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 only if the pleural cytology is positive. 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). NICE CG122 triages RMI ≥ 250 → high risk → refer to a specialist MDT / cancer centre (RCOG Green-top Guideline No. 34, the postmenopausal-specific guideline, uses RMI ≥ 200; GTG-62 is a premenopausal guideline and does not set the 250 threshold). 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.