OCEANIA: real-world study of ovarian cancer treatment patterns across multiple lines of therapy in Argentina and Brazil

Abstract

Aim: To evaluate real-world data on treatment patterns in Argentina and Brazil in patients with ovarian cancer.

Methods: This study evaluated de-identified antineoplastic exposure data from a private healthcare provider in Argentina and health claims database (Orizon) in Brazil from 2010 to 2019 and 2015 to 2020, respectively.

Results: Platinum-based chemotherapy was the most common first-line therapy (Argentina: n =311 [87.6%]; Brazil: n = 1142 [79.3%]). The proportion of patients receiving platinum-based chemotherapy declined across both populations from first- to second-line, while use of non–platinum-based, targeted, and hormone therapies increased. Duration of platinum-based treatment and time to next treatment decreased from first- to fourth-line.

Conclusion: There is an unmet need for effective therapies that can prolong time to next treatment in ovarian cancer in Argentina and Brazil.

Plain language summary

Summary points.

• Antineoplastic exposure data were captured from databases in Argentina (Hospital Italiano de Buenos Aires) and Brazil (Orizon) from 2010 to 2019 and 2015 to 2020, respectively, and treatment patterns were described across multiple lines of therapy for patients with ovarian cancer (OC).

• In total, 741 patients from Argentina and 2529 patients from Brazil were identified.

• Across both countries, around 50% of patients with OC did not receive antineoplastic agents postdiagnosis.

• Of the 355 patients from Argentina and 1441 patients from Brazil who received 1L therapy, the most commonly received treatment was platinum-based chemotherapy.

• Use of platinum-based chemotherapy declined from the 1L to the 4L in Argentina (1L: n = 311, 87.6%; 4L: n = 29, 37.7%) and Brazil (1L: n = 1142, 79.3%; 4L: n = 40, 33.3%).

• Median time to next treatment between a line of any antineoplastic agent and the next line of any antineoplastic agent was longest between the 1L and 2L settings, with a median of 5.3 months (interquartile range [IQR]: 10.7 months) and 3.0 months (IQR: 6.8 months) in Argentina and Brazil, respectively.

• Median progression-free survival was 6.0 months (IQR: 11.3) in patients from Argentina and 3.6 months (IQR: 9.6) in patients from Brazil.

• Median duration of treatment was the longest in 1L therapy across both countries (Argentina: 4.4 [IQR: 3.3] months; Brazil: 3.5 [IQR: 3.5] months) and became shorter across successive lines of therapies.

• These findings highlight the unmet need for effective therapies to optimize outcomes in patients in Argentina and Brazil with ovarian cancer, particularly those with advanced disease.

1. Background

Ovarian cancer (OC) is one of the most common gynecologic cancers worldwide and is associated with a poor prognosis and high mortality rate [1,2]. Approximately 70% of cases are diagnosed at advanced stages because of a lack of effective screening strategies and the absence of early, disease-specific symptoms [2,3]. Approximately 85% of patients who achieve remission following primary treatment experience disease recurrence, and for advanced cases the 5-year relative survival rate is only ∼30–40% [4,5].

Although the incidence of OC is lower in Latin America than in Europe, the mortality rate is higher [6]. Across 17 Latin American countries, Brazil has one of the highest OC incidence rates in patients aged 60–74 years, while Argentina has one of the highest standardized OC mortality rates [6]. According to the WHO Global Cancer Observatory, the number of deaths related to OC in Argentina and Brazil in most age groups has increased significantly over the past two decades [7–9]. The increased mortality-to-incidence ratios across Latin American countries are mainly attributed to poor access to medical care, resulting in diagnosis at more advanced stages [10]. Healthcare insurance is also likely to affect access to treatment and therefore outcomes [10]. Studies in the USA have shown that lack of access to private health insurance can adversely affect the likelihood of early diagnosis [11]. Public healthcare systems in Argentina and Brazil are generally underfunded [12–14] and access to novel and more expensive treatments is often limited [10]. Furthermore, there are fewer clinical trial centers in South America compared with the USA and Western Europe, but this trend is slowly changing [15]. Access to treatment is better within private healthcare, but the proportion of patients with private healthcare is relatively low (around 16% of the population in Argentina and 28% in Brazil) [12,13], demonstrating the socioeconomic inequalities in these countries with respect to healthcare.

Cytoreductive surgery with platinum-based chemotherapies remains the standard treatment for OC [16]. For patients who receive treatment, multiple lines of therapy are often required due to disease progression and acquired resistance to platinum-based chemotherapies [17]. Novel therapies offer alternative treatment options to patients with OC and may reshape management strategies [4,18,19]. However, despite advances in treatment options for OC, access to newer cancer drugs remains an issue in Argentina and Brazil [10]. This highlights not only the disparities in cancer treatment in these countries [20,21] but also the missed opportunity for patients to potentially benefit from treatment.

Real-world studies investigating the burden of OC for patients and healthcare systems in Latin America are limited. It is important to gain insights into the real-world treatment patterns for OC in Latin America to help inform treatment decisions and improve the outcomes for these patients.

1.1. Objectives

The objective of this study was to evaluate real-world data on antineoplastic treatment patterns in patients in Argentina and Brazil with OC across multiple lines of treatment, focusing particularly on the use of platinum-based chemotherapy.

2. Methods

2.1. Data sources

‘Ovarian Cancer: drug utilization resEArch in LatiN AmerIcA’ (OCEANIA) was a retrospective database study that evaluated treatment patterns in patients with OC (including peritoneal and fallopian tube cancer) in Argentina and Brazil. Available databases with electronic medical records (EMRs) and healthcare administrative claims from Argentina and Brazil were used, respectively.

The Hospital Italiano de Buenos Aires (HIBA) in Argentina has a prepaid private healthcare maintenance program (Hospital Italiano Medical Care Program [HIMCP]) [22], which has >70,000 affiliated adult members with full coverage for medical care, and proportional coverage for prescribed medications. The medical and healthcare services provided by HIMCP are through two main hospitals and 24 peripheral outpatient centers, primarily covering urban and suburban populations in Buenos Aires. The database contains data on diagnosis, procedures, and medications.

The Orizon private health administrative claims database in Brazil consists of billing transactions between health insurance companies (or ‘operators’) and healthcare providers. Medical billing data from the 14 health plan operators in Orizon include authorizations and payment of private service providers. Nearly 70% of operators are located in the southeast region of Brazil [23,24]. Between 2015 and 2020, the Orizon database consisted of 9,577,482 patients (equating to 20% of private healthcare users, which is approximately 5% of the Brazilian population overall) [24]. The database is composed of medical invoices related to hospitalization and medical procedures (known as SADT, auxiliary diagnostic and therapeutic services forms) sent by hospitals, clinics, physicians' offices, and emergency/ambulatory services. The database also includes details of inpatient and outpatient medications dispensed but does not include medications purchased directly from pharmacies.

2.2. Patient population

The study included patients with OC (including peritoneal and fallopian tube cancer) who were at least 18 years old at the index date from 1 January 2010 to 30 June 2019 for Argentina and from 1 January 2015 to 31 December 2019 for Brazil (Supplementary Figure S1).

The index date for Argentina was defined as the first time a patient received a health term code, procedure (e.g., surgical intervention, imaging test, radiotherapy), or treatment (chemotherapy or radiotherapy) considered OC-related. For Brazil, the index date was defined as the first time a patient had a claim with an OC-related International Classification of Diseases version 10 (ICD-10) code or an OC-related procedure without an ICD-10 code or OC-related treatment was dispensed to the patient (whichever came first). Follow-up was defined as the period after the index date until loss of follow-up or end of study period (31 December 2019 for Argentina and 30 June 2020 for Brazil).

For the population in Argentina, several predefined medical health terms were used in the EMR. A list of structured health terms related to OC was prepared by oncologists and gynecologists from HIBA and keywords were searched in the free text through the Structured Query Language server retrieval from the database behind the EMR (Supplementary Table S1). After selection of all possible eligible patients using the medical health terms, oncologists and gynecologists from HIBA manually reviewed all EMR data to confirm each OC diagnosis. A biopsy was also required at HIBA as additional confirmation of the diagnosis. Patients who participated in any other clinical trial at HIBA from 1 January 2010 to 31 December 2019 were excluded from this study.

For the population in Brazil, patients with OC were identified through the presence of at least two ICD-10 codes during the study period: C56 (malignant neoplasm of ovary), C57.0 (malignant neoplasm: fallopian tube), C48.1 (malignant neoplasm: specified parts of peritoneum), or C48.2 (malignant neoplasm: peritoneum, unspecified). These ICD-10 codes had been selected after discussion with clinical oncologists.

2.3. End points

Treatment patterns relating to antineoplastic drug use were assessed. All antineoplastic agents evaluated in this study were based on dispensed drugs. Drugs were classified based on the National Administration of Drugs, Foods, and Medical Devices guidelines in Argentina and the Health National Regulatory Agency (Agência Nacional de Saúde) in Brazil. Therapeutic classes included chemotherapy, hormone therapy, targeted therapies (including but not limited to inhibitors of VEGF, PARP, RTK, EGFR, CDK, EGF, and TNF) and immunotherapy (anti-PD-1). Chemotherapies were classified as ‘platinum-based’ if the agents were either carboplatin, cisplatin, or oxaliplatin. Other chemotherapies were classified as ‘non–platinum-based’.

Time from diagnosis to first-line (1L) therapy was assessed for all patients. All antineoplastic treatments received by patients with OC were described by line of treatment, from the 1L to the fourth-line (4L) setting. After the index date, the first antineoplastic drugs dispensed within an interval of 30 days were classified as 1L therapy. The end date of a line of therapy was defined as the last day on which a drug in the regimen was infused or the oral drug was supplied (defined as the dispensing date plus 30 days); for regimens combining oral and infusions, the latest date using these definitions was used to define the end of the line of therapy. Maintenance therapy was recorded as part of the line of therapy and therefore not analyzed separately in this study. A switch in line of therapy was defined as the initiation of new agents or an interval of ≥120 days between two consecutive dispensation dates. As maintenance therapy was considered part of the line of therapy previously used, if bevacizumab or a PARP inhibitor were added within 120 days after the end of a line of therapy, this was considered an extension of the previous line and not considered a switch in line of therapy. Discontinuation and/or change of line was considered if there was an interval of >120 days from the date from one line of therapy to the addition of a new drug other than in the initial regimen or the addition of any chemotherapy or hormone therapy. Coadjuvant drugs that were excluded from the construction of the lines of therapy included folinic acid (leucovorin), filgrastim, medroxyprogesterone acetate, megestrol acetate, drospirenone plus ethinylestradiol, and gestodene plus ethinylestradiol. Initiation of these drugs did not indicate a switch in line of therapy.

Time to next treatment (TTNT) for any antineoplastic agent, in months, was considered as the time from the end of one line of antineoplastic agent to the start of the next antineoplastic agent. Progression-free survival (PFS) was also assessed in patients receiving platinum-based therapy in the 1L setting and was defined as time from the end of the first platinum-based therapy to the next platinum or non–platinum-based chemotherapy, surgery, or radiotherapy.

Duration of treatment (DOT), in months, of antineoplastic therapy from the 1L to 4L setting was also recorded.

Patients lost to follow-up were censored on the date of their last healthcare encounter in the database.

2.4. Data analysis

All analyses were carried out separately for each country. Results were analyzed descriptively, and country-specific differences in treatment availability, healthcare practices, and population composition were considered for each database. No country-level generalizations were performed. Analyses were performed using the statistical software R and Stata 17 (StataCorp LLC). Continuous variables were presented as descriptive statistics of central tendency (mean, median) and dispersion (standard deviation [SD], interquartile range [IQR], 25th percentile, and 75th percentile). Categorical variables were described as frequencies and percentages. Specifically, age at index date was described in years as a continuous variable and categorically, according to age group (18–39 years, 40–49 years, 50–59 years, 60–69 years, ≥70 years). Weight and BMI were described continuously. Nutritional status considered WHO classification based on BMI categories and was described as frequency and percentages according to category (BMI <18.5 kg/m²: underweight; 18.5–24.9 kg/m²: normal weight; 25.0–29.9 kg/m²: overweight; ≥30.0 kg/m²: obese). International Federation of Gynecology and Obstetrics (FIGO) staging, number of patients per index year, and number of patients with borderline tumors were described categorically as frequencies and percentages. Duration of follow-up period in months was described as median (IQR). The number of patients who received surgery, antineoplastic agents, and radiotherapy and the number of patients per line of systemic therapy were described as frequencies and percentages. Time from diagnosis to 1L systemic therapy was described as median (IQR). Antineoplastic therapy use by categorical therapeutic class and monotherapy, doublet, triplet, or quadruplet therapy use were described as frequencies and percentages. TTNT was described as median (IQR), as was duration of antineoplastic treatment. An independent quality control analyst reviewed each step of the process. Analyses were conducted using only the data from patients with data available; no other data imputation was carried out. Patients with data missing were reported for each study variable.

3. Results

3.1. Patient population

Overall, 823 patients with ovarian cancer from Argentina and 3570 patients from Brazil were identified, of which 741 and 2529 patients with ovarian cancer, respectively, were included in this study (Figure 1). Of the eligible populations, four (<1%) patients in the Argentina cohort had fallopian tube cancer and seven (<1%) had primary peritoneal cancer, while in the Brazil cohort, 120 (4.7%) had peritoneum cancer and 27 (1.1%) had an unspecified cancer of the female genital organs. Identification of patients with subsequent treatment within the eligible populations is shown in Figure 1. Patients had a mean (SD) age of 58.7 (14.9) and 52.1 (15.7) years in the Argentinian and Brazilian cohorts, respectively (Table 1). Based on the FIGO staging system, the majority of patients from Argentina were classified as FIGO stage III (n = 321 [62.5%]; data unavailable for Brazil).

Figure 1.

Flow diagrams of attrition in the Argentinian (A) and Brazilian (B) ovarian cancer cohorts.

^†See Supplementary Table 1 for a full list of health codes.

*ICD-10 codes were C56 (malignant neoplasm of ovary), C57.0 (malignant neoplasm: fallopian tube), C48.1 (malignant neoplasm: specified parts of peritoneum), or C48.2 (malignant neoplasm: peritoneum, unspecified).

EMR: Electronic medical record; ICD: International Classification of Diseases; OC: Ovarian cancer.

Table 1.

Demographics and clinical characteristics at index.

Category	Argentina (n = 741)	Brazil (n = 2529)
Age at index, years, mean (SD)	58.7 (14.9)	52.1 (15.7)
Age group at index, n (%)   18–39   40–49   50–59   60–69   ≥70   Missing	88 (11.9) 116 (15.7) 174 (23.5) 194 (26.2) 169 (22.8) 0	277 (27.5) 245 (24.3) 258 (25.6) 228 (22.6) 148 (14.7) 1373 (54.3)
Weight, kg, mean (SD)	68.1 (14.5)	–
BMI, kg/m2, mean (SD)	26.6 (5.4)	–
Index date, n (%)   2010   2011   2012   2013   2014   2015   2016   2017   2018   2019	73 (9.9) 59 (8.0) 63 (8.5) 84 (11.3) 84 (11.3) 105 (14.2) 96 (13.0) 67 (9.0) 71 (9.6) 39 (5.3)	– – – – – 680 (32.5) 516 (24.7) 476 (22.8) 419 (20.0) 438 (20.9)
FIGO staging, n (%) Yes   I   II   III   IV Not reported	514 (69.4) 91 (17.7) 46 (8.9) 321 (62.5) 56 (10.9) 188 (25.4)	– – – – – –
Borderline tumors	39 (5.2)	–

FIGO: International Federation of Gynecology and Obstetrics; SD: Standard deviation.

3.2. Treatments received during the study

Patients had a mean follow-up of 31.7 (SD: 30.6) and 20.5 (SD: 15.7) months and a median of 23.3 (IQR: 47.4) and 15.7 (IQR: 21.3) months in Argentina and Brazil, respectively (Table 2). Across both countries, most patients underwent surgery (Argentina: n = 496 [66.9%]; Brazil: n = 2286 [90.4%]) during the study. Around one-half of patients received antineoplastic agents (Argentina: n = 355 [48.0%]; Brazil: n = 1441 [57.0%]) and a small proportion of patients underwent radiotherapy either alone or in conjunction with other therapies (Argentina: n = 18 [2.4%]; Brazil: n = 86 [3.4%]). Among patients in Argentina receiving antineoplastic treatment, 321 (90.4%) received platinum-based chemotherapy at any point during the study; in Brazil, the corresponding proportion was 1253 (87.0%). Oxaliplatin was used by <5% of patients in Argentina and Brazil. Patients who received platinum-based chemotherapy received it during their first year following diagnosis.

Table 2.

Treatments received during the study period.

	Argentina (n = 741)	Brazil (n = 2529)
Follow-up, months, median (IQR)	23.3 (47.4)	15.7 (21.3)
Treatments received, n (%)   Surgery   Antineoplastic agents   Radiotherapy	496 (66.9) 355 (48.0) 18 (2.4)	2286 (90.4) 1441 (57.0) 86 (3.4)
Patients receiving each line of antineoplastic therapy, n (%)   1L   2L   3L   4L	355 (47.9) 187 (25.2) 121 (16.3) 77 (10.4)	1441 (57.0) 593 (23.4) 271 (10.7) 120 (4.7)
Time from diagnosis to 1L antineoplastic therapy, months, median (IQR)	1.9 (2.1)	2.3 (3.9)

1L: First-line; 2L: Second-line; 3L: Third-line; 4L: Fourth-line; IQR: Interquartile range.

The number of patients receiving each line of antineoplastic therapy decreased with successive lines. The mean (SD) time from diagnosis to the 1L of any antineoplastic therapy in Argentina and Brazil was 4.3 (9.1) and 3.3 (4.0) months, respectively.

3.3. Antineoplastic treatment selection by treatment line

Among patients receiving antineoplastic therapies, platinum-based chemotherapy was the most common 1L therapy (Argentina n = 311 [87.6%]; Brazil n = 1142 [79.2%]; Figure 2). The proportion of patients receiving platinum-based chemotherapy declined notably between the 1L and 2L, dropping to 47.6% (n = 89) in Argentina and dropping to 55.7% (n = 330) in Brazil. This trend for decline continued between the 2L and 4L setting, where the proportion of patients receiving platinum-based chemotherapy decreased to 37.7% (n = 29) in the 4L setting in Argentina and to 33.3% (n = 40) in the 4L setting in Brazil.

Figure 2.

Antineoplastic therapy use by therapeutic class in ovarian cancer in Argentina (n = 355) (A) and Brazil (n = 1441) (B).

Each line of therapy does not equate to 100%, as patients may have been treated with more than one class of antineoplastic treatment in each line of therapy; additionally, a small percentage of patients (<3%) who received combinations of immunotherapy/hormone therapy and targeted therapy were not included. Any agent in a combination regimen with a platinum chemotherapy agent was classified as platinum-based chemotherapy.

*Immunotherapy/targeted therapy/hormone therapy (monotherapy or combination) includes combinations of hormone therapy, immunotherapy, and targeted therapy but does not include combinations with platinum-based or non–platinum-based chemotherapy.

L: Line of therapy.

Treatment patterns in the first through fourth lines showed that a combination of platinum-based and non–platinum-based therapies was the most frequently used therapy class in the 1L setting but was less common in subsequent lines of therapy in both cohorts (Figure 3). Across treatment lines, an increase in the proportion of patients receiving non–platinum-based chemotherapy, immunotherapy/targeted therapy, and hormone therapies was observed (Figure 2 & Supplementary Tables S2 & S3). The greatest increase in the use of non–platinum-based chemotherapy occurred between the 1L and 2L settings (rising from 8.7% [n = 31] to 35.3% [n = 66] in Argentina and 12.6% [n = 181] to 35.4% [n = 210] in Brazil). Use of targeted therapy/immunotherapy in the 2L–4L settings was more common in Brazil than in Argentina, particularly in the 2L and third-line (3L) settings.

Figure 3.

Treatment classes used to treat ovarian cancer in first-line to fourth-line therapy in Argentina (A) and Brazil (B).

HT: Hormone therapy; IO: Immunotherapy; NPT: Non–platinum-based chemotherapy; PT: Platinum-based chemotherapy; TT: Targeted therapy.

Patients received monotherapy and a combination of antineoplastic therapies across all four lines of therapy (Figure 4). Across both countries, drug combinations, mostly platinum doublets, were most commonly received in the 1L setting. The most frequent combination in the 1L setting was carboplatin and paclitaxel (Argentina: n = 178 [50.1%]; Brazil: n = 639 [44.3%]; Supplementary Tables S4 & S5). In the 2L setting, a similar number of patients in Argentina received either monotherapy or combination therapy, while a majority of patients in Brazil received combination therapy. Carboplatin plus paclitaxel remained the most common 2L antineoplastic therapy in Argentina (n = 21 [11.2%]) alongside doxorubicin (n = 19 [10.2%]) and tamoxifen (n = 16 [8.6%]). In Brazil, doxorubicin (n = 67 [11.3%]), carboplatin plus doxorubicin (n = 63 [10.6%]), and carboplatin plus paclitaxel (n = 60 [10.1%]) were the most frequently used regimens. At the 3L and 4L settings, the majority of patients in both Argentina (n = 77 [63.6%]; n = 43 [55.8%]) and Brazil (n = 114 [42.1%]; n = 64 [53.3%]) received monotherapy. The most common 3L therapies were doxorubicin in Argentina (n = 14 [11.6%]) and gemcitabine in Brazil (n = 28 [10.3%]). In the 4L setting, the most common therapies were carboplatin in Argentina (n = 8 [10.4%]) and paclitaxel in Brazil (n = 16 [13.3%]). The number of patients who received triplet therapy declined across the 1L–4L setting in Argentina (n = 68 [19.2%] to n = 4 [5.2%]) and across the 2L–4L setting in Brazil (n = 123 [20.7%] to n = 16 [13.3%]).

Figure 4.

Antineoplastic therapy use by regimen type in ovarian cancer in Argentina (n = 355) (A) and Brazil (n = 1441) (B).

Each line of therapy does not equate to 100%, as patients receiving a regimen consisting of quadruplet agents (<1%) were not included.

L: Line of therapy.

3.4. Time to next treatment

The TTNT between one line of any antineoplastic agent to the next line of any antineoplastic agent was longest between the 1L and 2L, with a median of 5.3 months (IQR: 10.7 months) and 3.0 months (IQR: 6.8 months) in patients from Argentina and Brazil, respectively. Overall, TTNT decreased gradually across treatment lines in both countries (3L–4L 1.4 [IQR: 5.2] months in Argentina and 1.3 [IQR: 2.5] in Brazil; Figure 5).

Figure 5.

Time to next treatment between one line of any antineoplastic agent and the next line of any antineoplastic agent for ovarian cancer in Argentina (n = 355) (A) and Brazil (n = 1441) (B).

IQR: Interquartile range; L: Line of therapy; P25: 25th percentile; P75: 75th percentile.

Median PFS, measured as the time between the end of the first platinum-based line of therapy to the next platinum- or non–platinum-based chemotherapy, surgery, or radiotherapy was 6.0 months (IQR: 11.3 months) for Argentina and 3.6 months (IQR: 9.6 months) for Brazil.

3.5. Duration of treatment

The DOT for any antineoplastic therapy was the longest in the 1L setting (Argentina: 4.4 [IQR: 3.3] months; Brazil: 3.5 [IQR: 3.5] months) and the shortest in the 4L setting (Argentina: 2.3 [IQR: 3.4] months; Brazil: 2.0 [IQR: 3.2] months). In both countries, the median DOT per line decreased between the 1L and 4L treatment settings (Argentina: 4.4 [IQR: 3.3] to 2.3 [IQR: 3.4] months; Brazil: 3.5 [IQR: 3.5] to 2.0 [IQR: 3.2] months). However, in Brazil, the median DOT for antineoplastic treatment was similar in the 2L and 3L settings (2L: 2.8 [IQR: 4.5] months and 3L: 2.8 [IQR: 4.6] months; Figure 6).

Figure 6.

Duration of antineoplastic treatment for ovarian cancer in Argentina (n = 355) (A) and Brazil (n = 1441) (B).

IQR: Interquartile range; L: Line of therapy; P25: 25th percentile; P75: 75th percentile.

4. Discussion

The results of this study highlight the unmet need for treatment in patients with OC in Argentina and Brazil, as only approximately 50% of the patients received any antineoplastic agents after diagnosis, despite more than 70% being considered advanced cases (disease stage data available for Argentina only). In patients receiving antineoplastic treatment, platinum-based chemotherapy was most commonly used in the 1L setting in both countries, but its usage decreased in the 2L–4L treatment settings. Exposure to non–platinum-based chemotherapy, immunotherapy/targeted therapy, and hormone therapy generally increased as the lines of therapy progressed. This is to be expected, given the gradual development of platinum resistance and the practice of introducing other classes of drugs in later lines.

A similar study using medical and pharmacy claims databases described the treatment patterns and survival of >12,000 patients with advanced OC in the USA and reported a higher percentage of patients (∼70%) receiving antineoplastic agents (including radiotherapy) after primary surgery [25]. The most commonly used therapies in the 1L setting were carboplatin (51%) and paclitaxel (50%), which are in line with what were reported for Argentina (carboplatin/paclitaxel: 50%) and Brazil (carboplatin/paclitaxel: 44%) [25]. Similar observations were observed in this study where a decline in platinum-based chemotherapy use was also evident up to the 3L treatment setting; the largest decrease was between the 1L and 2L settings (51 to 36%) [25].

TTNT has been shown to correlate with PFS and is considered a real-world surrogate for disease progression [26,27]. Median PFS, assessed as time between the end of the first platinum-based line of therapy to the next platinum- or non–platinum-based chemotherapy, surgery, or radiotherapy, was relatively short in both Argentina (6.0 months) and Brazil (3.6 months). The observed PFS rates in this study are in line with PFS rates reported in clinical trials of chemotherapy monotherapy (median PFS: 6–9 months) [28,29] and are somewhat lower than PFS rates for clinical trials of chemotherapy combinations; in the GOG-0218 trial of newly diagnosed advanced OC, median PFS for the carboplatin and paclitaxel arm was 10 months [28]. Given that combination therapy with carboplatin plus paclitaxel was the most common 1L regimen in both Argentina and Brazil, these findings indicate that while standard-of-care treatments for patients continue to be chemotherapy-based, long-term survival outcomes of most chemotherapy-only regimens are poor and highlight the need for new effective therapies in patients with OC [30]. Finally, DOT by line of therapy also decreased during the follow-up period, potentially due to disease progression, acquired resistance, or reduced effectiveness. It may also be reflective of a change in antineoplastic therapy and differences in treatment duration and cycles of treatment (e.g., platinum-based chemotherapy is typically administered for three to six cycles every 3–4 weeks, while targeted therapies such as bevacizumab can be administered every 2–3 weeks for 15 months) [31,32].

Economic status is a major factor when considering patient access to novel targeted therapies in Argentina and Brazil. In comparison with higher-economic countries, many low- and middle-economic countries have limited access to next-generation antineoplastic agents. However, access to high-cost drugs is better in private insurance systems within Argentina and Brazil [10,33]. Since most individuals in these countries rely on public-funded healthcare, the results of this study, which utilized data from private healthcare databases, may not be generalizable to the wider population and remains a factor restricting the comparison with populations with a lower socioeconomic status. In addition, interpretation of results from Argentina and Brazil should factor the specific populations that each database represents and local clinical practice; thus, comparisons between countries should be made with caution. Although the populations of Argentina and Brazil make up a significant proportion of Latin America, the results cannot be extrapolated into a regionwide estimate, since the data are not completely representative of all healthcare settings, owing to the limitations of this study.

In terms of study design, missing data are a common issue in administrative claims databases and medical records, which can yield biased results if the data were not missing at random and not dealt with appropriately in the study [34]. The use of ICD-10 codes in Brazil's database (Orizon) permits easy identification of OC patients; however, health claims may be inaccurately coded, leading to misclassification bias [34,35]. Additionally, records of any treatment received by patients in treatment centers other than HIMCP/Orizon would not have been documented in either database and is therefore unknown. This may have led to an overestimation of undertreatment levels; however, given that both databases are from private settings, treatment records are likely to be accurate, as this information is required for reimbursement purposes. Another limitation of this study is the lack of histological subtyping and incomplete stage data available for patients. As such, analysis of treatment by disease stage was not possible. Despite these limitations, real-world data from health claims databases, such as Orizon and HIBA, provide supportive evidence to clinical trial data using retrospective analyses of large, more diverse populations of patients, which is not possible with randomized, controlled trials [34,35]. The single methodology used to analyze the data from the private healthcare databases across Argentina and Brazil demonstrates the consistency of this approach and its applicability to different countries and databases, standardizing the evidence generation approach for future real-world observational studies.

Overall, this study provides insights into the burden of OC in Argentina and Brazil and generates real-world evidence of antineoplastic treatment patterns across multiple lines of therapy in these countries. Importantly, this study highlights the missed opportunity for treatment for these patients, given the high proportion of patients who received no antineoplastic treatment, which likely contributes to the increased mortality-to-incidence ratio for patients with OC in Latin America compared with other countries. Increased awareness of OC and improved screening for earlier diagnosis are crucial to improve patient outcomes. Optimizing evidence-based use of antineoplastic therapies, including novel agents, may also help prolong the time to disease recurrence. Currently, several phase III trials are under way to explore new approaches in OC. These include the combination of immunotherapies and targeted therapies for 1L treatment in combination with platinum-based chemotherapy (e.g., FIRST, KEYLYNK-001) [36,37] and in platinum-resistant disease (e.g., EPIK-O, KEYNOTE-B96, NRG-GY005) [38–40], which could enable the addition of alternative treatment strategies to the OC landscape. Therefore, it will be important for future real-world observation studies to assess the treatment patterns of novel therapies/combinations for patients with OC in Latin America, in order to support improvements in treatment decision-making and survival outcomes.

5. Conclusion

Almost half of the patients with OC in Argentina and Brazil did not receive any antineoplastic treatment after diagnosis, highlighting an undertreatment issue in this patient population. In patients who received antineoplastic treatment, platinum-based chemotherapy was the most common therapy in the 1L setting. Exposure to non–platinum-based chemotherapy, targeted/immunotherapy, and hormone therapy increased after the 1L setting. This study highlights the unmet need for effective therapies to optimize outcomes in patients in Argentina and Brazil with OC, particularly those with advanced disease.

Funding Statement

This study was funded by GSK (study 213815).

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/14796694.2024.2343650

Author contributions

G Abreu, J Queiroz, TLN da Silva, and C Soares contributed to conception and/or design, acquisition of data, and data analysis and/or interpretation. P Menezes and ALAR de Souza were involved in acquisition of data and data analysis and/or interpretation. T Pires, M Carrizo, MC Riggi, and F Cravero contributed to conception and/or design and data analysis and/or interpretation. G Bernardino, R Felice, and L Jotimliansky were involved in data analysis and/or interpretation.

Financial disclosure

This study was funded by GSK (study 213815). The authors have no other relevant affiliationsor financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Competing interests disclosure

C Soares, P Menezes, M Carrizo, R Felice, G Bernardino, and L Jotimliansky are employees of and hold stocks in GSK. G Abreu, J Queiroz, TLN da Silva, and T Pires are employees of GSK. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing disclosure

Medical writing support was provided by Priyanka Vaz and Claire Kelly of Fishawack Indicia, (part of Avalere Health) and funded by GSK.

Ethical conduct of research

There was no direct contact with the patients or primary data collection of individual patient data. This study was evaluated and approved (IRB00010193) by the Ethics Committee of the Hospital Italiano in Argentina to extract the anonymous database. Owing to the anonymous data collection of medical records of patients in which the confidentiality of the information was handled according to the local laws of protection of personal data and according to resolution 1480/11 of the Argentine National Ministry of Health, it was not necessary to obtain informed consent from patients.