Cervical cancer (CC) is one of most common cancer-related deaths among women worldwide (1). China also has a high burden of disease and a relatively high prevalence of advanced-stage disease (2). Locally advanced cervical cancer (LACC) refers to stage IB2, IIA, and IIB carcinomas, classified by the International Federation of Gynecology and Obstetrics (FIGO) staging system (3). The treatment for women with LACC remains challenging. The standard treatment for LACC is cisplatin-based concurrent chemoradiotherapy (CCRT) (4, 5). However, surgical treatment, including radical hysterectomy (RH) followed by adjuvant treatment is also a preferred treatment option for LACC in some circumstances (6, 7) and could offer potential benefits, including reducing the burden of tumor, preserving ovarian function and precisely determining the postoperative stage on the basis of histopathologic findings, thereby allowing quality of life improvements in young patients and individualizing postoperative treatment (4).

Recently, a phase III trial (8) and an epidemiological study (9) revealed that women undergoing minimally invasive surgery (MIS) through RH for early CC had a higher recurrence and worse survival than those who underwent abdominal radical hysterectomy (ARH). Nevertheless, only a few studies have compared the surgical and oncologic outcomes of MIS through RH to those of ARH for LACC. In the report from Zanagnolo et al. (10), ARH and robotic RH after neoadjuvant chemotherapy (NAC) in women with LACC were associated with similar perioperative and oncologic outcomes. Type B RH after NAC in well-selected patients is a safe procedure that could potentially reduce the operative time and late postoperative morbidity, without detrimental effects to survival (11). However, these trials lack sufficient sample sizes, long-term follow-ups and good comparators, which limit the generalizability of their conclusions.

In this study, we aimed to compare the survival outcomes of LRH with those of ARH for patients with FIGO stage IB2 to IIB CC at a tertiary institutional hospital in China from 2001 to 2015. All major procedures, consisting of parametrium resection and systematic lymphadenectomy, were performed by the corresponding authors.

In this study, we reported a large cohort of LACC patients who underwent RH with or without NAC. The surgical and survival outcomes between ARH and LRH could become the basis of decision making for physicians and patients, which could probably improve the quality of life of these relatively young patients. In the United States, CCRT is typically preferred over radical surgery for patients with LACC (17). However, to reduce the size of the primary tumor (19), improve the chances of operative curability and safety, and reduce long-term morbidity due to radiotherapy (20), NAC followed by RH had been attempted for patients with LACC. As previously reported, NAC followed by RH was suggested as a useful strategy (21), even for patients with nonsquamous cell carcinoma of the uterine cervix (22). Currently, no conclusive evidence is available regarding the relative benefits and drawbacks of primary RH vs. primary chemoradiotherapy for LACC (23). In a phase 3 randomized controlled study of the squamous subtype of LACC, NAC followed by chemoradiotherapy could improve DFS but not the OS compared with NAC followed by RH, and NAC had more toxicity (24). Judged by these findings, the impact of RH on the surgical or oncologic benefits in patients with LACC still requires more substantial evidence.

In our study, compared to LRH, ARH had a significantly superior survival outcome, especially in terms of DFS. The non-significant differences in OS between the two groups may be due to the insufficient follow-up periods in the LRH group. These findings are in accordance with previously reported results of early stage cervical cancer (8, 9). In a subgroup analysis, ARH for stage IB2 and the squamous carcinoma subtype led to a better DFS than LRH did, mainly because the ARH patients comprised the majority of the study population. However, despite several assumptions, the reasons for the inferior survival outcomes of MIS are unknown.

The learning curves of RH were most likely the most important reason, since almost all the LRH procedures were performed after 2011, and all the ARH procedures were performed before 2011. Due to the complex and complicated characteristics of radical surgeries, a lengthy learning process is needed to achieve comparable survival outcomes, which is equally essential for laparotomy or laparoscopy (25). Surgeons in the United States began to adopt minimally invasive RH for the treatment of cervical cancer in 2006 (26). In recent published study, the adoption of MIS was associated with a significant change in the survival trends and coincided with the beginning of a decline of survival rate between 2006 and 2010 (9). No definite reasons could explain this phenomenon except for the surgical route, and very few studies have considered factors that involve surgeons and their learning curves (25). Mastery of LRH requires experience with at least 20 cases and up to 50 cases (27–29), which shows the gradual slope of the learning curve (29). The learning curve for LRH and LN dissection would reach a turning point at 40 cases (30). These studies support the essential evaluation of the learning curves of a demanding surgical modality, both in early stage cervical cancer and LACC.

Other perspectives in our study need further consideration. A better description of the surgical procedure for the purpose of quality control is essential for future studies, since no conversions occurred in our study, which differs from other reports with rates of 1.5–3.5% (31). The significantly higher recurrence at the vaginal vault in the ARH group was comparable to the previously reported results (8). It is supposed that total laparoscopic/robotic intracorporeal colpotomy under CO₂ pneumoperitoneum might pose a risk for implantation in vaginal cuff margin and intraperitoneal spread (32). ARH may limit the chance of intra-abdominal seeding during cancer cell removal by placing two sets of clamps across the vagina below the cervix (33). However, all of these considerations lack the support of validated evidence and cannot explain the poor survival outcomes of LRH.

Studies suggest that histopathology is an independent prognostic factor that plays an essential role in the outcomes of LACC patients (34). Most reports suggest that patients with adenocarcinoma and adenosquamous carcinoma have a worse prognosis than patients with squamous carcinoma (34). In our study, there was a meaningful survival difference between the ARH and LRH groups in patients with squamous carcinoma, which is the primary subtype. The results indicate that further exploration of other less common histological subtypes is needed. At present, based on our findings, ARH should be the procedure of choice for LACC, at least for the squamous subtype.

In our study, as expected, LRH resulted in better surgical outcomes than ARH, including a shorter mean operating time, less blood loss, and shorter hospital stay, all of which are in accordance with previously reported findings (23). However, these surgical benefits still could not compensate for the low survival. A long, comprehensive evaluation of the surgical protocols is essential to benefit patients.

In our study, NAC did not provide survival benefits in the whole cohort or in the subgroups of various surgical routes or stages. Despite of numerous cohort studies guaranteeing the safety and effectiveness of NAC, randomized controlled trials (35) and meta-analyses (36) have proved that NAC had no positive impact on the survival outcomes of patients with LACC. Therefore, the applicability of NAC in LACC is controversial (17). However, current randomized studies of NAC did not rigorously follow uniform chemotherapy protocols (35). Currently, some expert opinions provided the appropriate selection of patients for the application of NAC (37).

The strengths of our study were the relatively large cohort and uniform RH procedures performed by experienced physicians. However, there are several major limitations. Similar to previous studies about LACC, our conclusion was also limited by inadequate power and probable residual confounding. As patients were all from a single center and a single team, selection bias would probably greatly interfere with the interpretation of the oncologic outcomes. The lack of uniform NAC protocols also limits the conclusions on the influence of NAC. Data on long-term survival assessed in randomized trials or in large cohort studies are needed for patients with LACC. The insufficient number of disease cases from stages other than IB2 and subtypes other than squamous carcinoma also limits the generalizability of our conclusions. Last, the inclusion of patients with stage IIB did not accord with current guideline (17). We have performed separate analysis for stage IB2 to IIA2, and have achieved similar conclusions. However, we wish our data, failure or possible experiences would provide knowledge and insight on the topic of treatment for LACC, the troublesome and confusing type.

Despite its favorable surgical outcomes, LRH was associated with a shorter DFS than ARH for patients with LACC (IB2 to IIB) as a whole population, and in patients with stage IB2 disease or the squamous subtype. However, in patients with stage IIA and IIB, and in patients of non-squamous subtypes, LRH hand no significant differences with ARH in the survival prognosis, probably due to the limited sample size in these subgroups. NAC had no significant impact on the survival outcomes.

All datasets generated for this study are included in the article/Supplementary Material.

The Institutional Review Board of Peking Union Medical College Hospital approved this study (No. ZS-1427). The registration No. is NCT03291236 (clinicaltrials.gov).

LL and MW conceived of the original idea for the study, interpreted the results, carried out the statistical analysis, edited the paper, and were the overall guarantors. MW completed all the major procedures of the radical hysterectomy, WW obtained ethical approval, contributed to the preparation of the data set, interpreted the results and contributed to drafts of the paper. SZ, SM, and XT contributed to the study design, interpretation of the results, and commented on drafts of the paper.