Real-world effectiveness of Glasgow prognostic score and tumor biomarkers in predicting post-relapse survival in patients with recurrent cervical cancer after minimally invasive surgery

Background Cervical cancer ranks the third most common malignancy of women worldwide, and recurrence of cervical cancer treatment is the major concern. Carcinoembryonic antigen (CEA), carbohydrate antigen 125 (CA125), squamous cell carcinoma antigen (SCCA) and Glasgow Prognostic Score (GPS) were potential prognostic indicator for cervical cancer. However, none of these markers have been evaluated to predict post-relapse survival in recurrent cervical cancer after treatment based on real-world clinical data. Aim To evaluate biomarkers CEA, CA125, SCCA and Glasgow Prognostic Score (GPS) in predicting post-relapse survival in recurrent cervical cancer after treatment based on real-world clinical data. Results Among the 1607 patients, the majority of patients (75.5%) were non-smokers, and the majority of histologic type (68.3%) was squamous cell carcinoma. Except CEA, there were significant difference between different GPS groups in these markers. Areas under the curves (AUC) for GPS, CEA, CA125 and SCCA were 0.632, 0.617, 0.641 and 0.628, respectively. All clinicopathologic characteristics were significantly correlated with CA125. Higher levels of biomarkers and GPS had lower survival and GPS=2 and SCCA was an independent prognostic factor for survival (P=0.008 and P=0.010, respectively). Conclusions In real-world settings, GPS and tumor biomarkers, especially SCCA to independently predict post-recurrence survival in patients with recurrent cervical cancer. antigen; SCC, squamous cell carcinoma.

each year [2,3]. Among all factors, recurrence contributes a lot, and it has been reported that more than 50% of women with cervical cancer experience cancer recurrence [4].
Long-term survival can be achieved in few patients with a cancer relapse [5], and most patients will suffer further disease progression or death within a median time of 11-16 months [6-8].
Laparoscopic hysterectomy was introduced to decrease the morbidity of the operative procedure, and has been adopted as the standard approaches in gynecologic oncology [9,10]. It has been proven that the laparoscopic approach is associated with fewer postoperative complications and recovery [11][12][13]. However, there have been concerns about the difficulties in achieving a sufficient resection margin by minimally invasive approaches. As a result, recurrence of cervical cancer after laparoscopic hysterectomy should be emphasized.
Serum tumor markers have long been used to predict prognosis and disease monitoring after primary treatment, of which carbohydrate antigen 125 (CA125) and squamous cell carcinoma antigen (SCCA) are two markers often used in cervical cancer. CA125 is membrane-bound glycoproteins expressed in some epithelial cancers [14,15]. Squamous cell carcinoma antigen (SCCA) can be overexpressed during the neoplastic transformation of the cervical squamous epithelium [16]. Apart from tumor markers, the inflammationbased Glasgow Prognostic Score (GPS) is a useful in predicting prognostic outcomes of cancer, which combines serum C-reactive protein (CRP) and albumin. Previous studies found that higher GPS at treatment time was independently associated with shorter survival of cervical cancer [17][18][19], suggesting that GPS is a potential prognostic indicator for cervical cancer. However, none of these markers have been evaluated to predict postrelapse survival in recurrent cervical cancer after minimally invasive surgery in real-world clinical data.
Here, we conducted a multicentered large-scale study to retrospectively compare the realworld effectiveness of these biomarkers in the survival of patients of cervical cancer.

Study cohort
The study was approved by the ethics committee of The Second Affiliated Hospital of Henan University of Chinese Medicine, China. As a retrospective study, informed consent was not required and the study protocol conforms to the ethical guidelines of the 1975

Clinical management and follow-up
Patients of FIGO IIa and before were managed with hysterectomy by means of minimally invasive approach. After surgery, patients were treated by adjuvant therapy if necessary according to FIGO guideline. Following initial surgery, patients underwent follow-up every suggested or there was an elevation of tumor markers at each follow-up time. Cancer recurrence was diagnosed by biopsy or imaging examinations, and the time of recurrence was defined as the baseline point. Survival time after recurrence was documented for each patient. The endpoint of follow-up was defined as either death or survival after recurrence.

Measurement of serum parameters
All serum parameters were assessed at relapse time. Serum albumin and CRP levels were assayed with bromocresol green and immunoturbidimetric test respectively using Hitachi 7600 chemistry autoanalyzer. Serum CA125, CEA and SCCA levels were assessed with ARCHITECT reagent kits using ARCHITECT i2000SR.

Statistical analysis
SPSS version 22.0 statistical software (IBM, USA) were applied for all statistical analysis.
Student T-test or Mann-Whitney test was applied to compare the differences between two groups. Receiver Operating Characteristics (ROC) Curves were depicted to evaluate the detecting efficiency of each marker, and DeLong test was applied to compare the different AUROC. For survival analysis, Kaplan-Meier analyses and log-rank test were used.
Univariate and multivariate cox proportional hazard models were used to calculate hazard ratios and 95% confidence intervals. Two-tailed P value less than 0.05 was defined to be statistically significant. Table 1 shows the characteristics of patient categorized by GPS scores. Among the 1607 patients, the majority of patients (75.5%) were non-smokers, and 68.3% of them were squamous cell carcinoma. By recurrence, the majority of post-surgery patients were recurrent within pelvis (86.3%). Patients in different GPS groups have significantly different (P=0.045) median tumor-free intervals (TFI) after surgery, which decreased with GPS increasing. Except CEA, there were significant difference between GPS groups in these markers.

ROC curves for NLR, PLR and parametrial involvement and cutoff values determination
ROC curves of predictive significance of post-relapse survival values were shown in Fig. 1.
Areas under the curves (AUC) for GPS, CEA, CA125 and SCCA were 0.632, 0.617, 0.641 and 0.628, respectively. DeLong test showed that there were no significant difference between every two of these four markers (P>0.05 for all). Youden indexes showed that the best cutoff values were 50 ng/ml for CEA, and 75 U/ml for CA125, and 8.5 ug/ml for SCCA.

Association between GPS, tumor biomarkers and clinicopathologic parameters
Patients were stratified into the high and low group according to the cutoff values above.
The clinicopathologic characteristics were compared, as showed in Table 2. Overall, no significant relationships were noted between age, BMI, smoke with markers. However, a statistically significant correlation was observed between smoke, median TFI, primary tumor histology, site of recurrence and GPS (P<0.001, P=0.038, P<0.001, P<0.001, respectively); Significant correlations between BMI, median TFI, primary tumor histology and CEA (P<0.001, P=0.012 and P<0.001, respectively); Significant correlations between smoke, primary tumor histology and SCCA (P=0.034 and P<0.001, respectively). All clinicopathologic characteristics were significantly correlated with CA125 (P<0.05 for all).

Survival analysis
Generally, the overall death within 50 months after cancer recurrence was detected in 508 (31.6%) patients, consisting of 419 (44.8%) patients in the GPS2 group, 58 (21.8%) patients in the GPS1 group, and 32 (7.9%) patients in GPS0 group. The median interval between recurrence and death was 21 months (range: 6-46 months) and 90% of death were identified within 2 years after recurrence. The estimated 1-, 2-and 3-year's survival rates for patients in the GPS0 group were 86.3%, 62.8%, and 58.2%, compared with 85.9%, 51.2%, and 30.4% for those in the GPS1 group, and 65.2%, 31.7%, and 18.4% for those in the GPS2 group, respectively. As shown in the figure 2, patients with a higher GPS scores had decreased survival (P < 0.001). Besides, different levels of tumor biomarkers could also predict the survival rates. Figure 2BCD showed that higher levels of all three biomarkers had lower survival. Our results demonstrate a strong relationship between higher GPS at recurrence time and shorter survival in patients with recurrent cervical cancer. And besides, higher GPS was an independent risk factor for shorter post-relapse survival.
Real-world researches often enrolled an abundant number of participants with a relatively less limited inclusion criterion, so as to provide a satisfactory external validity [25]. Our study for the first time analyzed the efficacy of four markers in predicting post-recurrence survival based on real-world clinical data from 5 centers. As a result, larger number was the main strength of our study. However, there are some limitations that deserve to be mentioned. Within this study, data were retrospectively analyzed, leading to shortcomings, such as patient selection and incomplete data acquisition. In addition, due to the relatively long study period and complex methods of treatment after cancer recurrence, different types of treatment were not analyzed in this study.

Conclusion
The present study suggests that GPS and tumor biomarkers, especially SCCA could independently predict post-recurrence survival in patients with recurrent cervical cancer.
In addition, distant relapse and shorter TFI after initial treatment were also associated with shorter post-recurrence survival. However, prospective and multi-centered study should be designed to validate this study.

Funding
None.

Availability of data and materials
The datasets and supporting materials generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions
JH Guo design the study, conducted the investigation and drafted the manuscript. YY Wang, XX Wang, SL Zhou and PM Liu collected and analyzed the data. All authors read and approved the final manuscript.

Ethics approval and consent to participate
The study protocol was approved by the Ethics Committee of The Second Affiliated Hospital of Henan University of Chinese Medicine.

Consent for publication
Not applicable