> For the complete documentation index, see [llms.txt](https://www.parapathology.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.parapathology.com/appendix/clippings/pancreatic-ductal-adenocarcinoma-and-its-subtypes-2026-who-classification-virchows-archiv.md).

# Pancreatic ductal adenocarcinoma and its subtypes: clinical relevance of histopathology and molecular characterization, integrating the key updates of the 2026 WHO classification

Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal human cancers, due to its late clinical presentation, early vascular, perineural, nodal and distant dissemination, and profound resist

### Summary

A single-author review (Claudio Luchini, University of Verona) in *Virchows Archiv* that synthesizes the histopathology and molecular pathology of **pancreatic ductal adenocarcinoma (PDAC)** and its subtypes, structured around the updates in the 2026 edition of the WHO Classification of Tumours ("Blue Book"). The digest below is written in my own words; the verbatim abstract is preserved in the `description:` frontmatter, and the full text (figures, tables, references) is in the open-access source.

> Keywords: Pancreatic ductal adenocarcinoma · PDAC · Pancreatic cancer · Pancreas · WHO.

### Why it matters

PDAC is the 4th (men) / 3rd (women) leading cause of cancer death; \~80% of patients present with locally advanced or metastatic disease, and outcomes have improved only minimally in decades. \~60–65% arise in the pancreatic head. The review argues that integrating histology with molecular and immunological profiling is now essential — while noting only a small fraction of PDACs harbor currently actionable alterations.

### Pathogenesis and molecular drivers

* **Multistep model:** acinar-to-ductal metaplasia → **PanIN** (low-/high-grade dysplasia) → invasive carcinoma. >80% develop from microscopic PanIN precursors; \~15–20% from macroscopic (>1 cm, often cystic/intraductal) precursor lesions. "Ductal" is a phenotypic label, not proof of an exclusively ductal cell of origin.
* **The four "genetic mountains":**
  * **KRAS** — mutated in >90% (codons 12/13/61); earliest driver, required for initiation and maintenance.
  * **CDKN2A** (p16) — inactivated in \~95% (homozygous deletion \~40%, LOH+mutation \~40%, promoter hypermethylation \~15%).
  * **TP53** — inactivated in 70–75%.
  * **SMAD4** (DPC4) — inactivated in 50–55%.
* **"Genetic hills":** homologous-recombination repair genes (BRCA1/2, PALB2), chromatin remodelers (ARID1A, KDM6A, PBRM1), TGF-β pathway (TGFBR2). **MTAP** loss (9p21, near CDKN2A) is a potentially actionable vulnerability under therapeutic exploration.

### Histopathology of conventional PDAC

Typically scar-like, highly fibrotic, hypocellular tumors dominated by a **desmoplastic stroma** (collagen, fibroblasts/myofibroblasts, scant immune cells); true neoplastic epithelium is often a minority of the mass. **Eight diagnostic hallmarks** of infiltrative glands help distinguish PDAC from reactive ducts:

1. Incomplete glandular borders / irregular lumina
2. Intraluminal necrotic debris (± granulocytes)
3. Marked nuclear size variation within one gland exceeding 1:4 — the **"4:1 rule"**
4. Mitoses, especially atypical figures
5. A gland immediately adjacent to a muscular vessel
6. A tubule/duct in an unequivocally abnormal location (e.g. "naked glands in fat")
7. Vascular invasion
8. Perineural invasion

Even isolated findings may support a diagnosis of invasive PDAC on small biopsies.

### Why PDAC is so deadly

* **Vascular tropism:** invades vessels even when very small; in veins, tumor cells often *replace the endothelium from within* (relining the wall), exploiting an immune-privileged lumen for local extension and hematogenous spread.
* **Perineural invasion** in >70% of resected PDACs; extends beyond the pancreas.
* **Immunologically "cold":** low T/B lymphocytes, FOXP3+ Treg enrichment, an "immune-escape" microenvironment associated with poorer survival.
* **Early metastasis:** nodal metastases in 70–80% of treatment-naïve resections, often with extranodal extension (\~60% of node-positive cases). Liver is the most frequent distant site (also lung, bone, adrenal).

### Grading

Well (G1) / moderately (G2) / poorly (G3) differentiated, integrating glandular differentiation, mucin production, mitotic activity, and nuclear features. **Highest-grade component rule:** the worst grade is assigned even if a minor fraction. Well-differentiated PDAC is uncommon; most are G2–G3. **Grading is not applicable after neoadjuvant chemotherapy** (treatment-induced changes obscure native morphology).

### Morphological patterns (not subtypes)

Three patterns lack independent molecular/clinical correlates and do not independently affect survival:

* **Large duct pattern** — glands >0.5 mm, bland cells, haphazard; mimics IPMN (major differential).
* **Foamy gland / clear cell pattern** — microvesicular "lace-like" cytoplasm, foveolar-like nuclei; a clear-cell variant can mimic metastatic clear cell RCC.
* **Micropapillary pattern** — small solid clusters in stromal lacunae, ± intraepithelial neutrophils. **Reclassified in 2026 from a distinct subtype to a morphological pattern** (still inconclusive survival data). Single-morphology tumors carry a better prognosis than those with coexisting multiple morphologies, even after neoadjuvant therapy.

### Histological subtypes and their correlates

* **Adenosquamous carcinoma** — ≥30% squamous component; up to 4% of PDACs; larger, more often poorly differentiated, tail-located, node/distant-metastasizing, worse prognosis. KRAS in nearly all; enriched SMAD4/CDKN2A/TP53; UPF1 reported. Higher PD-L1 (mainly squamous component) raises immunotherapy interest.
* **Colloid carcinoma** — ≥80% neoplastic epithelium suspended in extracellular mucin; \~3%; mostly from intestinal-type IPMNs. Intestinal phenotype (CDX2, MUC2); enriched **dMMR/MSI** → reflex testing + immunotherapy relevance; ATM germline variants reported. **Better prognosis** than conventional PDAC.
* **Medullary carcinoma** — very rare; pushing/expansile borders, syncytial solid sheets with prominent TILs; **dMMR/MSI**, high TMB, often KRAS/TP53 wild-type, JAK-family mutations; may be sporadic or Lynch-associated; more favorable prognosis; candidate for immune checkpoint inhibition. Exclude ampullary/duodenal primary.
* **Hepatoid carcinoma** — extremely rare; ≥50% hepatocellular differentiation (AFP, canalicular CD10, HepPar1); histology remains primary criterion. **2026 update:** hepatoid differentiation now also recognized in other pancreatic neoplasms (well-diff NETs, solid pseudopapillary neoplasms) — broadening the differential. Exclude metastatic HCC. Aggressive.
* **Poorly cohesive / signet-ring cell carcinoma** — very rare; ≥80% poorly cohesive cells (gastric-like); genomically similar to conventional PDAC but enriched for actionable EGFR/RET fusions and MSI; highly aggressive (lung a common metastatic site). Exclude gastric/breast primary.
* **Undifferentiated carcinoma (UC)** — ≥80% lacking definitive differentiation; hypercellular, scant stroma; median OS \~5 months. Three patterns: anaplastic, sarcomatoid (± heterologous cartilage/bone; UC with rhabdoid cells enriched for SWI/SNF alterations), carcinosarcoma (biphasic, each component ≥30%).
* **Undifferentiated carcinoma with osteoclast-like giant cells (UCOGC)** — non-neoplastic osteoclast-like giant cells + histiocytic component + atypical neoplastic mononuclear cells; KRAS/TP53 like conventional PDAC. Prognosis depends on an associated glandular component (pure forms can show prolonged survival). PD-L1 in tumor cells → worse outcome.

### A changing landscape: WHO 2026 novelties and emerging issues

* Expanded histological and molecular sections; integrates 2019–2025 data (consortium studies, meta-analyses).
* **Micropapillary** demoted subtype → pattern; **hepatoid** morphology broadened beyond PDAC.
* **Neoadjuvant response assessment remains unresolved** — treatment-naïve grading is inapplicable; no tumor regression grading system (CAP, MD Anderson) is yet sufficiently reproducible, especially for partial responders.
* **Emerging entities (not yet in the 2026 classification):** *metatypic carcinoma of the pancreas* (well-demarcated, centrally necrotic, squamous/basal, ARID1A-enriched, possibly less aggressive) and *intraductal papillary squamous neoplasm (IPSN)* precursor lesion (can progress to invasive PDAC incl. adenosquamous).

### Future perspectives

In the precision-medicine era, pancreatic pathologists are expected to lead multi-omic, single-cell and spatially resolved approaches, identify diagnostic/prognostic/predictive biomarkers, and build reproducible methods for assessing treatment response — converging on improved survival and quality of life.

### Source

* Luchini C. *Pancreatic ductal adenocarcinoma and its subtypes: clinical relevance of histopathology and molecular characterization, integrating the key updates of the 2026 WHO classification.* **Virchows Archiv** (2026). Received 10 Jul 2026 · Revised 4 Aug 2026 · Accepted 13 Aug 2026. DOI: [10.1007/s00428-026-04681-1](https://doi.org/10.1007/s00428-026-04681-1).
* **Open Access** under a Creative Commons Attribution 4.0 International License (CC BY 4.0).
* Competing interests (disclosed): C.L. is an Associate Editor of *Virchows Archiv*; honoraria from Astellas, Bayer, Gilead, MSD, Medica s.r.l.