Assess and manage ovarian malignancies

In one line

Ovarian malignancy is a histologically heterogeneous group dominated by high-grade serous epithelial carcinoma — usually diagnosed late, treated by maximal cytoreduction (to no visible residual disease) plus platinum–taxane chemotherapy, with outcome now driven as much by BRCA/HRD-directed maintenance as by the scalpel; the consultant's task is to triage the adnexal mass correctly, get the right patient to a gynae-oncology MDT before the first incision, and individualise primary surgery versus neoadjuvant chemotherapy.

This chapter assumes the triage groundwork in adnexal mass in pregnancy and endometrial carcinoma assessment; it develops the consultant-level appraisal and the judgement calls — subtype-specific biology, the surgical-pathway controversy, biomarker-directed maintenance, recurrent-disease decisions, and where the SA reality bends the algorithm — on top of that first-principles classification rather than re-deriving it.

Why this matters in South Africa

The SA controversy is access, not algorithm. The interventions that change ovarian-cancer outcome most — early referral, complete cytoreduction by a trained gynae-oncologist, and full-dose platinum–taxane delivery — are the cheap, structural, protocol-driven ones, and they are exactly the ones the SA system most often fails. A sub-Saharan survey found PARP inhibitors generally unavailable to the majority of physicians and bevacizumab to a similar majority, with very few prescribing any maintenance. The consequence for the consultant is a deliberate inversion of the high-income algorithm: you spend your energy on the things that are universally available and outcome-defining (get her staged on imaging and into a centre that can achieve R0, deliver six full cycles), and you treat molecularly-directed maintenance as an aspiration constrained by formulary — while still testing BRCA regardless, because it guides family risk and surgical candidacy even when the drug is not yet on the EML. The decisive moves are running the part of the pathway that saves the life and making the system-level referral call, not reciting a maintenance algorithm the public sector cannot yet fund.

Aetiology / pathophysiology — why the subtypes behave differently

"Ovarian cancer" is a misnomer for a family of biologically distinct diseases that share an anatomical address but not a mechanism, a natural history, or a treatment. The most consequential modern insight is the dualistic (Type I / Type II) model, and the corollary that most "ovarian" high-grade serous carcinoma is in fact tubal in origin. The subtype determines the whole management plan — it is not a pathology-report footnote.

• High-grade serous carcinoma (HGSC) — the Type II disease, ~70% of epithelial cancers. Arises not from ovarian surface epithelium but from the fimbrial end of the fallopian tube, from a precursor lesion — serous tubal intra-epithelial carcinoma (STIC) — that already carries the founding TP53 mutation. It is genomically unstable, almost universally TP53-mutated, frequently BRCA1/2- or HRD-driven, and grows fast with early transcoelomic spread — which is precisely why it presents late, with omental cake and ascites, and why it is exquisitely platinum-sensitive and now PARP-targetable. Consequence: this is the subtype that the entire platinum–taxane–maintenance machine was built for, and the reason opportunistic salpingectomy is reframed as primary prevention.
• Low-grade serous carcinoma (LGSC) — a Type I disease. A different animal: arises through a borderline-serous → LGSC continuum, driven by KRAS/BRAF/MAPK-pathway mutations, genomically stable, TP53 wild-type, indolent, and characteristically chemo-resistant to platinum. Consequence: you cannot judge an LGSC by HGSC rules — surgery matters even more (because chemo does less), it is often oestrogen-receptor positive so endocrine maintenance (e.g. an aromatase inhibitor) is rational, and MEK inhibition is the targeted lever. A "serous carcinoma" that barely responds to carboplatin is usually low-grade, not chemo-failed high-grade.
• Endometrioid and clear-cell carcinomas — Type I, endometriosis-associated. Both arise from endometriosis (atypical endometriosis is the precursor) and carry ARID1A, PIK3CA and (endometrioid) PTEN/CTNNB1 mutations. Endometrioid tumours are often low-grade, frequently co-exist with a synchronous endometrial primary, and are oestrogen-driven (sample the endometrium). Clear-cell carcinoma is the troublemaker: relatively platinum-resistant, prone to venous thromboembolism and paraneoplastic hypercalcaemia, and over-represented in certain populations — so a clear-cell histology re-frames prognosis and lowers the threshold for thromboprophylaxis. Consequence: the endometriosis link is mechanistic, not incidental — it explains the mutation set, the chemo-resistance and the associated cancers.
• Mucinous carcinoma — the impostor. Genuinely primary ovarian mucinous carcinoma is rare; most "mucinous ovarian" tumours are metastases from a GI primary (colon, appendix, stomach, pancreas, biliary). A large, unilateral, mucinous mass demands a hunt for a GI source (colonoscopy, gastroscopy, CEA/CA19-9) before it is labelled primary ovarian — because the management is the GI cancer's. Primary mucinous carcinoma is also relatively platinum-insensitive, so some units treat it with GI-type regimens. Consequence: mislabelling a Krukenberg/appendiceal metastasis as primary ovarian sends the patient down the wrong treatment road entirely; bilateral, multinodular, surface-involving mucinous tumours are metastatic until proven otherwise.
• Non-epithelial tumours. Malignant germ-cell tumours (dysgerminoma, yolk-sac, immature teratoma) arise in young women, are fast-growing but exquisitely chemo-curable, and secrete markers (AFP, β-hCG, LDH) that both diagnose and monitor. Sex-cord stromal tumours (adult granulosa-cell the commonest) are hormonally active (oestrogen → endometrial hyperplasia/carcinoma; inhibin/AMH as tumour markers), indolent, and relapse late — a decade-plus surveillance horizon. Consequence: the markers and the natural history are subtype-specific; a normal CA-125 never excludes these.

The mechanism→management links: HGSC's genomic instability is its platinum/PARP sensitivity; LGSC's genomic stability is its chemo-resistance and its case for surgery-plus-endocrine therapy; the endometriosis-associated tumours' ARID1A/PIK3CA biology is the clear-cell chemo-resistance and VTE risk; and the tubal origin of HGSC is the rationale for opportunistic salpingectomy as prevention.

Assessment

Build on the triage groundwork linked above; the advanced task is risk stratification that decides who needs an oncology surgeon and recognising the atypical presentation that does not fit the late-HGSC stereotype.

• Presentation. No reliable early symptom complex — persistent bloating, early satiety, pelvic/abdominal pain, urinary urgency and altered bowel habit in a woman >50 should prompt CA-125. In SA the modal presentation is advanced (FIGO III–IV) with ascites and an omental cake; a new pleural effusion implies stage IV. The atypical presentations that trip the unwary: a young woman with a rapidly-growing solid mass and abdominal pain (germ-cell — torsion or rupture may be the presenting event); post-menopausal bleeding with an adnexal mass and a thickened endometrium (oestrogen-secreting granulosa-cell tumour driving endometrial pathology); isolated hypercalcaemia in a young woman with a unilateral mass (small-cell carcinoma of the ovary, hypercalcaemic type — aggressive, distinct from epithelial disease); and a "GI cancer" picture (weight loss, altered bowel habit, a unilateral mucinous mass) that is actually a metastasis. Paraneoplastic subacute cerebellar degeneration or a new VTE can also herald an occult ovarian primary.
• Risk of Malignancy Index (RMI). RMI = U × M × CA-125. Ultrasound score U (multilocular cyst, solid areas, bilaterality, ascites, intra-abdominal metastases: 0/1/3); menopausal status M (premenopausal 1, postmenopausal 3). RCOG/BSGE GTG-62 and NICE CG122 both triage RMI ≥ 250 → high risk → refer to a specialist MDT / cancer centre. RMI fails in two settings: it underperforms in premenopausal women (CA-125 rises with endometriosis, fibroids, PID, pregnancy) and has poor sensitivity for the non-serous and borderline tumours whose CA-125 is often normal. IOTA Simple Rules / ADNEX (which model the probability of malignancy directly from ultrasound morphology, and crucially estimate the chance the mass is borderline or metastatic, not just benign-vs-malignant) or ROMA (CA-125 + HE4) sharpen specificity there. ADNEX is increasingly the preferred discriminator in expert hands because it triages which kind of malignancy, which changes the operative plan.
• Markers by likely histology. CA-125 (epithelial, less specific premenopausally); HE4 complements CA-125 in ROMA and is less elevated by endometriosis. In a woman <40 with a solid adnexal mass, add AFP, β-hCG, LDH (germ-cell) and inhibin/AMH (granulosa) — a normal CA-125 does not exclude a malignant germ-cell or sex-cord tumour. For a mucinous mass add CEA and CA19-9 and think GI primary. Interpret CA-125 as a trend and a context, never a yes/no: a level of a few hundred in a post-menopausal woman with a complex solid mass is ominous; the same number in a 28-year-old with known endometriosis is nearly meaningless.
• Imaging & staging work-up. Transvaginal ± transabdominal ultrasound first; CT chest/abdomen/pelvis for disease mapping and a resectability assessment — the upper-abdominal disease patterns that predict an unresectable or high-morbidity debulking (large-volume disease at the porta hepatis, mesenteric root retraction/encasement, lesser-sac and gastrosplenic disease, diffuse small-bowel serosal studding, parenchymal liver or extra-abdominal metastases). This radiological resectability map is what the MDT uses to choose PDS vs NACT — it is the heart of the modern assessment, not a box-tick. Image-guided core biopsy (preferred over cytology — it subtypes and grades, enabling HRD testing) where neoadjuvant chemotherapy is contemplated. Never aspirate or rupture a presumed early-stage tumour: spill upstages IA/IB disease to IC and worsens prognosis.
• Staging is surgical, but you stage on imaging first. FIGO 2014 staging (the version that incorporated tubal/peritoneal origin) is surgicopathological — stage I confined to ovaries/tubes (IC subdivided by how the capsule was breached: IC1 surgical spill, IC2 pre-operative rupture or surface tumour, IC3 malignant ascites/washings — a distinction with prognostic weight and a direct argument against careless intra-operative spill); stage II pelvic extension; stage III peritoneal spread beyond the pelvis and/or retroperitoneal nodes; stage IV distant (IVA malignant pleural effusion, IVB parenchymal/extra-abdominal). The consultant point: nodal disease alone now counts as stage IIIA1, but the LION trial showed removing clinically-normal nodes confers no survival benefit — so "stage" and "what to resect" have diverged.
• Genetics. Germline ± somatic BRCA1/2 and HRD testing in all high-grade non-mucinous epithelial cancers at diagnosis — it now directs maintenance, not just family counselling. Test reflexively at diagnosis, not at relapse: the result times the first-line maintenance decision and triggers cascade screening of relatives regardless of whether the drug is funded.

Differential diagnosis & the mimics that change management

The differential of a complex adnexal mass is not academic — the wrong label sends the patient to the wrong surgeon and the wrong chemotherapy.

• Metastasis masquerading as primary ovarian cancer. A unilateral (or bilateral, multinodular) mucinous or solid mass with normal-ish CA-125 but raised CEA/CA19-9 is a GI metastasis (Krukenberg from stomach, or colorectal/appendiceal) until a primary is excluded. Bilateral solid ovarian masses with signet-ring cells are metastatic by definition. Treating these as primary ovarian denies the patient the correct (GI) oncological pathway.
• Borderline (low malignant potential) tumour vs invasive carcinoma. A young woman with a large multilocular-solid mass and modest CA-125 may have a borderline tumour — excellent prognosis, fertility-sparing surgery, no chemotherapy. Over-treating a borderline tumour as carcinoma (radical surgery, adjuvant platinum) is a real harm; the distinction is histological (no stromal invasion) and demands a staging laparotomy with frozen-section judgement, not a guess.
• Tubo-ovarian abscess / endometrioma / pelvic TB. In SA, pelvic tuberculosis is a genuine mimic — peritoneal nodularity, ascites, a raised CA-125 and an "omental cake" picture in a young woman, which can be misread as carcinomatosis. Ascitic-fluid ADA, an appropriate clinical picture and tissue (peritoneal biopsy) prevent a catastrophic mislabelling. An endometrioma with a solid mural nodule raises the spectre of malignant transformation (clear-cell/endometrioid).
• Functional and benign masses. Haemorrhagic corpus-luteum cysts, dermoids and hydrosalpinges have characteristic morphology; the error is operating on a benign mass with the spillage and over-staging that a malignancy protocol entails, or — the more dangerous error — the reverse.
• The microangiopathy / paraneoplastic distractors. A new VTE, hypercalcaemia or cerebellar syndrome can be the first sign; chasing the syndrome without imaging the pelvis misses the cancer.

Management

Frame answers immediate → ongoing → long-term, and state explicitly that definitive surgery belongs in a gynae-oncology MDT (district → regional → tertiary referral in SA; the diagnostic generalist stabilises, stages on imaging and refers before operating).

Immediate. Resuscitate the patient with tense ascites/effusion or bowel obstruction; confirm tissue/cytological diagnosis (core biopsy preferred); complete CT staging; VTE prophylaxis (low threshold in clear-cell). Decide the surgical pathway at MDT:

The cytoreduction endpoint is the single modifiable prognostic driver, and the definitions matter. Outcome tracks with residual disease in three tiers — R0 (no macroscopic residual) is the goal and the only one associated with the best survival; R1 (≤1 cm "optimal") is the older threshold and is now second-best, not the target; R2 (>1 cm, "suboptimal") confers little surgical benefit. The consultant framing is that the goal has shifted from "optimal" to complete (R0) — and that a planned operation that cannot achieve R0 in a fit patient should prompt the NACT route rather than a heroic suboptimal debulking. Cytoreduction in advanced disease routinely means multivisceral surgery: total hysterectomy + bilateral salpingo-oophorectomy, omentectomy, peritoneal stripping (including diaphragmatic), bowel resection, splenectomy and appendicectomy as needed — which is exactly why it belongs with a trained gynae-oncologist in a centre resourced for it, and why the radiological resectability map drives the decision.

Lymphadenectomy is no longer routine. The LION trial established that in advanced disease completely resected to R0 with clinically normal nodes, systematic pelvic and para-aortic lymphadenectomy adds morbidity (more re-operation, more transfusion, higher post-operative mortality) with no survival gain — median OS 69.2 months without versus 65.5 months with lymphadenectomy. Resect bulky/suspicious nodes; do not strip clinically normal ones. This is a clean example of more surgery ≠ better surgery.

Ongoing — systemic therapy. Platinum–taxane is the chemotherapy backbone: carboplatin AUC 5–6 + paclitaxel 175 mg/m² 3-weekly × 6. Two consultant points on the named regimen:

• Dosing carboplatin by AUC, not body-surface-area, via the Calvert formula — dose (mg) = AUC × (GFR + 25). The "+25" approximates non-renal clearance; the AUC target (5–6) is the therapeutic intensity. Because the dose is GFR-driven, renal function must be measured/estimated before every cycle — this is why clear-cell tumours (VTE, possible obstructive uropathy) and elderly patients need careful GFR assessment, and why a falsely high estimated GFR over-doses and causes profound myelosuppression.
• Regimen variants and how they differ. Standard is 3-weekly carboplatin AUC 5–6 + paclitaxel 175 mg/m². Dose-dense weekly paclitaxel (80 mg/m² weekly) with 3-weekly carboplatin (the JGOG regimen) showed a PFS/OS signal in a Japanese population but was not replicated in Western trials (ICON8, GOG-262) — so it is an option in selected patients, not the default. Intraperitoneal chemotherapy (the GOG-172 IP regimen) improves survival in optimally-debulked stage III but at a heavy toxicity/logistics cost and has largely been displaced. Single-agent carboplatin is the pragmatic choice for the frail or those who cannot tolerate paclitaxel neuropathy. The backbone is fixed and the modifications are toxicity- and fitness-driven.

Add bevacizumab (concurrent + maintenance) for high-risk stage III–IV where funded — watch its specific toxicities (hypertension, proteinuria, impaired wound healing → avoid peri-operatively, and bowel perforation, the feared complication, especially with bowel-serosa disease). Then maintenance, biomarker-directed:

• BRCA-mutated → olaparib maintenance (SOLO-1).
• HRD-positive (incl. BRCA) receiving bevacizumab → olaparib + bevacizumab (PAOLA-1).
• HRD-negative → bevacizumab maintenance or observation.

Subtype-specific management — where the algorithm forks:

• Borderline (low malignant potential) tumours: typically younger women, excellent prognosis; cystectomy or unilateral salpingo-oophorectomy with peritoneal staging preserves fertility; no adjuvant chemotherapy. The only nuance is the serous borderline tumour with invasive implants, which behaves more like LGSC and warrants closer follow-up.
• Low-grade serous carcinoma: maximal surgery is more important because platinum does less; it is usually ER-positive, so endocrine maintenance (aromatase inhibitor) is rational, and MEK inhibition is the emerging targeted lever. Do not interpret poor platinum response as treatment failure — it is the biology.
• Clear-cell and mucinous carcinoma: relatively platinum-resistant; ensure the mucinous tumour is not a GI metastasis before committing to ovarian-type chemotherapy; aggressive surgery and thromboprophylaxis (clear-cell) matter.
• Malignant germ-cell tumours (dysgerminoma, yolk-sac, immature teratoma): chemosensitive — fertility-sparing unilateral salpingo-oophorectomy + staging at any stage when fertility is desired, preserving the contralateral ovary and uterus. Adjuvant BEP (bleomycin, etoposide, cisplatin) × 3–4 for all except stage IA dysgerminoma and stage IA grade-1 immature teratoma (surgery alone, with surveillance). Outcomes are excellent and most women resume menses and conceive. BEP toxicity to flag: bleomycin pulmonary fibrosis (baseline and monitored lung function; cap cumulative dose; caution with high-FiO₂ anaesthesia later), cisplatin nephro- and ototoxicity, and the etoposide late risk of secondary leukaemia.
• Sex-cord stromal (granulosa-cell): indolent, late relapse (sometimes >10 years, so surveillance is long-horizon); surgery is primary; inhibin/AMH track disease; sample the endometrium (oestrogen excess → hyperplasia/carcinoma — see endometrial-cancer).

Long-term — surveillance, recurrence and the recurrent-disease decision. Structured surveillance (symptoms, examination, CA-125; routine imaging only if clinically driven). Counsel that a CA-125 rise alone, without symptoms or radiological disease, does not mandate immediate chemotherapy — the MRC OV05/EORTC 55955 trial showed that starting chemotherapy on a biochemical (CA-125) relapse alone, before symptoms, gives no survival benefit and worse quality of life than waiting for clinical relapse. Do not commit a well woman to early chemotherapy on a number.

When relapse is established, classify by the platinum-free interval (PFI) — the central decision variable: platinum-sensitive (PFI ≥ 6 months, re-challenge with a platinum doublet) versus platinum-resistant (PFI < 6 months, single-agent non-platinum chemotherapy ± bevacizumab, palliative intent, poor prognosis). The PFI concept is increasingly seen as a spectrum rather than a cliff, but it remains the practical fork. Secondary cytoreductive surgery is selectively offered in platinum-sensitive relapse — and this is a live controversy worth arguing: DESKTOP III (patients selected by a positive AGO score — good performance status, no ascites, and complete resection at first surgery) showed secondary surgery improved median OS (53.7 vs 46.0 months), with the benefit confined to those achieving complete resection (61.9 vs 46.0 months), whereas GOG-0213 (a less stringently selected population) showed no survival benefit. The reconciliation: secondary surgery helps only the rigorously selected patient in whom complete resection is achievable — selection, not surgery per se, is the discriminator, exactly as in the primary setting. Maintenance in recurrence: a PARP inhibitor after response to platinum re-challenge prolongs PFS (SOLO2 in BRCA-mutated relapse; NOVA and ARIEL3 across broader populations) — though benefit is attenuated and resistance to prior PARP exposure is an emerging problem.

Genetic follow-up. Cascade BRCA testing of relatives; risk-reducing salpingo-oophorectomy (RRSO) by the late 30s–mid 40s (earlier for BRCA1 than BRCA2, reflecting the earlier onset in BRCA1) and breast surveillance ± risk-reducing mastectomy for carriers. Opportunistic salpingectomy at the time of benign hysterectomy or sterilisation is now offered to the general population as primary prevention, on the strength of the tubal (STIC) origin of HGSC.

The evidence & the controversy

The central modern surgical debate is PDS versus NACT. EORTC 55971 and CHORUS established that NACT–IDS is non-inferior to PDS for survival in bulky stage IIIC–IV disease, with markedly less perioperative morbidity. The viva nuance is patient selection: the pooled individual-patient analysis showed NACT is better in stage IV (HR 0.76) but that a fit patient whose disease can be cleared to R0 still does best with upfront surgery — the single strongest modifiable prognostic factor remains complete cytoreduction, not the sequence. Critics note CHORUS recruited lower-volume surgical centres with modest optimal-debulking rates; specialist high-complexity surgery may shift the balance back toward PDS where that capability exists. The honest synthesis: NACT legitimised a less-morbid pathway for the patient who cannot be cleared upfront, but it is not a licence to default away from surgery in the fit, resectable patient — and the choice should be made by the operating gynae-oncologist on the resectability map, not by the referring generalist.

Surgical extent is being de-escalated where it does not help and escalated where it does. LION removed routine lymphadenectomy from the completely-resected patient with normal nodes (no OS benefit, more morbidity). Conversely, HIPEC at the time of interval debulking earned a place: OVHIPEC-1 randomised stage III patients at IDS to cisplatin HIPEC and found a median OS gain (54.3 vs 45.8 months; HR 0.73, 95% CI 0.56–0.96), sustained on long-term follow-up. HIPEC remains controversial and is not standard everywhere (logistics, toxicity, generalisability beyond the trial's specific IDS setting), and it is firmly an aspirational/centre-of-excellence intervention in the SA context — but it is the counter-current to LION in the "how much surgery" debate.

The second major shift is maintenance therapy. Anti-angiogenesis (GOG-218, ICON7) improved PFS but the OS gain was confined to the high-risk subgroup, so bevacizumab is not universal. The transformative data are the PARP inhibitors: SOLO-1 showed maintenance olaparib cut progression/death by 70% (HR 0.30) in newly-diagnosed BRCA-mutated disease, and PAOLA-1 extended benefit to the broader HRD-positive population (median PFS 37.2 vs 17.7 months, HR 0.33) when added to bevacizumab. In the recurrent platinum-sensitive setting, SOLO2/ENGOT-Ov21 showed maintenance olaparib prolonged PFS (19.1 vs 5.5 months; HR 0.30) in BRCA-mutated relapse. This is why BRCA/HRD status is now a first-line decision, per the ESGO-ESMO-ESP consensus — though a counter-current of caution has emerged over PARP-inhibitor use in HRD-negative recurrent disease (signals of detriment in subgroups, plus the practical problem of post-PARP resistance), so the maintenance decision is biomarker-gated, not reflexive.

The recurrent-surgery debate (DESKTOP III positive vs GOG-0213 negative) and the "don't treat the CA-125 rise" lesson (MRC OV05) are covered under Management above; together they make the same consultant point that runs through this whole disease — selection beats intervention, and a number is not a reason to treat.

The SA controversy is access, not algorithm (developed in Why this matters in South Africa). Argue maintenance as an aspiration constrained by formulary, and test BRCA regardless — it still guides family risk and surgical candidacy even when the drug is not yet on the EML.

Landmark trials & key evidence

Worked viva — how to structure the answer

Examiners give a stem like "a 58-year-old presents with abdominal distension and early satiety; CT shows bilateral complex adnexal masses, an omental cake, ascites and a small pleural effusion; CA-125 is markedly raised." A high-scoring answer runs:

1. Frame it — "This is an advanced epithelial ovarian malignancy, radiologically FIGO IV given the pleural effusion, most likely high-grade serous of probable tubal origin; my job is to confirm tissue, map resectability and get her to a gynae-oncology MDT before any surgery."
2. Assess and risk-stratify — confirm the diagnosis with a core biopsy (not aspiration), pleural-fluid cytology for the stage-IV call, complete CT mapping of the upper-abdominal disease, baseline performance status/albumin/renal function, VTE prophylaxis, and reflex BRCA/HRD testing.
3. Decide the pathway — "Given stage IV with extensive upper-abdominal disease and an effusion, the EORTC 55971/CHORUS evidence supports NACT–IDS rather than upfront surgery here; carboplatin AUC 5–6 + paclitaxel 175 mg/m², reassess for interval debulking aiming for R0, and I would not strip clinically normal nodes (LION)."
4. Plan maintenance honestly — "If BRCA/HRD-positive I would want PARP maintenance (SOLO-1/PAOLA-1), but in our setting I flag this as formulary-dependent and prioritise full-dose chemotherapy and complete cytoreduction, which change outcome universally; I test BRCA regardless for family cascade."
5. Anticipate the subtype fork — "If histology comes back low-grade serous, I expect platinum-resistance and lean on surgery plus endocrine therapy; if mucinous, I exclude a GI primary before calling it ovarian; if she were young with a solid mass and raised AFP/β-hCG, this is a germ-cell tumour managed fertility-sparingly with BEP."
6. Close the loop — surveillance on symptoms and CA-125 (but don't treat an asymptomatic CA-125 rise — OV05), classify any relapse by platinum-free interval, consider secondary cytoreduction only in the rigorously-selected (DESKTOP III), and arrange cascade genetic testing and RRSO for carriers.

Exam traps & red flags

• Aspirating or morcellating a suspicious cyst — rupture upstages early disease (IA → IC1) and worsens prognosis; never enucleate a solid adnexal mass outside a staging operation.
• Operating before MDT referral. A generalist who debulks a stage III cancer incompletely denies the patient R0 by a gynae-oncologist; stage on imaging, refer, then operate.
• A normal CA-125 reassuring you in a young woman — germ-cell and sex-cord tumours need AFP/β-hCG/LDH and inhibin; a solid ovarian mass under 40 is a germ-cell tumour until proven otherwise.
• Calling a mucinous mass "primary ovarian" without excluding a GI primary — most are metastatic; check CEA/CA19-9 and scope the gut, because the treatment is the GI cancer's.
• Judging a low-grade serous carcinoma by high-grade rules — its platinum-resistance is biology, not treatment failure; surgery and endocrine therapy, not more carboplatin.
• Stripping clinically normal lymph nodes at debulking — LION showed no benefit and more harm; resect only bulky/suspicious nodes.
• Treating an asymptomatic CA-125 rise — OV05 showed early chemotherapy gives no survival benefit and worse quality of life; wait for clinical relapse.
• Offering secondary cytoreduction to an unselected relapse — only the AGO-score-positive patient in whom complete resection is achievable benefits (DESKTOP III vs GOG-0213); selection is everything.
• Giving bevacizumab to everyone — OS benefit is high-risk-restricted; watch bowel perforation (especially with bowel-serosa disease), hypertension and impaired wound healing (avoid peri-operatively).
• Forgetting BEP toxicity — bleomycin causes pulmonary fibrosis; check lung function, cap cumulative dose, and beware high-FiO₂ anaesthesia later.
• Mis-dosing carboplatin — it is AUC-dosed by the Calvert formula off GFR; a falsely high estimated GFR over-doses and causes profound myelosuppression — measure renal function before every cycle.
• Missing the BRCA conversation. Failing to test high-grade serous cancer forfeits both maintenance eligibility and cascade family screening — a defendable plan tests at diagnosis.
• Mistaking pelvic TB for carcinomatosis (or vice versa) in a young SA patient with ascites, peritoneal nodularity and a raised CA-125 — get tissue and ascitic ADA before committing to a cancer pathway.

Evidence anchors

• Vergote et al. NEJM 2010 — EORTC 55971
• Kehoe et al. Lancet 2015 — CHORUS
• Vergote et al. Lancet Oncol 2018 — EORTC/CHORUS pooled analysis
• Harter et al. NEJM 2019 — LION (lymphadenectomy in advanced ovarian cancer)
• van Driel et al. NEJM 2018 — OVHIPEC-1 (HIPEC at interval debulking)
• Burger et al. NEJM 2011 — GOG-218
• Perren et al. NEJM 2011 — ICON7
• Moore et al. NEJM 2018 — SOLO-1
• Ray-Coquard et al. NEJM 2019 — PAOLA-1
• Pujade-Lauraine et al. Lancet Oncol 2017 — SOLO2/ENGOT-Ov21 (PARP maintenance in relapse)
• Harter et al. NEJM 2021 — DESKTOP III (secondary cytoreduction)
• Coleman et al. NEJM 2019 — GOG-0213 (secondary cytoreduction, negative)
• Rustin et al. Lancet 2010 — MRC OV05/EORTC 55955 (CA-125-triggered treatment)
• Menon et al. Lancet 2021 — UKCTOCS (population ovarian-cancer screening shows no mortality benefit)
• Prat J, FIGO 2014 staging (abridged republication), J Gynecol Oncol 2015
• Ledermann et al. Ann Oncol 2024 — ESGO-ESMO-ESP consensus on ovarian cancer
• Sub-Saharan Africa epithelial ovarian cancer physician survey, ecancer 2026
• RCOG/BSGE Green-top Guideline No. 62 — Management of suspected ovarian masses; NICE CG122 — Ovarian cancer: recognition and initial management (both triage RMI ≥ 250 to a specialist MDT)
• SA NDoH / SAMF: platinum–taxane on the EML; gynae-oncology MDT referral via district → regional → tertiary pathways