The present study included a total of 163 stage IB-IIA cervical cancer patients treated between 1999 and 2007 with RH and accompanying pelvic and/or para-aortic lymphadenectomy.

The medical records of each patient were retrospectively reviewed and 7 patients with insufficient data, including 2 patients whose records lacked a date of recurrence and 5 patients with occult cervical cancer detected following simple hysterectomy, were excluded; the remaining 156 patients formed the study population.

Abdominal type III RH with pelvic lymphadenectomy was the standard surgery. Bilateral salpingo-oophorectomy and para-aortic lymphadenectomy were not always performed. Postoperative adjuvant therapy was considered based on histopathological prognostic factors. Patients with at least one of the three high-risk factors (microscopic parametrial invasion, positive pelvic LNs and positive tumour resection margins) received adjuvant RT or platinum-based concurrent chemoradiotherapy (CCRT), which has been the preferred treatment since the 2000's (12). Patients with at least two of the three intermediate-risk factors (LVSI, stromal invasion of >1/2 of the cervix or >the outer 1/3, and a largest tumour diameter of ≥4 cm) received adjuvant RT alone.

Pathological specimens retrieved from the archives were included in this study. We collected information on the following parameters: FIGO clinical stage, age, histological type, grade of differentiation, tumour size, depth of cervical stromal invasion, LVSI, parametrial invasion, resection margin status, LN metastasis, adjuvant therapy, date of recurrence, location of recurrent disease, treatment at recurrence, and date of death or last follow-up.

The patients were periodically inspected every 3 months for the first 2 years, every 6 months for the following 3 years, and every 12 months thereafter. At the time of the examination, imaging studies, including magnetic resonance imaging or computed tomography and chest radiography, were conducted annually. When tumour recurrence was suspected based on the clinical findings or other imaging studies, biopsies on the suspected lesions were performed. We defined progression-free survival (PFS) as the last follow-up day or the time from the initial treatment to relapse, and overall survival (OS) as the last follow-up day or the time from the initial treatment to death from cervical cancer.

OS and PFS were assessed using the Kaplan-Meier method. Group comparison in the univariate analysis was performed using the log-rank test for categorical factors and a Cox proportional hazards model for continuous factors. All the prognostic factors identified as being significant in the univariate analysis were included in the multivariate analysis using a Cox proportional hazards model. Stepwise backward elimination methods were used to select the factors included in the multivariate Cox proportional hazards model (inclusion criteria, P<0.05; exclusion criteria, P>0.1). Significance was defined as a P-value <0.05. Data analyses were performed using SPSS software for Windows, version 11.0 (SPSS Inc., Chicago, IL, USA).

The clinicopathological characteristics of the patients are summarised in Table I. The mean age was 49 years (range, 29–75 years). A total of 141 patients had stage IB and 15 patients had stage IIA disease. A total of 123 patients had squamous cell carcinoma, 24 had adenocarcinoma and 9 patients were diagnosed with other histological types. The depth of stromal invasion was <1/3 in 32 patients, between 1/3 and 2/3 in 50 patients, and ≥2/3 in 74 patients. The size of the tumour was ≤2 cm in 64 patients, 2–4 cm in 67 patients, and >4 cm in 25 patients. Of the 156 patients, 24 had parametrial invasion, 96 had LVSI, 15 had positive resection margins, and 27 patients had LN metastasis. Of the 27 patients with LN metastasis, 14 had 1 positive node and 13 had ≥2 positive nodes. A total of 68 patients received treatment with CCRT, 54 received RT alone, and 34 received no adjuvant therapy (Table I).

The 5-year OS and PFS rates were 92.9 and 87.2%, respectively. A total of 20 patients developed recurrence, including 8 patients with recurrence in the pelvis, 10 with recurrence at distant sites, and 2 with recurrence at local as well at distant sites. The most frequent sites in the pelvis were the vaginal stump (n=6), followed by the pelvic sidewall (n=2). The most frequent sites of distant metastases were the lung (n=5), followed by the liver (n=3), bone (n=2), peritoneum (n=1) and brain (n=1). Of the 156 patients included in our analysis, 42 had a follow-up duration of <60 months, with 31 patients lost to follow-up for >1 year at the time of the analysis. The median follow-up duration of patients lost to follow-up was 29 months (range, 12–54 months).

Univariate analysis. The results of the univariate analysis of the prognostic factors are presented in Table II. In the log-rank test, tumour stage (P<0.001), stromal invasion (P=0.031), the size of the tumour (P=0.032) and the number of positive LNs (P=0.006) were identified as significant prognostic factors for OS, favouring a lower tumour stage, stromal invasion of <2/3, a tumour size of <4 cm and a small number of positive LNs, respectively. Patients with metastatic LNs were stratified into two groups according to the number of positive LNs, with one group that had 1 positive node, and the other group that had ≥2 positive nodes. The 5-year OS rates for patients with 0, 1 and ≥2 positive nodes were 91, 80 and 47%, respectively (P=0.006). There was no evidence of an association between OS rates and age, histological type, parametrial invasion, LVSI, resection margin, LN metastasis, or type of adjuvant therapy.

Tumour stage (P=0.002), stromal invasion (P=0.013), and the number of positive LNs (P<0.012) were also identified as significant prognostic factors for PFS, favouring a lower tumour stage, stromal invasion of <2/3, and a small number of positive LNs, respectively.

The results of the multivariate analysis of the prognostic factors are presented in Table III. The Cox proportional hazards model was used for the multivariate analysis. In terms of OS, tumour stage (P=0.002) and the number of positive LNs (P=0.021) were identified as significant prognostic factors. In terms of PFS, tumour stage (P=0.049) was identified as a significant prognostic factor.