Diagnose and manage gestational trophoblastic disease, including molar pregnancy, gestational trophoblastic neoplasia and post-molar surveillance

In one line

Gestational trophoblastic disease is a spectrum of pregnancy-derived trophoblast disorders — from the premalignant complete and partial moles to the malignant gestational trophoblastic neoplasias (invasive mole, choriocarcinoma, placental-site and epithelioid trophoblastic tumour) — and it is the rare cancer you are expected to cure: get the mole evacuated by suction, register the patient for hCG surveillance, and the falling β-hCG itself diagnoses the malignant transformation early enough that single- or multi-agent chemotherapy delivers survival approaching 100% for low-risk and ~90% for high-risk disease.

This chapter assumes the Intermediate groundwork on how a mole presents and is recognised — revise molar pathology and the diagnosis of GTD if those are not solid. The consultant-level material is reading the genetics, building the surveillance plan, scoring the neoplasia, and defending the chemotherapy choice from the evidence.

Mechanism & pathophysiology

The trophoblast is the only human tissue that is genuinely foreign to the woman carrying it, and when its growth uncouples from a viable fetus the genetics define exactly what you are dealing with. The imaging, the hCG behaviour and the malignant potential all follow from that genetic origin.

The complete mole is paternal-genome-only — androgenetic diploidy. A complete hydatidiform mole (CHM) has a normal diploid chromosome count (46 chromosomes) but every chromosome is paternal in origin. The usual route is an "empty" ovum (no functional maternal nucleus) fertilised by a single sperm whose haploid set then duplicates, giving 46,XX — homozygous and entirely paternal. Less commonly two sperm fertilise the empty ovum (dispermy), giving 46,XX or 46,XY. A 46,YY conceptus is non-viable, so you do not see it. Because there is no maternal contribution and no embryo, there are no fetal parts, the villi are diffusely hydropic, and trophoblast proliferation is florid and circumferential — which is why the classic mole oversecretes hCG and why it carries the highest malignant potential of the molar pair (progression to neoplasia in roughly 15–20%).

The partial mole is triploid — two paternal sets and one maternal. A partial hydatidiform mole (PHM) has 69 chromosomes (triploid), the extra set being paternal: a normal ovum fertilised by two sperm, or by one sperm that duplicates. Because a maternal genome is present, a PHM can contain identifiable fetal or embryonic tissue (often with the stigmata of triploidy — growth restriction, syndactyly), the villous hydrops is focal rather than diffuse, and the trophoblast proliferation is more modest. Its malignant potential is correspondingly low (progression around 0.5–5%). The distinction is not academic: it changes the counselling, the anti-D decision, and the intensity of surveillance.

p57 is the immunostain that makes the genetics visible — and you must know which way it runs. p57^KIP2^ is a cyclin-dependent kinase inhibitor encoded by a gene that is paternally imprinted and maternally expressed — that is, the protein is produced only from the maternal allele. Apply that rule to the genetics above and the result is mechanical:

• Complete mole has no maternal genome, so there is nothing to express p57 → villous cytotrophoblast and stromal nuclei stain NEGATIVE.
• Partial mole has a maternal genome, so p57 is expressed → POSITIVE staining.

A negative p57 confirms a complete mole and excludes a partial mole or a hydropic non-molar abortion; a positive p57 tells you it is not a complete mole but cannot by itself separate a partial mole from a non-molar hydropic gestation (both have a maternal genome). When p57 leaves the question open, microsatellite (short tandem repeat) genotyping resolves it by quantifying the parental genetic contributions — absence of any maternal contribution = complete mole; diandric (two-paternal) triploidy = partial mole. This molecular pathway, not villous morphology alone, is the modern diagnostic standard, because early first-trimester moles increasingly lack the textbook gross appearance.

From mole to neoplasia — the malignant members of the spectrum. When trophoblast acquires invasive or metastatic behaviour, the disease is collectively gestational trophoblastic neoplasia (GTN):

• Invasive mole — molar villi that burrow into the myometrium (and occasionally embolise to lung or vagina). It is the commonest cause of persistently raised hCG after evacuation and usually needs no separate tissue diagnosis; the rising hCG is the diagnosis.
• Choriocarcinoma — a frankly malignant tumour of abnormal trophoblast with no chorionic villi, bulky, haemorrhagic and necrotic, with early haematogenous spread to lung, brain, liver, kidney and bowel. Crucially, choriocarcinoma can follow any pregnancy — a term delivery, a miscarriage or an ectopic, not just a mole — which is why an undelivered or postnatal woman with unexplained metastatic disease and a high hCG has choriocarcinoma until proven otherwise.
• Placental-site trophoblastic tumour (PSTT) and epithelioid trophoblastic tumour (ETT) — the dangerous outliers. Both arise from intermediate trophoblast (PSTT from implantation-site intermediate trophoblast, ETT from chorionic-type), grow as nodular masses in the myometrium or lower segment/cervix, and have two properties that overturn the usual rules: they secrete relatively little hCG for their bulk (so hCG is a poor monitor), and they are relatively chemoresistant. That combination makes surgery (hysterectomy) the primary treatment for PSTT/ETT, the opposite of the chemo-first logic that governs the rest of GTN.

Across the spectrum — paternal-only diploid CHM, diandric triploid PHM, villous-free choriocarcinoma, low-hCG chemoresistant intermediate-trophoblast tumours — the diagnosis predicts the behaviour, so management need not wait for the disease to declare itself.

Assessment

The modern presentation is sonographic, not clinical. Earlier ultrasound diagnosis has largely replaced the textbook second-trimester picture, so a consultant should expect the diagnosis to arrive before the classic syndrome:

• The hCG itself. A markedly raised β-hCG, often disproportionate to gestation, is the signature; very high levels drive the associated syndromes below. Always use an assay that detects all hCG isoforms (intact, free β, core, C-terminal, nicked and hyperglycosylated forms), because a tumour producing predominantly one variant can read falsely low or falsely negative on a narrow assay.
• Ultrasound. The complete mole gives the "snowstorm" or vesicular/"bunch-of-grapes" appearance with no fetus and often bilateral theca-lutein cysts (ovarian cysts driven by hCG stimulation). The honest caveat is that the classic honeycomb pattern is frequently absent in the first trimester — many moles are diagnosed only on histology of products evacuated for an apparent miscarriage, which is exactly why all products of conception from a non-viable pregnancy should go for histology.
• The hCG-mediated syndromes. When hCG is very high you may see hyperemesis gravidarum, early (pre-20-week) pre-eclampsia — itself almost pathognomonic of a mole, since pre-eclampsia essentially never appears that early in a normal singleton — and hCG-mediated hyperthyroidism. The thyroid effect is worth understanding mechanistically: hCG and TSH share a common α-subunit and have structurally similar β-subunits, so at extreme concentrations hCG cross-reacts at the TSH receptor, producing a biochemical (and occasionally clinical) thyrotoxicosis that resolves once the mole is evacuated and hCG falls.
• Bleeding and vesicle passage. Irregular first- or second-trimester bleeding is the commonest symptom; passage of grape-like vesicles is now uncommon.

Investigations once a mole is suspected or confirmed. FBC and group-and-save (these patients bleed at evacuation), thyroid function if clinically thyrotoxic or hCG extreme, U&E, and a baseline quantitative hCG. Histology of the evacuated tissue is mandatory and is the step that secures the diagnosis and the CHM/PHM distinction (with p57 and, if needed, STR genotyping as above). Chest imaging and cross-sectional staging are not routine for an uncomplicated mole — they belong to the GTN work-up, once surveillance flags persistence (covered below).

Management

Organise it the way you must speak it: immediate (evacuate safely) → ongoing (surveillance) → long-term (when she may conceive again, and treating neoplasia if it declares itself).

Immediate — evacuating the mole safely

Suction (vacuum) evacuation is the method of choice, regardless of uterine size, whenever fertility is to be preserved. The technique detail is examinable and clinically real:

• Use suction curettage under ultrasound guidance where available; a 12–14 mm cannula is typically needed for the bulky molar tissue. Have blood cross-matched and available if the uterus is larger than ~16 weeks' size — these uteri bleed.
• Avoid medical induction of labour and avoid pharmacological cervical preparation/oxytocics before the uterus is emptied where you can. The concern is twofold: forceful uterine contraction against an undelivered molar bulk risks trophoblastic embolisation/dissemination, and induction/medical methods are associated with higher subsequent rates of postmolar GTN than suction evacuation. An oxytocin infusion may be started at the onset of evacuation and continued after to control bleeding once the bulk is being removed — the principle is not "never use oxytocics" but "do not drive contractions against an intact mole."
• Anti-D is required for the partial mole (a PHM has fetal tissue and therefore RhD antigen on the trophoblast), and is given to RhD-negative women at evacuation. The complete mole has no fetal RhD antigen, so anti-D is not strictly indicated for a confirmed CHM — but because the precise type is often unknown at the time of evacuation, many units give anti-D to all RhD-negative women and let histology refine future counselling. State the principle and the pragmatic default.
• Send everything for histology and register the patient with a trophoblast/GTD centre. Centralised surveillance is the single intervention most consistently associated with good outcomes, because it standardises hCG assays, the monitoring schedule and the threshold to treat.
• Hysterectomy is a reasonable alternative only when childbearing is complete; it removes the molar tissue and reduces (but does not eliminate) the need for later chemotherapy, so hCG surveillance must continue after hysterectomy because metastatic GTN can still arise.
• Prophylactic chemotherapy at evacuation is not routine. It reduces postmolar GTN but is reserved for the unusual situation where the risk of GTN is high and reliable hCG follow-up cannot be guaranteed — a scenario more relevant to resource-limited or hard-to-reach populations than a blanket policy.

Ongoing — hCG surveillance, and when she may conceive again

The mole is evacuated; the surveillance now is the management, because it detects malignant transformation while it is still curable with single-agent therapy.

• Serial quantitative hCG every 1–2 weeks until normal, then confirmatory measurements. Historically a complete mole was followed for 6 months after hCG normalisation; the partial mole, whose malignant potential is far lower, can be followed far more briefly. Charing Cross data on over 20,000 moles quantified how low the residual risk is — the chance of developing postmolar GTN after the first normal hCG was about 1 in 406 for a complete mole and only about 1 in 3,195 for a partial mole, and almost all of the complete-mole risk fell within the first six months — which is why the partial mole now needs only a single confirmatory normal hCG roughly one month after the first normal result. Local protocols still vary, so register and follow the centre's schedule rather than improvising.
• Reliable contraception throughout surveillance — a new pregnancy raises hCG and destroys the surveillance signal. Barrier methods or the combined/progestogen pill are acceptable; the practical rule is to avoid an intrauterine device until hCG has normalised (risk of perforation/bleeding into a possibly still-involved uterus). Modern data show that hormonal contraception, including the combined pill, does not increase the risk of postmolar GTN, so a woman may use the pill once you are confident there is no high baseline GTN risk.
• When may she conceive again? Once surveillance is complete and hCG has been normal for the required interval. If a woman conceives during surveillance and the hCG had already normalised, termination is not mandated — the pregnancy may continue with reassurance — but every future pregnancy warrants an early ultrasound and a postpartum hCG check to exclude recurrence, and she should know her recurrence risk in a subsequent pregnancy is low (around 0.6–2%) but real, and higher after two consecutive molar pregnancies. Recurrent/familial biparental moles are linked to NLRP7 and KHDC3L mutations — a point worth recognising when a woman presents with repeated molar pregnancies.

Long-term — diagnosing and treating gestational trophoblastic neoplasia

Surveillance is doing its job when the hCG misbehaves. The diagnosis of postmolar GTN is made on the hCG trajectory or histology — not on symptoms — using the FIGO criteria. Once GTN is diagnosed, it is staged and scored, and the score (not the stage alone) decides the chemotherapy.

FIGO criteria for diagnosing postmolar GTN — any one of:

• an hCG plateau (four values within ±10% across at least 3 weeks — days 1, 7, 14, 21);
• an hCG rise of ≥10% across three values over at least 2 weeks (days 1, 7, 14);
• a histological diagnosis of choriocarcinoma.

(The older "hCG persisting ≥6 months after evacuation" criterion has fallen away as a stand-alone trigger in current practice, since most such cases now self-resolve — recognise it but lead with the plateau/rise/histology triad.)

Once GTN is diagnosed, work it up to stage and score it: examination, quantitative hCG, and imaging for metastases. FIGO anatomical stage: I = confined to the uterus; II = spread to adnexa/vagina but limited to genital structures; III = lung metastases (± genital involvement); IV = all other metastatic sites (brain, liver). The modified WHO prognostic score then sums risk factors:

The patient is recorded as stage (Roman numeral) : score (Arabic) — e.g. III:4. A total score of ≤6 is low-risk; ≥7 is high-risk — and that single dichotomy chooses single-agent versus multi-agent chemotherapy. Imaging to score: chest X-ray is used to count lung metastases for the score (CT may find more lesions but is not used for counting, to keep the score calibrated to the system on which it was validated); liver metastases by ultrasound or CT; brain by MRI or CT.

Low-risk GTN (score ≤6) → single-agent chemotherapy. The two agents are methotrexate (with folinic-acid rescue) and actinomycin D (dactinomycin). The 8-day methotrexate–folinic acid regimen (methotrexate 50 mg IM on days 1, 3, 5, 7 with folinic acid 15 mg orally 24 h after each MTX dose on days 2, 4, 6, 8, repeated every 2 weeks) is the common first choice, with actinomycin D (e.g. pulsed 1.25 mg/m² IV every 2 weeks, or a 5-day course) as the alternative. Chemotherapy continues until hCG normalises and then 2–3 consolidation cycles to prevent relapse. If the first agent stalls (hCG plateaus or rises, or toxicity prevents adequate dosing) switch to the alternative single agent before escalating to multi-agent therapy — many will still be cured with the second single agent. The complete remission rate is close to 100%. Note that scores of 5–6 sit at the upper end of low-risk and carry more single-agent resistance, so some centres lower the threshold to multi-agent therapy in that band. First-line hysterectomy is an option in selected low-risk non-metastatic disease in a woman with completed childbearing — French-centre data show hCG normalised without any salvage chemotherapy in about 82% of such patients — but it is not generally recommended over chemotherapy because postoperative chemotherapy and hCG monitoring are still often needed, and single-agent chemotherapy already cures ~100% while preserving fertility.

High-risk GTN (score ≥7) → multi-agent chemotherapy, EMA-CO. The standard regimen is EMA-CO — Etoposide, Methotrexate, Actinomycin D alternating weekly with Cyclophosphamide and Oncovin (vincristine). Overall survival in high-risk disease is around 90–95% in experienced centres. Two refinements matter at consultant level:

• Avoid sudden tumour collapse in very-high-burden disease. Patients with massive tumour burden risk catastrophic haemorrhage, metabolic acidosis, myelosuppression and multi-organ failure — early death — if hit with full-dose combination chemotherapy. Gentle induction with low-dose etoposide–cisplatin (EP) before starting EMA-CO markedly reduces these early deaths.
• Ultra-high-risk (FIGO score ≥13) behaves as a distinct, far more lethal group — much higher early-death and overall mortality — and is the strongest indication for induction EP and management in a specialist centre.

Salvage and the chemoresistant tumours. EMA-CO failures are usually salvaged with platinum-containing regimens — EMA-EP (substituting etoposide–cisplatin for the CO arm), or TP/TE (paclitaxel-based), with high-dose chemotherapy and autologous stem-cell support, or pembrolizumab, in resistant disease. Brain metastases are managed with intrathecal/high-dose methotrexate ± stereotactic or whole-brain radiotherapy and neurosurgery for bleeding or isolated resistant deposits. PSTT and ETT are the exception to all of the above: because intermediate-trophoblast tumours are relatively chemoresistant and secrete little hCG, hysterectomy is the primary treatment, with chemotherapy reserved for metastatic disease and an interval from the antecedent pregnancy of >48 months marking a particularly adverse prognosis.

South African context

There is no national GTD registry in South Africa, so the "register with a centre" principle has to be operationalised through the gynae-oncology units at the tertiary hospitals (Charlotte Maxeke/Chris Hani Baragwanath in Johannesburg, Groote Schuur and Tygerberg in the Western Cape, Inkosi Albert Luthuli in Durban). The consultant tasks at district/regional level are: evacuate the mole safely, secure histology through NHLS, start serial hCG, give anti-D where indicated, and refer early for surveillance and any chemotherapy — single-agent methotrexate/actinomycin D and EMA-CO are deliverable in the public oncology services, whereas pembrolizumab and high-dose chemotherapy with stem-cell support are largely tertiary/limited-access. The recurring SA failure mode is loss to follow-up: a woman evacuated at a district hospital, never registered, never surveilled, who re-presents with metastatic choriocarcinoma — turning a near-100%-curable disease into a lethal one. Building a tracked referral and recall pathway is therefore not administrative tidiness but the core of curing this disease in a fragmented system.

Guidelines compared

The bodies agree on the spine — suction evacuation, hCG surveillance, FIGO scoring, single- versus multi-agent chemotherapy by the ≤6/≥7 cut — and diverge mainly on the duration and intensity of surveillance and on score-band nuances.

The genuinely recent change to flag is the evidence-driven shortening of partial-mole surveillance to a single confirmatory hCG — a direct product of the large Charing Cross cohort showing the post-normalisation pGTN risk after a partial mole is on the order of 1 in 3,000. That is a guideline moving because of a primary study.

The evidence & the controversy

GTD is the cleanest example in O&G of a disease where the evidence base is almost entirely cohort and registry data, not randomised trials, and that limitation should be stated openly. Three threads are worth holding.

First, the one place a randomised question has actually been settled is low-risk first-line chemotherapy. The Cochrane review (seven RCTs, 667 women) found actinomycin D more likely to achieve primary cure than methotrexate (RR 0.65, 95% CI 0.57–0.75) and methotrexate more likely to fail (RR 3.55, 95% CI 1.81–6.95) — yet methotrexate–folinic acid remains many centres' first choice because it is well-tolerated, fertility-sparing and easily salvaged by switching to actinomycin D, with the certainty of cure preserved across the sequence. The defensible position is therefore not "actinomycin D is superior so use it first," but "actinomycin D has a higher single-agent cure rate, methotrexate has a gentler toxicity profile, and because second-line single-agent therapy rescues most methotrexate failures the sequence matters more than the first choice" — a reading of the trial itself rather than its headline.

Second, everything in high-risk and salvage disease rests on single-centre cohorts — overwhelmingly Charing Cross and the French Reference Centre. The case for induction low-dose EP before EMA-CO is cohort evidence (early-death rate falling from around 7% to under 1% in large-burden disease), and the recognition of FIGO ≥13 as an ultra-high-risk group with markedly higher mortality comes from the same kind of data. These are not RCTs and cannot ethically be — the disease is rare and the effect sizes are large — but they are consistent, biologically coherent and centre-replicated, which is the appropriate standard for a rare curable cancer. Saying "this is cohort, not trial, evidence, but it is the best obtainable and it is consistent across centres" is a stronger answer than pretending it is level-1.

Third, the live system-level controversy in a country like South Africa is centralisation and access, not drug choice. GTD outcomes track most strongly with organised, centralised surveillance — standardised assays, a fixed monitoring schedule, a low threshold to treat, and active recall — and the principal cause of avoidable death is loss to follow-up in a fragmented service, not lack of a fancier chemotherapy. The defensible consultant argument is that the highest-yield intervention here is organisational — build the registration-and-recall pathway — because it converts the disease back into the near-universally-curable one it should be, at almost no drug cost.

Landmark trials & key evidence

Worked viva — how to structure the answer

A typical stem: "A 26-year-old woman is referred from a district hospital two weeks after suction evacuation of a complete mole. Her hCG was 180,000 at evacuation; the values over the last three weeks are 4,200 → 4,500 → 4,400. She is well. What is your assessment and plan?" A high-scoring answer runs:

1. Name the diagnosis from the numbers. "These three values are a plateau within 10% over at least three weeks, which meets the FIGO criterion for postmolar gestational trophoblastic neoplasia. This is no longer surveillance — she has GTN and needs treatment."
2. Stage and score before treating. "I would examine her, repeat a quantitative hCG, and image for metastases — chest X-ray to count lung lesions for the score, with liver ultrasound/CT and brain imaging if indicated. I then assign a FIGO stage and a modified WHO score."
3. Let the score choose the regimen. "If her score is ≤6 she is low-risk and I would start single-agent methotrexate with folinic-acid rescue, switching to actinomycin D if she plateaus or can't be dosed, and continue 2–3 consolidation cycles after normalisation — cure approaches 100%. If her score is ≥7 she is high-risk and needs multi-agent EMA-CO; with very high tumour burden I would use induction low-dose etoposide–cisplatin first to prevent early death, and a score of ≥13 would prompt induction EP and specialist-centre management."
4. Run it through a centre and protect follow-up. "I would register her with a gynae-oncology/trophoblast unit so the assay, schedule and threshold to treat are standardised, ensure reliable contraception throughout, and put a recall system in place — loss to follow-up is the main avoidable cause of death from this otherwise curable disease."
5. Justify from evidence. "The framework is the FIGO 2021 update; the single-agent choice is informed by the Cochrane review showing actinomycin D's higher cure rate but methotrexate's tolerability and salvageability; and induction EP rests on cohort data cutting early deaths from ~7% to <1%."

Exam traps & red flags

• Inducing labour or using oxytocics to "deliver" a mole. Forceful contraction against an intact mole risks trophoblastic embolisation/dissemination and is associated with more postmolar GTN. Suction evacuation is the method, with oxytocin started as evacuation proceeds, not before.
• Forgetting anti-D for the partial mole. The PHM has fetal tissue and RhD antigen — RhD-negative women need anti-D. (The complete mole has no fetal RhD antigen; know the distinction and the pragmatic "give if type unknown" default.)
• Reading p57 backwards. p57 is negative in the complete mole (no maternal genome to express the maternally-expressed gene) and positive in the partial mole. Saying it the other way round inverts the diagnosis.
• Trusting a "normal" hCG on a narrow assay. Use an assay detecting all hCG isoforms; a tumour secreting predominantly one variant can read falsely low or negative.
• Treating PSTT/ETT like ordinary GTN. Intermediate-trophoblast tumours secrete little hCG and are chemoresistant — hysterectomy is primary, and hCG is a poor monitor. Pouring on EMA-CO and watching a low hCG is undertreatment.
• Forgetting choriocarcinoma can follow any pregnancy. Unexplained metastatic disease (lung, brain) with a high hCG after a term delivery, miscarriage or ectopic is choriocarcinoma until excluded — not just a post-molar event.
• Hitting massive high-burden disease with full-dose EMA-CO. Sudden tumour collapse causes haemorrhage, metabolic catastrophe and early death; induction low-dose EP first, especially with FIGO score ≥13.
• Mis-scoring by counting CT lung lesions. The prognostic score is calibrated to chest X-ray lung counts; counting extra lesions seen on CT can wrongly upstage the score.
• Missing early-onset pre-eclampsia or hyperthyroidism as a clue. Pre-eclampsia before 20 weeks or unexplained thyrotoxicosis with very high hCG should make you think mole, via the hCG–TSH receptor cross-reactivity.
• Losing the patient to follow-up. In a fragmented service this is the commonest route from a curable disease to a fatal one — registration, surveillance and recall are treatment, not paperwork.

Evidence anchors

• Ngan HYS et al. — Diagnosis and management of gestational trophoblastic disease: 2021 update, Int J Gynecol Obstet 2021;155(Suppl 1):86–93
• Lawrie TA, Alazzam M et al. — First-line chemotherapy in low-risk gestational trophoblastic neoplasia, Cochrane Database Syst Rev 2016;(6):CD007102
• Osborne RJ et al. (GOG-174) — Pulsed dactinomycin beats weekly methotrexate on complete response (70% vs 53%) in low-risk GTN, J Clin Oncol 2011;29(7):825–831
• Coyle C et al. — Optimal duration of hCG surveillance after a molar pregnancy, Gynecol Oncol 2017;148:254–257
• Alifrangis C et al. — EMA/CO for high-risk GTN with induction low-dose etoposide-cisplatin, J Clin Oncol 2012;31:280–286
• Bolze PA et al. — Mortality rate of GTN with a FIGO score ≥13, Am J Obstet Gynecol 2015;214:390.e1–8
• Bolze PA et al. — First-line hysterectomy for low-risk non-metastatic GTN, Gynecol Oncol 2018;150:282–287
• FIGO 2000 anatomical staging and modified WHO prognostic score (reproduced in Ngan et al. 2021, Tables 1 & 2).
• RCOG Green-top Guideline on the Management of Gestational Trophoblastic Disease (UK trophoblast-centre practice).
• South Africa: GTD is managed through tertiary gynae-oncology units (no national registry); NHLS histology and serial quantitative hCG underpin district-to-tertiary referral.