BRCA screening, treatment patterns and response among patients with ovarian cancer in the second line treatment setting: results from a real world survey

Objective : Although BRCA gene testing of patients diagnosed with ovarian cancer (OC) is now recommended, the level of testing undertaken in various countries is largely unknown. This study describes BRCA mutation screening patterns and results, demographics, clinical characteristics and the use of poly (ADP-ribose) polymerase inhibitors (PARPi) as maintenance treatment within a real-world sample of patients with advanced OC in the second-line (2L) treatment setting. Methods : Data were collected using the Adelphi Real World OC Disease Specific Programme (DSP)™, a point-in-time survey of physicians and their consulting patients with OC in clinical practice, undertaken across Europe (EU5: France, Germany, Italy, Spain and United Kingdom) and the United States (US) between December 2017–March 2018. Physicians completed a detailed patient record form for their next eight consecutively consulting patients, capturing data on their clinical history and treatment. All analysis was descriptive. Results : Of 1315 patients identified, 1096 (83%) were receiving 2L treatment and 219 (17%) were receiving 2L maintenance treatment; either PARPi (olaparib, rucaparib or niraparib, n = 103) or a non-PARPi (n = 116). BRCA screening rates varied between countries, increased with each line of therapy and were higher in the EU5 (55%) than the US (44%). 28% of patients receiving 2L treatment had a BRCA1/2 mutation. Patients receiving PARPi maintenance treatment had better Eastern Cooperative Oncology Group (ECOG) performance status, higher BRCA screening rates and higher proportions of serous epithelial OC than those receiving 2L treatment or non-PARPi maintenance. Common reasons for choosing 2L treatments were progression-free/overall survival benefit and improvement of quality of life. Conclusions : Despite guidelines recommending BRCA testing in patients with OC, many OC patients undergoing 2L treatment were not screened for BRCA mutations. Decisions related to PARPi use in 2L maintenance appeared to be driven by BRCA status, histology and response to first-line treatment.


Introduction
Ovarian cancer (OC) most commonly affects older, postmenopausal women, with >80% of women diagnosed with OC being >50 years old [1]. Globally, OC is the fifth most common type of cancer in women and the fourth most common cause of cancer death in women [1]. Worldwide, there were an estimated 314,000 new cases of OC and 207,000 OC-related deaths in 2020 [2].
BRCA mutation status is important when considering treatment options in advanced OC, particularly in those patients with a familial history of breast and OC [3]. Around 20% of high-grade serous OCs harbour germline or somatic BRCA mutations [4]. Although BRCA testing of patients diagnosed with OC is recommended [5,6], the level of testing undertaken in various countries is largely unknown.
The recommended first-line (1L) treatment for advanced OC in 2018 was surgical cytoreduction followed by carboplatin/paclitaxel chemotherapy [5,7]. However, dis-ease progression occurs in around 70% of patients within 3 years of receiving treatment, with the probability of response to second-line (2L) and subsequent lines of therapy depending largely on the progression-free period during or following the previous course of chemotherapy [1]. A further course of chemotherapy, usually single-agent carboplatin, may be prescribed as 2L therapy in patients who are platinum-sensitive at relapse. Treatment options for platinum-resistant or platinum-refractory patients are more limited; these patients usually receive non-platinum chemotherapy.
In recent years, several poly (ADP-ribose) polymerase inhibitors (PARPi) have received approval in Europe and the United States (US) [8], with the National Comprehensive Cancer Network guidelines recommending considering maintenance therapy with a PARPi in platinumsensitive patients with relapsed OC, regardless of BRCA mutation status [6].
In the NOVA trial, the PRIMA trial and the QUADRA study, niraparib was shown to improve progression-free survival (PFS) in advanced OC across a broad range of patient groups [9][10][11]. In the ARIEL3 trial, rucaparib significantly improved PFS in platinum sensitive, recurrent (PSR) patients, indicating its utility as maintenance treatment in platinum-sensitive OC following a complete or partial response to 2L or later platinum-based chemotherapy [12]. Both rucaparib and niraparib have been approved for treatment of OC across a range of clinical settings in both the US and Europe [7].
The PSR setting has become an important focus of research. Understanding current BRCA screening rates in OC on a wider geographical scale, and how these relate to real-world treatment patterns in the context of an increasing variety of newly-introduced therapies could provide interesting insights. Therefore, the objective of this study was to describe BRCA screening patterns and results, patient demographics and clinical characteristics, the use of PARPi as maintenance treatment and subsequent treatment response within a real-world sample of patients with advanced OC in the 2L treatment setting.

Study design
Data were collected using the Adelphi Real World Advanced OC Disease Specific Programme (DSP)™ in Europe (EU5 -France, Germany, Italy, Spain and United Kingdom (UK)) and the US between December 2017 and March 2018. DSPs are large, point-in-time, noninterventional multi-country surveys of physicians and their consulting patients conducted in real-world clinical practice. This methodological approach is well established and has been used extensively across multiple therapy areas. The full DSP methodology has been previously published in detail [18], has been validated against external data sources [19], and has demonstrated the power of trend data over time [20].
The DSP collects data from physicians, including perceptions and workload using physician surveys, and data related to individual patients via patient record forms. Data collected includes both subjective (opinion-based) and objective variables, clinical information about individual patients, their disease and treatment.

Ethics approval and consent to participate
Data collection was undertaken in line with European Pharmaceutical Marketing Research Association [21] guidelines, and as such did not require ethics committee approval. Each survey was performed in full accordance with relevant legislation at the time of data collection, including the US Health Insurance Portability and Accountability Act 1996 [22], and Health Information Technology for Economic and Clinical Health Act legislation [23], as well as in accordance with the Declaration of Helsinki. Using a check box, physicians and patients provided informed consent to take part in the survey. Data were collected in such a way that patients and physicians could not be identified directly; all data were aggregated and de-identified before receipt.

Data collection
A geographically diverse sample of physicians were recruited by field-based interviewers, and were identified from publicly available lists of healthcare professionals. Physicians were eligible to participate if they qualified as a medical/clinical oncologist or gynecologist between 1983 and 2013, saw at least 10 patients with advanced OC in a typical month and were personally responsible for treatment decisions for patients with advanced OC. Physician participation was financially incentivized, with reimbursement upon survey completion according to fair market research rates.
Once recruited to the study, participating physicians completed a detailed electronic patient record form for their next eight consecutively consulting patients who met the eligibility criteria, to mitigate against recruitment bias. Data were recorded at time of consultation, to mitigate against recall bias. All patients were ≥18 years old, with histologically confirmed epithelial ovarian, fallopian tube or peritoneal cancer (including malignant Müllerian tumors with high grade serous component) at stage II-IV OC and receiving systemic treatment at the point of data collection.
Of the eight consecutive patients included in the study, physicians were asked to ensure that the proportion of patients recruited met the following criteria; three patients were receiving 1L treatment (consolidation or maintenance) at point of data abstraction, one patient was receiving 2L or later treatment at point of data abstraction having previously received bevacizumab at 1L maintenance, and four patients were known to be on 2L or later treatment at data abstraction having received a platinum-based regimen at 1L. Physicians were asked to recruit the next sequential patient who met the various inclusion criteria above, in order to recruit as broad a patient cohort as possible whilst minimizing selection bias. Consulting patients were enrolled in the study, therefore alive at the point of data capture. As a result, the DSP did not capture survival data.

Study variables
Information captured in the patient record form included demographics, clinical characteristics, BRCA testing details and findings (BRCA testing, either germline or somatic, undertaken at any time point), current treatment for OC and treatment history (including reasons for treatment choices and treatment response). Physicians were asked to record whether patients were complete responders, partial responders or had stable disease following completion of treatment regimen, as well as recording any reasons for treatment failure.
Patients were classified as platinum-sensitive if disease progression was noted >6 months after 1L platinum treatment, platinum-resistant if disease progression was noted 0-6 months after 1L platinum treatment or platinumrefractory if disease progression was noted during 1L platinum treatment.

Analysis
Patients were included in the main analysis if they were receiving 2L treatment at the time of data collection, including 2L maintenance treatment. All analyses were descriptive, with statistics calculated for patient demographics, clinical characteristics, treatment patterns, 1L treatment response and reasons for choice of 2L maintenance therapy. Mean, standard deviation and range were calculated for continuous variables, and frequency counts and percentages for categorical variables. All analyses were conducted in Survey Reporter.
Just over three-quarters of patients (76%) were from the EU5, with the largest EU sample from Germany (17%) and the smallest from the UK (12%), and around a quarter of patients were from the US (24%). Patients were of similar age (mean age ± standard deviation (SD): 63.2 ± 9.6 years), regardless of the 2L treatment regimen they were receiving (Table 1).
The most common OC histology in all treatment groups was serous epithelial OC (overall, 63%). There were differences across treatment groups in the proportions of patients with each histology, with the highest proportion of patients receiving PARPi maintenance treatment having serous epithelial OC (Table 1). A higher proportion of patients receiving PARPi maintenance treatment had a family history of OC (27%) compared to patients from the other two treatment groups (2L treatment, 8%; non-PARPi maintenance treatment, 9%) ( Table 1).

Treatment at 1L setting
Of all patients receiving 2L treatment at the time of data collection, the majority (95%) of patients had received platinum-based chemotherapy in the 1L treatment setting; almost all (99%) patients receiving PARPi at 2L maintenance received a platinum-containing regimen at 1L compared with around three-quarters (72%) of patients who were receiving 2L maintenance with a treatment other than a PARPi (Table 2).
1L maintenance treatment was received by just under half (44%) of all patients undergoing 2L treatment, with the most common maintenance treatment overall being a bevacizumab-containing regimen (87%), although this was only received by just over half of the patients (57%) receiving non-PARPi 2L maintenance treatment. At 1L treatment stage, 28 (5%) patients received a PARPi-based maintenance treatment, with all but three of these receiving olaparib (Table 2).

1L treatment response
The most common response to 1L treatment across patients at 2L treatment stage in both the EU5 and US was completion of the treatment course with the patient showing a complete response to 1L treatment. Patients who were receiving a PARPi at 2L maintenance included a higher proportion of patients showing complete response to 1L treatment (EU5: 79%, US: 55%) compared to those receiving non-PARPi 2L maintenance treatment (EU5: 33%, US: 7%) ( Table 3).
For those patients receiving 2L maintenance treatment with therapy other than a PARPi, the most common response to 1L treatment was partial response on completion of the treatment course (Table 4). Country differences were observed between patients from the EU5 and US in 1L treatment responses, with 93% and 51% of EU5 and US patients, respectively, who were receiving a non-PARPi maintenance      treatment at 2L, having a complete/partial response or stable disease following 1L treatment (Table 3). Overall, 35% of all patients included in this analysis were defined as platinum-sensitive, 14% as platinum-resistant and 14% as platinum-refractory. Platinum sensitivity differed markedly between patients receiving PARPi and non-PARPi maintenance treatment at 2L, with 53% and 16% of these patients classified as platinum-sensitive, respectively.

2L treatment and reasons for choice of 2L maintenance treatment
Of patients receiving 2L non-maintenance treatment (n = 1096), at the time of data collection, 57% of patients were receiving a platinum-containing regimen, compared with 85% of patients receiving PARPi-based 2L maintenance treatment (n = 103) ( Table 2).
When physicians were asked to record their reasons for treatment choices at different lines of treatment, the most common reasons for choosing maintenance treatment were progression-free/overall survival benefit (60%/40%) and maintenance or improvement of quality of life (32% and 19%). The proportions of physicians indicating reasons for their choice of 2L maintenance treatment were generally similar between PARPi-containing and non-PARPi 2L regimens (predominantly either platinum-based chemotherapy or bevacizumab, as summarized in Table 2); however, differences were observed for PFS benefit (69% and 53%, respectively), manageable side effect profile (42% and 21%, respectively) and high response rate (35% and 24%, respectively) (data not shown).

BRCA screening rates and test results in patients at 2L treatment stage
Of all 2L patients (n = 1315), 692 patients had their BRCA status tested, with screening rates slightly higher in the EU5 than the US (55% vs 44%, respectively). Higher proportions of patients from both the EU5 and the US receiving a PARPi at 2L maintenance were tested for BRCA (87% and 76%, respectively) compared to those who were receiving 2L treatment (54% and 43%, respectively) or 2L maintenance that was not PARPi-based (38% and 23%, respectively) ( Table 4). Of all 2L tested patients (n = 692), 28% of patients receiving 2L treatment tested positive for a BRCA1/2 mutation, while test results for those receiving maintenance treatment varied between treatments and regions. Of those tested for BRCA and who were receiving PARPi 2L maintenance treatment (n = 103), positive results were obtained for 80% and 44% of EU5 and US patients, respectively. Of those tested for BRCA and who were receiving non-PARPi 2L maintenance treatment (n = 106), 11% and 100% of EU5 and US patients, respectively, had a positive test (Table 4). Of 2L maintenance patients receiving a PARPi, 27% of patients had a family history of OC and 93% of these patients had a positive BRCA status.

BRCA screening rates and test results by line of therapy at data collection and country (all patients)
When the BRCA screening rates by line of therapy and country were examined, it was found that the BRCA screening rate increased as the line of therapy progressed from 1L to 2L and to third-line treatment and beyond (3L+) ( Table 5). The highest rates of testing were observed in Spain (60% at 1L and 66% at 2L). Both Italy (39% at both 1L and 2L) and the UK (24% at 1L and 40% at 2L) were below the EU5 average (45% at 1L and 55% and 2L, respectively). When comparing the EU5 and the US, there was little difference in BRCA testing rates at 1L (45% vs 46%, respectively); however, a greater proportion of EU5 patients were tested at 2L compared to the US cohort (55% vs 44%, respectively).

Discussion
In this real-world study, we explored demographics, clinical characteristics, BRCA screening patterns and results, treatment patterns and response in a sample of 1315 patients with advanced OC receiving 2L treatment from five European countries and the US. At the time this study was undertaken (Dec 2017-Mar 2018), the recommended 1L treatment for advanced OC was surgical cytoreduction followed by carboplatin/paclitaxel chemotherapy [5,7], with PARPi maintenance therapies only recently being approved in certain settings. Compared to patients receiving non-PARPi-based 2L maintenance treatment (n = 116), the 103 patients receiving 2L maintenance treatment containing one of the approved PARPi therapies (olaparib, rucaparib or niraparib) included in our study had a more advanced stage of OC. Those receiving PARPi more frequently presented with serous epithelial OC and with a family history of OC, and were more frequently tested for, and had, BRCA mutations; had received platinum-based chemotherapy at 1L and 2L treatment stage; and were classified as platinum-sensitive.
PARPi have been shown to improve PFS in PSR OC, as demonstrated in clinical trials for niraparib [9][10][11], rucaparib [12] and olaparib [14][15][16], and are of particular benefit in patients with BRCA mutations who are at higher risk of developing OC [24], with the combination of PARPi in patients with BRCA1/2 deficiency resulting in tumour cell death [25]. In this study, a higher proportion of patients from both the EU5 and the US receiving a PARPi at 2L maintenance were tested for BRCA (87% vs 76%, respectively) compared to those who were receiving 2L treatment (54% vs 43%, respectively) or non-PARPi-based or 2L maintenance (38% vs 23%, respectively). Of those tested for BRCA and who were receiving PARPi 2L maintenance treatment (n = 103), positive results were obtained for 80% and 44% of EU5 and US patients, respectively. Of those tested for BRCA and who were receiving non-PARPi 2L maintenance treatment (n = 106), 11% and 100% of EU5 and US patients, respectively, had a positive test. Tested, n (%) 119 (67) 29 (78) 18 (64) 21 (60) 11 (69) 22 (73) 101 (69) 18 (58) Not tested, n (%) 58 (33) 8 (22) 10 (36) 14 (40) 5 (31) 8 (27) 45 (31) 13 (42) Of those tested -Positive, n (%) 42 (35) 9 (31) 9 (50) 7 (33) 1 (9) 8 (36) 34 (34)  8 (44) Of those tested -Negative, n (%) 72 (61) 17 (59) 8 (44) 14 (67) 9 (82) 14 (64) 62 (61) 10 (56) Of those tested -Results unknown, n (%) 5 (4) 3 (10) 1 (6) -1 (9) -5 (5) -Percentages rounded to nearest whole number, therefore percentages may not add up to 100%. BRCA testing can inform treatment decisions in OC. While it is well established that carriers of BRCA mutations are at increased risk of OC, there is also evidence for better prognosis in OC patients who carry germline BRCA mutations compared with those without [26]. However, as this study was a point-in-time survey, we did not capture information relating to long-term survival and prognosis. It has been previously reported that only one-third of 6,001 women diagnosed with OC in California and Georgia between 2013 and 2014 underwent genetic testing [27]. In this study, we found that BRCA screening rates were higher than previously reported; with rates slightly higher in the EU5 than the US (55% vs 44%, respectively). We also found that the BRCA screening rate increased as the line of therapy progressed from 1L to 2L (45% vs 53%, respectively) and 28% of patients receiving 2L treatment tested positive for a BRCA1/2 mutation, in line with the values previously reported in the literature [4]. In 2L maintenance patients receiving a PARPi, 27% of patients had a family history of OC, and of these, 93% of patients had a positive BRCA status. BRCA testing of all patients diagnosed with OC is now recommended in clinical practice [5,6]. It is important to acknowledge the role that financial burden may play in relation to BRCA testing rates and choice of 2L maintenance treatment. While the cost of BRCA testing may be relatively inexpensive, treatment options available at 2L and beyond may not be accessible or affordable to patients without insurance coverage, or where reimbursement is onerous. While this is not expected to be a major factor in Europe, In the US where insurance coverage status can vary, physicians and their consulting patients may decide not to undertake BRCA testing, despite updated clinical guidance, due to the potential cost implications for the patient with respect to future treatment options.
These findings are of particular relevance as they demonstrate the issue of testing inertia. Despite updated testing recommendations and the availability of new treatment regimens, there is a lag in physician uptake of these principles, indicating the importance of physician education surrounding new clinical practices. With further advances in biomarker-driven treatments, access to testing needs to be improved in order to enable patients to benefit from the most effective and appropriate treatments. As acceptance of this guidance filters down to clinical practice and treatment options in advanced OC develop, it would be expected that the screening rates reported in this study will increase further, and should be monitored closely in future studies.
In addition to testing for BRCA mutations, recent research has suggested HRD+ as an alternative biomarker for OC that might benefit from PARPi treatment. Homologous recombination is a DNA repair pathway, and approximately 13% of OCs harbor HRD+ attributable to germline BRCA1/2 mutations and 50% harbor HRD+ in the absence of germline BRCA1/2 mutations [28,29]. Further investigations to verify if an HRD+ test may be useful to select platinum-resistant tumors that may benefit from PARPi therapy are needed [29].
Our study had a number of limitations. The study was descriptive and not designed to formally compare treatment options. The sample was not a truly random sample but a convenience sample from physicians likely to practice in specialized centers. While minimal inclusion criteria governed the selection of the participating physicians, participation was influenced by willingness to complete the survey. Physicians are asked to provide data for a consecutive series of patients to avoid selection bias, but no formal patient selection verification procedures are in place. As data were only included for patients who consulted during the data collection period, patients who consulted more frequently might have been over-represented, and patients with very severe disease being treated as hospital inpatients not represented. Recall bias is a common limitation of surveys, however the data for these analyses were collected at the time of each patient's appointment which is expected to reduce the likelihood of this occurring. As with all studies of this type, findings depend on the accurate reporting of the participating physicians. Despite such limitations, realworld studies provide insights into current clinical practice. The methodological approach reported in this study is well established and has been used extensively across multiple therapy areas. The full DSP™ methodology has been previously published in detail [18], has been validated against external data sources [19], and has demonstrated the power of trend data over time [20].

Conclusions
In conclusion, our study showed low levels of testing for BRCA mutations in this population of patients with advanced OC receiving 2L treatment, despite recent guidelines recommending testing. BRCA screening rates varied between countries, increased with each line of therapy and were higher in the EU5 (55%) than the US (44%). PARPi are of benefit in OC patients with platinum-sensitive relapse, and in particular those with BRCA mutations. In this study, 28% of patients receiving 2L treatment had a BRCA1/2 mutation. 94% of patients who were receiving a PARPi in the 2L maintenance setting had a complete response, partial response or stable disease as a result of their 1L treatment. Patients who were receiving a PARPi in the 2L maintenance setting were more often tested for BRCA mutations compared to other treatment groups, and a higher proportion of those tested were positive, suggesting that decisions regarding PARPi use in the 2L maintenance setting were driven by patients' BRCA status, histology and response to 1L treatment.