Prognostic value of CA125, HE4, ROMA, and CPH-I in advanced epithelial ovarian cancer.

A total of 261 patients were included in this study, comprising 80 patients with benign ovarian tumors, 84 patients with early-stage EOC, and 97 patients with advanced-stage EOC. The baseline clinical and pathological characteristics of the three groups are summarized in Table  1 . No significant differences were observed among the three groups with respect to age, family history of ovarian cancer, comorbid hypertension or diabetes, smoking status, or alcohol consumption ( P  > 0.05). In contrast, menopausal status and pathological subtype distribution differed significantly among groups (both P  < 0.05), with a higher proportion of postmenopausal patients and serous carcinoma observed in the advanced-stage EOC group. Table 1 Baseline characteristics of patients in the benign, early-stage, and advanced-stage groups Benign (80) Early-stage (84) Late-stage (97) F/χ 2 P Age (x ± S) 22 ~ 78(54.20 ± 10.08) 33 ~ 77(55.45 ± 10.60) 25 ~ 79(57.20 ± 10.66) 1.840 0.162 Menopausal status (n, %) 41 (51.25) 51 (60.71) 70 (72.16) 8.242 0.016 Family history of ovarian cancer (n, %) 4 (5.00) 3 (3.57) 5 (5.15) 0.300 0.861 History of hypertension (n, %) 16 (20.00) 18 (21.43) 24 (24.74) 0.616 0.735 History of diabetes (n, %) 11 (13.75) 12 (14.29) 15 (15.46) 0.616 0.735 Smoking history (n, %) 4 (5.00) 3 (3.57) 4 (4.12) 0.210 0.900 Alcohol consumption history (n, %) 2 (2.50) 2 (2.38) 2 (2.06) 0.018 0.991 F values are reported for continuous variables and χ² values for categorical variables Baseline characteristics of patients in the benign, early-stage, and advanced-stage groups 41 (51.25) 51 (60.71) 70 (72.16) 4 (5.00) 3 (3.57) 5 (5.15) 16 (20.00) 18 (21.43) 24 (24.74) 11 (13.75) 12 (14.29) 15 (15.46) 4 (5.00) 3 (3.57) 4 (4.12) 2 (2.50) 2 (2.38) 2 (2.06) F values are reported for continuous variables and χ² values for categorical variables Serum CA125 and HE4 levels, as well as the derived indices ROMA and CPH-I, demonstrated a stepwise increase from the benign tumor group to the early-stage EOC group and further to the advanced-stage EOC group (all P  < 0.001; Table  2 ). This trend was consistently observed in both premenopausal and postmenopausal patients. These findings indicate that elevated CA125 and HE4 levels and higher ROMA and CPH-I scores are associated with increasing disease severity in EOC, supporting their relevance as markers reflecting tumor burden and disease progression. Table 2 Comparison of serum CA125, HE4, ROMA, and CPH-I levels among late-stage EOC, early-stage EOC, and Benign groups n CA125(U/mL) HE4 (pmol/L) ROMA(%) CPH-I(%) (Pre/Post) Pre Post Late-stage 97 (27/70) 725.70 (251.85, 1428.00) 360.00 (184.05, 745.40) 83.43d (57.52, 96.08) 95.37 (83.71, 98.55) 88.67 (64.74, 98.41) Early-stage 84 (33/51) 106.54(32.69, 370.93) 123.90 (61.34, 256.55) 23.88 d (9.75, 71.79) 59.20 (38.41, 90.68) 25.87 (3.31, 75.90) Benign 80 (39/41) 15.15 (10.33, 20.94) 52.78 (44.77, 64.05) 6.77 d (5.45, 9.17) 12.93 (9.16, 19.26) 1.40 (0.99, 2.71) H value 151.93 133.19 49.30 95.56 156.44 P value < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 Pre indicates premenopausal status, and Post indicates postmenopausal status. Data are presented as median (P25, P75) Comparison of serum CA125, HE4, ROMA, and CPH-I levels among late-stage EOC, early-stage EOC, and Benign groups 97 (27/70) 725.70 (251.85, 1428.00) 360.00 (184.05, 745.40) 83.43d (57.52, 96.08) 95.37 (83.71, 98.55) 88.67 (64.74, 98.41) 84 (33/51) 123.90 (61.34, 256.55) 23.88 d (9.75, 71.79) 59.20 (38.41, 90.68) 25.87 (3.31, 75.90) 80 (39/41) 15.15 (10.33, 20.94) 52.78 (44.77, 64.05) 6.77 d (5.45, 9.17) 12.93 (9.16, 19.26) 1.40 (0.99, 2.71) Pre indicates premenopausal status, and Post indicates postmenopausal status. Data are presented as median (P25, P75) Among the 97 patients with advanced-stage epithelial ovarian cancer, 70 patients achieved optimal cytoreductive surgery, whereas 27 patients underwent suboptimal cytoreduction. Comparisons of serum CA125, HE4, ROMA, and CPH-I levels between the two groups are presented in Table  3 . Patients in the suboptimal cytoreduction group exhibited significantly higher preoperative serum CA125 and HE4 levels compared with those in the optimal cytoreduction group ( P  < 0.05). Similarly, ROMA values in both premenopausal and postmenopausal patients, as well as CPH-I values, were significantly elevated in the suboptimal cytoreduction group (all P  < 0.05). These results indicate that higher preoperative levels of CA125, HE4, ROMA, and CPH-I are associated with an increased likelihood of suboptimal cytoreductive surgery in patients with advanced EOC, suggesting their potential utility in preoperative risk stratification. Table 3 Comparison of serum CA125, HE4, ROMA, and CPH-I levels between optimal and suboptimal cytoreduction groups in advanced EOC n CA125(U/mL) HE4 (pmol/L) ROMA(%) CPH-I(%) Pre Post Optimal cytoreduction 70 570.90 (156.40, 1113.50) 273.70 (164.80, 524.15) 76.89(34.02, 89.99) 93.40(79.44, 97.24) 84.57(49.19, 94.76) Suboptimal cytoreduction 27 1476.00 (691.70, 4431.00) 676.80 (382.50, 1051.00) 97.42(96.07, 98.87) 98.40(96.22, 99.21) 98.56 (93.93, 99.14) U value 434.00 550.50 6.000 248.00 378.00 P value < 0.001 0.001 0.001 0.001 < 0.001 Data are presented as median (P25, P75) Comparison of serum CA125, HE4, ROMA, and CPH-I levels between optimal and suboptimal cytoreduction groups in advanced EOC 570.90 (156.40, 1113.50) 273.70 (164.80, 524.15) 1476.00 (691.70, 4431.00) 676.80 (382.50, 1051.00) 98.56 (93.93, 99.14) Data are presented as median (P25, P75) Among the patients with advanced-stage epithelial ovarian cancer who experienced disease recurrence during follow-up, 36 patients were classified as platinum-sensitive recurrence and 22 patients as platinum-resistant recurrence according to the platinum-free interval. Comparisons of preoperative serum biomarker levels between the two groups are summarized in Table  4 . Although preoperative CA125, HE4, ROMA, and CPH-I levels tended to be higher in the platinum-resistant recurrence group than in the platinum-sensitive recurrence group, no statistically significant differences were observed between the two groups ( P  > 0.05). Table 4 Comparison of preoperative serum CA125, HE4, ROMA, and CPH-I levels between platinum-sensitive and platinum-resistant patients with advanced EOC n CA125(U/mL) HE4 (pmol/L) ROMA(%) CPH-I(%) Pre Post Platinum-sensitive 36 774.85 (231.6, 1907.70) 481.85(189.95, 1056.50) 89.13 (54.60, 97.42) 97.12(85.48, 98.86) 93.49 (64.63, 98.85) Platinum-resistant 22 767.2 (635.50, 3825.5) 511.50(289.68, 903.13) 93.90 (89.30, 98.91) 96.76(89.69, 98.76) 95.93 (89.16, 98.71) U value 317.00 347.50 9.00 227.00 318.00 P value 0.109 0.259 0.107 0.555 0.112 Data are presented as median (P25, P75) Comparison of preoperative serum CA125, HE4, ROMA, and CPH-I levels between platinum-sensitive and platinum-resistant patients with advanced EOC 774.85 (231.6, 1907.70) 89.13 (54.60, 97.42) 93.49 (64.63, 98.85) 767.2 (635.50, 3825.5) 93.90 (89.30, 98.91) 95.93 (89.16, 98.71) Data are presented as median (P25, P75) During the follow-up period, disease recurrence was observed in a subset of patients with advanced-stage epithelial ovarian cancer. Comparisons of preoperative serum CA125, HE4, ROMA, and CPH-I levels between patients with and without recurrence are presented in Table  5 . Patients who experienced recurrence exhibited significantly higher preoperative CA125 and HE4 levels than those without recurrence (both P  < 0.05). Similarly, ROMA values in both premenopausal and postmenopausal patients, as well as CPH-I values, were significantly elevated in the recurrence group (all P  < 0.05). Table 5 Comparison of preoperative serum CA125, HE4, ROMA, and CPH-I levels between patients with and without recurrence in advanced EOC n CA125 (U/mL) HE4 (pmol/L) ROMA (%) CPH-I (%) Pre Post Recurrence 58 767.20 (325.63, 2145.75) 505.80 (215.08, 960.50) 91.36 (75.00, 97.42) 96.93 (85.51, 98.82) 95.01 (69.76, 98.76) Non-recurrence 39 488.70 (158.10, 1094.00) 245.30 (158.10, 478.70) 74.79 (22.76, 91.14) 92.56 (82.25, 96.28) 78.37 (50.11, 92.47) U value 833.00 744.50 50.00 379 701.00 P value 0.028 0.004 0.047 0.025 0.002 Data are presented as median (P25, P75) Comparison of preoperative serum CA125, HE4, ROMA, and CPH-I levels between patients with and without recurrence in advanced EOC 767.20 (325.63, 2145.75) 505.80 (215.08, 960.50) 91.36 (75.00, 97.42) 96.93 (85.51, 98.82) 95.01 (69.76, 98.76) 488.70 (158.10, 1094.00) 245.30 (158.10, 478.70) 74.79 (22.76, 91.14) 92.56 (82.25, 96.28) 78.37 (50.11, 92.47) Data are presented as median (P25, P75) Multivariable logistic regression analyses showed that all four preoperative biomarkers, including CA125, HE4, ROMA, and CPH-I, were positively associated with suboptimal cytoreduction, platinum resistance, and recurrence in patients with advanced EOC (Fig.  1 ). Among these biomarkers, ROMA and CPH-I showed relatively stronger associations, particularly for suboptimal cytoreduction, while CA125 and HE4 demonstrated more moderate but consistently positive associations across all outcomes. Fig. 1 Forest plot of multivariable logistic regression analyses for the associations of CA125, HE4, ROMA, and CPH-I with suboptimal cytoreduction, platinum resistance, and recurrence in advanced EOC Forest plot of multivariable logistic regression analyses for the associations of CA125, HE4, ROMA, and CPH-I with suboptimal cytoreduction, platinum resistance, and recurrence in advanced EOC ROC curve analysis was performed to evaluate the discriminatory ability of preoperative serum CA125, HE4, ROMA, and CPH-I for disease recurrence in patients with advanced EOC (Table  6 ; Fig.  2 ). All four biomarkers demonstrated statistically significant predictive performance, with AUCs significantly greater than 0.5 (all P  < 0.05). The optimal cutoff values for each biomarker were determined using the maximum Youden index. Based on these cutoff values, patients were stratified into high-level and low-level groups for subsequent exploratory survival analyses. Among the four biomarkers, ROMA and CPH-I exhibited relatively higher AUC values, indicating superior discriminatory performance compared with CA125 and HE4, although all markers showed only moderate overall accuracy. Table 6 Parameters at the optimal cutoff values determined by the maximum Youden index Biomarker AUC P Cutoff Sensitivity, % Specificity,% Accuracy, % Suboptimal cytoreductive surgery  CA125 0.770 < 0.001 1456.00 51.85 87.14 76.29  HE4 0.709 < 0.001 437.9 74.07 70.00 70.10  ROMA 0.807 < 0.001 93.97 88.89 68.57 73.20  CPH-I 0.800 < 0.001 91.44 85.19 67.14 71.13 Platinum resistance  CA125 0.666 0.005 665.00 76.67 56.72 61.86  HE4 0.636 0.032 503.3 56.67 70.15 63.92  ROMA 0.677 0.003 94.56 70.00 64.18 64.95  CPH-I 0.692 0.001 85.06 83.33 56.72 63.92 Advanced EOC with recurrence  CA125 0.632 0.020 665 63.79 61.54 61.86  HE4 0.671 0.002 605.9 44.83 89.74 61.86  ROMA 0.691 < 0.001 93.97 62.07 74.36 65.98  CPH-I 0.690 < 0.001 88.18 65.52 71.79 67.01 P values indicate the statistical significance of the ROC curve analysis compared with the null hypothesis of AUC = 0.5. Parameters at the optimal cutoff values determined by the maximum Youden index P values indicate the statistical significance of the ROC curve analysis compared with the null hypothesis of AUC = 0.5. Fig. 2 Receiver operating characteristic curves of serum biomarkers in advanced epithelial ovarian cancer. A ROC curves of CA125, HE4, ROMA, and CPH-I for predicting platinum resistance. B ROC curves of CA125, HE4, ROMA, and CPH-I for predicting disease recurrence. C ROC curves of CA125, HE4, ROMA, and CPH-I for predicting suboptimal cytoreductive surgery Receiver operating characteristic curves of serum biomarkers in advanced epithelial ovarian cancer. A ROC curves of CA125, HE4, ROMA, and CPH-I for predicting platinum resistance. B ROC curves of CA125, HE4, ROMA, and CPH-I for predicting disease recurrence. C ROC curves of CA125, HE4, ROMA, and CPH-I for predicting suboptimal cytoreductive surgery Kaplan–Meier survival analysis was performed to evaluate PFS in patients with advanced EOC stratified by preoperative serum CA125, HE4, ROMA, and CPH-I levels using the optimal cutoff values derived from ROC analysis. Patients in the high-level groups of CA125, HE4, ROMA, and CPH-I exhibited significantly shorter PFS compared with those in the corresponding low-level groups (Fig.  3 ). Log-rank tests demonstrated statistically significant differences in PFS between the high- and low-level groups for all four biomarkers (all P  < 0.05). These findings indicate that elevated preoperative serum CA125, HE4, ROMA, and CPH-I levels are associated with an increased risk of disease progression in patients with advanced EOC. Fig. 3 Kaplan–Meier analysis of progression-free survival according to serum biomarker levels in advanced epithelial ovarian cancer. A Progression-free survival stratified by preoperative serum CA125 levels. B Progression-free survival stratified by HE4. C Progression-free survival stratified by preoperative ROMA. D Progression-free survival stratified by CPH-I Kaplan–Meier analysis of progression-free survival according to serum biomarker levels in advanced epithelial ovarian cancer. A Progression-free survival stratified by preoperative serum CA125 levels. B Progression-free survival stratified by HE4. C Progression-free survival stratified by preoperative ROMA. D Progression-free survival stratified by CPH-I Univariate Cox regression analysis demonstrated that elevated levels of all four biomarkers were significantly associated with shorter PFS in patients with advanced-stage EOC (Fig.  4 a). After adjustment for age, FIGO stage, histological subtype, menopausal status, cytoreductive surgery outcome, smoking status, alcohol consumption, hypertension, and diabetes mellitus, these associations remained statistically significant in multivariable Cox regression analyses. CA125 (HR = 1.22, 95% CI: 1.04–1.43, P  = 0.017), HE4 (HR = 1.30, 95% CI: 1.05–1.61, P  = 0.014), ROMA (HR = 1.02, 95% CI: 1.00–1.04, P  = 0.033), and CPH-I (HR = 1.01, 95% CI: 1.00–1.03, P  = 0.018) remained independently associated with poorer PFS (Fig.  4 b). Fig. 4 Univariate and multivariable Cox regression analyses of CA125, HE4, ROMA, and CPH-I for progression-free survival in patients with advanced-stage epithelial ovarian cancer. A Univariate Cox regression analyses evaluating the associations between preoperative biomarker levels and PFS. B Multivariable Cox regression analyses Univariate and multivariable Cox regression analyses of CA125, HE4, ROMA, and CPH-I for progression-free survival in patients with advanced-stage epithelial ovarian cancer. A Univariate Cox regression analyses evaluating the associations between preoperative biomarker levels and PFS. B Multivariable Cox regression analyses

This study was designed as a retrospective cohort study. Female patients aged 18–80 years who were initially diagnosed and treated at Tengzhou Central People’s Hospital between January 2018 and June 2024 were eligible for inclusion. Patients with incomplete clinical information or follow-up records were excluded from the retrospective analysis. Patients were included if they met the following criteria: (1) newly diagnosed ovarian tumors confirmed by postoperative pathological examination; (2) histopathological diagnosis of EOC or benign ovarian tumors; (3) no prior surgery, chemotherapy, radiotherapy, targeted therapy, or other anticancer treatment before admission; (4) availability of preoperative serum CA125 and HE4 measurements before initiation of treatment; and (5) availability of complete clinical, pathological, treatment, and follow-up data. For survival analyses, patients with advanced-stage EOC were additionally required to achieve complete remission after first-line treatment before entering follow-up evaluation. Exclusion criteria were as follows: (1) incomplete chemotherapy or loss to follow-up; (2) death from causes unrelated to ovarian cancer; (3) coexistence of other malignancies or endocrine disorders; and (4) special conditions such as pregnancy, menstruation at the time of blood sampling, or a history of severe psychiatric disorders. Based on pathological diagnosis and FIGO stage, patients were categorized into three groups: benign ovarian tumor group ( n  = 80), early-stage EOC group ( n  = 84), and advanced-stage EOC group ( n  = 97). This study was approved by the Ethics Committee of Tengzhou Central People’s Hospital (Approval No. 2023-15). Written informed consent was obtained from all participants or their legal guardians. Peripheral venous blood samples (3.0 mL) were collected from all patients after overnight fasting before initial treatment or surgery. Blood samples were allowed to clot at room temperature and then centrifuged at 3500 rpm for 15 min. Serum was separated and immediately used for laboratory analyses. Serum CA125 (U/mL) and HE4 (pmol/L) levels were measured using an automated electrochemiluminescence immunoassay system (cobas 8000 e602; Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. The ROMA and the CPH-I were calculated based on serum CA125 and HE4 levels and patient characteristics. For ROMA calculation, predictive index (PI) formulas were applied as follows. For premenopausal women: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$PI\:=\:-12.0\:+\:2.38\:\times\:\mathrm{ln}\left(HE4\right)+\:0.0626\:\times\:\mathrm{ln}\left(CA125\right)$$\end{document} For postmenopausal women: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:PI\:=\:-8.09\:+\:1.04\:\times\:\mathrm{ln}\left(HE4\right)+\:0.732\:\times\:\mathrm{ln}\left(CA125\right)$$\end{document} ROMA (%) was calculated using the following formula: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:ROMA=\:\frac{\mathrm{exp}\left(PI\right)}{1\:+\mathrm{exp}\left(PI\right)}\times\:\:100$$\end{document} Menopausal status was determined according to internationally accepted clinical criteria. Natural menopause was defined as spontaneous cessation of menstruation for at least 12 consecutive months in the absence of physiological, pathological, or treatment-related causes. The date of the final menstrual period was retrospectively confirmed after 12 months of amenorrhea. Patients who did not meet these criteria were classified as premenopausal. The CPH-I score was calculated as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:CPH-I\:=-14.0647\:+\:1.0649\:\times\:{log}_{2}\left(HE4\right)+\:0.6050\:\times\:{log}_{2}\left(CA125\right)+\:0.2672\:\times\:\:\frac{age}{10}$$\end{document} Predicted probability (PP) was calculated as: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:PP\:=\:\:\frac{\mathrm{e}\mathrm{x}\mathrm{p}(CPH-I)}{1\:+\:\mathrm{e}\mathrm{x}\mathrm{p}(CPH-I)}$$\end{document} All patients with advanced-stage EOC underwent primary cytoreductive surgery, followed by platinum-based chemotherapy administered intravenously or via combined intravenous and intraperitoneal routes. Chemotherapy was delivered for 6–8 cycles at intervals of 21–28 days. The main chemotherapy regimens included paclitaxel plus carboplatin (TC), paclitaxel plus cisplatin (TP), docetaxel plus carboplatin (DC), and cisplatin plus cyclophosphamide (PC). Advanced-stage EOC patients were further classified according to surgical outcomes, platinum sensitivity, and recurrence status. Optimal cytoreduction was defined as a maximum postoperative residual tumor diameter ≤ 1 cm, while suboptimal cytoreduction was defined as residual disease > 1 cm. Platinum sensitivity was defined as no disease progression within 6 months after completion of the last platinum-based chemotherapy cycle, whereas platinum resistance was defined as disease progression within 6 months. Patients with advanced-stage EOC routinely underwent postoperative clinical follow-up according to institutional practice. Follow-up information was retrospectively obtained from hospital electronic medical records, outpatient records, imaging reports, and documented telephone follow-up records when available. The follow-up period extended until June 2025, with a median follow-up duration of 11 months (range: 1–48 months) and a follow-up rate of 92%. All patients included in the survival analyses achieved complete remission after first-line treatment, which consisted of primary cytoreductive surgery followed by platinum-based adjuvant chemotherapy. Complete remission was defined as disappearance of all detectable lesions on imaging, absence of new lesions, normalization of CA125 and HE4 levels, and no clinical evidence of active disease. During follow-up, patients classified as non-recurrent maintained sustained remission, characterized by persistently normal tumor marker levels, absence of new or enlarging lesions on imaging, and no clinical evidence of recurrence. Disease progression or recurrence was primarily determined based on radiological evidence of new or progressive lesions together with clinical assessment by the treating physicians. Elevation of CA125 and/or HE4 alone was not considered sufficient for diagnosis, but was used only as supportive evidence according to GCIG-related criteria. PFS was defined as the interval from the date of initial hospitalization for primary diagnosis to the first documented disease progression, recurrence, or last follow-up in patients without progression or recurrence. Patients without documented progression or recurrence were censored at the date of the last available follow-up. Separate models were constructed for suboptimal cytoreduction, platinum resistance, and disease progression/recurrence. CA125 and HE4 levels were log2-transformed before analysis because of their skewed distributions. To reduce potential multicollinearity arising from the mathematical relationships among CA125, HE4, ROMA, and CPH-I, each biomarker was entered into a separate multivariable model. The models were adjusted for available clinicopathological variables, including age, menopausal status, FIGO stage, histological subtype, smoking status, alcohol consumption, hypertension, and diabetes mellitus. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to estimate the associations between biomarkers and clinical outcomes. Univariate and multivariable Cox proportional hazards regression analyses were performed to evaluate the associations between serum biomarkers and PFS in patients with advanced-stage EOC. Multivariable models were adjusted for age, FIGO stage, histological subtype, menopausal status, cytoreductive surgery outcome, smoking status, alcohol consumption, hypertension, and diabetes mellitus. To reduce potential multicollinearity, variance inflation factor (VIF) values were assessed before model construction, and variables with VIF < 5 were considered acceptable for inclusion in the multivariable models. The proportional hazards assumption was evaluated using Schoenfeld residual tests, and no significant violations were observed. Hazard ratios (HRs) and 95% CIs were calculated. Statistical analyses were performed using SPSS software (version 25.0; IBM Corp., Armonk, NY, USA). Continuous variables were assessed for normality. Variables with non-normal distribution were expressed as median (interquartile range) and compared using the Kruskal–Wallis test or Mann–Whitney U test, as appropriate. Normally distributed continuous variables were expressed as mean ± standard deviation and compared using analysis of variance (ANOVA). Categorical variables were expressed as frequencies and percentages and compared using the chi-square test. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the discriminatory ability of CA125, HE4, ROMA, and CPH-I for suboptimal cytoreduction, platinum resistance, and disease recurrence. The optimal cutoff values were determined using the maximum Youden index. P values for ROC analyses were calculated based on comparisons against the null hypothesis of AUC = 0.5. Progression-free survival was analyzed using the Kaplan–Meier method, and median PFS with corresponding 95% confidence intervals was calculated using the Brookmeyer–Crowley method. Survival differences between groups were assessed using the log-rank test. All statistical tests were two-sided, and a P value < 0.05 was considered statistically significant.

In this retrospective cohort study, elevated preoperative levels of CA125, HE4, ROMA, and CPH-I were associated with unfavorable clinical outcomes in patients with advanced epithelial ovarian cancer, including suboptimal cytoreductive surgery, disease progression or recurrence, and shorter progression-free survival. Compared with CA125 and HE4 alone, ROMA and CPH-I showed relatively improved performance in predicting surgical outcomes, although their predictive abilities for platinum resistance and recurrence remained moderate. These findings suggest that routinely available serum biomarkers may provide additional prognostic information for preoperative risk stratification and postoperative surveillance in advanced ovarian cancer. However, given the retrospective single-center design and the moderate discriminative performance observed, the findings of the present study should be interpreted cautiously. Further large-scale prospective studies are needed to validate these findings and clarify the potential clinical utility of these biomarkers in advanced ovarian cancer.

In the present retrospective cohort study, we evaluated the prognostic value of preoperative CA125, HE4, ROMA, and CPH-I in patients with epithelial ovarian cancer, particularly those with advanced-stage disease. Our findings showed that elevated baseline levels of these biomarkers were associated with suboptimal cytoreductive surgery, recurrence risk, and shorter progression-free survival. In multivariable analyses, CA125, HE4, ROMA, and CPH-I remained independently associated with unfavorable PFS after adjustment for clinicopathological confounders. Although the discriminative performance observed in ROC analyses was moderate, these findings suggest that routinely available serum biomarkers may provide additional prognostic information for risk stratification and postoperative surveillance in advanced-stage EOC. CA125 and HE4 are widely used tumor markers in the clinical management of ovarian cancer and play an important role in diagnosis [ 8 ]. In addition, both biomarkers have been reported to possess prognostic value [ 9 , 10 ], with elevated preoperative levels often indicating poor outcomes. Optimal cytoreductive surgery and sensitivity to platinum-based chemotherapy are key prognostic factors in advanced ovarian cancer [ 11 ], and Chudecka-Głaz et al. demonstrated that preoperative CA125 and HE4 levels were associated with surgical outcomes and platinum sensitivity [ 12 ]. Matsuhashi et al. further reported that patients with lower CA125 levels ( 100 U/mL) [ 13 ]. Consistent with previous studies, our results showed significantly higher preoperative CA125 levels in patients who underwent suboptimal cytoreductive surgery compared with those who achieved optimal cytoreduction. In ROC analyses, CA125 demonstrated relatively high specificity for predicting suboptimal cytoreduction, suggesting that elevated CA125 levels may reflect increased tumor burden and more extensive disease involvement in advanced ovarian cancer. Previous studies have also reported that failure of postoperative CA125 normalization is associated with an increased risk of recurrence in EOC [ 14 , 15 ]. Moreover, previous studies observed that biochemical recurrence preceded clinical recurrence by approximately 4.8 months [ 16 ], highlighting the importance of monitoring peri-treatment fluctuations in serum CA125 for disease surveillance. Nevertheless, CA125 has well-recognized limitations, including suboptimal sensitivity and specificity for early detection and uncertain predictive value for platinum resistance [ 17 ]. In the present cohort, although CA125 levels differed significantly between platinum-sensitive and platinum-resistant recurrence groups, its discriminative performance for platinum resistance remained moderate (AUC = 0.666) and was not significantly superior to that of HE4, ROMA, or CPH-I. These findings suggest that CA125 alone may be insufficient for precise prognostic stratification and may require combination with other clinical or molecular indicators. HE4 has been shown to exhibit higher specificity for early ovarian cancer diagnosis and has also been identified as an independent factor associated with disease recurrence [ 18 ]. As a secretory protein, HE4 may directly participate in tumor progression and metastasis by regulating cell adhesion, proliferation, and invasion [ 19 , 20 ]. Previous studies have reported that a decline in HE4 levels after the third chemotherapy cycle and CA125 regression after the first cycle were strongly correlated with progression-free survival, and suggested that HE4 outperformed CA125 in predicting platinum sensitivity [ 21 , 22 ]. In the present study, both CA125 and HE4 levels were significantly higher in patients with disease progression or recurrence compared with those who remained progression-free during follow-up. Notably, HE4 demonstrated relatively higher specificity for recurrence prediction, suggesting a potential advantage in identifying patients at increased risk of unfavorable outcomes. In addition, Kaplan–Meier analyses stratified by ROMA levels showed earlier disease progression in the high-ROMA group compared with HE4-based stratification, with progression events occurring approximately one month earlier. These findings suggest that HE4- and ROMA-based assessments may provide additional prognostic information in advanced ovarian cancer, although their clinical utility requires further validation in larger prospective studies. Although CA125 and HE4 are widely used in gynecologic oncology, CA125 levels can be elevated in various physiological or pathological conditions, including menstruation, pregnancy, endometriosis, and inflammatory diseases of the peritoneum [ 23 ]. Furthermore, several studies have questioned the diagnostic performance of CA125 for early ovarian cancer detection [ 24 ]. To address these limitations, composite indices such as ROMA and CPH-I were developed by integrating CA125 and HE4 with age or menopausal status to improve risk stratification in ovarian cancer screening. Subsequent studies confirmed their value in early diagnosis, and ROMA has been widely adopted in clinical practice; however, evidence regarding their prognostic utility and ability to predict recurrence remains limited. In this single-center retrospective study, we aimed to clarify the prognostic performance of ROMA and CPH-I in advanced ovarian cancer and to compare their clinical utility with that of traditional biomarkers. Our results demonstrated that ROMA and CPH-I provided superior predictive value for cytoreductive surgery outcomes compared with CA125 and HE4, while their performance in predicting platinum resistance and recurrence was comparable to that of single biomarkers. These observations may partially reflect differences in tumor burden and tumor biology. CA125 primarily reflects tumor burden and peritoneal inflammatory response, whereas HE4 is more closely related to tumor invasiveness. ROMA integrates patient physiological status to distinguish benign from malignant disease, while CPH-I emphasizes the interaction between tumor aggressiveness and the host peritoneal environment [ 25 ], which is a critical determinant of surgical feasibility and prognosis in advanced disease. However, given the retrospective design and moderate discriminative performance observed in the present study, these findings should be interpreted cautiously and require further validation in larger prospective cohorts. Failure to achieve optimal cytoreduction is associated with poor clinical outcomes in advanced ovarian cancer. In this context, preoperative noninvasive assessment using ROMA and CPH-I may provide additional information for identifying patients at higher risk of suboptimal cytoreduction. These patients may be more suitable candidates for neoadjuvant chemotherapy. Therefore, ROMA- and CPH-I-based assessments may provide additional value for preoperative risk stratification and treatment planning, rather than functioning only as passive prognostic indicators. Several limitations of this study should be acknowledged. First, this was a single-center retrospective cohort study with a relatively limited sample size, which may introduce selection bias, reduce statistical power, and limit the generalizability of the findings. Second, although multivariable regression analyses were performed to adjust for available clinicopathological factors, some potentially important confounders, including performance status, surgical complexity, imaging-based tumor burden, homologous recombination deficiency status, and circulating tumor DNA, were not consistently available and therefore could not be incorporated into the models. Third, the cutoff values for CA125, HE4, ROMA, and CPH-I were determined using the maximum Youden index within the same dataset, without external validation. Given the limited sample size and multiple outcomes analyzed, these data-driven thresholds may be subject to overfitting and should be interpreted as exploratory rather than clinically established cutoff values. Fourth, conventional ROC analyses were applied to outcomes with time-dependent features, such as disease progression/recurrence and platinum resistance, in a cohort with heterogeneous follow-up durations. Therefore, the discriminatory performance may not fully account for censoring or time-to-event information and should be interpreted cautiously. Fifth, although disease progression or recurrence was primarily determined based on radiological and clinical evidence, CA125 and HE4 elevations were used as supportive information in outcome assessment. Therefore, potential circular reasoning bias cannot be completely excluded, particularly for analyses involving CA125, HE4, ROMA, and CPH-I. Finally, the follow-up duration may be insufficient to fully evaluate long-term outcomes, including late recurrence and overall survival. Future large-scale, prospective, multicenter studies with standardized follow-up, independent validation cohorts, time-dependent predictive analyses, and integration of imaging and molecular biomarkers are needed to confirm these findings.

Epithelial ovarian cancer (EOC) accounts for approximately 90% of ovarian malignancies and remains one of the leading causes of gynecological cancer–related mortality worldwide [ 1 , 2 ]. Due to the lack of specific early symptoms and effective screening strategies, most patients are diagnosed at an advanced stage, resulting in high recurrence rates and poor long-term survival despite advances in surgical and systemic therapies. Primary cytoreductive surgery followed by platinum-based chemotherapy is the standard first-line treatment for advanced EOC [ 3 ]. Achieving optimal cytoreduction and maintaining platinum sensitivity are the most important determinants of patient prognosis. Patients with suboptimal debulking or early platinum resistance experience significantly worse outcomes. However, the ability to accurately identify, before surgery, patients who are unlikely to achieve optimal cytoreduction or who are at high risk of platinum resistance remains limited, highlighting the need for reliable preoperative risk stratification tools. Serum cancer antigen 125 (CA125) and human epididymis protein 4 (HE4) are widely used biomarkers in the management of ovarian cancer and have demonstrated value in disease monitoring and prognostic assessment [ 4 ]. To improve the limitations of single biomarkers, composite models such as the Risk of Ovarian Malignancy Algorithm (ROMA) and the Copenhagen Index (CPH-I), which integrate CA125, HE4, and patient characteristics, have been developed and shown improved diagnostic performance for distinguishing malignant from benign ovarian masses [ 5 – 7 ]. Nevertheless, evidence regarding the clinical utility of these biomarkers and models in predicting surgical outcomes, platinum resistance, and prognosis in advanced EOC remains limited and inconsistent. Therefore, this study aimed to systematically evaluate serum CA125, HE4, ROMA, and CPH-I in patients with epithelial ovarian cancer, with a particular focus on advanced-stage disease. We investigated their predictive value for cytoreductive surgery outcomes, platinum resistance, and disease recurrence, as well as their association with progression-free survival. Our findings may provide clinically relevant evidence to support preoperative risk assessment and individualized treatment strategies in advanced EOC.