For this retrospective study, the data used were obtained from the Project 1538 developed through a clinical trial (Project 1538; Ethics Clearance NFEC-2017-135; Clinical trial registration number: CHiCTR1800017778, http://apps.who.int/trialsearch/), which was authorized by the Ethics Committee of Southern Hospital. The database includes 63,926 cases of CC, collected across 47 hospitals in China. Included are patients’ clinical information, pretreatment biopsy results, laboratory and imaging information, treatment plans, treatment complications, and postoperative pathology reports. Data were collected by two gynecologists, who received specific training for the clinical trial, using EpiData 3.1 (EpiData Association, Odense, Denmark) for dual data entry and standard interviews for follow-up data by telephone calls or outpatient visits. Details of the data collection and follow-up methods have been previously described [23, 24].

For our study, the selection criteria for eligible cases were as follows: age $\ge$18 years; CC detected through cervical biopsy; histological diagnosis of squamous cell carcinoma (SCC), adenocarcinoma (AC), or adenosquamous cell carcinoma (ASC); FIGO 2018 stage IIIC classification; and availability of follow-up data. With regards to initial treatment, the inclusion criteria were as follows: R-CT, external radiation therapy (RT) at a dose of $\ge$45 Gy/brachytherapy at a dose of $\ge$40 Gy, on or off concurrent platinum-based chemotherapy (an example of the dose of R-CT: the doses of the external radiation therapy is 1.8–2 Gy per fraction, the doses of the brachytherapy (high dose rate) is 6–7 Gy per fraction, and the total dose of external radiation therapy and brachytherapy is usually $\ge$85 Gy); ARH, Q-M type-B or type-C radical hysterectomy pelvic lymphadenectomy, with or without para-aortic lymphadenectomy; and NACT, consisting of platinum-based neoadjuvant therapy (2–3 cycles) and radical surgery; and patients with available follow-up data. The patients of stage IIIC2 received extended-field external irradiation radiotherapy, concurrent platinum-based chemotherapy, and brachytherapy. Exclusion criteria consisted of violation of selection criteria, cancer of the uterine cervix stump, and CC combined with other malignancies or pregnancy.

The outcomes for this study were overall survival (OS) and disease-free survival (DFS), with a cutoff point of 5 years post-treatment. OS is the last point in time from diagnosis to valid follow-up or death for any reason. DFS is the last point in time from diagnosis to follow-up, relapse, or death.

Continuous variables were described as the mean $\pm$ standard deviation, while the independent samples t-test was used for between-group comparisons. Categorical values were described as their percentages, and the chi-squared test or Fisher’s exact probability test was used, as appropriate, for between-group comparisons. The Kaplan-Meier method was applied for survival analysis. Independent risk factors were established with Cox proportional hazards models along with hazard ration (HR) and 95% confidence interval (CI). Propensity score matching (PSM) was used to minimize the influence of baseline differences between groups. All analyses were performed using SPSS (version 29; IBM Corp., Armonk, NY, USA), with significance set at a p-value 0.05.

Based on the inclusion and exclusion criteria, 4086 CC cases were selected from the Project 1538 database, with no missing values. Fig. 1 exemplifies the patient screening process. The findings from baseline profiling for stages IIIC1 and IIIC2 are presented in Supplementary Table 1. The findings from baseline profiling for the R-CT and ARH groups, R-CT and NACT groups, and ARH and NACT groups for stage IIIC1 are described in Supplementary Tables 2,3,4, respectively. The findings from baseline profiling for the ARH and NACT groups for stage IIIC2 are depicted in Supplementary Table 5.

Fig. 1.

Patient selection flow diagram. Based on the inclusion and exclusion criteria, 4019 stage IIIC1 cases (R-CT, n = 1913; ARH, n = 1493; and NACT, n = 623) and 57 stage IIIC cases (ARH, n = 36; and NACT, n = 21) were screened from the Chinese 1538 project database. FIGO, international federation of gynecology and obstetrics; R-CT, radical chemoradiotherapy; ARH, abdominal radical hysterectomy; NACT, neoadjuvant chemotherapy and radical surgery.

Kaplan-Meier analysis revealed a marked distinction in 5-year DFS between stages IIIC1 and IIIC2 (73.8% vs. 70.2%, p = 0.638) but not in OS (65.6% vs. 38.5%, p 0.001) before PSM; the same result was obtained after PSM (DFS: 68.3% vs. 39.9%, p 0.001; OS: 79.9% vs. 73.1%, p = 0.583) (Fig. 2).

Fig. 2.

OS and DFS of stages IIIC1 and IIIC2. Direct comparison of 5-year OS (A) and 5-year DFS (B) for stage IIIC1 and IIIC2. Comparison of 5-year OS (C) and 5-year DFS (D) for stage IIIC1 and IIIC2 after PSM. PSM, propensity score matching; OS, overall survival; DFS, disease-free survival.

Cox regression analyses (post-PSM) showed that stage IIIC2 was correlated with worse DFS (HR = 2.295, 95% CI 1.470–3.581, p 0.001) compared with stage IIIC1 but not with worse OS (p = 0.716). Additionally, the ARH group was not correlated with 5-year OS (p = 0.076) or DFS (p = 0.371) as opposed to the R-CT group; moreover, the NACT group was not correlated with 5-year OS (p = 0.630) or DFS (p = 0.817). Finally, age was not correlated with 5-year OS or DFS (p $>$ 0.05); compared with SCC, AC was correlated with worse 5-year OS and DFS (p 0.05); however, ASC was not correlated with 5-year OS or DFS (p $>$ 0.05; Table 1).

Kaplan–Meier analysis revealed notable variations in 5-year OS (pre-PSM: 68.9% vs. 79.6%, p 0.001; post-PSM: 72.0% vs. 80.0%, p 0.001) and DFS (pre-PSM: 65.8% vs. 70.5%, p 0.001; post-PSM: 67.2% vs. 71.0%, p 0.001) between the R-CT and ARH groups (Fig. 3). A marked distinction in 5-year DFS (pre-PSM: 65.8% vs. 55.5%, p = 0.012; post-PSM: 67.7% vs. 55.3%, p = 0.002) was found between the R-CT and NACT groups but not in 5-year OS (pre-PSM: 68.9% vs. 71.2%, p = 0.634; post-PSM: 70.7% vs. 70.7%, p = 0.613; Fig. 4). Furthermore, there were notable variations in the 5-year OS (pre-PSM: 79.6% vs. 71.2%, p 0.001; post-PSM: 80.9% vs. 70.5%, p 0.001) and DFS (pre-PSM: 70.5% vs. 55.5%, p 0.001; post-PSM: 70.7% vs. 54.1%, p 0.001) between the ARH and NACT groups (Fig. 5).

Fig. 3.

OS and DFS of the R-CT and ARH groups (stage IIIC1). Direct comparison of 5-year OS (A) and 5-year DFS (B) for R-CT and ARH in stage IIIC1. Comparison of 5-year OS (C) and 5-year DFS (D) for R-CT and ARH in stage IIIC1 after PSM. PSM, propensity score matching; OS, overall survival; DFS, disease-free survival; R-CT, radical chemoradiotherapy; ARH, abdominal radical hysterectomy.

Fig. 4.

OS and DFS of the R-CT and NACT groups (stage IIIC1). Direct comparison of 5-year OS (A) and 5-year DFS (B) for R-CT and NACT in stage IIIC1. Comparison of 5-year OS (C) and 5-year DFS (D) for R-CT and NACT in stage IIIC1 after PSM. PSM, propensity score matching; OS, overall survival; DFS, disease-free survival; R-CT, radical chemoradiotherapy; NACT, neoadjuvant chemotherapy and radical surgery.

Fig. 5.

OS and DFS of the ARH and NACT groups (stage IIIC1). Direct comparison of 5-year OS (A) and 5-year DFS (B) for ARH and NACT in stage IIIC1. Comparison of 5-year OS (C) and 5-year DFS (D) for ARH and NACT in stage IIIC1 after PSM. PSM, propensity score matching; OS, overall survival; DFS, disease-free survival; ARH, abdominal radical hysterectomy; NACT, neoadjuvant chemotherapy and radical surgery.

Cox regression analyses (post-PSM) showed that the ARH group was correlated with better 5-year OS (HR = 0.603, 95% CI 0.493–0.738, p 0.001) and DFS (HR = 0.727, 95% CI 0.616–0.857, p = 0.001) than the R-CT group. Moreover, the study found that compared with SCC, both AC and ASC were correlated with worse 5-year OS and DFS (p 0.001); however, age was not correlated with 5-year OS or DFS (p $>$ 0.05; Table 2). Moreover, the NACT group was correlated with worse 5-year DFS compared with the R-CT group (HR = 1.308, 95% CI 1.088–1.573, p = 0.004) but not with 5-year OS (p = 0.586). Both AC and ASC were correlated with worse 5-year OS or DFS compared with SCC (p $\mathrm{\le }$ 0.017); however, age was not correlated with 5-year OS or DFS (p $>$ 0.05; Table 2). In addition, the 5-year OS (HR = 1.750, 95% CI 1.306–2.345, p 0.001) and DFS (HR = 1.653, 95% CI 1.312–2.083, p 0.001) were worse in the NACT group than in the ARH group. Nevertheless, other factors, including age, histological type, hysterectomy type, tumor diameter, lymphovascular space invasion (LVSI), cervical stromal invasion, parametrial involvement, and vaginal margin were not correlated with 5-year OS or DFS (p $>$ 0.05; Table 2).

PSM could not be performed because of the small number of cases in both groups. Kaplan-Meier analysis revealed a marked distinction in 5-year DFS (45.4% vs. 30.1%, p = 0.025) but not in 5-year OS (69.7% vs. 73.9%, p = 0.750) between the ARH and NACT groups (Fig. 6).

Fig. 6.

OS and DFS of the ARH and NACT groups (stage IIIC2). Direct comparison of 5-year OS (A) and 5-year DFS (B) for ARH and NACT in stage IIIC2. OS, overall survival; DFS, disease-free survival; ARH, abdominal radical hysterectomy; NACT, neoadjuvant chemotherapy and radical surgery.

Cox regression analyses (pre-PSM) showed that the NACT group was related to worse 5-year DFS (HR = 2.526, 95% CI 1.012–6.301, p = 0.047) but not to worse 5-year OS (p = 0.825) versus the ARH group. Age was also correlated with worse 5-year DFS (p = 0.040) but not with 5-year OS (p = 0.868). Compared with SCC, AC was correlated with worse 5-year DFS (p = 0.001) but not with worse 5-year OS (p = 0.075), whereas ASC was not correlated with either 5-year OS or DFS (p $>$ 0.05). Overall, hysterectomy type, tumor diameter, LVSI, cervical stromal invasion, parametrial involvement, and vaginal margin were not correlated with 5-year OS or DFS (p $>$ 0.05; Table 3).

Substantial evidence supports para-aortic LNM as a clear adverse prognostic factor [5, 18, 19, 25, 26, 27, 28, 29]. Cho et al. [25] analyzed stage IIIC CC patients and observed that patients with para-aortic LNM exhibited noticeably worse 5-year OS and DFS than those without para-aortic LNM (p 0.001), which was similar to the results of the studies of Guo et al. [26] and Yan et al. [5]. In the present study, the 5-year DFS was found to be superior in stage IIIC1 compared with that in stage IIIC2, and stage IIIC2 was related to worse 5-year DFS. As such, this research further classified stage IIIC CC patients into stages IIIC1 and IIIC2 subgroups for individualized treatment studies.

Although R-CT is used in patients with all stages of CC to improve the oncological outcomes, it may also damage nearby organs and cause adverse effects [17, 18, 19]. Contrastingly, ARH offers several advantages, including avoidance of the adverse effects of R-CT; removal of the primary tumor, infiltrating tissues, and LN; and the possibility of personalized postoperative treatment based on pathological findings. As such, ARH is the preferred treatment option for early CC [10]. However, it was demonstrated by Wu et al. [30] and Landoni et al. [31] that the therapeutic efficacies of ARH and R-CT are similar. Furthermore, investigations by Yan et al. [32] and Jang et al. [33] indicated that the prognosis after ARH was significantly better than that after R-CT, which is in keeping with the outcomes of the present research. In the present study, compared with the R-CT group, the ARH group had superior 5-year OS (post-PSM: 71.0% vs. 80.0%, p 0.001) and DFS (post-PSM: 67.2% vs. 71.0%, p 0.001) and was related to better 5-year OS and DFS (post-PSM: p 0.05) for stage IIIC1. The patients treated with R-CT may develop serious immediate or long-term complications, including impaired ovarian function, damaged vaginal structure and function, and inflammation of the bladder and rectum [17, 18, 19], which may affect their physical condition. For tumor treatment, whether it is initial treatment or treatment after the occurrence of recurrence and metastasis, the patient’s physical condition is a key component influencing the outcome [34, 35]. This may account for the poor treatment outcome in the R-CT group. Considering the undesirable effects of R-CT and the paucity of radiotherapy equipment in developing countries, ARH may be a suitable alternative option, especially for patients who desire to preserve their fertility.

Interestingly, NACT positively lowers tumor stage, reduces tumor burden, and enhances the chance of surgical tumor removal; however, conclusive evidence on whether NACT has a positive effect on the prognosis of CC is still lacking [36, 37, 38, 39]. A meta-analysis by Ye et al. [40] revealed that oncological outcomes were significantly superior in the NACT group compared with those in the direct surgery group (p 0.05). Similar findings were confirmed by Hu et al. [41] and Rydzewska et al. [42]. However, a phase III randomized controlled study by Katsumata et al. [43] in Japan that included 134 FIGO 2009 patients with stage IB2, IIA2, and IIB (including LNM) CC displayed a distinctively inferior OS in the NACT group compared with that in the radical surgery group (58% vs. 80%, p = 0.015), leading to the early termination of the study. In the present study, compared with the ARH group, the NACT group experienced the worse 5-year OS and DFS and was correlated with worse 5-year OS and DFS for stage IIIC1 (post-PSM: p 0.001); simultaneously for stage IIIC2, NACT had a worse 5-year DFS and was correlated with a worse DFS (post-PSM: p 0.05). For both stages, NACT showed no oncological advantages over ARH.

Duenas-Gonzalez et al. [44] also observed that, compared with R-CT, NACT did not have an improved prognosis for CC. Similar findings were showed in this research, where for stage IIIC1, R-CT had a better 5-year DFS than NACT, and NACT was also correlated with a worse 5-year DFS (post-PSM: p 0.05). Owing to the lack of R-CT cases, in-depth study is required to investigate the effect of R-CT on stage IIIC2. Nwankwo et al. [45] suggested that NACT may introduce confounding pathological factors in surgical specimens, influence postoperative pathological findings, complicate the assessment of the need for postoperative adjuvant therapy, and lead to missed or overtreatment. This may provide an explanation for the inferior effectiveness of NACT compared with ARH and R-CT in this research. As such, NACT should be used with caution for stage IIIC CC.

This current study is notable, as it is one of the largest studies based on the 2018 FIGO staging system for stage IIIC1 and IIIC2 CC from less developed regions. This study is innovative, as it stratified patients based on different LNM locations and compared the oncological prognosis of three commonly used CC treatments. In spite of the wide-ranging application of minimally invasive approaches, abdominal hysterectomy remains a general surgical procedure of intervention [46]. Furthermore, the Laparoscopic Approach to Cervical Cancer (LACC) confirmed that the prognosis of minimally invasive surgery was worse than that of ARH [47]. Therefore, only patients treated with ARH were included in this study. Nevertheless, this study has a few limitations. First, being retrospective, this study may have involved data imbalance between groups. Second, the lack of information on the specific treatment regimen, dose and duration of postoperative adjuvant therapy, R-CT, and NACT was also a shortcoming of the current study. Third, only 57 stage IIIC2 cases and no R-CT cases were available, which may have affected the reliability of the results. Fourth, stratification was performed based on only two LNM locations and did not include LN size and number and other LNM locations.