Patients who underwent staging laparotomy at the Severance and Gangnam Severance Hospital between March 2004 and March 2013 were included in this study. Inclusion criteria were as follows: 1) serous histological type ovarian cancer; 2) advanced International Federation of Gynecology and Obstetrics (FIGO) stage IIC–IV; 3) more than six cycles of taxane+platin-um-based first-line chemotherapy after primary debulking surgery; 4) serially measured CA125 values; and 5) clinical and radiological assessment during follow-up or recurrences. Finally, the study included 223 patients. Elevated CA125 levels are found in >90% of patients with advanced-stage disease, but in only 50% of patients with stage I disease.16 Normal serum CA125 levels are occasionally found in mucinous or endometrioid ovarian cancers, whereas elevated CA125 levels are most frequently observed in serous cancer.17 In this study, advanced stage patients with serous EOC were included to ensure a homogeneous study population. Patients with borderline malignancies treated with neoadjuvant chemotherapy and patients who did not receive staging laparotomy at Severance or Gangnam Severance Hospitals were excluded.

All patients received standard staging and debulking surgeries from gynecologic oncological specialists based on FIGO recommendations.18 The first cycle of a taxane plus platinum-based chemotherapy regimen was started an average of 17±10 days after surgery. Six to ten courses of postoperative adjuvant chemotherapy (POAC) were administered at approximately three-week intervals.

This study was approved by the Institutional Review Boards of Severance Hospital and Gangnam Severance Hospital. Some sera for CA125 assay were provided by Korea Gynecologic Cancer Bank through Bio & Medical Technology Development Program of the Ministry of Education, Science and Technology, Korea.

The following clinicopathological data were obtained from medical records: age, FIGO stage, histological type and grade, pathological result, type of primary treatment (type of surgery and adjuvant chemotherapy), surgical findings (mass size, ascites volume, residual mass size), CA125 level, surgery date, chemotherapy cycle and date, recurrence date, date of last visit, and date of death.

Histologic slides were reviewed by gynecologic pathologists, and histologic types were identified according to World Health Organization (WHO) criteria.19 The histologic grades of well-differentiated, moderately differentiated, and poorly differentiated tumours were classified as grades 1, 2, and 3, respectively.

Operative mass size was classified as either <5 cm or ≥5 cm, and ascites volume was classified as either <500 or ≥500 mL. A residual mass <1 cm after primary surgery was defined as optimal debulking surgery, whereas a residual mass of ≥1 cm was defined as suboptimal debulking surgery.

The treatment start date was defined as the date on which surgical staging was performed. The date of the last visit was defined as the date of the last recorded visit for the patient to either the gynaecology department or any other department. The recurrence date was defined as the date on which newly developed lesions were confirmed by computed tomography (CT), positron emission tomography (PET)-CT, or magnetic resonance imaging (MRI) based on elevated CA125 levels or clinical exams.

Overall survival (OS) and progression free survival (PFS) rates were calculated from the date of the first treatment to the date of death (last visit for surviving patients) and from the date of the first treatment to the date of recurrence (last visit in the absence of recurrence), respectively.

Pre-operative serum CA125 levels were measured within two weeks of the first surgery and approximately three weeks after each chemotherapy cycle, which was 1–3 days before the beginning of the subsequent chemotherapy cycle. CA125 immunoassays were performed using a Beckman DXI (Beckman Coulter, Inc., Brea, CA, USA) at Severance Hospital and a Modular Analytics E170 (Roche Diagnostics, Indianapolis, IN, USA) at Gangnam Severance Hospital according to manufacturer recommendations. All assays were performed in duplicate, and the reference range for normal CA125 levels was defined as ≤35 U/mL.

The CA125 nadir was defined as the absolute lowest CA125 value after the start of chemotherapy. The time to nadir was defined as the duration from the date of the first treatment to the date of the minimum value.

CA125 half-life (T1/2) was calculated based on mono-compartment models and exponential regression in accordance with earlier studies.202122 For the regression line calculation, at least three serial and aligned ln CA125 concentrations were required from the first six blood samples.

Half-life=ln 2/s,

in which was the slope of the regression curve that reflected the effects of surgery and chemotherapy.

S=ln C1-ln C2/t2-t1=ln C1/C2/t2-t1.

Therefore, half-life=ln 2×(t2-t1)/ln C1/C2.

C1 was the CA125 value before the first treatment, and C2 was the value at which CA125 decreased to <35 U/mL for the first time. If the CA125 level never decreased to <35 U/mL within three months of starting chemotherapy, the lowest value during that period was selected. The times (in days) for C1 and C2 were given as t1 and t2, respectively.

Univariate and multivariate Cox model analyses were used to investigate the effects of each clinicopathological prognostic factor on OS and PFS. For the statistical analysis of CA125-related prognostic factors, univariate Cox regression analysis was first performed with each variable and statistically significant variables were selected. Multivariate Cox regression analysis using stepwise selection was performed, and the most significant indicator of OS and PFS was then determined. Two methods were utilized for determining optimal cut-off values of each variable. First, the cut-off values that showed statistically significant outcomes regarding OS and PFS in earlier studies were used for the variables of pre-operative CA125, CA125 levels after the first and second chemotherapy cycles, time to normalization, and relative percentage change in CA125.4102023 Second, for the variables of nadir, time to nadir, CA125 half-life, and CA125 levels after the sixth chemotherapy cycle, the median was used to determine the cut-off value, because it was useful for dividing the group into two parts. Receiver operating characteristic analysis revealed a low AUC without significance when death or recurrence was selected as an event. Consequently, the cut-off values were as follows: pre-operative CA125 levels, 500 U/mL; CA125 levels after the first and second chemotherapy cycles, 35 U/mL; CA125 levels after the sixth chemotherapy cycle, 9 U/mL; CA125 levels at nadir, 8 U/mL; time to nadir, 160 days; time to normalization, third cycle; CA125 half-life, three days; and CA125 relative percentage change, 71%.

Statistical analyses were performed with the software package IBM SPSS Version 20.0 (SPSS Inc., Chicago, IL, USA). For all tests, all p values were two-sided, and p<0.05 was considered statistically significant.

Two hundred and twenty-three patients who met the inclusion criteria were included in this study. Patient characteristics and the results of survival analysis regarding the clinicopathological variables are shown in Table 1. The median patient age was 53 years (range, 25–81 years). Most (75.8%) patients had stage IIIC cancer; 80 patients (35.9%) were grade 2, and 107 patients (48.0%) were grade 3 at the time of diagnosis. One hundred and sixty-six patients (74.4%) received optimal debulking surgery, and 45 patients (20.2%) had residual tumours >1 cm. One hundred and seventy-eight patients (79.8%) received six cycles of taxane+platinum-based POAC, and 45 patients (20.2%) received seven to ten cycles. Univariate analysis showed that stage, largest diameter of residual tumour, ascites volume, and number of POAC cycles were prognostic factors for OS. Multivariate analysis showed that stage, ascites volume, and number of POAC cycles were prognostic factors for OS.

In the present study, the median pre-operative CA125 value in patients with stage IIC–IV serous ovarian cancer was 787.8 U/mL (range, 13.0–12001.0 U/mL), and CA125 median values after the first, second, and sixth chemotherapy cycles were 43.4 (range, 6.6–2677.2), 17.1 (range, 5.1–3954.5), and 8.8 (range, 2.2–9845.5) U/mL, respectively. The median CA125 nadir level was 7.9 U/mL (range, 2.0–2677.2 U/mL), median time to nadir was 165.0 days (range, 37.0–929.0 days), and median CA125 half-life was 3.1 days (range, 1.5–188.4 days). The median relative percentage change in CA125 compared to that at baseline after the first and second chemotherapy cycles were 91.7% and 96.8%, respectively.

Survival analyses were performed for the following 10 criteria: 1) pre-operative CA125 level; CA125 levels after the 2) first, 3) second, and 4) sixth chemotherapy cycles; 5) nadir; 6) time to nadir; 7) time to normalization; 8) half-life; and the relative percentage change from baseline to the 9) first and 10) second cycles of chemotherapy.

Each variable was divided into two groups based on the previously mentioned method.

Univariate Cox regression showed that seven variables [pre-operative CA125 level, CA125 levels after the first, second, and sixth chemotherapy cycles, nadir, time to normalization, and relative percentage change (baseline to after first chemotherapy cycle)] were significant prognostic factors for OS; six variables [pre-operative CA125 level, CA125 levels after the first and second chemotherapy cycles, time to normalization, half-life, relative percentage change (baseline to after first chemotherapy cycle)] were significant prognostic factors for PFS (Table 2).

To determine the most significant prognostic factors for OS and PFS, multivariate Cox regression, using stepwise regression techniques was performed with the variables that showed significant outcomes in univariate analysis. As a result, the CA125 level measured after the first chemotherapy cycle, time to CA125 normalization, and relative percentage change of CA125, compared to that at baseline after the first chemotherapy cycle, were sequentially selected for OS; time to CA125 normalization followed by CA125 level after the first chemotherapy cycle was selected for PFS (Table 3).

The CA125 level measured after the first chemotherapy cycle was selected as the most significant factor for OS and the second most significant prognostic factor for PFS, and was divided into two groups by the cut-off value of 35 U/mL. Kaplan-Meier analysis was then performed. The five-year OS rates were 87.4% and 56.9% for patients with CA125 levels ≤35 U/mL and >35 U/mL, respectively, after the first chemotherapy cycle (p<0.001). The prognosis for two-year PFS manifested differences of 69.7% and 36.7%, respectively (p<0.001) (Fig. 1). Kaplan-Meier analysis of the time to normalization, which was selected as the second highest factor for OS and the most significant prognostic factor for PFS, determined that patients who did not show normal CA125 values within three chemotherapy cycles had significantly unfavourable OS and PFS (five-year OS: 50.1%, two-year PFS: 22.2%) compared to that in patients who demonstrated normal CA125 values (five-year OS: 73.4%, two-year PFS: 55.0%) (p<0.001) (Fig. 2).