Interspecies comparative features of cholangiocarcinoma and histologic mimics

Step 1: determine presence or absence of mass effect (gross or histologic)

Mass effect, defined as an expansile nodule, served as the initial branching point. Lesions without mass effect typically were ductal plate malformations (DPMs) or responses to chronic biliary injury. When present, mass effect prompted further evaluation to distinguish among DPMs, benign proliferations, or true neoplasia.

Step 2: evaluate histomorphologic pattern (cystic vs. solid)

Among mass-forming lesions, cystic, non-invasive patterns favored benign entities, such as polycystic liver disease, hamartomas, or adenomas; solid masses required further distinction between epithelial and neuroendocrine tumors.

Step 3: immunohistochemical confirmation

IHC was essential for lineage determination of solid masses. Epithelial neoplasms had positivity for pan-cytokeratin (pan-CK), CK7, or CK19, whereas neuroendocrine tumors were synaptophysin-positive and CK-negative, allowing resolution of cases with pseudoglandular morphology. Although we evaluated only synaptophysin, additional neuroendocrine markers may be used depending on laboratory validation and availability. When lesions do not meet morphologic and/or immunohistochemical criteria for a defined entity, cases should exit the flowchart and be designated descriptively (e.g., “unclassified” or “indeterminate biliary carcinoma”) with documentation of differential diagnoses.

Step 4: evaluate features of malignancy

Epithelial lesions were assessed for malignancy using cytologic atypia, stromal response, and invasive growth, including parenchymal, lymphovascular, or perineural invasion.

Step 5: exclude metastatic disease, if possible

In hepatic tumors with atypical or mixed features, further evaluation of lesion distribution and IHC profiling was used to exclude metastatic adenocarcinoma, particularly from the pancreaticobiliary system, gastrointestinal tract, or other extrahepatic sites. Cases with widespread involvement or discordant primary-site morphology were classified as metastatic disease.

Of the 24 cases, 12 were confirmed as CCA, 10 were reclassified following histologic and immunohistochemical evaluation, and 2 remained unclassified ( Table 1 ). Relevant clinical data were recorded when available and cases are referenced by species and case number ( Suppl. Table 1 ).

OPEN IN VIEWER

Table 1.

Distribution of cholangiocarcinoma (CCA) subtypes and histologic mimics across taxonomic groups including morphologic subtypes and documented metastasis. Cases were retrospectively reviewed and reclassified based on histomorphology and immunohistochemistry.

Diagnostic category	Cases	Mammals	Reptiles	Avian	Metastasis documented
Large-duct CCA	7	4	2	1	2
Small-duct CCA	5	3	2	0	1
Biliary adenoma	1	0	1	0	0
Biliary hamartoma	3	0	3	0	0
Polycystic liver disease	1	1	0	0	0
Primary neuroendocrine tumor	2	2	0	0	1
Metastatic adenocarcinoma	3	3	0	0	3
Unclassified	2	0	0	2*	1
Total	24	13	8	3	8

*

Unclassified cases did not meet definitive morphologic and/or immunohistochemical criteria for a defined entity and were designated descriptively as indeterminate biliary carcinoma.

Cholangiocarcinoma

Most CCA cases were intrahepatic, with one extrahepatic case identified in the perihilar region of an olive baboon (Papio anubis; case 1). Grossly, intrahepatic CCAs were commonly well-demarcated, yellow-tan masses with umbilicate or fluid-filled centers, extending into adjacent liver parenchyma. In contrast to human cases, which are typically solitary and solid, ⁵ veterinary cases often had numerous lesions, likely reflecting delayed detection and secondary spread. In our cases, we classified CCA in animals using the 2 major subtypes recognized in humans: large-duct and small-duct.

Large-duct CCAs (cases 1–4, 8, 11, 12)

In large-duct CCAs, mucin-producing (PAS-positive) columnar epithelial cells formed lakes of intraluminal mucin ( Fig. 2A ). As outlined in the flowchart, these cases were identified at the decision point combining cystic or solid mass effect, CK positivity, and features of malignancy. Perineural and lymphovascular invasion was common (Fig. 2A, inset, 4 of 7 cases), and mitotic activity was higher (5–11 per 2.37 mm²) as was metastatic potential. This subtype was observed in a hamadryas baboon (Papio hamadryas), knob-tail gecko (Nephrurus amyae), Mertens’ water monitor (Varanus mertensi), speckled mousebird (Colius striatus), and 3 domestic shorthair (DSH) cats (Felis catus). In humans, large-duct CCAs are associated with more aggressive behavior and poorer prognosis, ² although this has not been validated in veterinary species. In our cohort, metastasis was identified in 2 of the 7 cases, both in domestic cats (cases 3, 4).

OPEN IN VIEWER

Figure 2.

Histologic features of cholangiocarcinoma (CCA) and histologic mimics. H&E, unless otherwise noted. A. Large-duct CCA, domestic shorthair cat, case 3. Moderately differentiated neoplastic ducts lined by mucin-producing columnar epithelial cells with intraluminal mucin and periductal neutrophilic infiltrates. Inset: higher magnification of malignant features, including perineural (arrowhead) and lymphovascular invasion (star). B. Small-duct CCA, red-tailed guenon, case 6. Narrow, well-differentiated ducts lined by cuboidal epithelial cells within a dense fibrous stroma. Inset: cytoplasmic cytokeratin 7 (CK7) positivity in neoplastic cells. Immunohistochemistry (IHC). C. Cholangiolocarcinoma-like CCA, savannah monitor, case 10. Anastomosing cords of small neoplastic ducts embedded in a loose fibrous stroma. D. Biliary adenoma, western hognose snake, case 19. Well-circumscribed, subcapsular lesion of uniform bile ducts within a pale fibrotic stroma. E. Von Meyenburg complex, blotched blue-tongued skink, case 20. Well-demarcated subcapsular area containing irregularly shaped, dilated biliary structures surrounded by abundant fibrous connective tissue. F. Polycystic liver disease, Ragdoll cat, case 13. Diffuse cystic biliary profiles lined by flattened biliary epithelium with minimal fibrous stroma and marked hepatocyte atrophy. G. Neuroendocrine tumor, mixed-breed dog, case 15. Monomorphic cells with round-to-oval nuclei arranged in tubular or acinar structures. Inset: neoplastic cells with strong cytoplasmic synaptophysin positivity. IHC. H. Metastatic adenocarcinoma, olive baboon, case 16. Stacked columnar epithelial cells resembling neoplastic glandular-like structures in the liver, with metastatic lesions observed in the mesentery, small intestine, and lymph nodes (not shown).

Small-duct CCAs (cases 5–7, 9, 10)

In small-duct CCA, well-differentiated, non–mucin-producing (PAS-negative) cuboidal epithelial cells were within a hyalinized or desmoplastic stroma, causing a mass effect and irregular invasion into adjacent liver parenchyma ( Fig. 2B ). These cases were classified based on a solid mass effect, CK positivity (inset), and histologic features of malignancy. This morphology was identified in a DSH cat, red-tailed guenon (Cercopithecus ascanius), Seba’s short-tailed bat (Carollia perspicillata), knob-tail gecko (Nephrurus amyae), and savannah monitor (Varanus exanthematicus). Notably, the CCA in the savannah monitor resembled cholangiolocarcinoma, a rare histologic variant of small-duct CCA described in humans, ² with anastomosing cords of neoplastic cells within a loose fibrous stroma ( Fig. 2C ). Whether such variation is prognostically relevant in animals remains uncertain. Three of 5 cases had multiple masses, and 1 DSH cat (case 5) had extrahepatic metastasis.

Biliary adenoma (cholangiocellular adenoma, case 19)

The case of biliary adenoma was a mass-forming lesion of cystic or solid, non-invasive bile ducts. A western hognose snake (Heterodon nasicus) had 2 discrete adenomas in subcapsular and intrahepatic locations. These were benign, well-circumscribed masses of uniform bile ducts with a fibrotic stroma ( Fig. 2D ). Unlike CCA, they lacked mitotic activity, nuclear pleomorphism, and invasive behavior.

Biliary hamartoma (cases 20–22)

Also known as von Meyenburg complex, biliary hamartoma is generally a small, non–mass-forming, and solitary malformation of abnormally formed ducts embedded in a fibrous mass, isolated from the biliary tree. When found in later stages, lesions may have a mass effect. In the diagnostic flowchart, these lesions are defined by cystic ductular profiles without features of malignancy. A solitary biliary hamartoma is typically considered an incidental lesion, whereas multiple biliary hamartomas can be compatible with a DPM. The prevalence of this lesion in veterinary species is unknown, but they have been reported in cats and reptiles.^12,14 Our cases included a blotched blue-tongued skink (Tiliqua nigrolutea), Mexican milk snake (Lampropeltis annulata), and green vine snake (Oxybelis fulgidus). Lesions were microscopic, well-demarcated, frequently subcapsular, variably cystic, irregularly formed ducts lined by attenuated cuboidal epithelium and supported by moderately hyalinized collagenous trabeculae ( Fig. 2E ).

Polycystic liver disease (case 13)

In polycystic liver disease (PCLD), mass-forming, multiloculated cystic lesions arise from aberrant ductal plate remodeling. PCLD follows the non-neoplastic, cystic branch but is distinguished from biliary hamartomas by the size, expansile nature, and marked mass effect of the lesions. Our case was in a Ragdoll cat (Felis catus), with large, fluid-filled cystic biliary masses that had a pronounced mass effect on adjacent liver parenchyma. Histologically, cysts were lined by flattened, monomorphic biliary epithelium and supported by thin, collagenous trabeculae with notable compression and atrophy of adjacent hepatocytes ( Fig. 2F ). ¹³ In humans, a subset of PCLD is accompanied by renal or pancreatic cysts or genetic associations such as PKD1 mutations. In our case, concurrent fibrocystic lesions in other organs were not identified, and genetic testing was not performed.

Primary neuroendocrine tumor (carcinoid, cases 14, 15)

Primary neuroendocrine tumors of the liver can closely mimic CCA, given the pseudoglandular or acini-like architecture, creating significant overlap on routine H&E evaluation. In the diagnostic flowchart, these lesions follow a mass-forming, solid architecture, but diverge at the immunohistochemical decision point. Our cases involved a DSH cat and a mixed-breed dog with hepatic neuroendocrine tumors that closely mimicked CCA with pseudoglandular architecture and malignant histologic features ( Fig. 2G ). Both tumors formed expansile, densely cellular masses of monomorphic neoplastic cells arranged in trabecular, nested, or pseudorosette-forming patterns, with high mitotic activity. Areas lacking pseudorosettes were particularly challenging and could be mistaken for well-differentiated CCA. Both neuroendocrine tumors were negative for cytokeratin markers (e.g., CK7, CK19, pan-CK) and were strongly synaptophysin positive (Fig. 2G, inset), confirming neuroendocrine origin.

Metastatic adenocarcinoma (cases 16–18)

Metastatic adenocarcinomas involving the liver can be morphologically and immunohistochemically indistinguishable from primary CCA, particularly when lesions are identified in the pancreaticobiliary system, major duodenal papilla, ampulla, or gastrointestinal tract. In the diagnostic flowchart, these lesions follow the mass-forming, epithelial branch with features of malignancy, but diverge based on lesion distribution in extrahepatic sites. Metastatic lesions were identified in an olive baboon, a rhesus macaque (Macaca mulatta), and a DSH cat. All 3 cases had invasive tubular or gland-forming architecture, mass effect, and IHC CK positivity, closely resembling CCA ( Fig. 2H ). However, extensive dissemination in multiple organs precluded identification of a definitive primary site, underscoring the diagnostic challenge posed by these tumors. CCA remains an important differential diagnosis, particularly when hepatic involvement is prominent.

Unclassified cases

Two avian cases remained unclassified (cases 23, 24) because of CK negativity and equivocal synaptophysin expression. Interpretation in these species is complicated by variable cross-reactivity of commonly used mammalian antibodies, the lack of species-validated markers limiting further classification, and the limitations of using a restricted number of markers. Morphologic findings were insufficient to support a definitive diagnosis without additional IHC confirmation, and the results were therefore interpreted with caution.

Our study offers a descriptive, multispecies overview of CCA and its major histologic mimics, highlighting both the diagnostic challenges and conserved morphologic patterns that emerge across taxa. Despite the broad species range, CCAs consistently had recognizable histomorphology, including division into large-duct and small-duct subtypes, characteristic stromal responses, and shared features of malignancy.

The recognition of large-duct and small-duct CCA morphologies in our cohort parallels subclassification schemes described in human medicine. In humans, cholangiocarcinogenesis is a multistep process shaped by various genetic, epigenetic, environmental, and inflammatory influences, with the cell of origin further contributing to morphologic heterogeneity. ² Large-duct tumors arise from biliary epithelial cells and peribiliary glands, whereas small-duct tumors arise from hepatic progenitor or cholangiolar cells. ¹⁰ Although metastatic disease was observed across both morphologic subtypes in our cohort, the small number of cases within each group and limited clinical data precluded meaningful conclusions regarding prognostic differences. Larger species-specific studies will be required to determine whether subclassification in veterinary species has prognostic significance comparable to that described in humans.

Across species, non-cystic tubular neoplasms posed the greatest diagnostic difficulty, overlapping with benign biliary proliferations, DPMs, and neuroendocrine tumors. Neuroendocrine tumors frequently had pseudoglandular patterns that were indistinguishable from well-differentiated CCA, underscoring the necessity of synaptophysin or other neuroendocrine markers to resolve cellular lineage. Metastatic adenocarcinomas from pancreaticobiliary or gastrointestinal tracts had tubular and infiltrative features nearly identical to primary hepatic CCA, reinforcing the importance of lesion distribution in classification.

The recurring decision points highlighted in our diagnostic flowchart, such as mass effect, cystic vs. solid architecture, and immunohistochemical markers (cytokeratin, synaptophysin), formed the basis for a structured diagnostic flowchart. The flowchart is intended as a structured interpretive aid rather than a definitive classification algorithm. Although descriptive, this framework is intended to promote a more consistent interpretation of proliferative biliary lesions across species.

Several limitations warrant consideration. Our small, taxonomically diverse caseload limits species-specific conclusions regarding behavior, prognosis, or grading schemes. IHC interpretation was also constrained in non-mammalian species because of variable antibody cross-reactivity. Ancillary testing, such as expanded IHC panels or ultrastructure evaluation, may be considered, if available, to further evaluate lineage differentiation. Moreover, we did not assess interobserver agreement nor formally validate the diagnostic flowchart, both of which will require larger, more homogeneous datasets.

Within these constraints, we identified shared diagnostic features and pitfalls that complicate interpretation of biliary proliferations. Together, our findings emphasize the comparative value of biliary tumors, particularly given the rarity of CCA in most veterinary species. By organizing recurring decision points into a structured approach, we offer a practical foundation upon which future comparative and species-focused investigations may build.