Real-world patient characteristics and survival outcomes in patients with advanced or recurrent endometrial cancer in England: a retrospective, population-based study Authors: Susana Banerjee, Alvaro Ingles Russo Garces, Jamie Garside, Tameera Rahman, Clare Pearson, Kiera Heffernan Published: 2024 ABSTRACT AbstractObjectiveThis study defined a retrospective cohort of patients in England with primary advanced or recurrent (A/R) endometrial cancer (EC) who may have been eligible for clinical trials evaluating immune checkpoint inhibitors (ICIs) in the first-line (1L) setting within a real-world dataset, and described the characteristics, treatment patterns and outcomes within this cohort.DesignThis was a retrospective, population-based study.SettingRoutine population-level data from the National Cancer Registration and Analysis Service in England were used. Patients diagnosed with A/R EC between 1 January 2013 and 31 December 2019 were included (follow-up until 23 August 2021). ICI-eligible patients who received any 1L therapy (defined as first systemic treatment for A/R EC with or without radiotherapy) and met key eligibility criteria for the RUBY trial (NCT03981796; 1L cohort) were included. A subpopulation who solely received carboplatin–paclitaxel at 1L (carboplatin–paclitaxel subcohort) was identified.MethodsDemographics, characteristics and therapy received were reported. Overall survival (OS), time to next treatment (TTNT) and time to treatment discontinuation (TTD) from 1L chemotherapy initiation were assessed using Kaplan-Meier methodology.ResultsOf 13 954 patients identified, 2376 ICI-eligible patients were included in the 1L cohort (median [range] age: 67.9 [26.7–94.0] years); 902 patients received solely carboplatin–paclitaxel at 1L. Demographics and disease characteristics were generally similar between cohorts. Median (95% CI) OS, TTNT and TTD from 1L chemotherapy were 27.2 (24.7, 30.2), 16.9 (15.8, 18.5) and 3.4 (3.4, 3.4) months, respectively, in the 1L cohort, and 17.2 (15.5, 19.0), 12.4 (11.6, 13.5) and 3.4 (3.4, 3.4) months, respectively, in the carboplatin–paclitaxel subcohort.ConclusionLong-term outcomes were poor for both cohorts, particularly the carboplatin–paclitaxel subcohort, where patients did not receive radiotherapy and had predominantly metastatic disease. This reflects the unmet need for more durable treatment options to prevent relapse and prolong survival in this patient population. This real-world study will help contextualise outcomes from ongoing phase III clinical trials investigating 1L ICI treatments. Strengths and limitations of this study A key strength of this study is the use of real-world, population-level data from the National Health Service England’s National Cancer Registration and Analysis Service (NCRAS), which fully captures the cancer population in England. A key strength of this study is the use of real-world, population-level data from the National Health Service England’s National Cancer Registration and Analysis Service (NCRAS), which fully captures the cancer population in England. To ensure real-world applicability of the findings from this study, a cohort of patients who may have been eligible for inclusion in phase III immune checkpoint inhibitor (ICI) trials were identified through applying key eligibility criteria from the RUBY trial (evaluating ICI dostarlimab plus carboplatin–paclitaxel in patients with primary advanced or first-recurrent EC; NCT03981796). To ensure real-world applicability of the findings from this study, a cohort of patients who may have been eligible for inclusion in phase III immune checkpoint inhibitor (ICI) trials were identified through applying key eligibility criteria from the RUBY trial (evaluating ICI dostarlimab plus carboplatin–paclitaxel in patients with primary advanced or first-recurrent EC; NCT03981796). As with all retrospective population-level studies, limitations in data availability and quality of data recording within source datasets may have introduced biases within results and led to differences between study and real-world populations. As with all retrospective population-level studies, limitations in data availability and quality of data recording within source datasets may have introduced biases within results and led to differences between study and real-world populations. Although the eligibility criteria applied to the ICI-eligible population in this study were similar to the eligibility criteria from the RUBY trial, limitations in data availability within source datasets resulted in some differences. Although the eligibility criteria applied to the ICI-eligible population in this study were similar to the eligibility criteria from the RUBY trial, limitations in data availability within source datasets resulted in some differences. Because disease progression and recurrence data were not captured within the NCRAS dataset, time to next treatment was used as a proxy for progression-free survival and a treatment-based algorithm was used for estimating probable disease recurrence. Because disease progression and recurrence data were not captured within the NCRAS dataset, time to next treatment was used as a proxy for progression-free survival and a treatment-based algorithm was used for estimating probable disease recurrence. Introduction Endometrial cancer (EC) is the most common gynaecological cancer in the UK.1 Within the UK, incidence rates have increased by 12% from 2006–2008 to 2016–2018, while mortality rates have increased by 24%.2 Increased incidence is likely related to increasing prevalence of obesity, metabolic syndrome and the ageing population, while increasing mortality may be related to increasing incidence of more aggressive high-risk histology types.3 4 Although EC incidence rates are projected to decline by 2% in the UK between 2023–2025 and 2038–2040, mortality rates are predicted to increase by 12%.2 Approximately 20% of patients with EC have advanced disease at diagnosis.5 Recurrence rates are also high, with approximately 13% of all ECs and 38%–67% of advanced ECs recurring.6 7 Patients with advanced or recurrent (A/R) EC follow a similar treatment strategy, with primary treatment involving a combination of surgery, radiotherapy and chemotherapy; hormone therapy may be an option for patients with low-grade, hormone receptor-positive A/R EC.810 The recommended systemic standard of care (SoC) therapy for primary A/R EC is combination carboplatin–paclitaxel.8 9 However, long-term outcomes are poor and subsequent treatment options limited, with a 5-year survival of 48% in patients with stage III and 15% in patients with stage IV disease at diagnosis.11 Median overall survival (OS) for patients with A/R EC who receive primary SoC ranges from 13 to 37 months.1215 These poor outcomes highlight the need for improved therapeutic options to enhance durability of treatment outcomes by preventing or delaying recurrence. Until recently, there has been little progress in new therapy development. Although multiple treatment modalities have been evaluated in the primary treatment of EC, few have demonstrated meaningful benefit beyond standard chemotherapy.16 Multiple phase III trials investigating the benefit of innovative immune checkpoint inhibitors (ICIs), alone or in combination with other agents, versus SoC for primary treatment of patients with A/R EC are currently ongoing.1721 There is a need for contemporary real-world evidence on characteristics and clinical outcomes in this patient population, including patients receiving SoC, to help contextualise recent and anticipated results from ongoing trials. The objectives of these analyses were to define a cohort of patients with primary A/R EC in England who may have been eligible for ICI trials within a real-world dataset, and to describe the characteristics, treatment patterns and outcomes within this cohort, including a focus on a subpopulation of patients who received first-line (1L) SoC carboplatin-–paclitaxel. Methods Study design and data sources This descriptive, non-interventional, retrospective cohort study used routine population-level data obtained through National Health Service (NHS) England’s National Cancer Registration and Analysis Service (NCRAS).22 NCRAS pools data from a range of sources to create the Cancer Analysis System, which can track patient progression from first point of contact to end of treatment or death (online supplemental figure 1). Constituent datasets include National Cancer Registration Dataset (NCRD; diagnostic and pathological data for every primary cancer diagnosed in England22); Hospital Episode Statistics datasets (hospital admissions, accident and emergency attendances and outpatient appointments at NHS hospitals23); Systemic AntiCancer Therapy dataset (systemic treatment information for treatments prescribed or funded by the NHS24); National Radiotherapy Dataset (treatment information for every radiotherapy appointment in England25); NHS England’s Data Access Tracing Service (routine death registration data26). From the NCRAS data, adult patients diagnosed with A/R EC in England between 1 January 2013 and 31 December 2019 (index period), with follow-up until 23 August, 2021, were identified; index was date of first diagnosis of A/R EC. Patients were followed until the occurrence of events, or until censored at study exit (lost to follow-up, death or moving outside of England) or end of study period. Patients and the public were not involved in the design, conduct or reporting of the study. Study population Primary advanced disease was defined as International Federation of Gynecology and Obstetrics (FIGO) stage III/IV disease. Recurrent EC disease was defined as patients with FIGO stage I/II disease who received their first subsequent treatment after initial treatment for stage I/II disease, with a gap of >90 days between consecutive treatments. To avoid duplication of patient data, recurrent stage III/IV disease was not included. A sensitivity analysis was performed, altering the gap between consecutive treatments to >180 days. Key eligibility criteria from the RUBY trial (evaluating ICI dostarlimab plus carboplatin–paclitaxel in patients with primary advanced or first-recurrent EC; NCT03981796) were applied where possible17 to identify patients who may have been eligible for ICI trials (ICI-eligible patients). Inclusion criteria were known Eastern Cooperative Oncology Group performance status (ECOG PS) 0 or 1; FIGO stage I/II with probable recurrence (first recurrent patients, as defined above); and FIGO stage III/IV (primary advanced patients) with at least one of: IIIa–IIIc1 with measurable disease and biopsy; IIIc1 with carcinosarcoma, clear cell, serous or mixed histology; IIIc2 or IV; first recurrent disease and no previous systemic anticancer therapy (SACT) within 90 days of index. Exclusion criteria included FIGO stage I/II with more than one recurrence (only first recurrence was included), previous cancer diagnosis within 36 months of EC diagnosis, evidence of anti–programmed cell death protein/ligand-1 therapy or anti–programmed cell death protein/ligand-2 therapy, receiving any drug as part of a clinical trial, or central nervous system metastases. From these patients, a 1L cohort and carboplatin–paclitaxel subcohort were formed. The 1L cohort included ICI-eligible patients (as defined above) who received 1L therapy (defined as first SACT after index, regardless of surgical or disease status), including those who received chemoradiation at 1L (defined as radiotherapy received within 90 days before, after or during chemotherapy; online supplemental figure 2). The carboplatin–paclitaxel subcohort was a subpopulation of the previous cohort, who solely received carboplatin–paclitaxel at 1L. Within the carboplatin–paclitaxel subcohort, patients who received radiotherapy as part of their 1L treatment were excluded. Data analysis Baseline demographics, disease characteristics and treatments received were summarised descriptively. OS, time to next treatment (TTNT; a proxy for progression-free survival (PFS)) and time to treatment discontinuation (TTD) from 1L chemotherapy initiation were assessed using Kaplan-Meier methodology. OS, TTNT and TTD from 1L were calculated from initiation of 1L therapy to the earliest of death or censoring, the start of a new line of therapy (TTNT only), or treatment discontinuation (TTD only, defined as a gap of >90 days from end date of 1L treatment to either resumption or no resumption of treatment). In accordance with the journal’s guidelines, we will provide our data for independent analysis by a selected team by the Editorial Team for the purposes of additional data analysis or for the reproducibility of this study in other centres if such is requested. Results Patient population Of the 13 954 patients with A/R EC identified, 3051 were defined as ICI-eligible, and of these, 2376 were included in the 1L cohort (figure 1). The sensitivity analysis revealed that defining recurrent disease using a 180-day gap between treatments resulted in a smaller A/R EC population of 11 907 patients (online supplemental table 1). Baseline demographics and disease characteristics are shown in table 1. Within the 1L cohort, median age was 67.9 years (range: 26.7–94.0) and most patients were white (86.3%, n=2051). Most patients had FIGO stage III/IV at primary diagnosis (primary advanced EC; 83.1%, n=1975), and 16.9% (n=401) had FIGO stage I/II (recurrent EC). The most common histology types were endometrioid (44.1%, n=1049), serous (28.1%, n=667) and carcinosarcoma (13.0%, n=309). Most patients had ECOG PS of 0 at registry diagnosis (68.7%, n=1633). Within the ICI-eligible population (figure 1), baseline demographics were similar between patients with advanced and recurrent disease, with no unexpected differences from a clinical perspective (online supplemental table 2). Cohort flow diagram. *A sensitivity analysis revealed that defining recurrent disease using a 180-day gap between the treatments resulted in an A/R EC population of 11 907 patients (22.2% of the total EC population), while a 90-day gap resulted a population of 13 954 patients (26.1% of the total EC population, which was used in this analysis); †probable recurrence was defined as patients with FIGO stage I/II disease who receive their first subsequent treatment following initial treatment for stage I/II disease, with a gap of >90 days between consecutive treatments; ‡patients may be in ≥1 of the exclusion criteria categories; §where chemotherapy and radiotherapy were administered concomitantly or within 90 days of each other. 1L, first-line; A/R, advanced or recurrent; CNS, central nervous system; EC, endometrial cancer; ECOG PS; Eastern Cooperative Oncology Group performance status; FIGO, International Federation of Gynecology and Obstetrics; ICI, immune checkpoint inhibitor; PD-(L)1, programmed cell death protein (or ligand) 1; PD-L2, programmed death ligand 2; SACT, systemic anticancer therapy. Table 1Baseline demographics and disease characteristicsDemographic or characteristic1L cohortn=2376Carboplatin–paclitaxel subcohortn=902Year of diagnosis, n (%) 2013160 (6.7)57 (6.3) 2014253 (10.6)84 (9.3) 2015329 (13.8)121 (13.4) 2016388 (16.3)157 (17.4) 2017403 (17.0)161 (17.8) 2018419 (17.6)156 (17.3) 2019424 (17.8)166 (18.4)Median age, years (range)*67.9 (26.7–94.0)67.5 (34.4–88.9) Age at diagnosis, n (%) <65943 (39.7)351 (38.9) 65–74928 (39.1)400 (44.3) ≥75505 (21.3)151 (16.7)Ethnicity, n (%) White2051 (86.3)775 (85.9) Black104 (4.4)51 (5.7) Asian103 (4.3)39 (4.3) Other52 (2.2)19 (2.1) Missing66 (2.8)18 (2.0)Disease stage at diagnosis, n (%)† FIGO I321 (13.5)146 (16.2) FIGO II80 (3.4)36 (4.0) FIGO III1262 (53.1)281 (31.2) FIGO IV713 (30.0)439 (48.7)Histology, n (%) Endometrioid1049 (44.1)338 (37.5) Serous667 (28.1)294 (32.6) Carcinosarcoma309 (13.0)132 (14.6) Non-specific carcinoma105 (4.4)41 (4.5) Clear cell carcinoma103 (4.3)50 (5.5) Other143 (6.0)47 (5.1)ECOG PS at diagnosis, n (%)† 01633 (68.7)627 (69.5) 1743 (31.3)275 (30.5)Charlson comorbidity score 02060 (86.7)778 (86.3) 1157 (6.6)61 (6.8) 249 (2.1)19 (2.1) 314 (0.6)5 (0.6) ≥45 (0.2)2 (0.2) Unknown91 (3.8)37 (4.1)Follow-up (months) from index, median (range)22.2 (0.0–98.8)17.4 (0.0–98.8)*At advanced diagnosis or recurrence.†aAt registry diagnosis.ECOG PSEastern Cooperative Oncology Group performance statusFIGOInternational Federation of Gynecology and Obstetrics1Lfirst-line Baseline demographics and disease characteristics At advanced diagnosis or recurrence. aAt registry diagnosis. ECOG PSEastern Cooperative Oncology Group performance statusFIGOInternational Federation of Gynecology and Obstetrics1Lfirst-line In the 1L cohort, 902 patients received solely carboplatin–paclitaxel at 1L (carboplatin–paclitaxel subcohort, figure 1). Baseline demographics and disease characteristics were generally similar between cohorts; however, a higher proportion of patients in the carboplatin–paclitaxel subcohort had FIGO stage IV disease (48.7%, n=439) than in the 1L cohort (30.0%, n=713), and median duration of follow-up was 17.4 months (range: 0.0–98.8) in the carboplatin–paclitaxel subcohort and 22.2 months (range: 0.0–98.8) in the 1L cohort (table 1). Treatments received Within the 1L cohort, 67.4% (n=1601) of patients received surgery, 60.3% (n=1433) received radiotherapy and 98.7% (2345) received SACT (table 2). Of those who underwent surgery, 87.9% (n=1405) received it before 1L therapy initiation. Carboplatin–paclitaxel was the most common SACT (77.8%, n=1824, including patients who received concomitant radiotherapy). Table 2Most common 1L and 2L treatments received by 1L cohortTreatment, n (%)n=2376Surgery*1601 (67.4) Patients receiving surgery before LOT1 initiation†1408 (87.9)Radiotherapy‡1433 (60.3) 1L SACT‡§2345 (98.7) Carboplatin–paclitaxel1824 (77.8) Carboplatin302 (12.9) Cisplatin49 (2.1) Other170 (7.2) 2L SACT‡§n=661 Carboplatin–paclitaxel206 (31.2) Paclitaxel82 (12.4) Liposomal doxorubicin76 (11.5) Carboplatin–liposomal doxorubicin73 (11.0) Carboplatin68 (10.3) Cisplatin–doxorubicin36 (5.4) Cisplatin27 (4.1) Other93 (14.1)Main category percentage values are shown as percentages of the overall cohort size; subcategory percentage values are shown as percentages of the main category total. For SACT, treatments received in <20 patients are grouped in Other.*Relevant resections received 31 days before and no time restriction after index date.†Relevant resections received before the initiation of 1L.‡After advanced or recurrent diagnosis.§Patients who received a systemic anticancer regimen which could have been solely systemic treatment or in combination with radiotherapy.1Lfirst-line2Lsecond-lineLOTline of therapySACTsystemic anticancer therapy Most common 1L and 2L treatments received by 1L cohort Main category percentage values are shown as percentages of the overall cohort size; subcategory percentage values are shown as percentages of the main category total. For SACT, treatments received in <20 patients are grouped in Other. Relevant resections received 31 days before and no time restriction after index date. Relevant resections received before the initiation of 1L. After advanced or recurrent diagnosis. Patients who received a systemic anticancer regimen which could have been solely systemic treatment or in combination with radiotherapy. 1Lfirst-line2Lsecond-lineLOTline of therapySACTsystemic anticancer therapy Of the ICI-eligible 1L patients who received 1L SACT, 28.2% (n=661) moved to a second-line (2 L) SACT (table 2). The most frequent 2 L therapy was carboplatin–paclitaxel (n=206; 31.2%). Clinical outcomes Among patients who received 1L SACT (n=2345), median OS from 1L initiation was 27.2 (95% CI: 24.7, 30.2) months for the 1L cohort and 17.2 (15.5, 19.0) months for the carboplatin–paclitaxel subcohort (figure 2). Median TTNT from 1L initiation was 16.9 (95% CI: 15.8, 18.5) months for the 1L cohort, and 12.4 (11.6, 13.5) months for the carboplatin–paclitaxel subcohort (figure 3). OS from initiation of 1L SACT for (A) 1L cohort (n=2345) and (B) carboplatin–paclitaxel subcohort (n=902).1L, first-line; OS, overall survival; SACT, systemic anticancer therapy. TTNT from initiation of 1L SACT for (A) 1L cohort (n=2345) and (B) carboplatin–paclitaxel subcohort (n=902). 1L, first-line; ICI, immune checkpoint inhibitor; SACT, systemic anticancer therapy; TTNT, time to next treatment. Within the 1L cohort, 43.1% of patients (n=1011) had a TTNT from 1L initiation greater than the median TTNT (ie, >16.9 months, late progressors), while 43.7% of patients (n=1024) had a TTNT from 1L of less than the median (ie, <16.9 months, early progressors). Patients who were censored before the median TTNT (n=304) or had unknown vital status (n=6) did not contribute to either group. Late progressors were younger and were more commonly primary advanced patients (FIGO stage III/IV), ECOG PS 0 and of endometrioid histology (online supplemental table 3). Among patients who received 1L SACT, median TTD from 1L initiation was 3.4 (95% CI: 3.4, 3.4) months for the 1L cohort and 3.4 (3.4, 3.4) months for the carboplatin–paclitaxel subcohort. Discussion Summary of main results This study described real-world characteristics, treatment patterns and outcomes in patients with primary A/R EC in England who may have been eligible for ICI trials, including a focus on those who received SoC carboplatin–paclitaxel. The data from this study help contextualise recent and anticipated results172127 from ongoing phase III clinical trials investigating 1L use of ICI, alone or in combination, versus SoC carboplatin–paclitaxel. Baseline demographics and disease characteristics between cohorts were similar. Of note, within the 1L cohort, most patients were FIGO stage III at diagnosis, whereas within the carboplatin–paclitaxel subcohort most were FIGO stage IV. Median follow-up time was shorter for the carboplatin–paclitaxel subcohort. The number of patients diagnosed per year consistently increased across both cohorts in consecutive years; this may be due to improved capture of staging data, and the increasing incidence rates observed for EC during the study duration.2 Most patients received surgery and/or radiotherapy at 1L, and almost all received SACT. Carboplatin–paclitaxel was the most frequently administered SACT, confirming it as the SoC for A/R EC and reinforcing the rationale for its use as a comparator arm in ongoing clinical trials.1721 In the 2L setting, carboplatin–paclitaxel was still the most commonly administered SACT; however, its use was less frequent than in the 1L setting. Of the patients who received 1L therapy, fewer than half received subsequent 2L therapy, which may reflect the lack of SoC in the 2L setting and the limited efficacy of available treatments.810 Combination SACT and radiotherapy is recommended for use in patients with stage III EC with no residual disease post-surgery and has demonstrated clinical benefit in this specific setting.9 31 Thus, concomitant radiotherapy with SACT was excluded from the carboplatin–paclitaxel subcohort to allow focus on patients who may not be suitable for radiotherapy, and to better align with the study populations in the ongoing ICI trials.17 Consistent with guideline-recommended management, most patients within the 1L cohort had FIGO stage III disease at diagnosis, whereas within the carboplatin–paclitaxel subcohort there was a lower proportion of stage III disease and higher proportion of FIGO stage IV.9 OS and TTNT outcomes were shorter within the carboplatin–paclitaxel subcohort than in the complete 1L cohort. This finding may reflect differences in the disease characteristics between the two cohorts, particularly the enrichment of patients with primary stage IV disease in the carboplatin–paclitaxel subcohort versus the 1L cohort; with a higher proportion of patients with stage IV disease, outcomes observed here are consistent with the poorer prognosis for stage IV versus stage III patients.11 Further, clinical guidelines recommend the use of chemotherapy alone for A/R EC when surgery or radiotherapy alone or in combination are not sufficient to treat disease.810 Therefore, by guideline recommendation, the carboplatin–paclitaxel subcohort may exclude a population of patients amenable to radiotherapy and/or surgery with curative intent who may exhibit improved outcomes compared with those who receive chemotherapy alone. The most common EC histology type was serous among early progressors but endometrioid among late progressors. Patients with serous EC are known to have worse clinical outcomes than other common histology types, with serous EC accounting for nearly half of EC-related deaths despite accounting for just 10–20% of diagnoses.32 Results in the context of published literature Compared with the RUBY trial population, this real-world population of patients in England (1L cohort) was older (median: 65 vs 68 years) and had a higher proportion of patients with FIGO stage III/IV at primary diagnosis (60% vs 83%) and patients with non-endometrioid histology (45% vs 56%).27 The 24-month OS was similar between the RUBY placebo group (54.3%) and the 1L cohort (53%) but lower in the carboplatin–paclitaxel subcohort (40%).33 Additional phase III trials evaluating the benefit of ICIs in the treatment of patients with primary advanced or first-recurrent EC include NRG-GY018 (NCT03914612),20 28 DUO-E (NCT04269200),34 35 AtTEnd (NCT03603184)18 30 and LEAP-001 (NCT03884101).19 29 Eligibility criteria for these trials are generally comparable to those of this real-world study; thus, as with RUBY, the results of this study support the ongoing investigations of novel ICI treatment combinations in these patients, and help contextualise the results of these trials. Strengths and weaknesses This study has several strengths and weaknesses. A key strength of this study is the population-based cohort design. The NCRAS dataset captures the entire population with cancer in England and allows linkage of healthcare events for surgeries, chemotherapies, and radiotherapy. Study weaknesses include limitations in the quality of recording within source datasets, which may have introduced biases within results and led to differences between the full 1L patient population and a real-world ICI-eligible population. A second potential weakness is that the cohort eligibility criteria used within this study, while similar, differ from exact ICI clinical trial (ie, RUBY) eligibility criteria. RUBY criteria allow for some prior radiotherapy: only patients who received prior radiotherapy within 21 days or <5 times the half-life of the most recent therapy (whichever is shorter) of study day 1 are excluded (palliative radiotherapy to a small field ≥1 week before study day 1 may be allowed).17 Within this real-world study, where nuanced information on radiotherapy use was not available, the exclusion was higher level patients who received any radiotherapy 90 days before, 90 days after or during 1L SACT were excluded from the carboplatin–paclitaxel subcohort. Consistent with patient populations reported from phase III ICI clinical trials, data presented for advanced and recurrent patients are combined to allow for contextualisation of results; baseline demographics and characteristics were similar between advanced and recurrent patients. Finally, there were some limitations in data availability and patient inclusion that should be acknowledged. Staging data were absent for 9.8% of all patients with EC and ECOG PS was missing for 49.2%; these patients were excluded from analyses. ECOG PS at point of diagnosis only is available within the NCRD and thus may not represent ECOG PS experienced at recurrence. Only first recurrent and primary advanced patients were included; advanced recurrent disease was not captured to prevent duplication of patient data. Because the NCRD dataset does not include disease progression or recurrence data, TTNT was used as a proxy for PFS, and a treatment-based algorithm was used for estimating probable disease recurrence. Only SACTs prescribed within secondary care are captured within this study; treatments otherwise prescribed, including oral therapies such as hormone therapy, are poorly captured in this dataset. Implications for practice and future research These results provide an overview of outcomes under the current treatment paradigm that serves as a benchmark for potential improved outcomes with the introduction of innovative 1L and 2L therapies. Overall, these data are robust and generalisable to ongoing phase III clinical trials investigating innovative ICIs. Conclusions This real-world study provides demographics and disease characteristics, treatment patterns and clinical outcomes in patients from England with primary A/R EC who may have been eligible for inclusion within an ICI trial, with a focus on the subpopulation of patients receiving solely carboplatin–paclitaxel, the current systemic SoC 1L treatment in this patient group. Outcomes for this patient population were poor, in particular those in the carboplatin–paclitaxel subcohort, where patients did not receive radiotherapy and had predominantly metastatic disease. These data highlight the unmet need for more durable 1L treatment options to prevent relapse and prolong survival in this patient population, as well as provide important insights to patients with A/R EC in the 1L setting to help contextualise anticipated results from ongoing phase III clinical trials investigating novel ICI treatments. supplementary material