Quick Reference
RxDrug of choicePlatinum–paclitaxel + pembrolizumab 200 mg IV q3w (± bevacizumab 15 mg/kg IV q3w) for PD-L1 CPS ≥1 disseminated or recurrent disease
AltAlternativesTisotumab vedotin 2 mg/kg IV q3w (post-progression); endostar + gemcitabine/docetaxel-cisplatin (anti-angiogenic rechallenge); non-platinum single agents for platinum-resistant disease; HER2/NTRK-directed therapy for actionable targets
⊘AvoidBevacizumab in patients with serum albumin <3.0 g/dL, active fistula, unhealed surgical/anastomotic site, or recent major surgery (<28 days); non-dihydropyridine caution with fistulising disease; rechallenge of platinum within 6 months of last exposure
DxTest of choice18F-FDG PET/CT for restaging plus image-guided core biopsy with reflex PD-L1 CPS, MSI/MMR, TMB, HER2, and NTRK testing
ScKey scorePD-L1 combined positive score (CPS) — CPS ≥1 qualifies for pembrolizumab; HR for death 0.58 at CPS ≥10, with median OS extended to ~26 months
→When to referRefer all patients with potentially resectable central pelvic recurrence to a high-volume pelvic exenteration centre for multidisciplinary review; refer to clinical trials at first progression on first-line therapy; integrate palliative care at recurrence diagnosis
Pattern of recurrence + PD-L1 CPS dictates the algorithm: central disease with DFI >12 months → exenteration; disseminated PD-L1+ disease → platinum–paclitaxel + pembrolizumab ± bevacizumab; post-progression → tisotumab vedotin with ocular prophylaxis
Recurrent cervical cancer develops in 10%–30% of patients within five years of curative-intent treatment [2]. Management hinges on anatomic pattern of relapse (local, regional, oligometastatic, or disseminated) and biomarker status, with curative salvage possible only in select central pelvic recurrences. Systemic immunotherapy and antibody-drug conjugates dominate the metastatic landscape.
Overview and Recommendations
Background
- •Recurrent cervical cancer is defined as disease reappearance after a disease-free interval following definitive primary therapy — most often within 5 years of completing chemoradiation or radical hysterectomy — and 10–30% of treated patients fall into this category, with persistent disease comprising a smaller subset that never achieved remission.
- •Anatomic pattern, not merely recurrence per se, is the dominant branch point in the treatment algorithm, dividing disease into local (central vs sidewall pelvis), regional (pelvic or para-aortic nodes), oligometastatic (≤5 lesions in ≤2 organs), and disseminated (lung, liver, bone, peritoneum) categories that each route patients to fundamentally different curative-intent, locally palliative, or systemic pathways.
- •Histology shapes both distribution and biology: accounts for ~80% of cases and preferentially seeds lung and bone, while shows higher rates of peritoneal and ovarian spread, and neuroendocrine/small cell variants signal aggressive platinum-resistant biology with markedly worse outcomes.
- •Prognosis turns sharply downward at recurrence, with median overall survival of 10–17 months on active therapy and 3-year survival below 20% in unselected disseminated disease; the dominant prognostic variables are disease-free interval (>12 months favours indolent biology), site of relapse (central pelvis is resectable), performance status (0–1 favours immunotherapy benefit), and platinum-free interval (>6 months defines platinum-sensitive disease).
- •The treatment paradigm shifted decisively with , which established pembrolizumab added to platinum-based chemotherapy (± bevacizumab) as the first-line standard for PD-L1-positive metastatic or recurrent cervical cancer, replacing the prior chemotherapy-alone default and making reflex PD-L1 CPS testing mandatory on every biopsy.
- •Biomarker-driven therapy now extends beyond PD-L1 to include pembrolizumab and dostarlimab for dMMR/MSI-H disease, -high status, HER2-directed antibody-drug conjugates, NTRK inhibitors, and the tissue factor-targeted ADC, making comprehensive molecular profiling a baseline requirement rather than an optional add-on.
Evaluation
- •Suspect recurrence in any post-treatment patient presenting with vaginal bleeding, malodorous discharge, pelvic or sciatic pain, lower-extremity lymphoedema, cough, or new bone pain — these symptoms should trigger formal restaging even when surveillance imaging is unremarkable.
- •A rising (SCC-Ag), abnormal vault cytology, or a new surveillance imaging abnormality all warrant full restaging, because early mapping preserves curative-intent options for central pelvic disease where exenteration can still be offered.
- •Order 18F-FDG as the cornerstone of restaging once recurrence is suspected — it confirms the lesion, maps nodal and distant spread in a single study, and provides a metabolic baseline for response assessment using PERCIST or 1.1 criteria.
- •Add pelvic MRI to better delineate central recurrences, parametrial invasion, and fistulae — MRI is particularly essential before -containing regimens, where baseline serum albumin <3.0 g/dL predicts fistula formation and may alter candidacy.
- •Obtain a chest CT when pulmonary metastases are suspected, as small parenchymal nodules may be below PET resolution and frequently seeds the lung as the first distant site.
- •Mandate image-guided core biopsy of the most accessible lesion before any systemic therapy — fine-needle aspiration is inadequate for immunohistochemistry, which requires preserved architecture and sufficient tumour cellularity for accurate combined positive score (CPS) reporting.
- •Reflex PD-L1 CPS testing on every biopsy, because CPS thresholds dictate first-line regimen choice: CPS ≥1 qualifies for per KEYNOTE-826, with the magnitude of benefit scaling with CPS (HR for death 0.58 at CPS ≥10, extending median OS to ~26 months vs 16.5 months with chemotherapy alone).
- •Test concurrently for (MSI)/mismatch repair (MMR) deficiency, (TMB), HER2 amplification, and NTRK fusions, as these direct checkpoint inhibition (pembrolizumab/dostarlimab), antibody-drug conjugates, or tumour-agnostic targeted therapy regardless of CPS.
- •Assess performance status ( 0–1 is the immunotherapy-eligible range), renal function (eGFR for platinum dosing and pemetrexed eligibility), hepatic function (transaminases, bilirubin for ICI safety), and complete blood count (cytopenias limit chemotherapy density and signal marrow involvement).
- •Evaluate fistula risk pre-emptively with serum albumin (target ≥3.0 g/dL) and clinical exam for vaginal, urinary, or rectal communication, as prior pelvic radiation, bevacizumab, and hypoalbuminemia form a triad for non-healing pelvic wounds.
- •Calculate disease-free interval from completion of primary therapy — intervals >12 months indicate indolent biology favouring exenteration or curative-intent re-irradiation, while intervals <12 months signal aggressive biology better served by systemic therapy.
- •Refer all patients with potentially resectable central pelvic recurrence to a high-volume exenteration centre for multidisciplinary review before any local therapy is committed, because preoperative candidacy depends on cross-sectional imaging, exam under anaesthesia, and histologic confirmation.
- •Apply a validated prognostic tool when available — the immunotherapy nomogram derived from 204 patients integrates clinicopathologic variables selected by LASSO regression to stratify PFS and OS into risk tiers, outperforming FIGO staging on decision-curve analysis and supporting shared decision-making about line of therapy.
Management
- •Route management by anatomic pattern: central pelvic recurrence → consider exenteration; pelvic sidewall or unresectable disease → re-irradiation with IMRT/3D-RT ± concurrent chemotherapy; regional nodes only → targeted nodal RT or completion dissection; oligometastatic → SBRT or metastasectomy plus systemic therapy; disseminated disease → systemic first-line.
- •Offer as the only curative-intent option for centrally recurrent disease confined to the pelvis, classified by the viscera removed (anterior, posterior, or total) based on pattern of relapse; candidates require disease-free interval ≥12 months, central (not sidewall) disease, no extrapelvic metastasis, and ECOG 0–1.
- •Choose modern re-irradiation with 3D-RT or ± concurrent chemotherapy for unresectable or medically inoperable recurrences, achieving 3-year overall survival of 56.1% and locoregional control of 59.7% in the Shanghai Cancer Hospital series of 50 patients, with median interval from initial treatment of 12 months (range 6–51).
- •Consider (intracavitary or interstitial) for small central volumes in patients with prior external beam radiation, as it delivers high conformality with steep dose fall-off that spares previously irradiated normal tissue.
- •Initiate first-line systemic therapy for disseminated disease as platinum–paclitaxel + ± ; give pembrolizumab 200 mg IV every 3 weeks for up to 24 months (or per protocol), continuing until progression or unacceptable toxicity.
- •Add 15 mg/kg IV every 3 weeks when no contraindication exists, on the strength of GOG-240, which demonstrated a 3.7-month overall survival gain when added to cisplatin–paclitaxel (median OS 16.8 vs 13.3 months; HR 0.77); the agent is particularly favoured for rapidly progressive, symptomatic, or bevacizumab-naïve disease.
- •In resource-appropriate settings, consider recombinant human endostatin (endostar) added to gemcitabine–cisplatin or docetaxel–cisplatin, which raised objective response rate from 22.7% to 42.9% (NNT ≈ 5) without significant grade 3–4 toxicity escalation in a 44-patient Chinese randomized experience.
- •Switch to 2 mg/kg IV every 3 weeks after first-line failure, on the strength of the innovaTV 301 trial showing median OS 11.5 vs 9.5 months versus investigator-choice chemotherapy (HR 0.70, 95% CI 0.54–0.91); apply ocular toxicity prophylaxis at every infusion.
- •Prevent tisotumab vedotin ocular toxicity with mandatory baseline ophthalmologic evaluation, refrigerated preservative-free artificial tears, topical vasoconstrictor eye drops (e.g., brimonidine or phenylephrine) immediately before infusion, topical corticosteroid drops for 72 hours after each dose, and avoidance of contact lenses throughout treatment — this protocol mitigates the conjunctivitis, keratitis, and dry eye that anchor the FDA black-box warning.
- •Refer for consideration of HER2-directed therapy (trastuzumab deruxtecan, trastuzumab emtansine) or NTRK inhibition (larotrectinib, entrectinib) when molecular profiling reveals actionable targets, as these tumour-agnostic options can deliver durable responses in heavily pre-treated patients.
- •Optimize serum albumin to ≥3.0 g/dL preoperatively and before bevacizumab, because hypoalbuminemia is a reproducible predictive factor for fistula formation in women receiving salvage therapy for recurrent cervical cancer, and correction is one of the few modifiable risk markers.
- •Use caution with bevacizumab in heavily pre-irradiated pelvic fields, post-exenteration cavities, or with concurrent vaginal/rectal wall involvement, as these scenarios compound fistula risk and may warrant deferring the agent until local healing is confirmed.
- •Apply palliative radiotherapy (typically 8 Gy × 1, 20 Gy × 5, or 30 Gy × 10) for symptomatic bleeding, pain, or obstructive nodal disease; it remains the most effective local tool for bleeding control and is delivered in days to weeks without interfering with systemic therapy timing.
- •Reassess response with imaging (PET/CT or MRI) every 2–3 cycles during systemic therapy, applying 1.1 or PERCIST and correlating with symptom change; early progression on first-line immunotherapy is a strong signal to switch to or trial enrolment rather than persist with the same regimen.
- •Avoid re-introducing platinum within 6 months of last platinum exposure (platinum-resistant disease), where response rates fall below 10% and non-platinum single agents, tisotumab vedotin, or clinical trials are preferred.
- •Address survivorship needs proactively: chronic pelvic pain, lower-extremity after salvage therapy, fatigue persisting 6–12 months beyond treatment, and high rates of anxiety and depression all warrant early referral to palliative care, lymphoedema services, and psycho-oncology — these are integral to prognostic care, not optional add-ons.
- •Counsel patients with disseminated disease that true cure is rare outside isolated pelvic disease treated with exenteration, that a 'second remission' is most often measured in months to a few years rather than decades, and that goals-of-care conversations should run in parallel with systemic therapy rather than be deferred to end-of-life.
Board Review — High Yield
- •KEYNOTE-826 — established pembrolizumab + platinum-based chemotherapy (± bevacizumab) as first-line standard for PD-L1 CPS ≥1 recurrent/metastatic cervical cancer, with HR for death 0.58 at CPS ≥10
- •Pelvic exenteration — the only curative-intent option for central pelvic recurrence, requiring disease-free interval ≥12 months, central (not sidewall) disease, and no extrapelvic metastasis; classified as anterior, posterior, or total
- •GOG-240 — established bevacizumab 15 mg/kg q3w in recurrent cervical cancer, adding 3.7 months of overall survival (16.8 vs 13.3 months) to cisplatin–paclitaxel; caution with fistula risk
- •Tisotumab vedotin — tissue factor–targeted antibody-drug conjugate at 2 mg/kg q3w; innovaTV 301 showed OS 11.5 vs 9.5 months; black-box warning for ocular toxicity mandates prophylactic eye drops and ophthalmology evaluation
- •PD-L1 CPS — a CPS ≥1 qualifies for pembrolizumab; the magnitude of benefit scales with CPS, with the strongest effect at CPS ≥10
- •10–30% — proportion of cervical cancer patients who develop persistent or recurrent disease within 5 years of completing primary therapy
- •Disease-free interval <12 months — marks aggressive biology, generally precluding exenteration and shifting intent to systemic therapy; >12 months suggests indolent biology and candidacy for curative-intent salvage
- •Hypoalbuminemia — serum albumin <3.0 g/dL is a reproducible predictive factor for fistula formation during salvage therapy and should be corrected before exenteration or bevacizumab
- •Squamous vs adenocarcinoma — squamous favours lung and bone metastases; adenocarcinoma shows higher rates of peritoneal and ovarian spread, informing surveillance strategy
- •Endostar — recombinant human endostatin added to platinum-based chemotherapy raised ORR from 22.7% to 42.9% in a 44-patient Chinese randomized trial, positioning anti-angiogenic rechallenge as a rational option in resource-appropriate settings
Deep Dive — Evidence Details
References
- [1]
Moore DH. “Chemotherapy for advanced, recurrent, and metastatic cervical cancer.” Journal of the National Comprehensive Cancer Network : JNCCN (2008). PMID: 18267059 ↗
L5REVIEW_NARRATIVECited in: Patterns of Recurrence - [2]
Hinojosa-Gonzalez DE, Escarcega-Bordagaray JA, Murillo-Torres RM et al.. “Immune checkpoint therapy in locally advanced, persistent, recurrent, and metastatic cervical cancer: A systematic review and Bayesian network meta-analysis.” International journal of gynaecology and obstetrics: the official organ of the International Federation of Gynaecology and Obstetrics (2026). PMID: 41239828 ↗
L1SR_OBSCited in: Patterns of Recurrence - [3]
Gass P, Thiel FC, Häberle L et al.. “Primary results of the AGO-Zervix-1 Study: A prospective, randomized phase III study to compare the effects of paclitaxel and topotecan with those of cisplatin and topotecan in the treatment of patients with recurrent and persistent cervical cancer.” Gynecologic oncology (2024). PMID: 38490057 ↗
L1RCTCited in: Patterns of Recurrence, Prognosis of Recurrent Disease - [4]
Xiao Y, Cheng H, Wang L et al.. “Clinical response and safety of apatinib monotherapy in recurrent, metastatic cervical cancer after failure of chemotherapy: a retrospective study.” Journal of gynecologic oncology (2020). PMID: 31788992 ↗
L4COHORTCited in: Patterns of Recurrence, Prognosis of Recurrent Disease - [5]
Hsiao SM, Chen CA, Hsu C et al.. “Weekly cisplatin, infusional high-dose 5-fluorouracil and leucovorin for advanced, recurrent and metastatic cervical carcinoma.” Anticancer research (2008). PMID: 18630476 ↗
L4TRIAL_NONRANDOMCited in: Patterns of Recurrence, Prognosis of Recurrent Disease - [6]
Guo F, Chen C, Liang Y et al.. “Efficacy of the combination of endostar with chemotherapy on stage IVb and recurrent metastatic cervical cancer.” Zhong nan da xue xue bao. Yi xue ban = Journal of Central South University. Medical sciences (2020). PMID: 33472996 ↗
L1RCTCited in: Patterns of Recurrence, Distant Metastatic Disease, Prognosis of Recurrent Disease - [7]
Bose CK. “Balstilimab and other immunotherapy for recurrent and metastatic cervical cancer.” Medical oncology (Northwood, London, England) (2022). PMID: 35092506 ↗
L5REVIEW_NARRATIVECited in: Patterns of Recurrence - [8]
Tao X, Hu W, Ramirez PT et al.. “Chemotherapy for recurrent and metastatic cervical cancer.” Gynecologic oncology (2008). PMID: 18533239 ↗
L5REVIEW_NARRATIVECited in: Patterns of Recurrence - [9]
Park KJ, Braschi-Amirfarzan M, DiPiro PJ et al.. “Multimodality imaging of locally recurrent and metastatic cervical cancer: emphasis on histology, prognosis, and management.” Abdominal radiology (New York) (2016). PMID: 27357415 ↗
L5REVIEW_NARRATIVECited in: Patterns of Recurrence, Workup of Suspected Recurrence, Prognosis of Recurrent Disease - [10]
Liu Z, Zhang S, Wang Y et al.. “Development and validation of immunotherapy nomogram for predicting the efficacy and prognosis of recurrent and metastatic cervical cancer.” Frontiers in immunology (2025). PMID: 41438759 ↗
L2OTHERCited in: Patterns of Recurrence, Workup of Suspected Recurrence, Distant Metastatic Disease, Prognosis of Recurrent Disease - [11]
Palavalli Parsons LH, Roane B, Manders DB et al.. “Hypoalbuminemia is a Predictive Factor for Fistula Formation in Recurrent Cervical Cancer.” American journal of clinical oncology (2018). PMID: 28817390 ↗
L4OTHERCited in: Workup of Suspected Recurrence, Local-Regional Recurrence - [12]
Liu SP, Huang X, Ke GH et al.. “3D radiation therapy or intensity-modulated radiotherapy for recurrent and metastatic cervical cancer: the Shanghai Cancer Hospital experience.” PloS one (2012). PMID: 22768273 ↗
L3TRIAL_NONRANDOMCited in: Local-Regional Recurrence, Prognosis of Recurrent Disease