Clinical overview
Gestational trophoblastic disease (GTD) is a spectrum of abnormal proliferations of placental trophoblast that runs from the benign-but-premalignant hydatidiform mole (complete and partial) through to frankly malignant gestational trophoblastic neoplasia (GTN): invasive mole, choriocarcinoma, placental-site trophoblastic tumour (PSTT) and epithelioid trophoblastic tumour (ETT). What makes GTD a registrar's recurring clinical problem is that it almost always presents as something far more ordinary — a "miscarriage", a heavier-than-usual early-pregnancy bleed, persistent vomiting, or simply a uterus that feels too big for dates. The unifying diagnostic thread is the syncytiotrophoblast's production of human chorionic gonadotrophin (βhCG): the more abnormal and proliferative the trophoblast, the higher and more abnormal the hormone signal. Diagnosis therefore rests on three pillars — βhCG, pelvic ultrasound, and histopathology of evacuated products — and the diagnosis of the malignant arm (GTN) is, uniquely in oncology, frequently made biochemically and radiologically without a tissue diagnosis at all.
For the South African registrar this is not a rare-disease footnote. GTD is more common in our population than in high-income European settings, often presents late with bulky disease and marked anaemia, and is managed without the luxury of a single national trophoblast registry. The unusual but reassuring counterpoint is that GTN, even in advanced and metastatic forms, is one of the most chemo-curable solid tumours in medicine — cure rates exceed 90% — provided the diagnosis is made, the disease is correctly risk-stratified, and βhCG surveillance is rigorous. Getting the diagnosis right and getting it early is the whole game.
Core knowledge
The spectrum and its cytogenetics
Figure D2.1 — Molar pregnancy: the clinical clues (bleeding, large-for-dates uterus, hyperemesis, very high β-hCG), the ultrasound 'snowstorm', and complete vs partial mole.
The two molar pregnancies are distinguished by their genetics, which in turn predicts their behaviour and is the reason histology and ploidy matter so much.
- Complete hydatidiform mole (CHM) — diploid and entirely androgenetic (all chromosomes paternal). Most commonly an empty ovum is fertilised by a single sperm that duplicates (monospermic, 46,XX); a minority are dispermic (46,XX or 46,XY). There is no fetal tissue. All chorionic villi are abnormal, with diffuse hydropic swelling and circumferential trophoblastic hyperplasia. Risk of progression to GTN is in the order of 15–20%.
- Partial hydatidiform mole (PHM) — triploid, classically 69,XXX or 69,XXY, arising when a normal ovum is fertilised by two sperm (or one diploid sperm). There is a fetus or fetal tissue (usually non-viable, often with triploid stigmata), focal villous hydrops and focal, mild trophoblastic hyperplasia. Risk of progression to GTN is low — roughly 0.5–5%.
GTN itself comprises invasive mole (molar tissue invading myometrium, usually post-CHM), choriocarcinoma (a highly malignant, haemorrhagic, early-haematogenous-spreading tumour of cyto- and syncytiotrophoblast with no villi — can follow any pregnancy event including a normal term birth), and the intermediate-trophoblast tumours PSTT and ETT (rare, slower-growing, comparatively hCG-low for their bulk, and relatively chemoresistant — surgery-dominant). Pathological detail is developed in gtd-pathology; this chapter concentrates on how each is recognised and diagnosed.
Why βhCG is the perfect tumour marker
Syncytiotrophoblast secretes hCG in proportion to its mass and proliferation. This gives clinicians a marker that is quantitative, exquisitely sensitive, and present in essentially 100% of disease (with the caveat that PSTT/ETT produce relatively little, and that human placental lactogen (hPL) is a better intermediate-trophoblast marker). A βhCG that fails to fall as expected after evacuation, plateaus, or rises is the single most important alarm in the whole field — it defines GTN. Two laboratory pitfalls must be understood:
- The "hook effect" — at extremely high hormone concentrations (as in a bulky CHM) the immunometric assay is saturated and reports a falsely low result. If clinical and ultrasound features scream mole but the βhCG looks unimpressive, ask the NHLS laboratory to re-run on serial dilutions.
- Phantom hCG / false-positive hCG — heterophile antibodies cross-react in serum assays. The discriminator is that phantom hCG is not detectable in urine (heterophiles are not renally excreted), so a negative urine hCG with a positive serum hCG points to assay interference, not disease — a vital check before committing a woman to chemotherapy.
Assessment
History and presentation
The classic CHM presentation — early-pregnancy bleeding, a uterus large for dates, expulsion of grape-like vesicles, hyperemesis, and early pre-eclampsia (before 20 weeks) — is now seen less often where early ultrasound is routine, because most moles are picked up at the first scan before they become florid. Where access to early scanning is patchy, as it still is in much of the SA public sector, the textbook late presentation remains common: a markedly enlarged uterus, severe anaemia from recurrent bleeds, theca-lutein cysts (from ovarian hyperstimulation by very high hCG), and occasionally hyperthyroidism (the alpha-subunit of hCG weakly cross-stimulates the TSH receptor) or early-onset hypertensive disease. PHM more often masquerades as an incomplete or missed miscarriage and is frequently a histological surprise on evacuated products. Choriocarcinoma can present long after any pregnancy with the consequences of metastasis — haemoptysis or a chest lesion, neurological signs from cerebral deposits, or abnormal bleeding — and must be on the differential for any reproductive-age woman with metastatic disease of unknown primary and a positive βhCG.
Examination
Confirm the uterine size against dates, assess for the pallor and tachycardia of anaemia, screen for the signs of pre-eclampsia, and palpate for the often-bilateral, sometimes large theca-lutein cysts. A respiratory and neurological examination is part of the workup for any suspected GTN because lung and brain are the commonest metastatic sites. A speculum examination may reveal expelled vesicular tissue or, rarely, a vascular vaginal metastasis that must not be biopsied (risk of catastrophic haemorrhage).
Investigations
- Quantitative serum βhCG is the cornerstone — for diagnosis, for the staging score, and for surveillance. Always interpret an unexpectedly low value against the hook effect.
- Pelvic ultrasound (transvaginal) is the principal imaging test. The CHM gives the classic heterogeneous, vesicular "snowstorm" / "cluster of grapes" appearance with multiple small cystic spaces and no fetus; theca-lutein cysts may be seen on the ovaries. The PHM shows a placenta with focal cystic spaces and an abnormal or non-viable fetus, sometimes with an enlarged, "swiss-cheese" gestational sac. Ultrasound is suggestive, not diagnostic — early CHM in particular can look like a simple missed miscarriage, and first-trimester ultrasound sensitivity for molar pregnancy is only moderate (often quoted around 40–60%). This is precisely why histology of all products of conception is mandatory.
- Histopathology of evacuated tissue is the diagnostic gold standard for molar pregnancy. The morphological distinction between CHM, PHM and a hydropic non-molar abortion can be genuinely difficult; the modern adjuncts are p57^KIP2^ immunohistochemistry (a paternally-imprinted, maternally-expressed gene — p57 is LOST/negative in CHM because there is no maternal genome, but retained/positive in PHM and in non-molar miscarriage) and, where available, ploidy analysis / molecular genotyping to separate diploid androgenetic (CHM) from triploid diandric (PHM). NHLS histology turnaround and access to p57 IHC and genotyping vary by centre — flag a clinically suspected mole on the request form so the pathologist applies the right ancillary tests.
- Baseline bloods: full blood count (quantify the anaemia), group and save / crossmatch (evacuation of a bulky mole can bleed heavily), renal and liver function, and thyroid function if clinically hyperthyroid.
- Staging imaging when GTN is diagnosed: a chest radiograph is used for the metastasis count in the WHO prognostic score (CT chest detects micrometastases but lung micromets do not change the score). Where the chest film shows lung metastases or the picture suggests dissemination, proceed to CT chest/abdomen/pelvis and MRI brain to assess liver and central-nervous-system disease.
Diagnosing GTN — a biochemical and radiological diagnosis
Figure D2.2 — Serial β-hCG as the GTN detector: the FIGO/WHO criteria (plateau, rise, persistence) that diagnose GTN biochemically, plus the hook-effect and phantom-hCG pitfalls.
GTN is unusual in that the diagnosis is most often made without fresh histology. After evacuation of a hydatidiform mole, GTN is diagnosed on βhCG behaviour, using internationally adopted FIGO/WHO criteria: any one of —
- a plateau of βhCG (four values within ±10% over a 3-week period: days 1, 7, 14, 21);
- a rise in βhCG (a rise of ≥10% across three values over a 2-week period: days 1, 7, 14);
- persistence of detectable βhCG for 6 months or more after evacuation; or
- a histological diagnosis of choriocarcinoma.
A tissue diagnosis (e.g. choriocarcinoma) is sufficient on its own, but for the great majority of post-molar GTN the rising or plateauing hormone is what triggers treatment — and biopsy of vascular metastatic deposits is actively avoided because of bleeding risk. This is why a structured, registry-style βhCG surveillance programme after every molar evacuation is the diagnostic safety net for the whole disease.
Staging and risk stratification (FIGO 2000)
Figure D2.3 — Staging and risk: FIGO 2000 anatomical stage + the modified WHO prognostic score deciding single-agent (low-risk) vs multi-agent EMA-CO (high-risk) chemotherapy.
Once GTN is diagnosed, it is staged anatomically (FIGO stage I–IV) and scored prognostically with the modified WHO prognostic score, exactly as adopted in FIGO 2000. The score, not the anatomical stage alone, drives the choice between single-agent and multi-agent chemotherapy.
| FIGO anatomical stage | Extent of disease |
|---|---|
| I | Disease confined to the uterus |
| II | Extends outside uterus but limited to genital structures (adnexa, vagina, broad ligament) |
| III | Extension to the lungs ± genital tract |
| IV | All other metastatic sites (e.g. brain, liver, kidney, GI tract) |
| Modified WHO prognostic score | 0 | 1 | 2 | 4 |
|---|---|---|---|---|
| Age (years) | <40 | ≥40 | — | — |
| Antecedent pregnancy | mole | abortion | term | — |
| Interval from index pregnancy (months) | <4 | 4–<7 | 7–<13 | ≥13 |
| Pre-treatment βhCG (IU/L) | <10³ | 10³–<10⁴ | 10⁴–<10⁵ | ≥10⁵ |
| Largest tumour size (incl. uterus) | — | 3–<5 cm | ≥5 cm | — |
| Site of metastases | lung | spleen, kidney | GI tract | brain, liver |
| Number of metastases | — | 1–4 | 5–8 | >8 |
| Prior failed chemotherapy | — | — | single drug | ≥2 drugs |
The scores are summed. A total of ≤6 = low-risk (single-agent chemotherapy); ≥7 = high-risk (multi-agent chemotherapy). Any FIGO stage IV disease is treated as high-risk regardless of the numerical score, because of its propensity for resistant, multi-site disease.
Management
Detailed treatment belongs to the GTN management objective; what the diagnosing registrar must understand is how diagnosis flows directly into management so the right pathway is set in motion.
Confirming and evacuating a molar pregnancy
A suspected mole is managed by suction (vacuum) evacuation of the uterus, ideally under ultrasound guidance, with blood crossmatched and available — bulky CHM can bleed briskly. Send all evacuated tissue for histology; the diagnosis is not secure until the pathologist confirms it. Medical evacuation and sharp curettage as a primary method are generally avoided for CHM (theoretical concern about disseminating trophoblast and incomplete evacuation), though a single gentle re-evacuation may occasionally be appropriate for retained tissue. Anti-D prophylaxis is given to rhesus-negative women for PHM; it is not required for CHM (no fetal red cells), but where the histological diagnosis is not yet known many units give it pragmatically. Hysterectomy is an option for the woman who has completed her family or has heavy uncontrolled bleeding, and it reduces but does not eliminate the need for βhCG follow-up.
Surveillance — the diagnostic engine of GTN detection
After evacuation, serial quantitative βhCG is followed to normalisation and for a defined period beyond. This surveillance is what diagnoses GTN early, when it is most curable. In the UK the RCOG/national trophoblast registry model centralises this; South Africa has no single national registry, so the practical equivalent is a disciplined local protocol: serial βhCG at the referral gynae-oncology unit until normal, with clear escalation criteria, and counselling the woman to avoid pregnancy and use reliable contraception until follow-up is complete (a new pregnancy obscures the very marker we are watching). Combined hormonal contraception is acceptable once βhCG is falling/normalised per RCOG guidance; an intrauterine device is best deferred until βhCG is normal because of perforation risk into a possibly invaded uterus.
When GTN is diagnosed
A low-risk (WHO ≤6) patient is treated with single-agent chemotherapy — methotrexate (with folinic-acid rescue) or actinomycin-D (dactinomycin) — with cure rates approaching 100% (some needing a switch of single agent or escalation for resistance). A high-risk (WHO ≥7) or stage IV patient receives multi-agent chemotherapy, classically EMA-CO (etoposide, methotrexate, actinomycin-D, cyclophosphamide, vincristine), with overall cure rates around 90–95%. One important refinement in the FIGO 2021 update: a patient with an ultra-high-risk score (≥13) or massive metastatic burden (especially liver/brain) can die suddenly from tumour collapse, haemorrhage and multi-organ failure if hit with full-dose chemotherapy up front — so gentle induction with low-dose etoposide–cisplatin (EP, e.g. etoposide 100 mg/m² + cisplatin 20 mg/m² on days 1–2, weekly for 1–3 weeks) is given before escalating to EMA-CO / EP-EMA, which reduces early deaths. Brain metastases are managed with high-dose methotrexate (infusion to 1 g/m²) ± intrathecal methotrexate and/or radiotherapy. PSTT and ETT are diagnosed histologically, are relatively chemoresistant and hCG-low for bulk, and are managed primarily by hysterectomy with chemotherapy reserved for metastatic/adverse-feature disease. All GTN should be managed in, or in close consultation with, a gynae-oncology unit.
Red flags / pitfalls
- Treating the βhCG, not the woman. Before acting on a "rising hormone", exclude a new intrauterine pregnancy (the commonest cause of a rising hCG in a fertile woman) and exclude phantom/false-positive hCG with a urine hCG and, if needed, serial dilutions/heterophile-blocking. Starting chemotherapy on assay interference is a never-event.
- The hook effect. A "low" βhCG with a uterus full of snowstorm is an assay artefact until proven otherwise — request dilutions.
- Not sending all products of conception for histology. Every miscarriage and "incomplete abortion" curetting is a potential PHM/CHM. Histology is the diagnosis; skipping it is how moles are missed, particularly the bland early CHM and the easily-overlooked PHM.
- Biopsying a vascular metastasis. A vaginal or other vascular GTN deposit can exsanguinate if biopsied — diagnose GTN biochemically/radiologically instead.
- Losing the patient to follow-up. In a system without a central registry, the surveillance βhCG schedule is only as good as the unit running it. A defaulting patient is the highest-risk scenario in GTD; counsel explicitly and document contraception.
- Forgetting choriocarcinoma after a normal pregnancy. It can follow a term delivery, ectopic, or any abortion, sometimes months later. Postpartum abnormal bleeding, haemoptysis, or unexplained metastatic disease in a reproductive-age woman demands a βhCG.
- Misreading hyperthyroidism or early pre-eclampsia. Thyrotoxicosis or hypertensive disease before 20 weeks should prompt a hunt for a mole.
- HIV context. In our high-prevalence setting, immunosuppression and competing diagnoses (TB with a lung lesion mimicking pulmonary metastasis) complicate the picture; correlate the chest film with the βhCG before attributing lung shadows to GTN, and conversely do not dismiss a positive βhCG with lung lesions as TB.
Evidence anchors
- RCOG Green-top Guideline No. 38 — Gestational Trophoblastic Disease (4th edition, 2020) is the principal guideline for diagnosis, evacuation, βhCG-based diagnosis of GTN, surveillance and the centralised-registration model. It anchors the diagnostic criteria for post-molar GTN and the contraception-during-follow-up advice.
- FIGO — Diagnosis and management of gestational trophoblastic disease: 2021 update (Ngan HYS, Seckl MJ, Berkowitz RS, et al. Int J Gynecol Obstet 2021;155(Suppl 1):86–93) — the current FIGO management reference. It retains FIGO 2000 anatomical staging (I–IV) + the modified WHO prognostic score (low-risk ≤6 → single-agent methotrexate ± folinic acid, or actinomycin-D; high-risk ≥7 or stage IV → multi-agent EMA-CO; ~90–95% cure), and defines the three FIGO criteria for post-molar GTN: hCG plateau across 4 values over ≥3 weeks (days 1, 7, 14, 21); hCG rise across 3 values over ≥2 weeks (days 1, 7, 14); or a histological diagnosis of choriocarcinoma (the older "elevated hCG ≥6 months" criterion was dropped). It adds low-dose etoposide–cisplatin induction for ultra-high-risk (≥13) disease, EP-EMA/TP-TE and pembrolizumab as salvage, and hysterectomy-led management of PSTT/ETT.
- WHO Classification of Tumours of the Female Genital Tract, 5th edition (2020) governs histological typing across the spectrum (complete vs partial mole; invasive mole; choriocarcinoma; PSTT; ETT) and the use of ancillary p57 immunohistochemistry and ploidy/genotyping, with the cytogenetic distinction of complete mole = diploid androgenetic (46,XX) versus partial mole = triploid.
- South African context: there is no single national gestational-trophoblast registry; surveillance is delivered through referral gynae-oncology units. Diagnostic adjuncts (quantitative βhCG dilutions, p57 IHC, ploidy/genotyping) are accessed via NHLS, with variable turnaround. Chemotherapy agents used (methotrexate, folinic acid, actinomycin-D, and the EMA-CO components) are available on the SA hospital-level EML for oncology use. The high background HIV prevalence means pulmonary TB is an important mimic of lung metastases and must be reconciled against the βhCG before staging conclusions are drawn.