Clinical overview
Endometrial carcinoma is the most common gynaecological malignancy in high-income settings and a rising problem in South Africa as obesity, type 2 diabetes and an ageing population converge. It is the cancer that announces itself early and loudly: roughly nine in ten women present with postmenopausal bleeding, and a small but important minority present with abnormal uterine bleeding in the perimenopause. That early symptom is the reason most endometrial cancers are diagnosed at FIGO stage I, while it is still confined to the corpus, and the reason the disease as a whole carries a comparatively good prognosis. The registrar's job is to take postmenopausal bleeding seriously every single time, because the symptom is shared by atrophic endometritis (benign and common) and by an aggressive serous carcinoma that will kill within two years if missed.
For the FCOG(SA) examiner, the heart of this objective is pathology — what the disease looks like grossly, down the microscope, and now at the molecular level. The field underwent a genuine paradigm shift over the last decade. The old binary of "type I oestrogen-driven endometrioid" versus "type II oestrogen-independent serous" is being replaced by a four-group molecular classification that is reproducible, prognostic and increasingly therapy-defining. You must be able to describe both frameworks, explain how they relate, and connect the pathology to the molecular subgroups that FIGO 2023 and the ESGO–ESTRO–ESP 2025 guideline now place at the centre of staging and risk stratification. A clean understanding of the pathology is what lets you predict behaviour and choose treatment.
Core knowledge
Precursor lesions
Figure D10.1 — The dualistic model: Type I endometrioid (oestrogen-driven, from atypical hyperplasia/EIN) vs Type II serous/clear-cell (non-oestrogen, aggressive), now refined by molecular class.
Most endometrioid carcinomas evolve from endometrial hyperplasia under sustained unopposed oestrogen. The current (WHO 2020) binary scheme divides hyperplasia into hyperplasia without atypia and atypical hyperplasia / endometrioid intraepithelial neoplasia (EIN). Hyperplasia without atypia carries a low (~1–3%) progression risk and is fundamentally a hormonal disturbance. Atypical hyperplasia/EIN is a true neoplastic precursor: progression to carcinoma is reported around 25–30% over follow-up, and crucially, when a hysterectomy specimen follows an atypical-hyperplasia biopsy, a concurrent carcinoma is found in up to 40% of cases. That figure is the single most exam-relevant number in the precursor story, and it drives the surgical management of atypical hyperplasia. The serous lineage has a different precursor — serous endometrial intraepithelial carcinoma (SEIC) — typically arising in atrophic endometrium or an endometrial polyp, p53-mutant, and able to metastasise despite being intraepithelial.
Histological types (WHO 2020)
The WHO Classification of Tumours of the Female Genital Tract, 5th edition (2020) is the authoritative typing reference. The major endometrial carcinoma types are:
| Histotype | Approx. share | Key features |
|---|---|---|
| Endometrioid | ~75–80% | Gland-forming; oestrogen-related; usually low-grade; PTEN/PIK3CA/KRAS/CTNNB1 and MMR defects |
| Serous | ~5–10% | Papillary/glandular, marked nuclear atypia; TP53-mutant; aggressive |
| Clear cell | ~1–5% | Hobnail cells, clear cytoplasm; mixed molecular profile |
| Carcinosarcoma (MMMT) | ~2–5% | Biphasic carcinoma + sarcoma; now classed as a metaplastic/dedifferentiated carcinoma |
| Mixed / undifferentiated / dedifferentiated | small | Aggressive; undifferentiated component often MMRd |
Older terminology is worth noting because it recurs in exam stems: "type I" (endometrioid, low-grade, oestrogen-driven, good prognosis, obese/diabetic phenotype) versus "type II" (serous and clear cell, oestrogen-independent, high-grade, older atrophic uterus, poor prognosis). The dualistic model captures real biology but breaks down at the edges — some grade 3 endometrioid tumours behave like type II — which is exactly why the molecular classification was needed.
Gross (macroscopic) pathology
On opening the uterus, endometrioid carcinoma usually appears as an exophytic, shaggy, friable polypoid or sessile tan-grey mass filling the endometrial cavity, often with surface haemorrhage and necrosis. The examiner wants you to describe the assessment of myometrial invasion macroscopically: the tumour erodes into the myometrial wall, and the depth relative to the full wall thickness is estimated grossly and confirmed microscopically. A diffusely thickened endometrium, focal firmness, or a tumour reaching toward the serosa all matter. Lower uterine segment / cervical stromal involvement is sought because it upstages disease. Serous carcinomas may look deceptively bland or arise within a small polyp in an otherwise atrophic, often small uterus, which is a trap — minimal gross disease can already be metastatic.
Microscopic pathology
Endometrioid carcinoma is defined by malignant glands that recapitulate proliferative endometrium, lined by stratified columnar cells, with complex back-to-back gland crowding and loss of intervening stroma (the architectural hallmark distinguishing carcinoma from atypical hyperplasia). Squamous (morular) differentiation is common and benign in significance. Grading is by the FIGO three-tier architectural system based on the proportion of solid (non-squamous, non-morular) growth: grade 1 ≤5% solid, grade 2 6–50%, grade 3 >50%; marked nuclear atypia raises the grade by one. Increasingly tumours are reported in a binary low- vs high-grade scheme (grade 1–2 = low, grade 3 = high), which aligns better with prognosis and with the molecular groups.
Serous carcinoma shows complex papillary and glandular architecture with markedly pleomorphic, high-grade nuclei, prominent nucleoli and brisk atypical mitoses, often with psammoma bodies. Immunohistochemistry shows diffuse aberrant p53 (overexpression or complete null pattern) and frequently p16 block positivity; this is the morphological correlate of the TP53 mutation. Clear cell carcinoma shows hobnail cells, clear (glycogen-rich) cytoplasm and hyalinised papillary cores. Carcinosarcoma is biphasic — a high-grade carcinomatous component (usually serous or high-grade endometrioid) intimately admixed with a malignant mesenchymal/sarcomatous component (homologous or heterologous, e.g. chondrosarcoma/rhabdomyosarcoma); it is now understood as a carcinoma that has undergone epithelial–mesenchymal transition rather than a true sarcoma, and is managed as a high-risk carcinoma. Contrast this with the pure mesenchymal tumours covered in gynaecological-sarcomas.
Routine immunohistochemistry
A practical IHC panel underpins both typing and the molecular workup: endometrioid tumours are usually ER/PR positive, p53 wild-type, and may show PTEN loss; serous tumours are ER-variable with aberrant p53. Mismatch-repair IHC (MLH1, PMS2, MSH2, MSH6) is now performed on essentially all endometrial cancers, both to flag the MMRd molecular group and to screen for Lynch syndrome — loss of MLH1/PMS2 is most often due to MLH1 promoter hypermethylation (sporadic), whereas isolated MSH6 or MSH2 loss raises Lynch suspicion and prompts germline testing.
The molecular classification — FIGO 2023 / ESGO 2025
Figure D10.2 — The FIGO 2023/TCGA molecular classification: POLEmut, MMRd, NSMP and p53abn — the testing hierarchy, prognosis, and treatment implications (incl. immunotherapy in dMMR).
The decisive modern advance, derived from TCGA and made clinically tractable by the ProMisE/TransPORTEC algorithms, is the four-group molecular classification, which FIGO 2023 endometrial staging (Berek et al., IJGO 2023) advises (advises, not mandates) be applied to all endometrial carcinomas, and which the ESGO–ESTRO–ESP 2025 update (Lancet Oncology, superseding the 2021 edition) builds risk groups around:
| Molecular group | Defining test | Prognosis | Note |
|---|---|---|---|
| POLEmut | Pathogenic POLE exonuclease-domain mutation | Excellent | Even high-grade looks indolent |
| MMRd | Loss of MMR proteins on IHC / MSI-high | Intermediate | Lynch screen; checkpoint-inhibitor responsive |
| NSMP | No specific molecular profile (none of the above) | Intermediate (heterogeneous) | ER status further refines it |
| p53abn | Aberrant p53 IHC / TP53 mutation | Poor | Most serous; treated aggressively |
The algorithm is sequential: test POLE first (its excellent prognosis trumps everything, even high grade or p53 abnormality), then MMR IHC, then p53 IHC; tumours negative on all three are NSMP. A small group with more than one classifying feature ("multiple classifiers") is resolved by this hierarchy — POLEmut over p53abn over MMRd.
FIGO 2023 explicitly integrates molecular class into the stage in two situations. Stage IAmPOLEmut: a POLE-mutated carcinoma that would otherwise be stage I–II is downstaged to IAm because its outcome is so favourable. Stage IICmp53abn: a p53-abnormal carcinoma with any myometrial invasion is upstaged to IICm. These molecularly-defined substages are a favourite exam point precisely because they show pathology directly altering the anatomical stage.
Spread, FIGO anatomical staging and prognostic factors
Figure D10.3 — From slide to stage: histological type, grade and myometrial invasion, and the molecularly-defined FIGO 2023 stages (IAmPOLEmut downstaged, IICmp53abn upstaged).
Endometrial cancer spreads by direct myometrial and cervical extension, lymphatically to pelvic and para-aortic nodes, transtubally to peritoneal surfaces (especially serous histology), and haematogenously late (lung). Anatomical FIGO 2023 staging in outline: stage I confined to the corpus (IA <50% myometrial invasion, IB ≥50%), stage II cervical stromal involvement, stage III local/regional spread including adnexa, serosa and nodes (IIIC1 pelvic, IIIC2 para-aortic nodes), and stage IV bladder/bowel mucosa or distant metastasis. The dominant prognostic factors the examiner expects are: histotype, FIGO grade, molecular group, depth of myometrial invasion, lymphovascular space invasion (LVSI) — with substantial/diffuse LVSI now reported separately because it strongly predicts nodal disease and recurrence — cervical stromal involvement, and nodal status.
Assessment
Diagnosis rests on endometrial sampling. Postmenopausal bleeding warrants transvaginal ultrasound as the first-line triage: an endometrial thickness ≤4 mm has a very high negative predictive value for cancer and, in a woman with a single bleed and a clearly thin stripe, may permit conservative follow-up; a thickness >4 mm, persistent or recurrent bleeding, or an inability to visualise the endometrium mandates tissue. Histology is obtained by outpatient endometrial biopsy (Pipelle) or by hysteroscopy with directed biopsy / dilatation and curettage where the office sample is inadequate, the cavity is focally abnormal (polyp), or bleeding recurs despite a benign Pipelle — Pipelle samples a minority of the surface and a negative result does not exclude focal disease. Hysteroscopy also evaluates cervical involvement. The pathology report should state histotype, FIGO grade, and trigger the MMR/p53 (and reflex POLE where available) workup.
Staging investigations after a tissue diagnosis include examination, and MRI pelvis to assess depth of myometrial invasion and cervical stromal involvement, with CT chest/abdomen/pelvis (and increasingly PET-CT) for nodal and distant disease in high-grade or advanced cases. In the South African public sector, MRI and PET-CT access is rationed and concentrated in tertiary centres; transvaginal ultrasound and CT are more widely available, and molecular testing (POLE in particular) is not universally accessible through the NHLS — MMR and p53 IHC are far more feasible and are the pragmatic backbone of molecular classification where sequencing is limited. Document the molecular workup actually performed rather than assuming the full algorithm.
Management
This is a pathology objective, so management is given in principle. Surgery is the cornerstone: total hysterectomy with bilateral salpingo-oophorectomy, with surgical staging (peritoneal assessment and nodal evaluation — sentinel lymph node biopsy has largely replaced systematic lymphadenectomy for apparent early-stage disease). Minimally invasive surgery is preferred where expertise exists. The extent of nodal and adnexal surgery is tailored to histotype, grade and molecular group.
Adjuvant therapy is decided by the ESGO–ESTRO–ESP risk group, which now folds in molecular class: a POLEmut stage I tumour may need no adjuvant treatment even when high-grade, whereas a p53abn tumour is treated aggressively. Options span vaginal vault brachytherapy (for intermediate-risk local control), external-beam pelvic radiotherapy, and chemotherapy (carboplatin–paclitaxel) for high-risk and advanced disease. The molecular era has brought immunotherapy into front-line and recurrent settings: dostarlimab plus carboplatin–paclitaxel (ENGOT-EN6/RUBY) is approved (FDA, August 2024) for primary advanced or recurrent disease regardless of MMR status with the biggest benefit in dMMR; pembrolizumab plus chemotherapy (NRG-GY018) is a first-line option; and pembrolizumab plus lenvatinib (KEYNOTE-775) is established second-line, particularly for MMR-proficient (pMMR) tumours. Fertility-sparing progestin therapy (with strict hysteroscopic surveillance) may be offered in highly selected grade 1 endometrioid stage IA tumours without myometrial invasion in young women, mirroring the conservative pathway used for atypical hyperplasia. In South Africa, access to immunotherapy is currently very limited outside private cover and select trials; surgery, brachytherapy/EBRT and carboplatin–paclitaxel remain the realistic backbone of public-sector care.
Red flags / pitfalls
- Treating postmenopausal bleeding as benign without tissue. A thin endometrial stripe lowers but does not abolish risk, and serous carcinoma can arise in an atrophic uterus with a thin endometrium — recurrent bleeding always needs histology regardless of stripe.
- A negative Pipelle does not exclude cancer. It samples a fraction of the cavity; focal lesions and polyp-based serous carcinoma are missed. Persistent symptoms → hysteroscopy and directed biopsy.
- Atypical hyperplasia/EIN on biopsy hides carcinoma in up to 40% of hysterectomy specimens — do not under-grade the surgery or rely on the biopsy alone for prognosis.
- Misreading the molecular hierarchy. POLEmut trumps p53abn and MMRd in "multiple-classifier" tumours; failing to test POLE first can mislabel an excellent-prognosis tumour as high-risk and overtreat it.
- Forgetting Lynch screening. Isolated MSH6/MSH2 loss (or MLH1 loss without promoter methylation) should trigger germline referral — endometrial cancer is frequently the sentinel Lynch malignancy, preceding colorectal cancer.
- Underweighting substantial LVSI, which predicts nodal disease and recurrence even in otherwise low-stage tumours and shifts adjuvant decisions.
- Calling carcinosarcoma a sarcoma. It is a high-grade (metaplastic) carcinoma, staged and treated as carcinoma — see gynaecological-sarcomas for the true uterine sarcomas.
Evidence anchors
- WHO Classification of Tumours of the Female Genital Tract, 5th edition (2020) — authoritative histological typing of endometrial carcinoma and its precursors (hyperplasia without atypia vs atypical hyperplasia/EIN; serous, clear cell, carcinosarcoma, undifferentiated/dedifferentiated).
- FIGO 2023 endometrial staging (Berek et al., IJGO 2023) — advises molecular classification (POLEmut / MMRd / NSMP / p53abn) for all tumours, and defines the molecularly-integrated substages stage IAmPOLEmut (downstaged) and stage IICmp53abn (upstaged).
- ESGO–ESTRO–ESP guidelines for endometrial carcinoma — 2025 update (Lancet Oncology; supersedes 2021) — risk groups integrating molecular class; surgical, radiotherapy and systemic-therapy principles.
- ENGOT-EN6 / RUBY (dostarlimab + carboplatin–paclitaxel) — FDA-approved (August 2024) for primary advanced/recurrent disease regardless of MMR status, greatest benefit in dMMR.
- NRG-GY018 (pembrolizumab + chemotherapy) — first-line systemic option.
- KEYNOTE-775 (pembrolizumab + lenvatinib) — second-line, particularly in pMMR/MMR-proficient disease.
- South African context: NHLS provides MMR and p53 immunohistochemistry far more readily than POLE sequencing; MRI/PET-CT are tertiary-concentrated; immunotherapy access is largely confined to private cover and trials, so surgery, brachytherapy/EBRT and carboplatin–paclitaxel remain the public-sector backbone.