Validation of ESGO Quality Indicators in Advanced Ovarian Cancer Surgery

Abstract Background: Multiple studies have confirmed that centralization improves outcomes in women with advanced stage ovarian carcinoma (OC), in part due to complex treatment algorithms including extensive cytoreductive surgery. Therefore, the European Society of Gynaecological Oncology (ESGO) has established and published ten quality indicators for centers performing advanced ovarian cancer surgery. Based on these quality indicators, ESGO runs institutional accreditation processes for advanced ovarian cancer surgery in two categories, i.e. regular accreditation for advanced ovarian cancer surgery and centers of excellence for advanced ovarian cancer surgery. ESGO also runs institutional accreditation processes for endometrial cancer surgery and for training centers in gynaecological oncology. Of note, no large international multicenter study has validated the ESGO quality indicators of advanced ovarian cancer surgery as of yet. Study design/objectives: The present study is set out to assess the prognostic impact of the established ESGO quality indicators for advanced ovarian cancer surgery in a large international, multicenter study comprising ESGO accredited centers. All ESGO accredited centers will be invited to participate in the present study. Participating centers will provide clinical data on the 10 ESGO quality indicators in ovarian cancer surgery and follow- up data on progression-free (PFS) and overall survival (OS) for the past three years, i.e. 2020, 2021, and 2022. To avoid selection bias, the following two patient cohorts will be included in the present study 1) ALL consecutively incoming patients at the center with newly diagnosed advanced (stages III and IV) ovarian cancer (importantly, irrespective of whether they underwent surgery or not during their primary therapy), and 2) ALL recurrent OC patients who underwent secondary of tertiary cytoreductive surgery at the center (importantly, surgical cases only). The primary objective of this study is to evaluate the prognostic impact of the ten ESGO quality indicators in advanced ovarian cancer surgery on PFS. Secondary objectives will comprise (not exclusively) association between the ten ESGO quality indicators and OS, and prognostic implication of center's accreditation status. Conclusions: The present study will evaluate the ten published ESGO quality indicators with respect to PFS and OS to assess their prognostic implication. Background Overall, ovarian cancer (OC) is the 5th most frequent cause of cancer-related death in women in western industrialized countries. It is however the leading cause of mortality among gynecological malignancies. Referring to histological classification, 90% of ovarian tumors are epithelial ovarian cancers (EOC), of which high-grade serous carcinoma (HGSOC) is by far the most common subgroup. Due to the lack of sufficient diagnostic tools for early detection, the vast majority of patients are diagnosed at an advanced stage of disease (stage III or IV). The cornerstone of primary treatment is cytoreductive surgery followed by adjuvant carboplatin and paclitaxel combination chemotherapy and subsequent maintenance therapy. Complete macroscopic resection and response to chemotherapy seem to be the most important prognostic factors for these patients. Multiple studies have confirmed that centralization improves outcomes for women with advanced OC due to the complex treatment algorithms, including extensive radical cytoreductive surgery. Among many different factors, the benefit of centralization for these patients seems to be related to high-quality infrastructure as well as high levels of expertise and high surgical volume per center and per surgeon. This is reflected by publications repeatedly reporting that high-volume hospitals as well as high-volume surgeons are associated with improved outcomes for patients with OC. Interestingly, there is no established definition for a "center of excellence" and/or number of cases for a high-volume surgeon and high- volume hospital. This definition varies from study to study. In an attempt to implement quantifiable measures, ESGO has established ten quality indicators for centers performing advanced OC surgery based on available literature and expert opinion. Based on these quality indicators, ESGO runs institutional accreditation processes for advanced OC sugery in two categories, i.e. regular accreditation for advanced OC surgery and excellence centers for advanced OC surgery. The ten quality indicators have been assessed in several national studies. Of note, no large international multicenter study has assessed these quality indicators so far. The present study is set out to assess the ten published ESGO quality indicators for advanced OC surgery in a large international study based on ESGO accredited centers. Study objectives The main aim of the present study is to assess the prognostic impact of ten published ESGO quality indicators for patients with advanced OC undergoing cytoreductive surgery during primary therapy. Primary objective To evaluate the association between the ten ESGO quality indicators for advanced OC surgery and progression-free survival. Secondary objectives To assess the association between the ten ESGO quality indicators for advanced OC surgery and overall survival. To assess the percentage of centers adhering to all ten ESGO quality indicators, and to each individual quality indicator. To assess the association between adherence to various ESGO quality indicators and progression-free survival and overall survival. To assess the association of accreditation status (training center vs endometrial cancer surgery vs advanced OC surgery in 2 categories (regular accreditation vs centers of excellence)) and progression-free survival and overall survival. To assess the association between specific factors (e.g. Surgical Complexity Score, ECOG performance status, HIPEC etc, respectively) during primary surgery and progression-free survival and overall survival To assess the association between specific factors and the accreditation status of the centers (training center, endometrial cancer surgery, advanced OC surgery in 2 categories (regular accreditation vs centers of excellence)) To assess progression-free survival and overall survival for patients with advanced OC not undergoing cytoreductive surgery during primary therapy. To develop a nomogram based on clinical variables and ESGO quality indicators to predict progression-free survival. Study design The present study is an international multicenter, retrospective cohort study based on the ESGO consortium of accredited centers. The following patients treated at participating ESGO accredited centers during the past three years, i.e. 2020, 2021 and 2022 will be included into the present study: ALL consecutive incoming patients at the center with newly diagnosed advanced stage (FIGO III and IV) ovarian cancer (importantly, irrespective of whether they underwent surgery or not during their primary therapy), and ALL recurrent OC patients who underwent secondary or tertiary cytoreductive surgery at the center (importantly, surgical cases only). Participating centers will provide both, patient-level data (e.g. demographic, tumor-related, treatment-related and prognostic information per patient) and center-level data (e.g. infrastructure, processes in place). Data will be entered into a REDcap database maintained by ESGO. This protocol will be reviewed by the Institutional Review Board (IRB)/ethics committee (EC) of each participating center. Data will not be entered into the ESGO REDCap database until IRB/EC approval is obtained. Each center will be responsible of obtaining its own IRB/EC approval. Patients For this international retrospective cohort study, data will be documented in a REDCap database maintained by ESGO. Patient data will be entered in a pseudo anonymized form in this database by the respective center. Study IDs are unique for each patient and will be automatically assigned when a patient is registered into the database. As the study ID can only be linked to specific patients at the respective center and not within the database, identification of individual patients will only be possible at the respective center. Participating centers will only have access to their own patient data via the database. Inclusion criteria Patients presented at an ESGO accredited center from January 1st to December 31st of the years 2020, 2021 or 2022 with: 1) newly diagnosed FIGO stage III/IV OC (importantly, surgical and non- surgical cases) or 2) relapsed OC undergoing secondary or tertiary cytoreductive surgery (importantly, surgical cases only). Of note, ALL consecutive incoming OC patients in the specified time-frame must be documented, in order to avoid selection bias. Exclusion criteria Patient \< 18 years at first diagnosis Patient with relapsed OC undergoing surgery for palliative intent (no cytoreductive surgery) Patient presenting at first diagnosis with early-stage OC Parameters Appendix A: Data collection form and codebook for patient-level data and center-level data, respectively Appendix B: ESGO publications on the ten ESGO quality indicators in advanced ovarian cancer surgery (Querleu et al., 2016; Fotopoulou/Concin et al., 2020) Main target parameters (primary target parameter) Ten ESGO quality indicators for advanced ovarian cancer surgery Prognosis: progression-free survival and overall survival Secondary target parameters (secondary target parameters) Center: type of accreditation status Adherence to the ten ESGO quality indicators for advanced ovarian cancer surgery Treatment performed at the accredited center Exploratory Parameter Diverse parameters, such as (not exclusively): ECOG status Surgical complexity score HIPEC performed during surgery Statistical considerations 7.1 Sample Size We expect to enroll at least 3000 patients with primary cytoreductive surgery conducted at the participating ESGO accredited centers. Although being a real-world data project there is not a specific number of patients to be included in the analysis. The primary PFS analysis will be conducted after a minim follow-up of 20 months in the overall cohort. Marth et al (2022) report that 67% of patients with advanced ovarian cancer had a PFS event after 20 months of follow-up, so we expect to have about 2000 PFS event at the time of the primary analysis. This number of events provide a statistical power higher than 95% to detect any difference with a PFS hazard ratio of \< 0.85 in the univariable analysis using the log-rank test. Univariable and multivariable analysis Univariable Cox proportional-hazard models will be used to obtain hazard ratios (HRs) with 95% confidence intervals (CIs). To select variables with the highest prognostic impact for PFS and OS, we will perform a least absolute shrinkage and selection operator (LASSO) regression using package glmnet in R software to build the most parsimonious multivariable model. Demographic and Clinical Characteristics We will use descriptive statistics to summarize the demographic and clinical characteristics of patients overall and stratified by their respective center's accreditation status (training center, endometrial surgery, advanced ovarian cancer surgery - regular, advanced ovarian cancer surgery - excellence). If a center is accredited in more than one category including advanced ovarian cancer surgery, their advanced ovarian cancer surgery accreditation status will be used. Continuous variables will be summarized with means, standard deviations, medians, and ranges. Categorical variables will be summarized with counts and percentages. Efficacy endpoints Progression-Free Survival Progression-free survival (PFS) will be defined as the time (months) from the date of primary cytoreductive surgery to the date of progression, recurrence or death, whichever occurs before. Patients who are alive without progression or recurrence will be censored on the date of last clinical assessment. Patients without or unknown post baseline tumour assessments but known to be alive will be censored at the time of the study inclusion. Overall survival Overall survival (OS) will be defined as the time (months) from the date of primary cytoreductive surgery to the date of death. Patients who are alive will be censored on the date of last follow-up. Secondary Outcomes We will tabulate the number of centres adhering to each of the ten ESGO quality indicators stratified by type of accreditation. We will estimate the number of centres adhering to each of the ten ESGO quality indicators with an exact binomial 95% confidence interval for each stratum. We will also tabulate the number of centres adhering to all ten ESGO quality indicators stratified by type of accreditation. We will estimate the number of centers adhering to all ten ESGO quality indicators with an exact binomial 95% confidence interval for each stratum. Efficacy analyses Time-to-event variables (PFS, OS, etc) will be analyzed according to the Kaplan-Meier method. Kaplan-Meier survival curves and the median will be reported, along with associated 95% CI calculated using the standard error derived from Greenwood's formula. In addition, PFS and OS rates at 1, 2 and 3 years will be estimated. Cox regression on univariate models will be used to calculate the risk reduction. Hazard ratio (HR) with 95% CIs will be reported to evaluate the impact of several prognostic factors. These factors include (not exclusively) each of the ten ESGO quality indicators (met/not met) that can be assessed on a patient level. We will also group patients according to their respective centre's accreditation status. Other clinical factors may be considered as well, such as (not exclusively) stage of disease, ECOG status, surgical complexity score, BRCA mutation status, systemic therapy use and HIPEC. For categorical endpoints counts and percentages, with 95% CIs, will be calculated. For the univariate and multivariate analysis, the logistic regression will be performed. Potential treatment comparison or potential trial simulation analysis will be reported in a separate statistical analysis plan (SAP) to minimise bias and to better define the steps in each of these analyses. Missing data All available demographic, clinical and efficacy data will be included in data listings and tabulations. The percentage of missing data per each variable will be presented in a table. The relevance of missing sample data will be assessed. If nothing unusual is observed, the missing at random (MAR) approach will be used, and multiple imputation of random missing values will be carried out via the mice R package. Sensitivity analysis without data imputation will be performed. Ethical, legislative, and regulatory considerations Regulatory and privacy considerations There is no direct benefit for patients evaluated in this research study. Future patients with ovarian cancer will probably benefit from the information provided by this study, as it will validate quality indicators in advanced ovarian cancer surgery, and support identification of centers with profound expertise in the surgical treatment of advanced ovarian cancer, including objective criteria for categorization into centers of excellence. This could improve the strategy for health systems to identify and/or develop centers of excellence for ovarian cancer treatment and may drive centralization of ovarian cancer treatment. There is no direct risk for patients evaluated in this research study. The only possible risk, the disclosure of sensitive patient data, is appropriately and proactively minimized through data pseudonymization and access restrictions. The investigator will ensure that this study is conducted in full compliance with the principles of the "Declaration of Helsinki" or with the laws and regulations of the country in which the research is conducted, whichever affords the greater protection to the individual. All data will be collected on a customised, secure, password protected central REDCap database hosted by ESGO. REDCap is used internationally to securely collect research data. All collected data will be transmitted and held pseudo-anonymised. Identification of the patient will only be possible at the respective center and not based on the information in the database. Regulatory and privacy considerations All data collected from the patients' medical records during this study will be kept pseudonymized in accordance with Institutional policies and supervisory authority guidelines, Regulation (EU) 2016/679 of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data (the "GDPR"), as well as any national rules taken in particular on the basis of Article 9.4 of the GDPR relative to the processing of the personal data concerning health. The patients whose data is collected and processed in the framework of the study must be informed of the use of their data for scientific research purposes, in accordance with the requirements of the data protection laws and regulations applicable at the respective center. Only pseudonymized information strictly limited to what is necessary for the purpose of the study in accordance with the protocol will be stored in this database. No data other than those defined in the attached protocol may be collected as part of this study. In particular, no sensitive data such as religious origin, political or trade union opinions, philosophical views or morals will be collected. Study IDs are unique to each patient and will be automatically assigned when a patient is registered onto the database. Study IDs will be assigned centrally by the REDCap bespoke database. Only one study ID should be created for each patient and new data registered on the appropriate electronic case record file (eCRF). The database will be hosted on a secure server (https) and its access will be secured by a code unique and personalised to each authorised personnel. Only authorized personnel at the respective center listed in the local IRB will have access to data at the center and the database, in which information will be entered. Individual centres will retain access to their patient data and not provide identifiable information on the international database. Local sites will retain information on patient identifiable data in a separate file but not share this with the central audit team. This information/file linking the study ID to patient identifiers will be kept on the local centre/unit/hospital server. Thus, identification of the patient will be only possible at the respective center and not in the database itself.