Assess and manage cervical cancer, including recurrent disease and post-treatment care

In one line

Cervical cancer is an HPV-driven malignancy staged by FIGO 2018 (now incorporating imaging and nodal status); cure depends on getting the first treatment right — radical surgery for genuinely early disease, definitive cisplatin-based chemoradiation plus brachytherapy (never surgery plus radiation if avoidable) for everything locally advanced — and in South Africa most patients still present too late for surgery to be the question.

This chapter assumes the premalignant groundwork: HPV carcinogenesis, CIN and colposcopy. Revise those at cervical carcinogenesis and CIN pathophysiology, and the screening/colposcopy interface at cervical-premalignancy-colposcopy. The focus here is the consultant decision: defending a treatment plan.

Assessment

• History/examination: abnormal bleeding (postcoital, intermenstrual, postmenopausal), discharge, pelvic/back pain or leg oedema (advanced). Examine the cervix directly; a friable, contact-bleeding lesion needs a biopsy, not a smear — cytology is a screening tool and a normal smear never excludes a visible cancer.
• Tissue diagnosis: punch/wedge biopsy of the lesion; cone/LLETZ where microinvasion is suspected and depth of invasion must be measured. Squamous (~70%) and adenocarcinoma (~25%, rising); note neuroendocrine and rare histologies, which behave aggressively.
• Staging is FIGO 2018, now hybrid clinical–radiological–pathological. The 2018 revision removed lateral extension from stage IA (depth only), split IB into IB1 (<2 cm), IB2 (2–<4 cm), IB3 (≥4 cm), and created stage IIIC for nodal disease (IIIC1 pelvic, IIIC2 para-aortic) — any tumour size with positive nodes upstages to IIIC. Suffixes r (imaging) and p (pathology) record how nodal status was assigned.
• Imaging: pelvic MRI is best for local tumour size, parametrial and stromal invasion; PET-CT (or CT where unavailable) for nodal and distant disease. In SA, MRI access is uneven — at district/regional level, examination under anaesthesia with cystoscopy/proctoscopy and CT still carry staging weight, and the r notation is honest about that.
• Baseline: FBC, U&E/creatinine (hydronephrosis from ureteric obstruction = stage IIIB and a renal-function threat), and an HIV test in every patient — non-negotiable in SA, where HIV both drives incidence and shapes treatment tolerance.

The advanced assessment — subtype, severity and the judgement calls

The diagnosis is rarely the difficulty; the harder consultant task is reading the histological subtype, the stromal/nodal geometry and the host and letting those change the plan. With the "how to take a biopsy and read a smear" groundwork assumed, the advanced layer is what follows.

Histological subtype changes prognosis and the threshold for adjuvant treatment, not just the label. The trials that built modern practice (SHAPE, the Sedlis intermediate-risk criteria) were derived overwhelmingly from squamous disease, so the further the patient is from squamous, the more you should distrust a reassuring stage:

• Squamous (~70%) — radiosensitive, the reference subtype for every chemoradiation trial; the one where omitting brachytherapy is most clearly fatal to local control.
• Usual-type (HPV-associated) adenocarcinoma (~25% and rising) — behaves broadly like squamous, but be alert to skip lesions and a higher rate of ovarian and peritoneal spread; the cervix can look deceptively normal with an endocervical (barrel) tumour growing upward, so a normal-looking ectocervix with an abnormal smear and a bulky cervix on bimanual is adenocarcinoma until proven otherwise.
• Gastric-type / HPV-independent adenocarcinoma — the dangerous mimic: HPV-negative, vaccine-irrelevant, often cytology-negative (it is not picked up by an HPV-based screen at all), associated with Peutz–Jeghers/STK11, frequently presents with watery discharge and a deeply infiltrative tumour, and is relatively chemoradiation-resistant. The Silva pattern-based classification of endocervical adenocarcinoma (pattern A purely glandular/indolent → pattern C destructive/node-positive) is the modern way to stratify these and decide who can be spared lymphadenectomy.
• Small-cell / large-cell neuroendocrine carcinoma — a different disease: early haematogenous and nodal spread, treated like small-cell lung cancer (platinum–etoposide systemic chemotherapy integrated with local therapy) regardless of how "early" the stage looks. Staging it as an ordinary IB1 and doing a radical hysterectomy alone is a classic error — even small NEC needs systemic therapy.

Severity stratification beyond the FIGO number. Two patients can share a stage and need opposite plans. The features that move a borderline-operable tumour towards primary chemoradiation are: tumour ≥4 cm (IB3), parametrial reach on MRI, suspicious pelvic or para-aortic nodes on PET-CT, deep stromal invasion, and extensive LVSI — because any one of them predicts the patient will earn adjuvant radiation after surgery, and stacking surgery and radiation is the "double whammy" to be designed out before the first incision. Conversely, the SHAPE-eligible patient (genuinely ≤2 cm, <10 mm stromal invasion or <50% on MRI, node-negative) is the one in whom you can now de-escalate.

The host is part of the stage in South Africa. An immunosuppressed woman with a low CD4 count, anaemia and renal impairment from ureteric obstruction tolerates pelvic chemoradiation differently from a fit HIV-negative woman with the same FIGO stage. Quantify host reserve — CD4 and viral load, haemoglobin, eGFR, nutritional state — because these, not the FIGO digit, predict whether she will complete treatment within the time window that determines cure.

Management

Organise the answer immediate → ongoing → long-term, and the single most important decision is avoiding the "double whammy" of radical surgery followed by adjuvant radiation: if pre-operative features predict adjuvant chemoradiation, go straight to primary chemoradiation and spare the patient the morbidity of both.

• Immediate / definitive (surgery): for operable disease, open radical hysterectomy is mandatory — the LACC trial showed minimally invasive surgery roughly quadrupled recurrence and halved disease-free survival, reversing prior practice overnight. Sentinel-node mapping (indocyanine green) is replacing full lymphadenectomy in early disease to reduce lymphoedema. Fertility-sparing radical trachelectomy is an option for selected IB1 (<2 cm) tumours.
• Immediate / definitive (chemoradiation): for locally advanced disease, external-beam RT (45–50.4 Gy) with concurrent weekly cisplatin 40 mg/m² plus image-guided brachytherapy is the backbone. Brachytherapy is not optional — omitting it (a real SA access problem) sharply worsens local control and survival. Complete treatment within ~7–8 weeks; prolonged overall treatment time reduces cure.
• Ongoing: manage anaemia (transfuse to keep cisplatin and oxygenation effective), renal protection and antiemetics, and proactively address HIV — start/optimise ART (the SA term is now PVT-era HIV care, not the old PMTCT framing), because immune status affects both tolerance and outcome.
• Long-term / post-treatment care: surveillance is clinical (symptom-directed history, examination, vaginal vault assessment) — routine cytology of an irradiated vault is unreliable. Counsel on and treat the survivorship burden: vaginal stenosis (dilators, oestrogen), radiation cystitis/proctitis, lymphoedema, premature menopause (offer HRT — squamous and most adenocarcinomas are not hormone-driven), bowel/bladder dysfunction, and sexual/psychological sequelae. In SA, link to district-level follow-up so tertiary clinics are not the only safety net.

Surgery in depth — radicality, route and the adjuvant trap

The consultant decisions in operable disease are how radical, by what route, and what features mandate adjuvant therapy — and the three interact.

Radicality is a spectrum, named (Querleu–Morrow A–D / the older Piver–Rutledge I–V). A type A (extrafascial/simple) hysterectomy takes no parametrium; type B takes the medial parametrium and ureteric tunnel; type C (the classic radical, Wertheim–Meigs) takes the full lateral parametrium to the pelvic sidewall and the upper vagina, dividing the uterosacral and cardinal ligaments and unroofing the ureter — which is exactly why bladder dysfunction follows it. The SHAPE trial's contribution is that for a strictly defined low-risk early tumour (FIGO 2009 IA2–IB1, ≤2 cm, depth of stromal invasion <10 mm on cone or <50% on MRI, node-negative) a simple hysterectomy is non-inferior for pelvic recurrence (3-year extra-pelvic recurrence-free survival 98.1% vs 99.7%; overall survival 99.1% vs 99.4%) with markedly less urological morbidity and better sexual function — so you do less parametrial dissection in the right patient. The two de-escalation vectors must not be confused: SHAPE de-escalates radicality, LACC fixes route. The answer is "less radical, but still open" — never "less radical, therefore laparoscopic."

The adjuvant trap is governed by the Sedlis and Peters/GOG-109 criteria — keep the two distinct.

• Sedlis (intermediate-risk → adjuvant pelvic radiation): node-negative, margin-negative, parametrium-negative disease that still carries a high recurrence risk by a combination of LVSI, depth of stromal invasion and tumour size — e.g. positive LVSI + deep-third invasion (any size); positive LVSI + middle-third invasion + tumour ≥2 cm; positive LVSI + superficial-third invasion + tumour ≥5 cm; or LVSI-negative + middle/deep-third invasion + tumour ≥4 cm. These patients get adjuvant radiation.
• Peters / GOG-109 (high-risk → adjuvant chemoradiation): positive nodes, positive parametrium, or positive surgical margins. GOG-109 (Peters 2000) showed that adding concurrent cisplatin to adjuvant pelvic radiation improved overall and disease-free survival over radiation alone in these patients — so high-risk disease earns adjuvant chemoradiation, not radiation alone.

The clinical force of these criteria is preventive: if pre-operative MRI/PET-CT predicts the patient will meet Sedlis or Peters, she is already a chemoradiation patient and should not be operated on at all. Operating and then discovering positive nodes converts a single curative modality into two morbid ones.

Fertility-sparing and ovarian-sparing nuance. Radical (vaginal or abdominal) trachelectomy with pelvic node assessment is offered for selected IB1 <2 cm tumours in a woman wanting fertility, after confirming node-negativity (the nodes are assessed first — positive nodes abort the trachelectomy and convert to chemoradiation). The CONTESSA/NeoCON-F direction explores neoadjuvant chemotherapy then fertility-sparing surgery for 2–4 cm tumours, but this is not yet standard. In a premenopausal woman needing pelvic radiation, ovarian transposition (oophoropexy out of the radiation field) at the time of surgery preserves endocrine function — a cheap, high-value step that is easy to forget.

Chemoradiation in depth — the named regimens, the dose, and the time window

The backbone for locally advanced disease (IB3, IIA2–IVA, and any node-positive disease) is concurrent cisplatin-based chemoradiation followed by brachytherapy, and the consultant detail is in the regimen, the dose, and the calendar.

• The concurrent chemotherapy regimen. The SA/NDoH and international default is weekly cisplatin 40 mg/m² (capped near 70 mg) during external-beam radiation, typically for 5–6 cycles, with pre-hydration and renal monitoring. The alternative, derived from the original Morris/GOG trials, is cisplatin 50–75 mg/m² plus 5-fluorouracil ~1000 mg/m²/day over 4 days, every 3 weeks — more toxic, no proven superiority over weekly cisplatin, and rarely used now. Where cisplatin is contraindicated (eGFR too low, significant neuropathy or ototoxicity), carboplatin (AUC ~2 weekly) is the substitute, accepting that the platinum-sensitising evidence is strongest for cisplatin.
• The radiation dose and target. External-beam pelvic radiation delivers 45–50.4 Gy in 1.8 Gy fractions to the pelvis (extended to the para-aortic field for IIIC2/para-aortic nodal disease), with a nodal boost where nodes are involved. This is not curative on its own.
• Brachytherapy is the part that cures, and it is not interchangeable with more external beam. Image-guided (MRI-based) adaptive brachytherapy delivers a high dose to the residual central tumour that external beam cannot safely reach. The EMBRACE-I cohort showed what modern IGABT achieves — a 5-year local control of ~92% and overall survival ~74% across all stages (and 5-year OS in the order of ~64% for IIIB and ~52% for IVA, disease that was once considered incurable). Replacing brachytherapy with an external-beam boost (the temptation when a brachytherapy machine is down, a recurring SA reality) demonstrably worsens local control and survival; it is undertreatment, and naming it as such is part of the answer.
• The 56-day rule. Cervical cancer is one of the few solid tumours where overall treatment time is itself a prognostic variable: prolonging total external-beam-plus-brachytherapy time beyond ~8 weeks (≈56 days) loses local control and survival at roughly 1% per day. Accelerated repopulation of clonogens during a protracted course is the mechanism. The practical corollary in a resource-limited service is brutal: gaps caused by machine breakdown, transport failure, anaemia delaying a cisplatin dose, or a missed brachytherapy slot are not administrative inconveniences — each lost day is measurable lost cure. Protecting the treatment calendar (transfusing promptly, pre-booking brachytherapy, not pausing radiation for manageable toxicity) is oncological treatment, not logistics.

Intensifying chemoradiation — who gets more, and the subtype-specific exceptions

For high-risk locally advanced disease the two positive 2024 intensification strategies attach at different ends of the chemoradiation course (developed in the evidence section): short induction carboplatin–paclitaxel before (INTERLACE) or pembrolizumab during and after (KEYNOTE-A18). Neither is yet a universal SA public-sector standard; INTERLACE is the more deliverable in an LMIC because it uses cheap, available cytotoxics. The subtype exceptions override the generic pathway: neuroendocrine carcinoma is treated with platinum–etoposide systemic chemotherapy built around local therapy from the outset regardless of stage, and gastric-type adenocarcinoma responds poorly to chemoradiation, so its management leans harder on radical surgery where the tumour is resectable — both are reasons to insist on the precise histological subtype before committing to a pathway.

Recurrent and metastatic disease — the systematic approach

Recurrence is where judgement is sharpest, because the right answer ranges from curative pelvic exenteration to best supportive care, and the discriminator is the site of prior radiation.

• Central pelvic recurrence after primary radiation, no extrapelvic disease is potentially curable by pelvic exenteration (anterior, posterior or total). It is mutilating, high-morbidity surgery justified only after PET-CT excludes distant disease and the patient is fully counselled about stomas and reconstruction; in SA it is a tertiary-only undertaking.
• Pelvic recurrence after primary surgery (a radiation-naïve pelvis) is usually best salvaged by chemoradiation — the modality not yet spent.
• Distant metastatic or unresectable recurrence is systemic disease. First-line is platinum (cisplatin or carboplatin) + paclitaxel + bevacizumab (GOG-240 established that adding the anti-angiogenic bevacizumab improves survival), now intensified by pembrolizumab where PD-L1 CPS ≥1 (KEYNOTE-826). Second-line, the antibody–drug conjugate tisotumab vedotin (innovaTV-301) has a survival signal. The honest SA framing: bevacizumab and immunotherapy are largely out of reach in the public sector, so the resource-stratified plan names the evidence-based regimen and the deliverable one (platinum-doublet chemotherapy ± palliative radiation for bleeding/pain), and is explicit about which the patient will actually receive.

Palliation and the HIV interface

Much SA cervical cancer presents incurable, and palliation is core consultant work, not an afterthought: haemostatic pelvic radiation for bleeding (even a single fraction can stop catastrophic haemorrhage), nephrostomy or ureteric stenting for obstruction, opioid analgesia for the neuropathic sidewall pain of pelvic recurrence, and early palliative-care linkage. On HIV: every patient is tested and ART optimised, but the interaction is more than a checkbox — HIV-positive women are less likely to complete chemoradiation, carry more infectious and mucocutaneous toxicity, and face drug–drug interactions between ART and chemotherapy; survival signals across cohorts are heterogeneous (some show worse outcomes, some — particularly those on suppressive ART — show parity), so suppressive ART and aggressive supportive care to complete treatment on time matter more than HIV status as a reason to under-treat. Withholding curative chemoradiation on HIV status alone is a clinical error.

The evidence & the controversy

Three shifts define modern practice. First, surgical de-escalation has two opposing vectors: route (open beats minimally invasive — LACC) and radicality (simple may suffice for tiny low-risk tumours — SHAPE). Reconciling them: do less radical surgery in the right patient, but through the abdomen. The unresolved question is how far SHAPE generalises beyond its strict ≤2 cm, low-stromal-invasion criteria — most examiners want the eligibility caveats stated, not just "simple is fine now".

Second, intensification of chemoradiation for locally advanced disease is being rewritten in real time. For years, sequential adjuvant chemotherapy after chemoradiation was attractive but unproven — OUTBACK killed it (no survival benefit, more toxicity). Then two 2024 trials reopened intensification from different ends: INTERLACE added short induction carboplatin–paclitaxel before chemoradiation (5-year OS 80% vs 72%), and KEYNOTE-A18 added pembrolizumab during and after chemoradiation (33% reduction in risk of death). Both are positive, but neither is yet universal standard of care, and they have not been compared head-to-head — claiming either is "the" new standard overreaches. INTERLACE is cheap and deliverable in an LMIC; immunotherapy is not yet broadly affordable in SA public practice — the cost/access argument is decisive.

Third, recurrent and metastatic disease is now an immunotherapy story. KEYNOTE-826 established pembrolizumab plus platinum chemotherapy (± bevacizumab) as first-line for persistent/recurrent/metastatic disease, with the benefit concentrated in PD-L1 CPS-positive tumours. The tension is access: the data are robust, the drug is largely out of reach in the SA public sector, so the defensible plan names the evidence and the resource-stratified reality.

The foundations the modern trials are built on

The whole edifice rests on the five concurrent-chemoradiation trials of 1999, which together prompted the NCI clinical alert that made cisplatin-based chemoradiation — not radiation alone — the standard for locally advanced cervical cancer. Two are worth citing by name. Morris (NEJM 1999) randomised high-risk disease to pelvic-plus-para-aortic radiation alone versus pelvic radiation with concurrent cisplatin/5-FU, and the chemoradiation arm had a large survival advantage — establishing that the gain comes from concurrent chemotherapy, not from a bigger radiation field. GOG-120 (Rose 1999) showed cisplatin-containing chemoradiation beat hydroxyurea-based chemoradiation (3-year survival ~65% vs ~47%), pinning cisplatin as the sensitiser of choice. These are the foundation; EMBRACE-I (brachytherapy is what converts that backbone into ~92% local control) and the 2024 intensification trials build on top, forming a coherent lineage rather than a list.

A second appraisal thread worth holding: the Cochrane neoadjuvant-chemotherapy-before-surgery question was reopened and largely closed by INTERLACE's comparator design — induction chemotherapy helps when it precedes chemoradiation, but neoadjuvant chemotherapy before surgery (the EORTC 55994 line of work) has not shown superiority over standard chemoradiation, so "shrink it then cut it" is not a shortcut around the chemoradiation pathway for bulky disease.

Landmark trials & key evidence

A worked arithmetic: INTERLACE moved 5-year OS from 72% to 80%, an absolute risk reduction in death of 8% (ARR = 0.80 − 0.72 = 0.08), so the number needed to treat ≈ 1/0.08 ≈ 13 women given induction chemotherapy to prevent one death at 5 years — a striking yield from cheap, available cytotoxics, which is the core of the LMIC-relevance argument. Apply the same logic to KEYNOTE-A18 (36-month OS 82.6% vs 74.8%, ARR ≈ 7.8%, NNT ≈ 1/0.078 ≈ 13) and note the catch: near-identical NNTs, but one regimen costs a few vials of carboplatin/paclitaxel and the other costs immunotherapy — which is precisely why, in a resource-stratified SA argument, the cheaper equal-benefit intervention is the defensible default.

Screening and prevention — the consultant context

This is a cancer that screening and vaccination should largely abolish, so the consultant must hold the SA prevention picture. Primary HPV testing (more sensitive than cytology, with a longer safe screening interval) is the WHO-endorsed direction and the one SA is transitioning to; the NDoH Cervical Cancer policy offers HIV-negative women screening from age 30 (historically three free cytology screens at 10-year intervals), and screens women living with HIV at diagnosis and then 3-yearly, reflecting their accelerated HPV carcinogenesis. The WHO 90-70-90 elimination target (90% of girls HPV-vaccinated by 15, 70% of women screened twice with a high-performance test by 35 and 45, 90% with disease treated) is the framework SA's school-based bivalent HPV vaccination programme feeds. The consultant point is the inversion of effort: the disease in front of you in clinic is, epidemiologically, a prevention failure, and the same patient's daughters belong in the vaccination and screening programmes — closing that loop (vaccinate the family, screen the contacts) is part of managing the index case in a high-incidence setting.

Long-term and post-treatment care — survivorship in depth

Surveillance is clinical and symptom-directed: history, speculum and bimanual/vault examination, typically 3–4 monthly for the first two years (the window of highest recurrence) then tapering. Routine cytology of an irradiated vault is unreliable (radiation atypia confounds it) and routine imaging in an asymptomatic woman does not improve survival — investigate symptoms (new pelvic/leg pain, leg oedema, vault lesion, weight loss). The survivorship burden is a chronic-disease workload that district follow-up must own, not just tertiary clinics:

• Vaginal stenosis and sexual function — prophylactic dilator use and topical oestrogen after radiation; counsel early and explicitly, because sexual rehabilitation is routinely neglected and SHAPE's whole point was that morbidity matters.
• Radiation cystitis and proctitis — haematuria/rectal bleeding months to years on; investigate to exclude recurrence, then manage (sucralfate enemas, hyperbaric oxygen in refractory cases) rather than assume benign late effect.
• Lymphoedema — after lymphadenectomy or nodal radiation; early physiotherapy referral and compression.
• Premature menopause — offer HRT (squamous and most adenocarcinomas are not hormone-driven, so it is safe); a young woman rendered menopausal by treatment needs bone and cardiovascular protection, not stoic neglect.
• Renal function — late ureteric stricture from radiation can recapitulate the original obstruction; monitor and stent/nephrostomy as needed.
• Psychological and fertility loss — the woman cured of cervical cancer at 32 who is now infertile and in surgical/radiation menopause needs that addressed as a survivorship diagnosis in its own right.

Worked viva — how to structure the answer

A typical stem: "a 38-year-old, HIV-positive on ART (CD4 410), presents with a 5 cm cervical tumour, parametrial induration on the right, and right-sided hydronephrosis on ultrasound." A high-scoring answer runs:

1. Frame and stage it — "This is locally advanced cervical cancer; the hydronephrosis from ureteric obstruction makes this at least FIGO 2018 stage IIIB, and I would confirm nodal status with PET-CT or CT, which could upstage her to IIIC. This is a chemoradiation, not a surgical, problem — operating here would be the double whammy."
2. Make the patient safe and stage the host — relieve the obstruction (nephrostomy/stent) to protect renal function before cisplatin; baseline FBC, U&E/creatinine, confirm HIV suppression and optimise ART; correct anaemia.
3. State the definitive plan with the regimen — "Definitive concurrent chemoradiation: external-beam pelvic radiation 45–50.4 Gy with weekly cisplatin 40 mg/m², followed by image-guided brachytherapy, aiming to complete the whole course within 56 days because local control falls ~1% per lost day. Brachytherapy is not negotiable and not replaceable with more external beam."
4. Address HIV honestly — "I would not under-treat on HIV status; suppressive ART and supportive care to complete treatment on time matter more than the diagnosis. I'd anticipate higher mucocutaneous and infective toxicity and watch completion closely."
5. Justify from evidence — the 1999 concurrent-chemoradiation trials (Morris, GOG-120) for the backbone, EMBRACE-I for what brachytherapy buys, and — if asked about intensification — INTERLACE versus KEYNOTE-A18 with the explicit cost/access argument (NNT ≈ 13 for both; choose the affordable one).
6. Close the loop — survivorship (vaginal stenosis, menopause/HRT, lymphoedema, renal surveillance), district-level follow-up, and the prevention inversion (vaccinate and screen her family).

Exam traps & red flags

• Smearing a visible lesion. A friable cervical mass needs a biopsy; a "normal" cytology result does not exclude cancer and delays diagnosis.
• The double whammy. Radical hysterectomy in a patient destined for adjuvant chemoradiation stacks surgical and radiation morbidity. If pre-op features (size ≥4 cm, parametrial/nodal involvement, anything meeting Sedlis/Peters) predict adjuvant RT, choose primary chemoradiation.
• Confusing the two de-escalation trials. SHAPE de-escalates radicality (simple may suffice for ≤2 cm low-risk); LACC fixes route (open, not laparoscopic/robotic). "Simple and laparoscopic" is the wrong synthesis — it is "less radical but still open."
• Minimally invasive radical hysterectomy. Post-LACC, offering laparoscopic/robotic radical hysterectomy as equivalent is a clear error — open is standard.
• Omitting or substituting brachytherapy. External beam alone, or an external-beam "boost" in place of brachytherapy, is undertreatment; brachytherapy is the curative component (EMBRACE-I ~92% local control) and its absence is a genuine SA access failure to name, not normalise.
• Ignoring the 56-day clock. Treating breaks, delays and machine downtime as administrative rather than oncological costs cure at ~1% per lost day — protecting the calendar is treatment.
• Treating neuroendocrine carcinoma as an ordinary stage. Small/large-cell NEC needs platinum–etoposide systemic therapy from the outset regardless of how early the FIGO stage looks; radical hysterectomy alone is undertreatment.
• Trusting an HPV screen in gastric-type adenocarcinoma. It is HPV-independent and cytology-negative — a normal screen does not exclude it; watery discharge with a bulky infiltrative cervix should prompt biopsy.
• Missing hydronephrosis / renal failure. Ureteric obstruction is stage IIIB and an emergency — nephrostomy/stenting may be needed before any oncological treatment (and to make cisplatin safe).
• Under-treating on HIV status. Untested/untreated HIV worsens tolerance and completion; the answer is ART optimisation and supportive care to complete therapy, not withholding curative chemoradiation.
• Over-claiming the 2024 trials. Presenting INTERLACE or KEYNOTE-A18 as settled universal standard, or ignoring cost/access, reads as uncritical.

Evidence anchors

• Morris et al. — pelvic vs pelvic+para-aortic RT with concurrent chemotherapy, N Engl J Med 1999
• GOG-120 — Rose et al., concurrent cisplatin-based chemoradiation, N Engl J Med 1999
• GOG-123 — Keys et al., cisplatin-chemoradiation + adjuvant hysterectomy for bulky stage IB, N Engl J Med 1999
• GOG-109 — Peters et al., adjuvant chemoradiation after radical surgery, J Clin Oncol 2000
• LACC trial — Ramirez et al., N Engl J Med 2018
• SHAPE trial — Plante et al., N Engl J Med 2024
• EMBRACE-I — Pötter et al., Lancet Oncol 2021
• Revised FIGO 2018 staging — Bhatla et al., Int J Gynaecol Obstet 2019
• OUTBACK trial — Mileshkin et al., Lancet Oncol 2023
• INTERLACE trial — McCormack et al., Lancet 2024
• KEYNOTE-A18 overall survival — Lorusso et al., Lancet 2024
• KEYNOTE-A18 progression-free survival — Lorusso et al., Lancet 2024
• KEYNOTE-826 — Colombo et al., N Engl J Med 2021
• South Africa NDoH Cervical Cancer Prevention and Control Policy — HIV-negative women screened from age 30 (three free screens at 10-year intervals in the cytology era; transitioning to primary HPV testing); women living with HIV screened at HIV diagnosis and then every 3 years if negative.
• WHO cervical cancer elimination strategy — the 90-70-90 targets (vaccination, twice-lifetime high-performance screening, treatment of detected disease).