9.3 Organ-Specific Epithelial Markers

9.3 Organ-Specific Epithelial Markers

Quick Answer: After keratins, use lineage markers in the expected compartment. Lung adenocarcinoma: nuclear TTF-1 (know 8G7G3/1 versus SPT24 as a specificity/sensitivity trade) and granular napsin A; squamous: nuclear p40 and CK5/6. Prostate: NKX3.1/PSA/PSAP versus basal p63/HMWCK and AMACR. GI: CDX2 and SATB2. GYN: PAX8 and WT1. Thyroid: TTF-1, thyroglobulin, PAX8. Breast: GATA3, mammaglobin, GCDFP-15 as lineage—not as ER/PR scores.

This OpenExamPrep section is independent teaching on organ-associated epithelial IHC patterns covering published QIHC tissue-staining topics. It is not an ASCP publication and does not claim Board, CAP, or ASCO/CAP approval. Patterns below are classic teaching combinations; exceptions occur and unpublished sensitivity percentages are not used here. Predictive-marker scoring (ER, PR, HER2, PD-L1, MMR) is Chapter 10.

How to use organ markers without inventing a diagnosis

These antibodies support a lineage hypothesis that already exists on the H&E. A TTF-1-positive gland-forming tumor in lung is consistent with pulmonary (or thyroid) epithelial differentiation. The same nuclear stain in a colon resection with an intestinal H&E is a clone-caveat problem, not proof that the colon primary is a lung metastasis. Always name the compartment: TTF-1, p40, p63, CDX2, SATB2, PAX8, GATA3, NKX3.1, and WT1 (when used as a serous/mesothelial marker) are nuclear. Napsin A, PSA, PSAP, thyroglobulin, mammaglobin, and GCDFP-15 are cytoplasmic. AMACR is cytoplasmic granular in secretory-type cells. HMWCK in basal cells is cytoplasmic.

Lung: TTF-1, napsin A, p40, and p63

TTF-1 (thyroid transcription factor-1) is a nuclear marker of lung epithelial and thyroid follicular differentiation. Two clones dominate practical conversation:

You do not need memorized unpublished percentages. You need the trade: SPT24 catches more true lung/thyroid labeling and also more unwanted labeling; 8G7G3/1 is stricter and can miss some tumors. Put the clone on the interpretation. Do not mix clones mid-validation and pretend they are the same reagent.

Napsin A is a granular cytoplasmic aspartic proteinase used with TTF-1 for lung adenocarcinoma. Type II pneumocytes are a handy internal control. Napsin A also labels some renal tumors (especially papillary renal-cell carcinoma in many teaching series), so it is not lung-exclusive. Mucinous lung adenocarcinomas and some poorly differentiated tumors may be napsin-weak; that is a pattern, not a percentage law.

p40 is a nuclear marker used for squamous differentiation and is generally preferred over p63 when the question is squamous versus adenocarcinoma in lung, because p63 is more likely to label some adenocarcinomas. p63 remains useful as a basal-cell nuclear marker (prostate, some myoepithelial applications). CK5/6 supplies the cytoplasmic keratin half of a squamous/basal story (Section 9.2). A lung tumor that is TTF-1/napsin A-positive and p40-negative fits an adenocarcinoma-type pattern; the reverse fits squamous-type. Dual staining and adenosquamous carcinoma exist; do not force a single box.

SPT24 can show unexpected TTF-1 labeling in some squamous or non-pulmonary tumors. That is a clone problem plus morphology, not a reason to skip p40.

Prostate: NKX3.1, PSA, PSAP, and the PIN-4 idea

NKX3.1 is a nuclear transcription-factor marker of prostatic epithelial lineage and is often the most robust prostate identifier on poorly differentiated metastases in current practice. PSA and PSAP are cytoplasmic; they can be weak in high-grade or treated tumors. None of these stains is a license to ignore the H&E.

The PIN-4 idea is a cocktail logic, not a single vendor product you must name:

Benign glands: basal markers present, AMACR usually low or patchy. Atypical / neoplastic secretory proliferations: AMACR often up, basal layer lost in invasive carcinoma. High-grade PIN retains basal cells. This is a morphology aid on prostate chips and cores. It is not ER/PR scoring, and it is not a CUP panel by itself. NKX3.1 still belongs on a metastasis when the question is prostate origin rather than “is this gland invasive.”

GI tract: CDX2 and SATB2

CDX2 is a nuclear intestinal transcription factor. Colorectal-type epithelium is the classic strong positive. Gastric, pancreaticobiliary, and some ovarian mucinous tumors can be CDX2-variable. Some non-GI tumors are reported positive; treat CDX2 as intestinal-type differentiation, not as a colon stamp.

SATB2 is a nuclear marker used to support lower-GI / colorectal-type epithelium and is also a marker of osteoblastic differentiation in bone tumors. A SATB2-positive gland-forming metastasis supports a colorectal-type hypothesis more strongly than CDX2 alone in many teaching algorithms, but exceptions occur (other GI sites, and rare non-GI labeling). SATB2 in a bone-forming tumor is a different lineage story; read the H&E.

Pair these with CK20 (Section 9.2). A CK7-/CK20+/CDX2+/SATB2+ pattern is the classic colorectal-type cluster. A CK7+/CDX2-variable pancreaticobiliary tumor should not be forced into that cluster.

GYN tract: PAX8 and WT1

PAX8 is a nuclear marker of Mullerian epithelium (ovary, endometrium, fallopian tube), kidney, and thyroid. It does not separate those three families by itself. A PAX8-positive gland-forming tumor in a woman still needs CK7/CK20, WT1, TTF-1/thyroglobulin, and renal markers according to the H&E and the clinical site.

WT1 (Wilms tumor 1), when read as nuclear labeling of tumor cells, is a classic marker of many ovarian serous carcinomas and of mesothelium. Mucinous ovarian tumors are typically WT1-negative as a teaching pattern. Cytoplasmic WT1 in other tumors is a different, often non-specific, read—do not promote cytoplasmic blush to a serous-positive result. WT1 does not replace morphology for mesothelioma versus carcinoma; that differential uses a broader panel (Chapter 10).

Thyroid: TTF-1, thyroglobulin, and PAX8

Follicular-cell thyroid tumors classically show nuclear TTF-1, nuclear PAX8, and cytoplasmic/colloid thyroglobulin. A TTF-1-positive lung tumor should be thyroglobulin-negative; thyroglobulin is the thyroid-restricted member of this trio. Medullary thyroid carcinoma is a C-cell (neuroendocrine) tumor: thyroglobulin is typically negative, calcitonin is the lineage stain, TTF-1 is often still nuclear-positive, and PAX8 is often weaker or negative compared with follicular-cell tumors. Do not use TTF-1 alone to call a neck mass lung versus thyroid.

Breast: GATA3, mammaglobin, and GCDFP-15 (lineage only)

GATA3 is a nuclear marker of breast epithelial (and urothelial, among other) differentiation. Mammaglobin and GCDFP-15 are cytoplasmic supportive breast-lineage stains; they can be patchy and they are not required to be strong in every breast carcinoma. Use them to support a breast-type hypothesis in a keratin-positive metastasis, together with morphology and history.

Do not score ER, PR, or HER2 in this section. Those are predictive-marker pattern-reading topics in Chapter 10. GATA3-positive does not mean ER-positive. Mammaglobin does not replace an Allred or H-score. If a stem asks for lineage of a CUP, pick GATA3/mammaglobin/GCDFP-15. If a stem asks for hormone-receptor status, you are in the wrong chapter.

Worked example

A CK7+/CK20- gland-forming lung mass is TTF-1-positive with clone SPT24, napsin A-granular-positive, and p40-negative. Residual pneumocytes are TTF-1-positive (internal control). The pattern supports pulmonary adenocarcinoma-type differentiation. If the same morphology were TTF-1-positive only with SPT24, thyroglobulin-negative, and the patient had a pancreatic mass that was CDX2-focal, you would not treat SPT24 as infallible lung proof—you would name the clone caveat and add history and more lineage stains. Patterns first, laws never.

Organ markers name neighborhoods after keratins sort the map. Keep clone identity, compartment, and internal epithelium in the same sentence as the result.