Collecting duct carcinoma

• Rare, aggressive renal carcinoma (Bellini duct carcinoma): <1% of RCCs.
• High-grade cells in infiltrative tubular, tubulopapillary or glandular architecture.
• Medulla-centred, often extending into cortex, renal sinus and perinephric fat.
• Irregular angulated tubules in a prominent desmoplastic stroma, often with neutrophils.
• Cuboidal, columnar or hobnail cells with irregular nuclei and prominent nucleoli.
• No defining molecular alteration, a diagnosis of exclusion.
• Key mimics: FH-deficient RCC, SMARCB1-deficient medullary carcinoma, upper tract urothelial carcinoma.

Diagnostic pearl: a medulla-centred high-grade adenocarcinoma with angulated tubules and desmoplasia suggests CDC, but is not enough on its own.

Architecture

• Irregular, angulated, variably sized tubules and glands, elongated, branching or poorly formed, dissecting native structures.
• Tubulopapillary and solid areas may coexist.

Cells

• Cuboidal to columnar, amphophilic/eosinophilic cytoplasm; hobnail cells.
• Enlarged, hyperchromatic nuclei, prominent nucleoli, mitoses.

Stroma & other

• Prominent desmoplasia with oedema, hyalinisation and neutrophils.
• Mucinous or eosinophilic luminal secretions; necrosis; lymphovascular and perineural invasion; sinus extension.

Diagnostic pearl: infiltrating glands + hobnail cells + desmoplasia is characteristic but not pathognomonic.

• Infiltrative tubular: the most characteristic pattern.
• Tubulopapillary: may resemble FH-deficient RCC or high-grade PRCC.
• Complex glandular: crowded, fused glands with secretions or mucin.
• Solid / nested: demands careful exclusion of medullary carcinoma.
• Reticular / cribriform-like: extensive reticular or yolk sac-like growth with rhabdoid cells suggests SMARCB1-deficient medullary carcinoma.
• Sarcomatoid: aggressive dedifferentiation; document it.

Diagnostic pearl: reticular, solid/rhabdoid or unusual papillary morphology should broaden the differential.

• Hobnail cells: apically protruding nuclei; distal-nephron-like, not specific.
• High-grade eosinophilic cells: big nuclei and nucleoli.
• Clear cell change: focal; extensive clearing → exclude ccRCC.
• Rhabdoid-like cells: prominent rhabdoid cytology → check INI1.
• Desmoplasia: conspicuous at low power.
• Neutrophils: also prominent in medullary carcinoma; not a discriminator.

Diagnostic pearl: desmoplasia and inflammation flag an infiltrative high-grade carcinoma but cannot separate CDC from FH-deficient RCC or RMC.

• Medullary location: supportive, but large tumours obscure the origin.
• Infiltration of native structures: entrapped benign tubules between malignant glands.
• Renal sinus involvement: frequent; may mimic upper tract urothelial carcinoma, check the urothelial lining.
• Intraductal atypia: dysplastic adjacent collecting ducts are supportive, not required or specific.

Diagnostic pearl: a tumour in the medulla is not automatically a collecting duct carcinoma.

Establish renal lineage and exclude mimics, no marker identifies CDC.

• PAX8: usually positive (also in a minority of urothelial carcinomas).
• CK7: often positive, variable, nonspecific.
• GATA3, p63/p40: usually negative · against urothelial carcinoma.
• SMARCB1 / INI1: retained · loss → SMARCB1-deficient carcinoma.
• FH / 2SC: FH retained, 2SC negative · against FH-deficient RCC (FH may be retained despite deficiency).

18 CDC vs 25 upper tract UC: GATA3 1/18 vs 22/25; p63 2/18 vs 23/25; PAX8 18/18 vs 3/25.

Diagnostic pearl: PAX8, GATA3/p63, INI1 and FH/2SC matter more than a broad cytokeratin panel.

• No defining driver: heterogeneous alterations; old cohorts included tumours now called FH-deficient or SMARCB1-deficient.
• NF2 inactivation in a subset.
• SETD2 mutations in small cohorts.
• TP53, cell-cycle changes, homozygous CDKN2A deletion.
• SMARCB1: complete loss should prompt reclassification.
• KRAS codon 12 mutations in 3/13 (2024).

Diagnostic pearl: molecular testing is most useful to exclude defined mimics, not to confirm CDC.

• Shares medullary location, high grade, desmoplasia and inflammation.
• Associated with sickle cell trait; adolescents and young adults (CDC: older adults).
• Reticular, cribriform, microcystic or yolk sac-like growth; rhabdoid cells; sickled erythrocytes.
• Complete INI1 loss is the key finding; CDC retains INI1; PAX8 overlaps.
• SMARCB1-deficient tumours without known haemoglobinopathy are not automatically CDC.

Diagnostic pearl: a young patient with medullary carcinoma and rhabdoid or reticular growth needs INI1, whatever the history.

• Infiltrative tubules, papillae, high-grade nuclei and desmoplasia, a critical mimic.
• Heterogeneous architecture: papillary, tubulocystic, cribriform, solid.
• Viral inclusion-like nucleoli with perinucleolar halos, may be focal or absent.
• FH loss / 2SC positivity; some retain FH protein despite functional loss.
• HLRCC implications for the patient and family.

~25% of tumours initially called CDC in a multi-institutional cohort were FH-deficient RCC; a 2022 re-evaluation of 18 high-grade infiltrative carcinomas found 7 FH-deficient, 9 CDC, 2 SMARCB1-deficient.

Diagnostic pearl: actively exclude FH deficiency, especially with striking nucleoli, mixed patterns or young age.

• Can infiltrate medulla and sinus with glandular or poorly differentiated morphology.
• Continuity with pelvicalyceal urothelium, urothelial CIS or a conventional UC component favours urothelial origin.
• CDC: PAX8+, GATA3−, p63/p40−. UC: GATA3+, p63/p40+, uroplakin+, PAX8 usually negative.
• High-molecular-weight keratins are not discriminatory; both retain INI1.

Diagnostic pearl: PAX8 positivity does not exclude UC, and GATA3 negativity does not exclude poorly differentiated UC.

• ALK-rearranged RCC: infiltrative, cribriform or solid, sometimes with sickle trait; ALK IHC screens, fusion confirms (therapeutic relevance).
• High-grade MTSCC: residual elongated tubules, spindle cells, mucin; multiple chromosomal losses.
• Papillary RCC: a defined papillary component without widespread destructive desmoplastic infiltration.
• Metastatic adenocarcinoma: GI, pancreatobiliary, lung; imaging and site-directed IHC.
• Unclassified RCC: when no subtype fits after full work-up.

Diagnostic pearl: CDC is reached by exclusion, not by accumulating nonspecific features.

• 1 · Pattern: irregular tubules/glands in desmoplastic stroma; medulla- or sinus-centred?
• 2 · Spectrum: hobnail cells, tubulopapillary growth, secretions, inflammation, sarcomatoid or rhabdoid features.
• 3 · Renal vs urothelial: PAX8, GATA3, p63/p40.
• 4 · Exclude mimics: INI1 loss → SMARCB1-deficient · FH loss/2SC → FH-deficient · ALK → ALK-rearranged.
• 5 · Diagnose CDC only after exclusion.

On biopsy

• Medullary origin and architecture are hard to establish; absent desmoplasia does not exclude CDC.
• Prefer “high-grade renal carcinoma with infiltrative tubular/glandular morphology” until mimics are excluded.

Diagnostic pearl: CDC should not become a catch-all for unclassified high-grade renal adenocarcinoma.

• Narrowing boundaries: FH-deficient, SMARCB1-deficient and ALK-rearranged RCC have been carved out of historical CDC.
• KRAS: G12A/G12D/G12V in 3/13 (23%) in a 2024 EMBO Mol Med study, alongside TP53 and NF2.
• Aristolochic acid: SBS22 signature in 3/13 tumours; suggests an environmental contribution in some populations.
• Cell-cycle dysregulation: CDK9 inhibition slowed growth in preclinical models.
• IHC surrogates: INI1, FH/2SC and ALK remain central, because they establish alternative diagnoses.

• Rare, aggressive, medulla-centred; infiltrative tubules/glands in desmoplastic stroma.
• Hobnail cells with big nuclei and nucleoli; papillary, solid, cribriform-like and sarcomatoid patterns occur.
• Neutrophils are supportive, not specific.
• PAX8+, GATA3−, p63/p40−, but PAX8 does not exclude UC.
• INI1 retained; loss → SMARCB1-deficient carcinoma.
• FH/2SC to exclude FH-deficient RCC; FH protein may be retained.
• ALK-rearranged RCC can mimic CDC.
• ~25% of historical CDCs were FH-deficient RCC.
• KRAS, NF2, TP53, SETD2 reported; none diagnostic.
• A diagnosis of exclusion, especially on biopsy.