Disease characteristics, treatment, and overall survival outcomes of patients with locally advanced cervical cancer referred for radiation therapy in Johannesburg, South Africa: a 2-year retrospective study Authors: Oluwatosin A Ayeni, Daniel Mmereki, Himal Mistry, Okechinyere Achilonu, Megan Kassick, Sydnie Swanson, Wenlong Carl Chen, Maureen Joffe, Mpho Ratshikana, Katharine Rendle, Surbhi Grover, Duvern Ramiah Published: 2026 ABSTRACT AbstractPurposeWe assessed the association between baseline patient characteristics, disease stage, and treatment-related factors and overall survival among women with locally advanced, non-metastatic cervical cancer in South Africa.Patients and methodsWe retrospectively analysed data from 642 women with 2018 International Federation of Gynaecology and Obstetrics (FIGO) stage IB3–IVA cervical cancer who were referred for and deemed eligible for radiotherapy at a public tertiary hospital in South Africa between January 2021 and December 2022. All women discussed at the multidisciplinary team (MDT) meeting were included, irrespective of whether treatment was ultimately initiated. Demographic and clinical characteristics, treatment timelines, and survival outcomes were assessed. Multivariable Cox proportional‑hazards models were used to identify factors associated with overall survival.ResultsOf the 642 women, 54.1% were aged < 50 years and 60.3% were living with HIV. The stage distribution was: IB3, 1.7%; II, 43.5%; III, 50.1%; and IVA, 4.7%. The median time to treatment initiation was ~4 months; the median overall treatment duration was 63 days. The 2-year overall survival rate was 57.2% and varied by stage (IB3 and II, 70.0%; III, 49.3%; IVA, 12.4%; p < 0.001). On multivariable Cox regression analysis, worse survival was independently associated with Eastern Cooperative Oncology Group performance status 2–4 [vs. 0 and 1; adjusted hazard ratio (aHR), 1.75], stage III (vs. IB3 and II; aHR, 1.46) stage IVA (vs. IB3 and II; aHR, 2.02), and receipt of radiotherapy alone (vs. chemoradiation; aHR, 1.96). A significant interaction between EQD2 and overall treatment duration was observed. Patients receiving EQD2 <80 with treatment duration <56 days was associated with worse overall survival (aHR 2.65). HIV status was not associated significantly with survival.ConclusionsOverall Survival in women with locally advanced cervical cancer in South Africa was associated with disease stage, baseline patient characteristics, and treatment-related factors. Interventions targeting timely delivery of standard of care may improve survival in this setting.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12905-026-04611-y.Keywords: Cervical cancer, Treatment delay, Stage, Overall survival Introduction Cervical cancer remains the leading cause of cancer-related death among women in sub-Saharan Africa (SSA) [1]. The disease disproportionately impacts low- and middle-income countries (LMICs), where 84–90% of new cases occur [2]. Despite advancements in cervical cancer prevention, screening, and treatment, SSA continues to bear a disproportionately high burden of the disease, with the highest incidence and mortality rates worldwide [1, 3, 4]. In South Africa, cervical cancer is a significant public health challenge, especially among women living with HIV (WLWH) [5]. With an estimated 26.3% of women aged 15–49 years living with HIV, South Africa faces unique challenges in cervical cancer management [6]. More than 63% of new cervical cancer cases occur in WLWH [6, 7], as HIV increases the susceptibility to persistent human papillomavirus infection and accelerates precancerous cervical lesion progression to invasive cancer [8]. Despite cervical cancer being preventable, most women in South Africa and across SSA continue to present with locally advanced disease at diagnosis [9, 10]. Factors contributing to late presentation include poor awareness, delays in diagnostic pathways, barrier to healthcare access, and structural constraints within health systems [11]. Treatment of locally advanced cervical cancer is resource-intensive and time-sensitive, typically requiring coordinated delivery of external beam radiotherapy, brachytherapy, and concurrent chemotherapy [12, 13]. Delays in treatment initiation or interruptions in care have been consistently associated with poorer survival, outcomes, making timely access to high-quality treatment particularly critical [14]. Despite South Africa’s status as an upper-middle-income country with specialized oncology services, detailed data on patient profiles, stage at diagnosis, treatment patterns, and survival outcomes among women with locally advanced cervical cancer remain limited. Existing studies have largely relied on administrative or registry data or have focused on selected subpopulations. restricting insight into real-word treatment pathways and outcomes [15–17]. As a result, there is limited insight into real-world treatment pathways, delays, and outcomes among women receiving care within public-sector oncology settings. Addressing these gaps is essential to inform health-system strengthening and improve outcomes for women with cervical cancer. To address this gap, we examined the association between baseline patient characteristics, disease stage, and treatment‑related factors and overall survival among women with locally advanced, non‑metastatic cervical cancer treated at a public referral centre in Johannesburg, South Africa. Methods Design, setting, and patients Charlotte Maxeke Johannesburg Academic Hospital (CMJAH) is a public tertiary-care hospital in Johannesburg, Gauteng Province. Its Radiation Oncology Department is the only public radiation oncology referral facility in Johannesburg, receiving referrals for radiotherapy from all secondary and tertiary hospitals in Johannesburg. In this study, retrospectively collected data on patients with histologically confirmed locally advanced cervical cancer (2018 International Federation of Gynaecology and Obstetrics stages IB3–IVA [18]) at diagnosis who were referred for radiotherapy at CMJAH’s Radiation Oncology Department between January 2021 and December 2022 were analyzed. Referrals were made via multidisciplinary team (MDT) meetings. Women who did not initiate treatment were retained in the cohort to capture pre‑treatment attrition and early mortality following referral for radiotherapy. Ethics approval and consent to participate The University of the Witwatersrand Human Research Ethics Committee (HREC) Medical, Johannesburg, South Africa, approved the study (approval number M240862). The requirement for informed consent was waived by the HREC due to the study’s retrospective design. The study was conducted in compliance with the Declaration of Helsinki. Diagnosis and treatment A comprehensive approach to diagnosis, staging, and therapeutic intervention was taken. Diagnoses were based on detailed patient histories, physical examination, and laboratory studies [complete blood count, renal function testing, HIV status determination with the cluster of differentiation 4 (CD4) count and viral load, chest X-ray, and abdominal ultrasound]. Patients were treated according to the National Comprehensive Cancer Network guidelines [19]. “External beam radiotherapy (EBRT) was delivered using either a four-field box technique with a 3D image set-up and radiotherapy delivery or Volumetric Modulated Arc Therapy (VMAT) technique. Standard EBRT prescriptions consisted of 50Gy in 25 fractions of 2Gy daily, Monday to Friday for women treated with curative intent [20]. A hypofractionated schedule of 37.4Gy in 11 fractions was used in selected patients, primarily those with stage IIIB with hydronephrosis, poor performance status, or limited eligibility for concurrent chemotherapy, in the context of prolonged waiting times and constrained radiotherapy capacity. Patients with stage IVA disease were prescribed EBRT of ≤30Gy. High-dose-rate (HDR) intracavitary brachytherapy was planned using two-dimensional point-based prescription, which was the standard technique at the institution during the study period. Image guided brachytherapy was not available. Prescribed brachytherapy doses most commonly included 24Gy in three fractions, or 18Gy in two fractions, delivered weekly following EBRT. The total radiation dose was calculated as the equivalent dose in 2-Gy fractions (EQD2) using an ratio of 10. Treatment interruptions occurred due to machine downtime, intercurrent illness, poor performance status, or missed appointments. Overall treatment time was calculated from treatment initiation to completion and was included as a treatment-related variable in the survival analyses to capture the potential impact of treatment prolongation on outcomes.”Newly diagnosed WLWH were referred for antiretroviral therapy (ART) initiation before cancer treatment. Data extraction The research team was trained to collect information from patients’ files and radiation planning software, and enter it into the Research Electronic Data Capture (REDCap) database, maintained by the University of the Witwatersrand, Johannesburg. Data extracted from the REDCap database for this analysis included patients’ demographic characteristics, presenting symptoms, clinical characteristics [Eastern Cooperative Oncology Group (ECOG) performance status, HIV status, CD4 count, and viral load], baseline laboratory values (white blood cell, platelet counts, hemoglobin, and creatinine levels), and disease characteristics (stage and histology). The treatment regimen and number of chemotherapy cycles administered were recorded. The total radiation dose received was calculated using the radiobiological equivalent dose in 2Gy fraction (EQD2) formula using an ratio of 10.21 We also collected information on intervals between biopsy, pathology report finalization, initial case presentation in an MDT meeting, (precise, individualized) radiotherapy planning, and treatment initiation, as well as the overall treatment time (from treatment initiation to treatment completion). We defined the overall survival time as extending from the cervical cancer biopsy report date to the date of death, last confirmation that the patient was alive, or 30 September 2024. For patients lost to follow-up whose vital status could not be ascertained from hospital records (probably due to care continuation at a base hospital, foreign status, or death), we searched VerifyID (a publicly available administrative database in South Africa) to determine the vital status. Statistical analysis Exploratory data analysis was conducted to describe patients’ demographic and clinical characteristics. Continuous variables are expressed as means ± standard deviations (SDs) or medians with interquartile ranges (IQRs), and categorical variables are expressed as frequencies with percentages. The times to treatment initiation and completion are reported as both means ± SDs and medians (IQRs) to fully capture their distributions. Kaplan–Meier curves were generated to estimate and visualize group-specific overall survival, and differences in survival between groups were assessed using the log-rank test. The overall survival was analysed using the Cox proportional hazards model, which estimates hazard ratios (HRs) for time-to-event outcomes while adjusting for covariates. The associations of all candidate variables with mortality were initially assessed using univariable Cox proportional-hazards models. Variables with p values < 0.10 in the univariable analyses were included in a multivariable Cox regression model following confirmation of the proportional-hazards assumption validity. Significance in the multivariable model was defined as p < 0.05. Analyses were performed using available-case (complete-case) data, and no multiple imputation was undertaken. All statistical analyses were performed using Stata (version 16; StataCorp Ltd., College Station, TX, USA). Results In total, 642 women with locally advanced cervical cancer were determined to be eligible for radiation therapy through MDT consultation during the study period; 348 (54.1%) of these women were aged < 50 years, and 116 (18.1%) were aged < 40 years. Most [n = 578 (92.6%)] women had an ECOG performance status of 0 or 1. Of the 627 women with known HIV status, 378 (60.3%) were living with HIV, most of whom had CD4 counts of ≥200 cells/mL. Approximately 60% of the women had hemoglobin levels > 10 g/dL (Table 1). Table 1Demographic and Clinical Characteristics of Women with Locally Advanced Cervical Cancer Who Received Radiation Therapy at CMJAH, 2021–2022 (N = 642)CharacteristicValue (%)Median age (years)48.5 (41.8–56.7)Age group (years) < 40116 (18.1) 40–49232 (36.1) 50–59172 (26.8) ≥ 60122 (19.0)ECOG performance status 0–1578 (92.6) 2–446 (7.4)HIV status Negative249 (39.7) Positive378 (60.3)CD4 count (cells/mL) < 20041 (13.0) ≥ 200275 (87.0)Viral load (copies/mL) < 40234 (80.7) ≥ 4056 (19.3)Pain No490 (76.3) Yes152 (23.7)Vaginal bleeding No102 (15.9) Yes540 (84.1)Vaginal discharge No503 (78.3) Yes139 (21.7)Lymphedema No641 (99.8) Yes1 (0.2)Weight loss No638 (99.4) Yes4 (0.6)Histology Squamous cell carcinoma572 (89.1) Adenocarcinoma47 (7.3) Adenosquamous carcinoma11 (1.7) Small cell carcinoma2 (0.3) Other10 (1.6)Grade Well differentiated14 (2.4) Moderately differentiated487 (84.1) Poorly differentiated78 (13.5)StageIB311 (1.7)II279 (43.5)IIA18 (1.3)IIA215 (2.3)IIB256 (39.9)III322 (50.1)IIIA14 (2.2)IIIB291 (45.3)IIIC116 (2.5)IIIC21 (0.1)IVA30 (4.7)Treatment Chemoradiation348 (54.2) Radiation only274 (42.7) None20 (3.1)Chemotherapy cycle received (n = 347) 1 cycle156 (45.0) 2 cycles191 (55.0)EBRT dose (Gy) 48–50426 (68.5) 37.4140 (22.5) ≤ 3056 (9.0)Brachytherapy dose (Gy) 24315 (49.8) 1890 (14.3) 15–16.521 (3.3) < 1533 (5.3) None172 (27.3)Total EQD2 received (Gy) ≥ 80313 (50.7) < 80305 (49.3)Follow-up status Alive261 (40.7) Deceased307 (47.8) Lost to follow up74 (11.5)Baseline laboratory valuesHemoglobin concentration (g/dL) ≤ 10239 (37.2) > 10403 (62.8)White blood cell count (×10/L)8.7 ± 4.2Platelet count (/µL)390.6 ± 144.0Creatinine level (µmol/L)81.6 ± 53.4Data are reported as median (interquartile range), n (%), or mean ± standard deviation. Missing data: HIV status, n = 15; CD4 count, n = 62; viral load, n = 88; grade, n = 63; EBRT dose, n = 20 (no treatment); brachytherapy, n = 11; EQD2, n = 24; white blood cell count, n = 22; platelet count, n = 16; creatinine level, n = 52Abbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, ECOG Eastern Cooperative Oncology Group, CD4 Cluster of differentiation 4, EBRT External beam radiation therapy, EQD2 Equivalent dose in 2-Gy fractions Demographic and Clinical Characteristics of Women with Locally Advanced Cervical Cancer Who Received Radiation Therapy at CMJAH, 2021–2022 (N = 642) Data are reported as median (interquartile range), n (%), or mean ± standard deviation. Missing data: HIV status, n = 15; CD4 count, n = 62; viral load, n = 88; grade, n = 63; EBRT dose, n = 20 (no treatment); brachytherapy, n = 11; EQD2, n = 24; white blood cell count, n = 22; platelet count, n = 16; creatinine level, n = 52 Abbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, ECOG Eastern Cooperative Oncology Group, CD4 Cluster of differentiation 4, EBRT External beam radiation therapy, EQD2 Equivalent dose in 2-Gy fractions Presenting signs and symptoms Vaginal bleeding was the predominant presenting symptom across all stages (IB3, 81.8%; II, 84.2%; III, 83.2%; IVA, 93.3%). Other major symptoms were vaginal discharge and pain (Fig 1). Symptoms at presentation by stage at diagnosis in women with locally advanced cervical cancer referred for radiation therapy at Charlotte Maxeke Johannesburg Academic Hospital, 2021–2022 Disease and treatment characteristics Eleven (1.7%) women had stage IB3 disease, 279 (43.5%) had stage II, 322 (50.1%) had stage III, and 30 (4.7%) had stage IVA disease. Most (89.1%) women had squamous cell carcinoma, which was moderately differentiated in most (84.1%) cases. Of the 642 women referred for radiotherapy, 348 (54.2%) received concurrent chemoradiation therapy (CCRT), 274 (42.7%) received radiotherapy alone, and 20 (3.1%) did not initiate treatment. The women who did not receive treatment had stage II (10.0%), stage III (80.0%), and stage IVA (10.0%) disease, and the majority (80.0%) died before treatment initiation (Supplementary Table 1). The most common EBRT doses were 48–50 Gy in 24 or 25 fractions (68.5%), 37.4 Gy in 11 fractions (22.5%), and ≤30 Gy (9.0%). Brachytherapy was administered to 306 (87.9%) women who received CCRT and 165 (60.2%) women who received radiation therapy alone. The most commonly administered intracavitary brachytherapy regimen was 24 Gy in 3 fractions (49.8%), followed by 18 Gy in 2 fractions (14.3%) and 15–16.5 Gy in 3 fractions (3.3%). A small proportion (5.3%) of patients received total doses of <15 Gy. Of the women who received both EBRT and brachytherapy, 313 (50.7%) received total EQD2 ≥ 80 Gy and 305 (49.3%) received total EQD2 < 80 Gy (Table 1). Timeline to treatment The median time from biopsy to pathology report finalization was 6 (IQR, 4–10) days, and that from pathology report finalization to MDT consultation about radiation therapy was 43 (IQR, 23–72) days. The median time from MDT consultation to computed tomography (CT) simulation was 34 (IQR, 22–56) days, and that from CT simulation to treatment initiation was 41 (IQR, 27–70) days. The overall median time from biopsy reporting to treatment initiation was 146 (IQR, 111-199) days, and the median overall treatment time was 63 (IQR, 36–87) days (Table 2). Table 2Radiation treatment intervals for women with locally advanced cervical cancer treated at CMJAH, 2021–2022TimeInterval (days)Mean ± SDMedian (IQR)Diagnosis (biopsy) to pathology report completion (n = 641)8.4 ± 7.96 (4–10)Pathology report completion to MDT consultation (n = 630)65.8 ± 112.543 (23–72)MDT consultation to CT simulation (n = 632)46.9 ± 49.034 (22–56)CT simulation to treatment (n = 618)56.2 ± 48.141 (27–70)Diagnosis (biopsy) to treatment (n = 618)175.8 ± 125.8146 (111–199)Pathology report completion to treatment (n = 617)167.6 ± 125.0137 (104–189)MDT consultation to treatment (n = 616)103.7 ± 62.488 (67–119)Overall treatment (n = 618)67.4 ± 44.863 (36–87)Follow-up (months)22.1 ± 11.824 (11–31)n does not sum up to 642 due to missing dates; hence, the interval in days could not be calculated. All 642 patients were discussed at the MDTAbbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, SD Standard deviation, IQR Interquartile range, MDT Multidisciplinary team, CT Computed tomography Radiation treatment intervals for women with locally advanced cervical cancer treated at CMJAH, 2021–2022 n does not sum up to 642 due to missing dates; hence, the interval in days could not be calculated. All 642 patients were discussed at the MDT Abbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, SD Standard deviation, IQR Interquartile range, MDT Multidisciplinary team, CT Computed tomography Overall survival At the median follow-up timepoint of 24.0 (IQR, 11.0–31.0) months, 307 (47.8%) women had died, 261 (40.7%) were alive, and 74 (11.5%) women had been lost to follow up (Tables 2 and 3). The pooled 2-year overall survival rate was 57.2% [95% confidence interval (CI), 53.1–61.0%]. The Kaplan–Meier analysis and log-rank test demonstrated significant unadjusted differences in overall survival according to several clinical factors (Fig 2). Poorer survival was associated with worse ECOG performance status (2–4), advanced disease stages (III and IVA), and receipt of radiation treatment alone or no treatment (Fig 2A, C, D). Women with stage IVA disease had poorer overall survival at 2 years (12.4%; 95% CI, 3.1–28.4%) than did women with stage III (49.3%; 95% CI, 43.5–54.8%) and stages IB3 and II (70.0%; 95% CI, 64.2–75.1%) disease (log-rank p < 0.001; Fig 2C). Patients with lower CD4 counts tended to have reduced survival (Fig 2B). Survival was also poorer among women who received EQD2 < 80 Gy and those with hemoglobin concentrations ≤ 10 g/dL (Fig 3). Table 3Hazard ratios for factors associated with overall survival in women with locally advanced cervical cancer referred for radiation therapy at CMJAH, 2021–2022CharacteristicDeceased (N = 307; category %)*Univariate hazard ratio (95% CI) p Multivariate hazard ratio (95% CI) p Age (years) < 4056 (48.3)1.00 (ref.)1.00 (ref.) 40–49107 (46.1)0.96 (0.68–1.28)0.6691.11 (0.75–1.63)0.600 50–5989 (51.7)0.99 (0.71–1.38)0.9600.94 (0.62–1.44)0.790 ≥ 6055 (45.1)0.84 (0.58–1.22)0.3690.69 (0.42–1.14)0.145ECOG status 0–1266 (46.0)1.00 (ref.)1.00 (ref.) 2–432 (69.6)1.81 (1.25–2.61) 0.002 1.75 (1.14–2.70) 0.011 HIV status Negative115 (46.2)1.00 (ref.) Positive184 (48.7)1.11 (0.88–1.40)0.370HIV status by CD4 count (cells/mL) HIV+, CD4 < 20025 (61.0)1.00 (ref.)1.00 (ref.) HIV+, CD4 ≥ 200129 (46.9)0.63 (0.41–0.97) 0.034 0.74 (0.46–1.19)0.210 HIV–115 (46.2)0.62 (0.40–0.96) 0.032 0.75 (0.46–1.23)0.260Stage IB3, II102 (35.2)1.00 (ref.)1.00 (ref.) III181 (56.2)1.95 (1.53–2.49) < 0.001 1.46 (1.07–1.99) 0.016 IV24 (80.0)6.38 (4.06–10.01) < 0.001 2.01 (1.08–3.78) 0.028 Histology Squamous cell carcinoma273 (47.7)1.00 (ref.) Adenocarcinoma23 (48.9)1.01 (0.66–1.55)0.959 Adenosquamous carcinoma6 (54,6)1.43 (0.64–3.22)0.384 Small cell cancer2 (100.0)1.75 (0.43–7.02)0.432 Other3 (30.0)0.66 (0.21–2.06)0.477Grade Well differentiated5 (35.7)1.00 (ref.) Moderately differentiated231 (47.4)1.73 (0.71–4.20)0.226 Poorly differentiated41 (52.6)1.83 (0.72–4.63)0.204Treatment Chemoradiation120 (34.5)1.00 (ref.)1.00 (ref.) Radiation only171 (62.4)2.76 (2.18–3.49) < 0.001 1.95 (1.39–2.75) < 0.001 None16 (80.0)3.42 (2.02–5.76) < 0.001 Not retainedTotal EQD2 received (Gy) ≥ 80103 (32.9)1.00 (ref.) < 80185 (60.7)2.69 (2.11–3.42) < 0.001 Hemoglobin level (g/dL) ≤ 10125 (52.3)1.34 (1.07–1.68) 0.012 1.20 (0.89–1.62)0.226 > 10182 (45.2)1.00 (ref.)1.00 (ref.) White blood cell count (×10/L)1.04 (1.01–1.06) 0.004 0.99 (0.96–1.02)0.621 Platelet count (/µL)1.09 (1.01–1.18) 0.026 1.01 (0.91–1.13)0.862 Creatinine level (µmol/L)1.02 (1.00–1.03) 0.018 1.02 (1.00–1.04)0.086Overall Treatment Time (Days) < 56160 (61.8)1.00 (Reference) ≥ 56128 (35.7)0.41 (0.33–0.52) < 0.001 *EQD2 (Gy) + OTT (Days) EQD2 ≥ 80, OTT < 5619 (32.2)1.00 (Ref)1.00 (Ref) EQD2 ≥ 80, OTT ≥ 5684 (33.1)1.11 (0.67–1.83)0.6831.24 (0.70–2.22)0.462 EQD2 < 80, OTT < 56141 (70.5)4.08 (2.52–6.59) < 0.001 2.65 (1.50–4.71) 0.001 EQD2 < 80, OTT ≥ 5644 (41.9)1.55 (0.90–2.65))0.1121.07 (0.57–2.01)0.841Bold, p < 0.05., Deceased* (Row percentages shown), *Total EQD2 received (Gy) + Overall Treatment Time (OTT) in daysAbbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, CI Confidence interval, ECOG Eastern Cooperative Oncology Group, CD4 Cluster of differentiation 4, EQD2 Equivalent dose in 2-Gy fractionsNot retained: Not retained in the final multivariable model because of sparse and unstable estimatesMultivariate hazard model adjusted for age Hazard ratios for factors associated with overall survival in women with locally advanced cervical cancer referred for radiation therapy at CMJAH, 2021–2022 Bold, p < 0.05., Deceased* (Row percentages shown), *Total EQD2 received (Gy) + Overall Treatment Time (OTT) in days Abbreviations: CMJAH Charlotte Maxeke Johannesburg Academic Hospital, CI Confidence interval, ECOG Eastern Cooperative Oncology Group, CD4 Cluster of differentiation 4, EQD2 Equivalent dose in 2-Gy fractions Not retained: Not retained in the final multivariable model because of sparse and unstable estimates Multivariate hazard model adjusted for age Kaplan–Meier curves and log-rank test results showing unadjusted differences in survival for women with locally advanced cervical cancer according to (A) Eastern Cooperative Oncology Group performance status, (B) HIV status stratified by cluster of differentiation 4 (CD4) count (cells/mL), (C) disease stage, and (D) treatment modality Kaplan–Meier curves and log-test results showing unadjusted differences in survival for women with locally advanced cervical cancer according to (A) total equivalent dose in 2-Gy fractions received (Gy) and (B) hemoglobin concentration (g/dL) Univariate analysis The factors associated with poorer survival were ECOG status 2–4 [vs. 0 and 1; cHR, 1.81 (95% CI, 1.25–2.61)], HIV positivity with CD4 count of ≥200 cells/mL [vs. HIV negativity; cHR, 0.63 (0.41–0.97) vs. HIV positivity with CD4 count < 200 cells/mL: cHR, 0.62 (0.40–0.96)], stage III disease [vs. stages IB3 and II; crude hazard ratio (cHR), 1.95 (1.53–2.49)], no treatment [vs. CCRT; cHR, 3.42 (2.02–5.76)] and radiation therapy alone [vs. CCRT; cHR, 2.76 (2.18–3.49)], total EQD2 < 80 Gy [vs. ≥80 Gy; cHR, 2.69 (2.11–3.42)], baseline hemoglobin concentration ≤ 10 g/dL [vs. >10 g/dL; cHR, 1.34 (1.07–1.68)], baseline white blood cell count [cHR, 1.04 (1.01–1.06)], baseline platelet count [cHR, 1.09 (1.01–1.18)], baseline creatinine level [cHR, 1.02 (1.00–1.03), and overall treatment time ≥56 [cHR, 0.41 (0.33 – 0.52) Table 3]. Multivariable analysis On multivariable analysis, predictors of poorer overall survival were ECOG performance status 2–4 [vs. 0 and 1; adjusted hazard ratio (aHR), 1.75 (95% CI, 1.14–2.70)], stage III disease [vs. stages IB3 and II; aHR, 1.46 (1.07–1.99)], stage IVA disease [vs. stages IB3 and II; aHR, 2.02 (1.08–3.76)], and radiation therapy alone [vs. CCRT; aHR, 1.96 (1.39–2.75)]. A significant interaction between EQD2 and overall treatment duration was observed; therefore, a combined EQD2-treatment duration variable was included in the multivariable model. Compared with patients receiving EQD2 ≥80 with overall treatment duration <56 days (reference group), patients receiving EQD2 <80 with treatment duration <56 days had significantly increased mortality risk (aHR 2.65, 95% CI 1.50–4.71). No significant associations were observed among patients receiving EQD2 ≥80 with treatment duration ≥56 days (aHR 1.24, 95% CI 0.70–2.22) or EQD2 <80 with treatment duration ≥56 days (aHR 1.07, 95% CI 0.57–2.01). Table 3. The HIV status was not significantly associated with overall survival. Exploratory analyses were performed to assess whether the relationship between HIV/CD4 category and survival differed by treatment modality. In the multivariable Cox model, no statistically significant interaction between HIV/CD4 category and treatment modality was observed. In addition, exploratory univariable Cox regression analyses stratified by treatment modality showed no statistically significant association between HIV/CD4 category and survival within treatment strata (Supplementary Table 2). Discussion In this large public‑sector cohort of women with locally advanced cervical cancer in South Africa, we found that overall survival was strongly associated with disease stage, baseline clinical characteristics, and treatment‑related factors. Most women presented with advanced disease with substantial delays occurring between diagnosis and treatment initiation. Fewer than two‑thirds received concurrent chemoradiation, and survival outcomes remained poor, particularly among women with stage III and IVA disease. Although HIV infection was highly prevalent in this cohort, HIV status was not independently associated with overall survival after adjustment. The disease stage distribution in our cohort is consistent with that from a studies from SSA [10, 21]. This finding is not unique to SSA; Tekalign and Teshome [22] reported pooled prevalences of late-stage cervical cancer presentation of 69.3% for Asia and 50.2% for North America, including Mexico. The large proportion of women diagnosed with advanced-stage disease in our cohort reflects delays in early detection, suggesting inadequate awareness, screening program gaps such as low coverage, and limited healthcare access. Our study highlights significant delays at multiple stages of the cervical cancer treatment journey, from diagnosis to treatment completion. The CMJAH is the only public hospital providing radiation treatment in the Johannesburg area and continues to have long wait lists due to its limited number of treatment devices and shortage of trained radiation oncologists, medical physicists, and radiotherapists [23]. Additionally, during the study period, the hospital was recovering from the effects of the COVID-19 pandemic and a fire that led to departmental shutdown, further compounding delays. Prolonged wait times are not unique to CMJAH; similar delays have been reported in SSA [24] and other LMICs [25–27]. Although prolonged diagnostic and treatment intervals were prevalent in this cohort and remain clinically concerning, delay variables were not independently associated with overall survival in multivariable analysis. This may reflect the dominant prognostic influence of disease stage and baseline clinical condition at presentation, as well as limited power to detect delay effects in a setting where delays were uniformly long. Nevertheless, treatment delays remain important from a clinical and health‑system perspective, as they may contribute to disease progression, symptom burden, and pre‑treatment mortality, even if their independent survival effect could not be demonstrated in this analysis. Beyond their clinical implications, prolonged delays impose significant psychological and financial burdens on patients, increasing their anxiety, distress, and loss of confidence in healthcare systems [28, 29]. Addressing of these delays is critical to improve the stage at diagnosis, which impacts treatment outcomes and overall patient well-being. The 2-year overall survival rate in our cohort reflects the significant impact of cervical cancer on patient outcomes, highlighting the disease’s mortality burden and the challenges of long-term patient retention in cancer care. Survival outcomes varied significantly by the stage at diagnosis, reinforcing this variable’s well-established prognostic role for cervical cancer. The 2-year overall survival rate obtained in this study is comparable to other rates reported for SSA, where timely access to chemoradiation is often limited [13]. In a recent systematic review and meta-analysis of cervical cancer survival in SSA, the pooled 2-year overall survival rate was 60.0% [13]. These rates remain poor relative to the 5-year overall survival rate of 92% obtained for North America, Europe, and Asia in the EMBRACE-I study [30], which reflects early screening program implementation, better healthcare infrastructure, and the use of advanced treatment modalities. The survival disparity between stages IB3 & II disease and stages III and IVA disease in our cohort is consistent with the global trend of poor outcomes for cervical cancer cases diagnosed at advanced stages [31]. The cervical cancer treatment paradigm is evolving, with clinical trials performed to examine the use of induction chemotherapy [32] and immunotherapy [33] to improve outcomes. Cervical cancer survival is also influenced by multiple clinical and treatment-related factors. In this study, the ECOG performance status, disease stage, and treatment modality were independent predictors of overall survival. Poor functional status limits chemoradiation tolerance and prolongs treatment courses [34]. Thus, early intervention, nutritional support, and symptom management are important to improve treatment tolerance in patients with poorer functional status. A substantial proportion of women received hypofractionated EBRT (37.4 Gy in 11 fractions), a regimen not included in standard guidelines for locally advanced cervical cancer. At our centre, this schedule was used primarily for women with stage IIIB disease complicated by hydronephrosis, poor performance status, or limited chemotherapy eligibility, in the setting of prolonged waiting times and radiotherapy capacity constraints. This pragmatic approach aimed to reduce overall treatment time and mitigate disease progression while awaiting care and has been previously reported in South African and other LMIC settings with comparable outcomes [35–37]. The association of radiation treatment alone with significantly worse overall survival in this study aligns with global evidence supporting CCRT as the standard of care for locally advanced cervical cancer, as chemotherapy enhances tumor cell radiosensitivity, improving local control and survival outcomes [38, 39]. The large proportion of patients in our cohort who received radiation treatment alone reflects the poor performance status and comorbidities in our patients, particularly those with more advanced disease. Ensuring adequate supportive care for vulnerable patients may help to bridge this survival gap. Inclusion of women who did not initiate treatment in this study allowed capture of early mortality and healthcare system delays prior to treatment initiation, which would not have been observed in a treated‑only cohort. A significant interaction between EQD2 and overall treatment duration was identified, indicating that the association between radiation dose and survival was modified by treatment duration rather than acting as an independent fixed effect. Patients receiving EQD2 <80 with overall treatment duration <56 days had significantly poorer survival compared with the reference group (EQD2 ≥80 and treatment duration <56 days). However, this association was not observed among patients with prolonged treatment duration. This suggests that the prognostic impact of lower delivered radiation dose may differ according to treatment timing and delivery patterns. In the present cohort, EQD2 and treatment duration demonstrated a strong association, supporting the possibility that these variables do not act independently. Therefore, interpreting either variable in isolation could obscure clinically relevant relationships. Historically, prolonged treatment time has been linked to inferior outcomes in cervical cancer because accelerated repopulation during treatment may reduce tumor control. However, the impact of treatment prolongation may vary across patient populations and treatment settings. The absence of a significant independent association in this study suggests that overall treatment duration may not uniformly affect outcomes and that its effect may depend on treatment dose and other clinical factors. Limitations in brachytherapy technique availability and treatment interruptions during the study period likely contributed to suboptimal dose delivery in a substantial proportion of patients, highlighting the critical role of radiotherapy infrastructure and continuity of care in determining outcomes. All patients, apart from the stage IVA, had brachytherapy prescribed; some patients did not receive brachytherapy because they defaulted treatment, while some patients died before treatment completion. In our cohort, brachytherapy was planned to point A, which has historically yielded acceptable outcomes but is associated with higher local failure and treatment-related complication rates relative to modern techniques, especially for bulky or advanced tumors [40]. The EMBRACE-I study showed that image-guided brachytherapy achieved 5-year local control rates of >90% for locally advanced cervical cancer, with less morbidity than caused by historically applied two-dimensional techniques, benchmarking magnetic resonance imaging–guided brachytherapy as the gold standard [30], but such treatment has significant limitations in many settings. Given that radiation therapy quality is important for survival, CT-guided brachytherapy has been implemented in the Radiation Oncology Department of CMJAH. The absence of interstitial brachytherapy during the study period, coupled with the predominance of stage II or higher disease at presentation, highlights the need to implement hybrid intracavitary–interstitial brachytherapy. HIV was not associated with overall survival in the multivariable model. These findings suggest that HIV infection alone does not independently drive mortality in patients with cervical cancer. This lack of association may reflect the effectiveness of ART in the reduction of HIV-related immunosuppression and improvement of cancer care integration for HIV-positive patients, enabling them to receive standard treatment. Nevertheless, close monitoring for treatment-related toxicity in WLWH is necessary. Although HIV status was not independently associated with survival in the multivariable analysis, additional exploratory analyses were performed to assess whether this relationship differed by immune status and treatment modality. Neither the interaction analysis nor the treatment-stratified subgroup analyses showed significant differences in survival across HIV/CD4 groups. These findings suggest that, in this cohort, the relationship between HIV/CD4 status and survival was not substantially modified by treatment modality. Survival was measured from biopsy confirmation while treatment-related variables, including treatment modality, total EQD2, and overall treatment time, were determined subsequently during care, treatment-related hazard ratios may be susceptible to immortal time bias and confounding by indication and should therefore be interpreted cautiously. “Loss to follow‑up occurred in 11.5% of women, which may introduce bias into survival estimates. To mitigate this, we sought to ascertain vital status for patients with incomplete hospital follow‑up using VerifyID, a national administrative database, thereby reducing misclassification of deaths. Nevertheless, survival outcomes should be interpreted with caution, as some unascertained outcomes may persist, particularly for women who continued care at other facilities or outside the public sector. Such challenges are common in observational cancer studies in resource‑constrained settings and underline the difficulty of long‑term outcome ascertainment in routine clinical practice.” This study has several limitations. Its retrospective, single‑centre design may limit generalisability to other settings, particularly those with different referral patterns, radiotherapy capacity, or resource availability. Our findings are therefore most applicable to high‑volume public‑sector oncology centres in similar resource‑constrained contexts.In addition, several potentially relevant variables were not systematically available and could not be incorporated into the analysis, including detailed nodal burden, tumour size, chemotherapy dose intensity and completion, treatment‑related toxicity, socioeconomic barriers, travel distance to care, and patient‑reported factors influencing treatment adherence. Treatment interruptions were captured at a high level through overall treatment time but could not be fully characterised with respect to cause or duration. These unmeasured factors may contribute to residual confounding and limit causal inference regarding treatment delivery and survival. Nodal assessment under the FIGO 2018 staging system was constrained by limited access to advanced imaging modalities such as CT, MRI, or PET-CT. Consequently, reliance on clinical examination, ultrasound, and chest x-ray may have resulted in under-detection of nodal metastases and potential under staging of some patients. Few patients received nodal boost irradiation; however, documentation of its delivery was inconsistently captured in the patient’s medical record, and therefore these data could not be reliably included in the analysis. Despite these limitations, this study provides robust real‑world evidence from a large cohort of women with locally advanced cervical cancer treated in the public sector, highlighting substantial disease burden at presentation and the prognostic importance of treatment adequacy in a high‑HIV‑prevalence setting. Future prospective, multi‑centre studies incorporating detailed disease burden, treatment adherence, toxicity, and socioeconomic factors are needed to better characterise pathways to improved outcomes in this population.” Women who received radiotherapy alone experienced poorer overall survival than those treated with concurrent chemoradiation; however, this association should not be interpreted as causal. Treatment allocation in this cohort was determined by baseline clinical factors, including performance status, renal dysfunction, disease extent, and immunological status, which also influence prognosis. Women treated with radiotherapy alone therefore represent a clinically vulnerable subgroup with contraindications to chemotherapy rather than a comparable treatment group. Although radiotherapy alone was included as a treatment variable in multivariable analysis, residual confounding by baseline health status is likely and limits causal inference regarding treatment modality. This study demonstrated that 2-year survival in women with cervical cancer is highly stage dependent, with early-stage diagnosis enabling the achievement of relatively favorable outcomes and stage IVA disease carrying a poor prognosis. Given that nearly 60% of women in this cohort were living with HIV, strengthening linkages between HIV care and cervical cancer screening represents an important opportunity to reduce advanced‑stage presentation. WLWH have regular contact with healthcare services through ART programmes, and evidence from South Africa demonstrates improved screening uptake when cervical cancer prevention services are integrated into HIV care. WHO guidelines recommend more frequent HPV‑based screening for WLWH due to their elevated risk and accelerated disease progression. Leveraging HIV care platforms to deliver or fast‑track cervical cancer screening may facilitate earlier diagnosis and reduce the proportion of women presenting with locally advanced disease. This study highlights modifiable system‑level factors that may improve outcomes for women with locally advanced cervical cancer. Reducing delays between pathology, MDT discussion, and treatment initiation, strengthening patient navigation, optimising eligibility for concurrent chemoradiation, and safeguarding brachytherapy delivery are key priorities. Monitoring pathway intervals and treatment completion may further support improvements in survival in this public‑sector setting. The findings emphasize the need to improve early diagnosis and treatment access, expand cancer care services, and enhance patient follow-up strategies to improve survival outcomes. Supplementary Information Supplementary Material 1: Table S1: Stage at diagnosis of women who received no treatment, and the reason for no treatment in women with locally advanced cervical cancer referred for radiation therapy at Charlotte Maxeke Johannesburg Academic Hospital, 2021–2022. Table S2: Stratified univariable Cox regression analyses of HIV/CD4 category and survival by treatment modality in women with locally advanced cervical cancer referred for radiation therapy at Charlotte Maxeke Johannesburg Academic Hospital, 2021–2022. Acknowledgements We are grateful to the research staff of the Division of Radiation Oncology and all participants who made this work possible. Abbreviations Adjusted Hazard Ratio Antiretroviral Therapy Concurrent ChemoRadiation Therapy Cluster of Differentiation 4 Crude Hazard Ratio Confidence Interval Charlotte Maxeke Johannesburg Academic Hospital Computed Tomography External Beam Radiotherapy Eastern Cooperative Oncology Group Equivalent Dose in 2Gy fraction Human Immunodeficiency Virus Human Research Ethics Committee Interquartile Range Low and Middle-Income Countries Multidisciplinary Team Research Electronic Data Capture Standard Deviation Sub-Saharan Africa Women Living With HIV Authors’ contributions DR and SG conceptualized the original idea for the study. OAA, DM, and DR led the design of the core study. OAA, DM, and HM supervised data collection. MJ provided funding support for the ongoing data collection. OAA and OA conducted the formal analysis. OAA drafted the manuscript. HM, OA, MK, SS, WCC, MJ, MR, KR, SG, and DR reviewed and provided input on the manuscript. All authors contributed to the writing of the manuscript and read and approved the final manuscript. All authors read and approved the final manuscript. Funding This study was funded by the National Cancer Institute of the National Institutes of Health (R01 Grant Number CA192627). It was also supported by the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH, commissioned by the Research Networks for Health Innovations in sub-Saharan Africa Funding Initiative of the German Federal Ministry of Education and Research, German, 01KA2220B. It was further supported by Developing Excellence in Leadership, Training and Science in Africa (DELTAS Africa) grant made by the Alliance for Accelerating Excellence in Science in Africa (AESA) in partnership with the Wellcome Trust and the UK Foreign, Commonwealth and Development Office, and is part of the EDCPT2 programme supported by the European Union.” Data availability The dataset used/analysed in this study may be available from the corresponding author upon reasonable request and as allowable by regulatory processes. Declarations Ethics approval and consent to participate The University of the Witwatersrand Human Research Ethics Committee (HREC) Medical, Johannesburg, South Africa, approved the study (approval number M240862). The requirement for informed consent was waived by the HREC due to the study’s retrospective design. The study was conducted in compliance with the Declaration of Helsinki. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1: Table S1: Stage at diagnosis of women who received no treatment, and the reason for no treatment in women with locally advanced cervical cancer referred for radiation therapy at Charlotte Maxeke Johannesburg Academic Hospital, 2021–2022. Table S2: Stratified univariable Cox regression analyses of HIV/CD4 category and survival by treatment modality in women with locally advanced cervical cancer referred for radiation therapy at Charlotte Maxeke Johannesburg Academic Hospital, 2021–2022. Data Availability Statement The dataset used/analysed in this study may be available from the corresponding author upon reasonable request and as allowable by regulatory processes.