of these cases was an intraoperative FNA, which was proven to be pancreatic tuberculosis by AFB positivity. The pancreas is one of the few organs that are resistant to tuberculosis.[7,8] The unusually rare cases of pancreatic tuberculosis generally manifest like acute/chronic pancreatitis or pancreatic abscess.[9] Chronic pancreatitis is said to alter the enzymatic milieu, predisposing to the secondary infection of the organ.[7] Local immunosuppression due to chronic pancreatitis is the other likely cause of reactivation of the latent tuberculous foci. This case emphasizes not only the utility of pancreatic FNA but also the limitation of radiological techniques, at least in selective instances.
Under benign category cases, SPN had characteristic cytomorphologic and cytoarchitectural features with a high cellularity displaying papillary structures and a monotonous population of cells showing nuclear grooving; one case showed tumor cells with plasmacytoid morphology. The case of serous cystadenoma was suspected to be a retention cyst on radiological examination, but the scrape smears taken from the frozen section specimen were found to be useful in making the tissue diagnosis.
All seventeen pancreatic adenocarcinomas (true positives) diagnosed on cytology were cellular with clusters of malignant cells exhibiting hyperchromatic nuclei and irregular nuclear membranes, with a few cells displaying prominent nucleoli. Most of them were interpreted as ductal adenocarcinomas. One of the ductal adenocarcinomas showed a prominent stroma and also a rare fragment of perineural infiltration that was appreciable on FNA smears.
FNAC smears of one case were highly particulate with several spindle cell fragments having scanty intervening collagenous matrix; the cells showed spindly nuclei with a mild to moderate anisonucleosis and the background showed a few patches of magenta matrix. ICC performed for S100, CD117, and CD34 was non-contributory and therefore, was interpreted as a low-grade spindle cell lesion with a differential diagnosis of GIST and MPNST.
FNA smears in a case interpreted under the ‘ATYPICAL’ category showed significant crush artefact with some better preserved cells exhibiting a mild nuclear pleomorphism, owing to which a definitive opinion on the malignant or benign nature could not be made. Three cases reported under the ‘SUSPICIOUS’ category had insufficient cellularity but with marked nuclear pleomorphism and hyperchromasia. As the progression of malignancy in atypical and suspicious categories was 44% and 86% respectively, these were considered as true positives.[10]
Bile cytology for pancreatic malignancies is not a well-described diagnostic tool. However, as per some of the studies, about 20 to 35% pancreatic malignancies have been detected on bile cytology. In our study, two out of four cases were positive in bile cytology.[11]
Gall bladder
USG-FNA showed 82% sensitivity in the present study. Cases under positive for malignancy as ‘adenocarcinoma’ showed varied cellularity with clusters of cells exhibiting a glandular pattern, moderate to marked pleomorphism, vesicular nuclei, and prominent nucleoli. Few showed both intracytoplasmic and extracellular mucin and a papillary pattern.
This study recorded a single case of adenosquamous carcinoma in both FNA and bile cytology. Along with a glandular pattern, the smears showed well-preserved malignant cells exhibiting squamous differentiation in a necrotic/ hemorrhagic background. Squamous differentiation, though uncommon in the gall bladder, has been reported in literature to often occur as a result of ‘metaplastic’ change, or as a part of ‘dedifferentiation’.[12,13] We had also encountered a case of benign epithelial lesion which presented as a polyp, smears of which are highly cellular with several monolayered strips and fragments of cuboidal to columnar epithelium without atypia arranged in honeycomb and papillary patterns.
As compared to USG-FNA, the bile cytology showed only 25% sensitivity in our study. Smears were sparse to moderately cellular with occasional tiny clusters of benign epithelial cells and /or bile pigment and debris. Only two out of ten cases were positive for malignancy.
Exfoliation of malignant cells into the bile fluid may occur intermittently; hence, a daily or an alternate-day bile fluid examination is recommended to increase its sensitivity. We could not perform repeat bile fluid cytology in any of our cases due to difficulty in following up with patients. Enzymatic digestion and degeneration of cells in bile fluid were also responsible for scant cellularity.[14]
Studies have highlighted the diagnostic utility of USG FNAC in gall bladder lesions with a sensitivity ranging from 70 to 97% and a specificity range of 90 to 100% with a diagnostic accuracy between 86 and 88%. The sensitivity of USG FNAC in our study for gall bladder lesions was 92%. The underinterpretation of ‘suspicious for malignancy’ in some of our cases was due to scant cellularity attributable to sampling and technical errors.[15,16]
The low 25% sensitivity of bile cytology for gall bladder lesions in our study is in accordance with 4 to 50% sensitivity documented in the literature. Single bile samples are of inherently lower sensitivity than the brush or FNA cytology [17]. The advantage of FNAC is that the needle can be accurately directed to the precise site of abnormality.
Biliary tract
The mode of cytological sampling depended on the level of the biliary tract involved by the malignancy. Malignancies are more common in the intrahepatic and proximal bile ducts, and they present as hypodense mass lesions infiltrating into the liver. The malignancies involving the remaining part of the bile duct and distal CBD present as strictures, while most ampullary lesions present as smaller mass lesions with obstructive jaundice.
Due to the complex location and smaller size, an image-guided FNA is not feasible in most of these cases. Intraoperative FNAC was for ampullary mass lesions, and three out of eight cases were cytologically ‘positive for malignancy’. One more case was interesting; it was diagnosed as positive for malignancy due to highly cellular smears showing discrete as well as small clusters of highly pleomorphic cells exhibiting prominent nucleoli and a moderate amount of cytoplasm, along with several strips of intestinal epithelium. However, histopathology showed well-organized and back-to-back arrangement of epithelial units exhibiting tubular, papillary, and the combined tubular/papillary architecture with minimal intervening stroma with moderate dysplasia. This case was finally diagnosed as an ‘intra-ampullary tubulopapillary neoplasm’, which is a premalignant lesion. Though cytologically misinterpreted as ‘adenocarcinoma’ due to marked pleomorphism (owed to dysplastic cells), it was taken as ‘true positive’ owing to its neoplastic/ premalignant nature [18].
The sensitivity of intraoperative FNAC for ampullary lesions was only 50%. The low sensitivity can be attributed to the smaller size of ampullary lesions and contamination of the GI epithelium (while passing the needle through the duodenal wall to reach the ampulla) during the FNA procedure, resulting in a false negative diagnosis. We could not trace any literature concerning the diagnostic accuracy of intraoperative FNAC of the biliary tract lesions.
The sensitivity of bile cytology for the proximal bile duct lesions was 55%. As per the literature, the sensitivity of bile cytology ranges from 6 to 50%, with the repeat bile fluid sampling increasing it up to 74%. As stated earlier, the bile fluid sampling could not be repeated in our cases, and the reasons for low sensitivity can be due to the submucosal location of the lesion, low cell yield, tight stenosis due to fibrosis/desmoplasia/inflammation, and technical errors.[19]
The literature reports the sensitivity of brush cytology in the range of 40 to 70% and specificity up to 100% with a diagnostic accuracy of 70%. For malignant strictures, the sensitivity of bile cytology ranges from 50 to 60% and the specificity from 98 to 100%. The sensitivity of image-guided FNAC for intrahepatic lesions has been reported to be around 90% (range: 67% -100%) with 100% specificity. The sensitivity of FNAC (inclusive of image-guided FNA in the proximal biliary tract and intraoperative FNA in the distal biliary tract) was 50%, whereas bile and brush cytology showed a very low sensitivity of 25% and a specificity of up to 100%; the diagnostic accuracy of FNA was 37% .[20,21]
Diagnostic utility
Only 40 of the total 166 pancreaticobiliary lesions (inclusive of both neoplastic and non-neoplastic) had the histopathological gold standard, while in the remaining 126 cases, clinicoradiological impression /diagnosis was taken as the gold standard. Therefore, the diagnostic utility of cytological methods was calculated separately for cases with histopathology and clinicoradiological diagnoses as gold standards. When histopathological diagnosis was taken as the gold standard, the sensitivity was 54.5%, while the specificity was 85.7% with a PPV, NPV of 94.7%, 29% and 60% respectively. When clinicoradiological diagnosis was taken as the gold standard, the sensitivity was 53% and specificity was 100% respectively, with a PPV, NPV, and diagnostic accuracy of 100%, 66% and 75.3% respectively.
Limitations
Conclusion
Competing Interest Statement by All Authors
There is no competing interest from any of the authors.
Authorship Statement by All Authors
Dr. Vijayalakshmi-Collection of data and preparation of the manuscript including histopathology, imaging, and review of the manuscript.
Dr. Siddaraju, Dr. Sreerekha, Dr. Debasis Gochait- Review of the manuscript and cytodiagnosis.
All authors have contributed to the diagnosis, management and the manuscript preparation.
Ethics Statement by All Authors
Consent was obtained for FNAC and permission was obtained for publication. No identification details of the patients were disclosed.
Acknowledgment
We acknowledge the Departments of Medical and Surgical gastroenterology for their contribution to the management of these cases.
List of Abbreviations
AFB- Acid-fast bacilli
CBD- Common bile duct
CECT - Contrast Enhanced Computerized Tomography
EUS-FNA- Endoscopic ultrasound-guided fine needle aspiration
ERCP- Endoscopic Retrograde Cholangiopancreatography
FNAC - Fine needle aspiration cytology
GIST- Gastrointestinal stromal tumor
H&E- Hematoxylin and Eosin
IHC- Immunohistochemistry
MGG- May-Grünwald Giemsa
MPNST- Malignant peripheral nerve sheath tumor
PPV- Positive predictive value
NPV- Negative predictive value
USG- Ultrasonography
ZN- Ziehl Neelsen
References