1. Introduction

Advanced-stage epithelial ovarian carcinoma (EOC) is a primarily chemosensitive solid tumor that initially responds to first-line chemotherapy in 70–80% of cases. However, recurrence occurs in 62% of cases across all stages and in 80–85% of advanced-stage cases, with 75% of advanced-stage patients being incurable [1, 2, 3].

Disease recurrence in ovarian cancer is influenced by multiple factors, including the extent of disease at initial diagnosis, the amount of residual tumor after initial surgery, the rate of decline in CA-125 levels following treatment, and the response to initial therapy [4]. The recurrence pattern varies, ranging from isolated nodal disease to isolated peritoneal disease and peritoneal carcinomatosis. Most patients with recurrent ovarian cancer are managed with chemotherapy alone.

Over the past 30 years, the platinum-free interval (PFI) has played a crucial role in treatment selection and prognosis [5]. After achieving a complete clinical response to platinum-based therapy, disease recurrence within six months is classified as platinum-resistant disease, whereas recurrence beyond six months is considered platinum-sensitive. The response to second-line chemotherapy differs significantly between these groups.

The role of surgery in recurrent ovarian Cancer remains a subject of debate. The Descriptive Evaluation of Preoperative Selection Criteria for Operability in Recurrent Ovarian Cancer (DESKTOP) III study demonstrated that adding secondary cytoreductive surgery (SCS) to platinum-based chemotherapy in patients with recurrent ovarian cancer provided an overall survival benefit. This randomized controlled study selected platinum-sensitive (PFI >6 months) patients based on the AGO score system ((i) complete resection at first surgery, (ii) good performance status, (iii) <500 cc ascites). Selection patients using the AGO scores resulted in high efficacy and low morbidity. Several scoring systems have been validated to predict complete resection in secondary cytoreductive surgery, including the AGO score, Memorial Sloan Kettering (MSK) criteria, and a study from Korea published in August 2022. These models primarily focus on clinical and radiologic parameters to identify optimal surgical candidates. The AGO score, in particular, has demonstrated a 76% predictive accuracy for complete resection in previous studies. However, emerging models, such as the MSKCC and Korean scoring systems, highlight additional prognostic variables that may further refine patient selection [6, 7].

This study aims to investigate the parameters predictive of complete resection in patients undergoing SCS in our clinic and to evaluate of AGO score compliance on complete resection rates and survival outcomes.

2. Material and methods

This retrospective observational study included 326 patients diagnosed with epithelial ovarian carcinoma (EOC) between 2006 and 2020 at the Gynecologic Oncology Clinic of Etlik Zübeyde Hanım Women’s Diseases Training and Research Hospital. Only patients who had completed primary treatment and had documented first recurrence were included. Primary cytoreductive surgery (PCS) followed by adjuvant chemotherapy or neoadjuvant chemotherapy followed by interval cytoreductive surgery (ICS) were employed as first-line treatment strategies for EOC. After PCS, all patients received standard chemotherapy consisting of taxane and platinum within 2–3 weeks, for a total six cycles. The response to chemotherapy was assessed through gynecologic examination, computed tomography (CT), ultrasonography and CA-125 analysis.

Following initial cytoreductive surgery (PCS or ICS), the size of the residual tumor was recorded and categorized as: maximal (R = 0, no visible residual tumor), optimal (residual tumor ≤1 cm), suboptimal (residual tumor >1 cm).

Clinical recurrence was defined based on elevated serum CA-125 levels combined with radiologic evidence of relapse on CT and/or positron emission tomography-computed tomography (PET-CT) imaging, per our institutional protocol. Among these patients, 42 who underwent SCS for their first recurrence were included in the study. Patients who underwent surgery for other indications, had no documented first recurrence, received hyperthermic intraperitoneal chemotherapy (HIPEC) in the first-line setting, or received only chemotherapy or radiotherapy at relapse were excluded. Fig. 1 shows the flowchart of the study.

Fig. 1.Flowchart of the patients in the study. HIPEC: hyperthermic intraperitoneal chemotherapy.

Patients were divided into two groups based on their SCS outcomes: those who achieved complete resection (maximal cytoreductive surgery, no macroscopic residual tumor) and those with residual tumor (optimal or suboptimal cytoreductive surgery). We retrospectively collected demographic data, intraoperative findings from the initial surgery, surgery complexity scores [8] low (≤3 points), intermediate (4–7 points), high (≥8 points), histological subtype of EOC, CA-125 levels at recurrence, recurrence patterns and location, the number of recurrence sites, chemotherapeutic agents used at recurrence, and residual tumor volume after SCS. We also recorded the dates of initial EOC diagnosis, recurrence diagnosis, last follow-up, and death for survival analysis. Post-relapse survival was calculated from the date of first recurrence to the last follow-up or death, and the platinum-free interval (PFI) defined as the time (months) from the completion of initial treatment to first recurrence. To further evaluate the impact of surgical selection criteria, patients were categorized according to the AGO scoring system (PFI >6 months, (i) complete resection at the first surgery, (ii) good performance status, (iii) <500 cc ascites). The relationship between AGO score compliance, complete resection rates, and survival outcomes was analyzed.

The decision to perform SCS was made independently of the AGO score by the gynecologic oncology board, incorporating clinical expertise, surgical experience, and contemporary best practices.

Statistical analyses were conducted using SPSS 26.0 (Statistical Package for the Social Sciences for Windows, Version 26.0 SPSS Inc., Chicago, IL, USA), with a significance level of 0.05. The Kolmogorov-Smirnov test was used to assess data normality, and continuous variables were expressed as mean ± standard deviation, median (minimum–maximum) or frequency (percentage). Comparisons between categorical variables were performed using Fisher’s exact test or the Chi-square test. Variables were analyzed using the Cox regression model with stepwise variable selection to determine independent prognostic factors for complete resection. Survival analysis was conducted using the Kaplan-Meier method, and survival curves were compared using the log-rank test.

3. Results

Among the 42 patients included, 40 received adjuvant chemotherapy as first-line treatment, and 2 received neoadjuvant chemotherapy. Complete resection (CR) was achieved in 29 out of 42 patients (69%), while 13 patients (31%) had residual tumor following secondary cytoreductive surgery (SCS). There was no significant difference between the CR and residual tumor groups in terms of age, International Federation of Gynecology and Obstetrics (FIGO) stage, histopathology, intraoperative findings during the initial surgery, ascites volume or the platinum-free interval (PFI) (p > 0.05). At the time of initial diagnosis, 21% of patients had diffuse peritoneal carcinomatosis. In addition to hysterectomy and lymph node dissection, all patients underwent omentectomy. Diaphragmatic stripping was performed in 11 patients, large bowel resection in 4, splenectomy in 5, small bowel resection in 1, and Winslow tumor resection in 1. Splenectomy during the initial surgery was the only factor associated with a significantly lower CR rate (20% vs. 80% in patients without splenectomy; p = 0.005). When patients were classified according to surgical complexity scores, 34 underwent intermediate surgery; however, no differences in CR rates were observed across these categories (Table 1). Similarly, when patients were evaluated based on AGO score compliance, no significant differences in CR rates were observed (Table 1).

At the time of recurrence diagnosis, patients with a maximum of three nodules on imaging had a CR rate of 76.7%, which was significantly higher than in those with more than three nodules. Intraoperatively, patients with discrete lesions (up to three nodules) had higher CR rates compared to those with diffuse disease (more than three nodules and/or peritoneal carcinomatosis) (p < 0.05). When patients were grouped according to the number of recurrence sites, the rate of CR was lower in those with recurrence at more than three sites. While 100% CR was achieved in patients with recurrence confined to a single site (regardless of the number of nodules), the CR rate was 50% in those with recurrence at more than three sites (p < 0.05). In patients with a single-nodule recurrence, nodules were most frequently located in the colon, ileum or brain (Table 2).

Following secondary cytoreduction, 54.8% of patients received carboplatin-paclitaxel as second-line treatment. One patient was treated with liposomal doxorubicin alone, and another received bevacizumab in addition to carboplatin-paclitaxel. The remaining patients were administered carboplatin combined with docetaxel or gemcitabine. Although splenectomy, recurrence pattern, and the number of recurrence sites were significant predictors of CR in the univariate analysis, no parameter independently predicted CR in the regression analysis.

Patients who achieved CR had a median overall survivor of 102 months, demonstrating a 38-month survival advantage compared to those with residual tumor (p < 0.05). Survival outcomes were similar between patients receiving carboplatin-paclitaxel and those treated with alternative chemotherapy regimens. Subgroup analyses showed that different chemotherapy protocols did not significantly affect survival in patients who underwent complete resection. Regarding post-relapse survival, patients achieving CR had an 18-month survival benefit (p < 0.05). Fig. 2 shows the post-relapse survival according to the AGO score and SCS results. Moreover, compliance with the AGO score did not influence overall survival or post-relapse survival (Table 3). Postoperative complications occurred in one patient who had previously undergone splenectomy, and 27 patients died during the follow-up period.

Fig. 2.The impact of AGO score and surgical outcomes on post-relapse survival in recurrent ovarian cancer. (A) Post-relapse survival according to the AGO score. (B) Post-relapse survival according to the SCS. AGO: Arbeitsgemeinschaft Gynaekologische Onkologie; SCS: Secondary Cytoreductive Surgery.

4. Discussion

One of the key database in the management of recurrent ovarian cancer is whether surgical intervention provides a survival advantage over chemotherapy alone. In carefully selected patients with a platinum-free interval (PFI) of more than six months who have completed their initial treatment, the addition of secondary cytoreductive surgery (SCS) to systemic therapy has been shown to improve disease remission. Studies also indicate better outcomes of SCS in the neoadjuvant chemotherapy (NACT) group and even in platinum-resistant patients with limited regional recurrence [9, 10]. However, this advantage is limited if complete resection cannot be achieved. A meta-analysis of 80 studies reported that maximal effort in SCS significantly improved overall survival in platinum-sensitive recurrent patients, with an 8.9% and 7.4% increase in median overall survival (OS) for every 10% increase in optimal or maximal resection rates, respectively [11]. Complete resection is thus one of the most important prognostic indicators in recurrent ovarian cancer [12]. Scoring systems designed to identify optimal surgical candidates are therefore critical for balancing surgical morbidity with oncological benefit [13, 14].

Several landmark studies have evaluated the role of surgery in the treatment of recurrent epithelial ovarian cancer. The Gynecologic Oncology Group (GOG)-213 study reported no survival benefit from surgery over chemotherapy alone, whereas the DESKTOP III and Surgery or Chemotherapy in Recurrent Ovarian Cancer (SOC)-1 studies demonstrated a survival advantage in patients selected for surgery based on standardized scoring criteria [6, 15, 16]. The discrepancy in findings between these studies may be attributed to differences in treatment protocols, particularly the use of bevacizumab. In the GOG-213 study, 84% of patients who did not undergo surgery received maintenance bevacizumab, compared to only 23% in the DESKTOP III study and 1% in the SOC-1 study [15]. Since bevacizumab has been shown to improve both progression-free survival (PFS) and OS in relapsed ovarian cancer, its widespread use in the GOG-213 study may have masked any potential benefit of surgery [14, 17]. In contrast, the SOC-1 study, which used the iMODEL to select resectable patients, confirmed that cytoreductive surgery significantly improved PFS when added to chemotherapy [16].

In the DESKTOP III study, patients selected for surgery using the AGO score achieved a CR rate of 76%, closely aligning with the 69% CR rate observed in our study [6]. Although no universally accepted scoring system exists for identifying ideal SCS candidates, the primary objective of these models is to select patients who are likely to achieve complete resection, thereby maximizing survival benefit. Newer models, such as the MSKCC and Korean scoring systems, propose additional prognostic factors to further refine patient selection [7]. In the Korean study, an 87% complete resection rate was reported in patients with a prolonged progression-free interval and limited regional carcinomatosis. Similarly, the MSKCC group selected patients based on a disease-free interval (DFI) >6 months and single-site recurrence, demonstrating the importance of recurrence characteristics in achieving complete resection [7]. Evidence suggests that even platinum-resistant patients with limited regional recurrence may benefit from SCS [10]. Our findings align with these results, indicating that recurrence characteristics, tumor burden at the time of recurrence, and the number of recurrence sites are crucial factors in patient selection. We found that SCS results are limited in cases of peritoneal carcinomatosis and recurrence in more than three sites, while single-site or single-nodule recurrences significantly increase CR rates.

The key challenge in SCS is determining the appropriate surgical candidates. An artificial intelligence model identified DFI and retroperitoneal recurrence as key predictors of CR and overall survival [18]. Our institution is a high-volume center with an overall cytoreduction rate exceeding 75%, and we select candidates for SCS through a multidisciplinary gynecologic oncology board, independent of the AGO score. Consistent with previous literature, our results indicate a survival advantage in patients who achieve complete resection [19, 20, 21]. A meta-analysis by Faiza Gaba et al. [22] also confirmed that aggressive cytoreductive surgery can improve survival when complete resection is achieved. However, when our surgical cohort was stratified by AGO score, no significant differences were observed in CR rate or survival outcomes, suggesting that AGO scoring alone may be insufficient. Instead, recurrence pattern and the number of recurrence sites appear to be critical factors in identifying candidates for complete resection. In line with the findings of Rong-yu Zang et al. [23], patients with solitary tumor regions at recurrence were more likely to achieve optimal surgical outcomes.

A 2024 meta-analysis on splenectomy found no effect on overall survival or mortality, indicating that splenectomy is an acceptable and safe procedure to achieve optimal cytoreduction [24]. Rather than splenectomy alone, the presence of extensive upper abdominal disease—necessitating resection of the diaphragm, liver capsule, hepatorenal pouch or splenic hilum to achieve complete cytoreduction (CC-0) at primary surgery—appears to be a more reliable predictor of surgical outcomes in recurrent disease. In our study, patients requiring splenectomy due to widespread disease involvement had lower complete resection rates at secondary cytoreduction, suggesting that tumor burden and anatomical distribution play a more critical role in predicting resectability. Our study also highlights the importance of tumor biology in treatment selection. Patients who had undergone splenectomy during initial surgery due to parenchymal involvement had lower CR rates, whereas those with recurrence as discrete lesions had higher CR rates. This suggests that the biological behavior of the tumor should be carefully considered when determining treatment strategies. In line with our findings, Satyam et al. [25] reported that progression-free survival was significantly lower in patients with multiple metastases who required bowel resection for recurrence. Additionally, studies in patients with isolated lymph node recurrence have demonstrated that SCS can be successful in this setting and is associated with improved long-term survival outcomes [26, 27]. These findings underscore the need for further genomic studies to better understand tumor biology and its impact on surgical outcomes in recurrent ovarian cancer.

Another meta-analysis revealed that SCS, when performed by experienced teams, does not negatively affect quality of life, emphasizing the importance of high-volume centers and adequate surgical expertise [28, 29]. Standardized, mentor-supported training modules for SCS should be developed, as it is a challenging procedure. Our clinic is a high-volume center with a high cytoreduction rate, which likely contributes to our low recurrence rate and CR rates consistent with those reported in the literature.

The retrospective design is the primary limitation of our study, the small and bit old sample size is another constraint, although it is partly attributable to our high cytoreduction rates and low recurrence rates. Additionally, the long study period may introduce bias. Our retrospective analysis of AGO score compliance revealed no significant differences in CR rates, suggesting that clinical and surgical expertise and a multidisciplinary approach play pivotal roles in patient selection. We also propose that recurrence pattern and the number of recurrence sites—factors not included in standard scoring models—should be integrated into patient selection criteria for SCS. Although there is a risk of selection bias, as surgical outcomes depend on appropriate patient selection, our sample size is more than 10 times larger than the predictive number in a similar study [7]. Future prospective studies with larger cohorts are needed to refine selection parameters, clarify the role of recurrence patterns in predicting CR, and optimize outcomes for patients undergoing secondary cytoreductive surgery.

5. Conclusions

Complete resection remains the most significant predictor of improved survival in patients undergoing secondary cytoreductive surgery for recurrent ovarian cancer. While the AGO score serves as a valuable tool for preoperative assessment, our findings highlight the paramount importance of the recurrence pattern—particularly the number and anatomical location of recurrent lesions—in determining the likelihood of achieving optimal cytoreduction. Patients presenting with single or limited regional recurrence demonstrated significantly higher complete resection rates compared to those with peritoneal carcinomatosis or multifocal disease. These observations underscore the need to incorporate detailed recurrence characteristics into current selection models to enhance the identification of suitable surgical candidates.

A multidisciplinary approach within high-volume, experienced centers is critical for accurate patient selection and effective surgical planning. Future prospective studies should aim to validate integrated scoring systems that combine clinical, anatomical, and biological variables—including recurrence patterns—to optimize outcomes in the management of recurrent ovarian cancer.

Availability of data and materials

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author contributions

BŞ—performed the statistical analysis, and manuscript writing/editing and participated in the study design, coordination, protocol/project development. TBG—performed the manuscript editing. FK—participated in the study design, coordination, protocol/project development, and manuscript writing/editing.

Ethics approval and consent to participate

Ethics approval: The institutional review board obtained approval from AnkaraEtlik Zübeyde Hanım Women’s Health Training and Research Hospital on 24 October 2022, with the number 14. Informed consent: A verbal and written informed consent was obtained from all participants. Consent for publication: Data are openly available in a public repository that issues datasets with DOIs.

Acknowledgment

The authors would like to thank everyone who voluntarily dedicated their time and effort.

Funding

This research received no external funding.

Conflict of interest

The authors declare no conflict of interest.