Assess and manage benign ovarian and adnexal masses, including malignancy risk stratification

In one line

Most adnexal masses are benign, so the consultant task is not diagnosis but disciplined risk stratification — using transvaginal ultrasound morphology, age/menopausal status and the right marker — to separate the benign mass that can be watched or removed simply from the one that must reach a gynae-oncologist before a single incision, and to do that without either under-triaging a cancer or over-treating a follicular cyst.

The groundwork on simple cyst types and the differential of pelvic pain is assumed here; revise it at genital tract cysts. The focus is the exit-level judgement: stratifying risk and defending a plan.

Mechanism & pathophysiology

Adnexal masses are a spectrum, and reasoning about risk starts with knowing what each lesion is and how its biology shows on the screen.

Functional cysts are not neoplasms at all — they are exaggerated physiology. A follicular cyst is a dominant follicle that fails to rupture and keeps growing under FSH drive; a corpus luteum cyst is a corpus luteum that fails to involute and may bleed into itself (the haemorrhagic cyst, with its lace-like retracting clot on ultrasound). Both are oestrogen- or progesterone-supported transient structures, which is exactly why most resolve within two to three cycles once the hormonal stimulus passes. Theca-lutein cysts are the extreme of this logic — bilateral, multiloculated ovaries driven by very high β-hCG (multiple/molar pregnancy, ovarian hyperstimulation), which regress as hCG falls. Understanding functional cysts as physiology, not pathology, is what licenses watchful waiting.

Benign epithelial tumours arise from the ovarian surface epithelium (or, on the current pathogenesis model, from tubal/secondary müllerian epithelium). A serous cystadenoma is typically unilocular, thin-walled and anechoic; a mucinous cystadenoma is often large, multilocular, with low-level internal echoes from mucin. They matter because their malignant counterparts (serous and mucinous carcinoma) share the morphological vocabulary — locules, septa, solid areas — so the same ultrasound features that define the benign lesion are the ones that, when they tip over (thick septa, solid papillary projections, vascularity), signal malignancy.

Benign germ-cell tumours are dominated by the mature cystic teratoma (dermoid) — the commonest ovarian neoplasm in women under 30. It contains tissue from all three germ layers (skin, hair, sebaceous material, teeth, occasionally thyroid), which gives its pathognomonic ultrasound signs: the echogenic Rokitansky nodule, hyperechoic lines and dots (hair), and acoustic shadowing from fat/calcium. Its biology drives its complications — fat content makes it buoyant and torsion-prone, and rare malignant transformation (usually squamous carcinoma in older women) is the reason a large or atypical dermoid is not entirely a non-event.

Endometrioma is ectopic endometrium that bleeds cyclically into a cyst, producing the homogeneous low-level "ground-glass" echoes of altered blood. It links directly to the endometriosis story (assume that groundwork) and carries two specific consultant concerns: it distorts pelvic anatomy and impairs fertility, and a small but real subset undergoes malignant transformation to clear-cell or endometrioid carcinoma — a new solid nodule with vascularity inside a long-standing endometrioma is the warning.

Sex-cord-stromal tumours are hormonally active and that is the diagnostic clue. Granulosa-cell tumours secrete oestrogen and inhibin (presenting with postmenopausal bleeding, endometrial hyperplasia, or precocious puberty in girls); Sertoli-Leydig tumours secrete androgens (virilisation). Fibromas/thecomas are benign solid ovarian tumours; a fibroma with ascites and a pleural effusion is Meigs syndrome — a benign cause of the picture that otherwise screams ovarian cancer, and a classic trap.

Para-ovarian and paratubal cysts arise from mesonephric or paramesonephric remnants in the broad ligament; they sit beside a normal ovary, which is the key to recognising them (the "split sign" — separately identifiable ipsilateral ovary). They are almost always benign.

Tubo-ovarian abscess is the inflammatory mass that does not belong to the neoplastic spectrum but is mistaken for it: a complex, often bilateral, multiloculated adnexal mass with thick walls, fever, raised inflammatory markers and a pelvic-infection history. In SA, where HIV and tuberculosis are common, a complex adnexal mass with constitutional symptoms must prompt consideration of pelvic/genital TB before anyone reaches for an oncology operation — granulomatous tubo-ovarian disease can radiologically and biochemically (raised CA-125, ascites) mimic ovarian cancer precisely, and the management is antituberculous therapy, not laparotomy.

The malignant lesions to exclude run alongside this list: epithelial ovarian cancer (the bulk, including high-grade serous, which on the current pathogenesis model arises largely from a precursor lesion at the fimbrial end of the fallopian tube — serous tubal intraepithelial carcinoma — rather than from the ovary itself), the malignant germ-cell tumours of the young (dysgerminoma, yolk-sac, immature teratoma, choriocarcinoma), malignant sex-cord-stromal tumours, and secondary metastases (the Krukenberg tumour — bilateral solid ovarian metastases, classically signet-ring gastric primary). The tubal-origin model matters clinically because it underpins opportunistic salpingectomy as a prevention strategy and explains why removing the tubes, not just the ovaries, reduces future cancer risk. Risk stratification is the discipline of reading this whole spectrum and placing the patient in front of you on it. For the malignant end and its staging and treatment, see ovarian-malignancy.

Assessment

History and examination orient the probability before any test:

• Age and menopausal status are the single most powerful modifiers — the same morphology carries a far higher malignancy risk in a 65-year-old than a 25-year-old, and the marker that helps differs by age. In a woman under ~30 with a solid or mixed mass, think germ-cell tumour and send germ-cell markers; in a postmenopausal woman, think epithelial cancer and the CA-125–based tools.
• Symptoms: most masses are asymptomatic and incidental. Acute severe pain suggests torsion, rupture or haemorrhage; chronic pelvic pain with dysmenorrhoea suggests endometrioma; bloating, early satiety, urinary frequency, weight loss and a palpable mass — especially with ascites — are the sinister cluster of advanced ovarian cancer. Hormonal symptoms (postmenopausal bleeding, virilisation, precocious puberty) point to a sex-cord-stromal tumour.
• Examination: abdominal mass, ascites, an immobile/fixed or bilateral adnexal mass, and nodularity in the pouch of Douglas raise malignancy concern. A normal examination never excludes a mass — imaging is decisive.

Transvaginal ultrasound is the cornerstone investigation. It is far more than "cyst present"; the consultant reads the specific morphological features that the structured tools depend on:

• Reassuring (benign) features: unilocular, anechoic, thin smooth wall, ≤10 cm, no solid component, no vascularity, acoustic shadows (dermoid/calcification), ground-glass content (endometrioma), or a clear single classic pattern recognised by an experienced examiner.
• Concerning (malignant) features: solid components, irregular/thick septa (>3 mm), papillary projections (≥3), bilaterality, ascites, strong internal Doppler flow, irregular solid tumour, and large size.

Where ultrasound is indeterminate, MRI characterises soft tissue best (fat, blood, solid enhancement); CT is for staging suspected malignancy (omental/peritoneal/nodal disease), not for characterising the mass itself.

Tumour markers — interpret, do not just measure.

• CA-125 is the workhorse but its caveats are the exam point. It is a non-specific epithelial glycoprotein raised by many benign conditions — endometriosis, fibroids, pelvic inflammatory disease, menstruation, pregnancy, ascites of any cause, and peritoneal irritation — so it is far less specific in premenopausal women, where these conditions are common. Conversely it is insensitive for early-stage and for mucinous, clear-cell and germ-cell tumours, so a normal CA-125 never excludes cancer. Its real value is in the postmenopausal woman and as one input into a structured index, not as a standalone test.
• HE4 and the ROMA algorithm combine HE4 (less elevated by benign gynaecological disease, so more specific) with CA-125 and menopausal status. HE4/ROMA improve specificity over CA-125 alone, particularly in distinguishing endometriosis and fibroids from cancer, but are less sensitive in premenopausal women, and ROMA cut-offs are assay- and menopause-dependent. HE4/ROMA are not uniformly available in the SA public sector.
• In the young woman with a solid/complex mass, send germ-cell and stromal markers, not CA-125: AFP (yolk-sac tumour, some immature teratomas), β-hCG (choriocarcinoma, dysgerminoma), LDH (dysgerminoma), and inhibin/oestradiol (granulosa-cell tumour). Missing these in a young woman by reflexively ordering CA-125 is a recurrent error.

Structured risk stratification — the core skill

The exit-level competency is converting morphology, marker and menopausal status into a defensible risk category, using one of the validated tools rather than gestalt.

• RMI (Risk of Malignancy Index) is the workhorse triage tool, and its arithmetic is worth knowing exactly: RMI = U × M × CA-125, where U is the ultrasound score (one point each for multilocularity, solid areas, bilaterality, ascites, intra-abdominal metastases → scored 0, 1 or 3), M is menopausal status (1 premenopausal, 3 postmenopausal), and CA-125 is the absolute serum value in U/mL. At a cut-off of 200, a woman with RMI >200 carries roughly 42 times the background risk of cancer and should be referred to a gynae-oncology multidisciplinary team. Its strengths are that the inputs are cheap and universally available; its weakness is poor sensitivity in premenopausal women, where benign CA-125 elevation and the menopausal multiplier of 1 both pull the index down.
• IOTA Simple Rules apply five ultrasound features predicting malignancy (M-features: irregular solid tumour, ascites, ≥4 papillary projections, irregular multilocular-solid ≥10 cm, very strong Doppler) and five predicting benignity (B-features: unilocular, solid components <7 mm, acoustic shadows, smooth multilocular <10 cm, no/low Doppler). One or more M-features without a B-feature = malignant; the reverse = benign; conflicting or absent features = inconclusive (~20%), in which case an experienced examiner's subjective assessment is the best fallback.
• IOTA ADNEX model goes further, giving a probability of malignancy and discriminating benign vs four malignant subtypes (borderline, stage I, stage II–IV, secondary metastatic) from nine predictors — age, serum CA-125, type of centre, and six ultrasound variables. It is the most discriminating of the validated models and supports nuanced counselling, not just a binary triage.
• O-RADS (Ovarian-Adnexal Reporting and Data System) is the ACR's structured reporting and risk-category system (categories 0–5), standardising how a mass is described and which malignancy-risk band and management recommendation follow — increasingly the common language between sonographer and clinician.

The unifying point: in a premenopausal woman an IOTA tool (Simple Rules/ADNEX) outperforms RMI; in the SA public sector, where an experienced gynae-sonographer is not always at hand, RMI remains the deliverable workhorse, escalated to IOTA/ADNEX or a tertiary scan where expertise exists. The tools are not interchangeable across the menopausal divide, and stating that is half the answer:

The deeper interpretive habit is to ask why a number is what it is. A premenopausal RMI of 600 driven entirely by a CA-125 of 200 in a woman with a textbook ground-glass endometrioma is a false alarm; a postmenopausal RMI of 180 (just under the cut-off) with a growing solid component deserves more suspicion than its score, because the index is a triage tool, not a verdict, and morphology can override it in either direction.

Management

Management follows risk category and patient context, organised immediate → ongoing → long-term.

Immediate — triage and emergencies. The first decision is whether this is an emergency (torsion, rupture, haemorrhage) requiring urgent surgery, or an elective problem. The second is the malignancy risk category, which routes the patient: a low-risk benign-appearing mass stays in general gynaecology; a high-risk mass (RMI >200, malignant IOTA/ADNEX result, O-RADS 4–5) is referred to gynae-oncology before surgery so that, if cancer is found, the right operation (full staging, intact removal, appropriate surgeon) is done once.

Expectant management / surveillance. Benign-appearing asymptomatic masses can be watched, and the evidence supports this confidently:

Dermoids and endometriomas that are small, asymptomatic and stable can also be surveilled rather than reflexively excised, balancing the surgery against the ovarian-reserve cost in a woman wanting fertility.

Surgery — the route and the radicality both matter.

• Laparoscopy is preferred for benign-appearing masses (less pain, faster recovery, shorter stay). The proviso is risk-dependent: if malignancy is a genuine possibility, avoid intraperitoneal spillage — upstaging a contained borderline/stage I cancer by rupturing it during laparoscopic removal worsens prognosis, so a suspicious mass is removed intact (in a retrieval bag, or via laparotomy if it cannot be removed whole), and if frank malignancy is encountered, the operation converts to proper surgical staging.
• Ovarian cystectomy vs oophorectomy: in a young woman, cystectomy preserving ovarian tissue is the default for a benign mass; in a postmenopausal woman or where malignancy is suspected, salpingo-oophorectomy (intact) is appropriate. Opportunistic bilateral salpingectomy at the time of pelvic surgery is increasingly offered as ovarian-cancer risk reduction.
• Fertility preservation governs the young woman: spare ovarian cortex, avoid unnecessary oophorectomy, and send germ-cell markers pre-operatively so a malignant germ-cell tumour is not mismanaged. The young patient is a different calculus from the postmenopausal one because the malignancies she is at risk of — germ-cell and, less often, sex-cord-stromal tumours — are typically unilateral, chemosensitive and amenable to fertility-sparing surgery (unilateral salpingo-oophorectomy with staging, preserving the contralateral ovary and uterus). That changes the cost of getting it wrong: rupturing a dermoid spills sebaceous material and causes chemical peritonitis; removing both ovaries of a 22-year-old for what proves benign is an irreversible reproductive injury; but equally, treating a solid mass with a raised AFP as "just a complex cyst" and shelling it out laparoscopically squanders the chance to stage a curable cancer correctly. Pre-operative markers and an honest morphological read are what keep the young woman on the right side of all three errors.

The surgical-staging discipline when malignancy is possible. If a mass is genuinely suspicious, the operation is planned as a potential cancer operation from the outset: a vertical incision (or laparoscopy with a clear conversion plan), peritoneal washings taken on entry, systematic inspection of peritoneal surfaces and the contralateral ovary, intact removal of the mass, and frozen-section guidance where available — with the capacity to proceed to full staging (omentectomy, peritoneal biopsies, nodal assessment) if invasive cancer is confirmed. The reason to insist on this is the same as the reason to refer high-risk masses to gynae-oncology first: a cancer operated as a benign cyst, ruptured and incompletely staged in a general theatre, has had its prognosis worsened before histology is even back.

Ovarian torsion — the time-critical emergency. Torsion of the ovary/adnexa on its pedicle compromises venous then arterial supply; it is a surgical emergency where the ovary's viability is a function of time. The modern standard is laparoscopic detorsion with ovarian conservation, not reflexive oophorectomy: even a dusky, congested ovary frequently recovers function after untwisting, and a black appearance at surgery correlates poorly with irreversible necrosis. Removing the cyst (cystectomy) can be done at the same sitting or as an interval procedure once oedema settles. Oophoropexy is considered for recurrent torsion or where no causative cyst is found. Pre-operative imaging supports but never excludes the diagnosis — a normal Doppler flow does not exclude torsion (arterial flow can persist because the ovary has dual blood supply), so clinical suspicion plus a mass with severe pain mandates surgery, and waiting for imaging confirmation costs ovaries.

The adnexal mass in pregnancy is usually a corpus luteum or other functional/benign cyst found incidentally on the dating scan, and most resolve by the second trimester. The default is expectant management; intervention is reserved for torsion, rupture, a genuinely high malignancy suspicion, or a very large symptomatic mass, with surgery (when needed) best timed to the second trimester.

Ongoing and long-term. Surveillance of conserved masses follows the thresholds above; resolution or stability over interval scans is reassuring. After surgery, histology dictates the path — a benign result closes the episode; an unexpected borderline or invasive cancer reopens it as a gynae-oncology problem (staging, MDT). In SA, the practical long-term task is ensuring surveillance scans actually happen across the district–regional interface rather than defaulting every benign cyst to a tertiary clinic.

When to refer to gynae-oncology is the judgement the whole chapter builds towards — refer before operating when: RMI >200; an IOTA/ADNEX result or O-RADS 4–5 indicating malignancy; ascites with a complex mass; a raised germ-cell marker in a young woman; bilateral solid masses (query metastatic); or any complex mass in a postmenopausal woman. The twin failures are under-triage (operating on a cancer in a general theatre, rupturing it, leaving disease behind) and over-treatment (removing the ovary of a 24-year-old for a follicular cyst that would have resolved).

Guidelines compared

Where they diverge and why. The RCOG anchors triage on RMI (cheap, reproducible, deliverable) while the IOTA group's data show that Simple Rules/ADNEX discriminate better, especially in premenopausal women where RMI's sensitivity is poor — the tension is pragmatism (RMI works everywhere) versus performance (IOTA needs a trained sonographer). ACOG is deliberately less prescriptive about a single index, emphasising the referral decision over the number. The recent shift is towards probabilistic IOTA models and O-RADS standardised reporting as the lingua franca, and towards confident expectant management of benign-morphology masses on the strength of the IOTA5 cohort. For SA, RMI remains the public-sector default with IOTA/ADNEX where ultrasound expertise exists.

The evidence & the controversy

The central modern shift is away from surgery as the default. For decades a persistent adnexal mass earned an operation; the IOTA programme inverted that by showing how safely a benign-appearing mass can be watched. The IOTA5 expectant-management cohort found that, among masses judged benign by an experienced examiner and followed conservatively, the 5-year risk of torsion was effectively nil (0%), cyst rupture 0.1%, and invasive malignancy only 0.2%, while two-thirds resolved spontaneously. The clinical force is real — every avoided operation avoids anaesthetic risk, adhesions and lost ovarian reserve — but the caveat is equally important: that safety depends on the initial classification being made by a competent examiner, which is precisely the bottleneck in a resource-limited service. Watchful waiting is only as safe as the scan that licensed it.

A second, quieter controversy is the CA-125 trap in premenopausal women. Because endometriosis, fibroids and pelvic infection all raise CA-125, an RMI driven by a benign-cause CA-125 of, say, 300 can manufacture a falsely alarming index and route a young woman with an endometrioma towards oncology surgery she does not need. The defensible position is to weight morphology (an IOTA pattern, a classic ground-glass endometrioma) over an isolated CA-125 in the premenopausal woman, and to reserve CA-125's full triage weight for the postmenopausal setting.

Third, how hard to look sits in the current screening debate: population ovarian-cancer screening (the rationale that drove decades of CA-125/ultrasound trials) has not delivered a mortality benefit, which reframes the marker as a triage tool for a mass already found, not a screening tool — an important distinction, and a useful corrective to over-investigation.

Two further threads are moving the field now. Opportunistic salpingectomy — removing the tubes (rather than just performing tubal ligation, or doing nothing) at the time of any pelvic surgery the woman is already having — has gained traction as a primary-prevention strategy precisely because of the tubal-origin model of high-grade serous cancer; it adds little morbidity, preserves ovarian endocrine function, and is increasingly the default counsel at hysterectomy or sterilisation. The honest caveat is that the mortality benefit is inferred from pathogenesis and observational data rather than proven by a completed randomised trial, so it is offered as a low-risk, biologically-rational option, not an absolute. The second thread is machine-learning–assisted ultrasound: automated IOTA/ADNEX-style classifiers are being developed to bring expert-level morphological assessment to settings without a subspecialist sonographer — which, if validated, speaks directly to the SA bottleneck, since the IOTA5 safety of expectant management hinges on the quality of the initial scan. It is promising and unproven at scale; presenting it as ready for routine practice would overreach.

A genuinely contested topical question candidates should be able to frame rather than pronounce on is whether environmental endocrine-disrupting chemicals (bisphenols, phthalates and related compounds) contribute to ovarian pathology and functional-cyst dynamics. The mechanistic plausibility (oestrogenic/anti-androgenic activity perturbing folliculogenesis) is real and the topic recurs in public discourse, but the human epidemiological evidence for a causal link to specific adnexal pathology is weak and confounded. The defensible stance is to acknowledge the biological hypothesis, label the clinical evidence as not yet established, and decline to translate it into clinical advice — exactly the calibrated scepticism the discipline rewards.

Landmark trials & key evidence

A worked figure makes the premenopausal point concrete: in the Kaijser meta-analysis, RMI-1 missed roughly half of malignancies in premenopausal women (sensitivity 0.44), whereas IOTA Simple Rules caught 93% — so in a young woman, choosing the IOTA tool over RMI is not a stylistic preference but the difference between catching one cancer and missing it.

Exam traps & red flags

• Ordering CA-125 in a young woman with a solid mass. Under ~30, the likely malignancy is germ-cell — send AFP, β-hCG, LDH (± inhibin/oestradiol), not CA-125, which is insensitive here.
• Trusting CA-125 in the premenopausal woman. Endometriosis, fibroids, PID, pregnancy and menstruation all raise it; an alarming RMI may be a benign endometrioma. Weight morphology over an isolated marker.
• Rupturing a possible cancer during laparoscopy. Spillage upstages a contained borderline/stage I tumour — a suspicious mass comes out intact (bag or laparotomy); refer high-risk masses to gynae-oncology before operating.
• Removing the ovary in torsion. Detorsion with conservation is standard; a black, congested ovary often recovers — reflexive oophorectomy in a young woman is the error.
• Excluding torsion on a normal Doppler. Arterial flow can persist (dual blood supply); a normal scan does not exclude torsion — severe pain plus an adnexal mass warrants surgery.
• Operating on a follicular cyst. A simple <5 cm premenopausal cyst resolves over 2–3 cycles; surgery is over-treatment that costs ovarian reserve.
• Missing Meigs syndrome. A benign fibroma with ascites and pleural effusion mimics advanced cancer; the picture resolves after removing the tumour.
• Forgetting hormonal clues. Postmenopausal bleeding + adnexal mass → query granulosa-cell (oestrogen, inhibin); virilisation → Sertoli-Leydig (androgens). The endocrine presentation is the diagnosis.
• Bilateral solid ovarian masses. Think Krukenberg / metastatic disease and look for a gastrointestinal primary, not a primary ovarian operation.
• Under-triage vs over-treatment. Both fail the patient: operating on a cancer in a general theatre, and removing a young woman's ovary for a functional cyst.

Evidence anchors

• Jacobs et al. — Risk of Malignancy Index, Br J Obstet Gynaecol 1990
• Timmerman et al. — IOTA Simple Rules prospective validation, BMJ 2010
• Van Calster et al. — IOTA ADNEX model, BMJ 2014
• Pascual et al. — IOTA5 conservative-management cohort, Ultrasound Obstet Gynecol 2024
• Kaijser et al. — meta-analysis of presurgical models, Hum Reprod Update 2014
• Di Legge et al. — effect of lesion size on model performance, Ultrasound Obstet Gynecol 2012
• Sayasneh et al. — IOTA three-step strategy external validation, Gynecol Oncol 2013
• Andreotti et al. — O-RADS US Risk Stratification and Management System, Radiology 2020
• Menon et al. — UKCTOCS: population screening gives a stage shift but no ovarian-cancer mortality benefit, Lancet 2021
• External validation of O-RADS US, Radiology 2022
• RCOG/BSGE Green-top Guideline No. 62 — Management of Suspected Ovarian Masses in Premenopausal Women (2011)
• RCOG Green-top Guideline No. 34 — Ovarian Cysts in Postmenopausal Women
• ACOG Practice Bulletin No. 174 — Evaluation and Management of Adnexal Masses (2016)
• South Africa: adnexal-mass triage runs district → regional → tertiary; CA-125 and TVS widely available, HE4/ROMA and MRI unevenly so, so RMI remains the public-sector workhorse with IOTA Simple Rules/ADNEX used where an experienced sonographer is present.