Unmet need for patients with metastatic triple-negative breast cancer initiating first-line treatment: data from the prospective German tumor registry OPAL.

Introduction
Triple-negative breast cancer (TNBC) is a histopathological subtype of breast cancer (BC), defined by the absence of tumor cell expression of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 (HER2).1 TNBC accounts for ∼10%-15% of all BC cases, with a higher incidence among younger women.2 Compared with other BC subtypes, TNBC demonstrates a more aggressive tumor biology, which is associated with a high risk of developing distant metastases (especially visceral metastases), early recurrence, and a poor prognosis.2,3 A high proportion of patients diagnosed with TNBC, including those with initially localized disease, experience recurrence with distant metastasis,4,5 and the estimated 5-year relative survival rate for those patients is ∼15%.6
Given the lack of actionable targets, treatment options for TNBC are limited, as patients do not benefit from endocrine-based or anti-HER2-targeted therapies, available for other BC subtypes. Consequently, conventional chemotherapy (CTx) regimens, based on anthracyclines and taxanes, have been the only first-line (1L) treatment options for metastatic TNBC (mTNBC) for many years.2 However, with the approval of new therapeutic agents, the treatment landscape for mTNBC has evolved.
For patients whose tumors express programmed death-ligand 1 (PD-L1), 1L treatment with a PD-(L)1 inhibitor [PD-(L)1i] in combination with CTx is the preferred treatment strategy.7,8 In Europe, based on favorable outcomes observed in clinical trials, atezolizumab plus nab-paclitaxel (for PD-L1 immune cell score ≥1%)9 and pembrolizumab plus CTx (for PD-L1 combined positive score ≥10)10 were approved in 2019 and 2021, respectively.7,8 In addition, the poly (ADP-ribose) polymerase inhibitors (PARPi) olaparib and talazoparib, which have demonstrated improvements in progression-free survival (PFS) for patients with metastatic BC and germline breast cancer gene (BRCA1/2) mutation,11,12 now provide treatment options for mTNBC with such genetic alterations. In second (2L) and further lines, treatment options were expanded by antibody–drug conjugates (ADCs), such as sacituzumab govitecan13 as well as, in the case of HER2-low BC, trastuzumab deruxtecan.8,14
With the advent of these novel treatment approaches in recent years, it is of particular interest to examine their implementation in clinical routine and to observe the clinical outcomes for patients outside of randomized clinical trials. However, real-world data (RWD) addressing these questions are limited.
In this study, we present comprehensive RWD from the German tumor registry OPAL,15,16 focusing on patient characteristics, 1L treatment patterns, real-world PFS (rwPFS), and real-world overall survival (rwOS) of patients diagnosed with mTNBC who started their 1L treatment between January 2018 and August 2023.

Patients and methods

Study design and cohort definition
The OPAL registry, which was initiated in 2017, is a national, ongoing, open, non-interventional, prospective, longitudinal, multicenter clinical research platform focusing on changes in treatment reality and outcome of patients with early and advanced (locally advanced and/or metastatic) BC. Eligible patients were ≥18 years of age, diagnosed with histologically confirmed early or advanced BC, and had started their first systemic treatment. More than 2000 patients with metastatic BC have been enrolled by 189 study sites, including hospitals, medical care centers, and office-based oncologists/gynecologists located all over Germany. OPAL was approved by the responsible ethics committee and is registered at ClinicalTrials.gov (NCT03417115). Further details on the OPAL registry have been published previously.15,16 All aspects of study design, data collection, analysis, and reporting adhered to established standards for real-world evidence in oncology. This analysis followed the European Society for Medical Oncology (ESMO)-Guidelines for Real-World Data in Oncology (GROW) recommendations for RWD publications.17
All experiments comply with the current local laws and regulatory requirements to ensure the protection of patients’ personal data, including collection of patient informed consent. All procedures carried out in studies involving human participants were in accordance with the ethical standards of the national research committee and with the 1964 Helsinki declaration and its later amendments. The study protocol was reviewed and approved by the responsible ethics committee (F-2017-086, date of initial ethics approval: 6 December 2017; amendment F-2017-086#A1: 5 August 2019; amendment F-2017-086#A2: 22 April 2021). Written informed consent was obtained from all patients.
For the present analyses, female patients diagnosed with de novo or recurrent mTNBC who initiated their 1L treatment between January 2018 and August 2023 (index period) and whose last contact date was at least 1 day after the start of 1L treatment were included. Patients who started their 1L therapy outside the index period were not considered in this analysis, to ensure a minimum potential follow-up period of 16 months. Patients with clinical trial participation, treated with sacituzumab govitecan in 1L, or receiving endocrine-based treatment during metastatic setting and/or exclusively anti-HER2-targeted treatment during 1L were excluded. For patients to be considered evaluable, the following variables were required to be documented: year of birth and data on start of the initial treatment.
Patients were stratified according to PD-L1 status (as documented in clinical routine: PD-L1 positive, negative, or unknown). Furthermore, patients with PD-L1-positive mTNBC treated with PD-(L)1-based therapy in 1L [1L PD-(L)1i ± CTx] and patients within the PD-L1-negative subgroup receiving 1L CTx ± bevacizumab (1L CTx ± BEV) were analyzed individually.
Detailed information on patient and tumor characteristics, biomarker testing, (sequential) treatments, and clinical outcomes was collected. Patients were followed up until death or up to 5 years from start of 1L treatment. Database cut was on 31 December 2024.

Time-to-event analysis
Time-to-event analyses were conducted using the Kaplan–Meier method.18 rwPFS was defined as the interval between the start of 1L treatment and the date of first documented progression or death from any cause, whichever occurred first. Patients who started a 2L treatment without disease progression were censored at the start of 2L treatment. Patients without a PFS event and no 2L treatment were censored at last contact. rwOS was defined as the time from start of 1L treatment until date of death from any cause. Patients alive or lost to follow-up at database cut were censored at last contact. The complementary log–log transformation was used to calculate two-sided 95% confidence interval (CI) in the survival plots. Additional survival analyses including patients with endocrine-based treatment in later lines of therapy were carried out to exclude immortal time bias.

Statistical analysis
All analyses were calculated using R software, version 4.3.2.19

Results

Cohort description and tumor characteristics
Until database cut on 31 December 2024, a total of 448 evaluable patients diagnosed with mTNBC were recruited by 120 sites located across Germany into the OPAL registry. The present analysis included only patients who started 1L treatment between January 2018 and August 2023 (N = 406), were not participating in clinical trials (N = 391), were not treated with 1L sacituzumab govitecan (N = 387), and did not receive endocrine-based treatment during metastatic setting and/or exclusive anti-HER2-targeted treatment during 1L (N = 368) (Figure 1). Within this total cohort, 31.8% (n = 117/368) of patients were documented as PD-L1 positive, 35.9% (n = 132/368) as PD-L1 negative, and 32.3% (n = 119/368) had an unknown PD-L1 status. It should be noted that most of the patients whose tumors had unknown PD-L1 status started their 1L treatment before approval of PD-(L)1i. After PD-(L)1i approval, for most of the patients, a PD-L1 test result was documented (81.1%, n = 228/281).
The baseline patient and tumor characteristics of the total cohort as well as of the subgroups of patients based on PD-L1 status and 1L treatment are summarized in Table 1. In the total cohort, the median age at start of 1L treatment was 62.0 years, with 45.7% (n = 168) of patients being ≥65 years. A proportion of 10.6% (n = 39) of patients presented with an Eastern Cooperative Oncology Group (ECOG) performance status of ≥2. In 72.0% (n = 265) of patients, visceral metastases were documented at start of 1L. For 22.8% (n = 84) of patients, a BRCA1/2 test result was documented, and 4.3% (n = 16) of patients in the total cohort exhibited a BRCA1/2 mutation.
At initial BC diagnosis, 37.8% (n = 139) of patients had de novo metastatic disease. Details on patients with M0 at diagnosis (recurrent BC, 59.8%, n = 220) are shown in Table 2. Of these patients, 34.5% (n = 76/220) were initially diagnosed with a BC subtype other than TNBC, and the median time from initial BC diagnosis to mTNBC diagnosis was 39.8 months. The treatment-free interval was <6 months for 22.3% (n = 49/220) and ≥12 months for 55.9% (n = 123/220) of the patients with recurrent BC.

First-line treatment patterns
Most of the patients with PD-L1-positive tumors were treated with 1L PD-(L)1i ± CTx (81.2%, n = 95); of these, 88 patients received atezolizumab ± CTx and 7 received pembrolizumab ± CTx (Figure 2). A proportion of 13.7% (n = 16) of patients in the PD-L1-positive subgroup were treated with CTx monotherapy ± bevacizumab, and 5.1% (n = 6) of patients received CTx combination therapy ± bevacizumab.
Among patients with PD-L1-negative and PD-L1 unknown status, CTx monotherapy ± bevacizumab (59.1%, n = 78/71.4%, n = 85) followed by CTx combination therapy ± bevacizumab (26.5%, n = 35/26.1%, n = 31) were the most common treatment strategies. Anthracyclines were rarely used in 1L. A total of 12.9% (n = 17) of patients in the PD-L1-negative subgroup and 2.5% (n = 3) in the PD-L1 unknown subgroup received PD-(L)1i ± CTx.
PARPi and other therapies were rarely administered.

Second-line treatment
The percentage of patients receiving 2L treatment is shown in Figure 3. Patients marked as ‘2L treatment possible’ were either still in 1L therapy, had finished 1L but not started a 2L treatment, or had completed their observation period at database cut. These patients could potentially still receive a 2L treatment, while patients defined as ‘lost to follow-up’ (e.g. due to continuation of treatment in a hospice, by nursing service, or by primary care physician) will probably not receive a 2L therapy. This means, in the total cohort, 57.3% (n = 211/368) of the patients received 2L treatment, and 11.7% (n = 43/368) could potentially still receive a 2L treatment, while 25.8% (n = 95/368) of the patients died before 2L so far.
In the groups of patients based on PD-L1 status and 1L treatment, the percentage of patients who were treated in 2L therapy ranged from 49.6% to 65.5%, and between 14.7% and 28.6% of patients did not survive to receive 2L. However, the proportion of patients who could still receive a 2L treatment varies between 6.2% and 23.2%, resulting in comparable values for the maximum proportion of patients who potentially could be treated in 2L (66.4%-75.8%).

Clinical outcome
The median rwPFS from start of 1L therapy in the total cohort (Figure 4A) was 6.8 months (95% CI 5.9-7.6 months; 84.0% events), which was analogous across the different subgroups analyzed, with 7.3 months (95% CI 6.5-9.1 months; 76.9% events) in patients with PD-L1-positive tumors, 6.3 months (95% CI 5.0-7.6 months; 89.4% events) in patients with PD-L1-negative tumors, and 5.9 months (95% CI 4.6-8.0 months; 84.9% events) in patients whose tumors had an unknown PD-L1 status (Figure 4B). Patients with PD-L1-positive tumors who were treated with PD-(L)1i ± CTx in 1L had a median rwPFS of 7.2 months (95% CI 6.5-9.1 months; 73.7% events), and patients with PD-L1-negative tumors who received 1L CTx had a median rwPFS of 6.3 months (95% CI 4.9-7.6 months; 91.2% events), as depicted in Figure 4C and D, respectively.
Median rwOS from start of 1L therapy was 17.6 months (95% CI 15.6-19.7 months; 63.0% events) for the total cohort (Figure 4E). Patients with PD-L1-positive tumors had a median rwOS of 21.5 months (95% CI 16.4-24.8 months; 56.4% events), those with PD-L1-negative tumors had a median rwOS of 16.7 months (95% CI 14.0-20.7 months; 64.4% events), and those within the PD-L1 unknown subgroup had a median rwOS of 14.3 months (95% CI 12.1-19.7 months; 68.1% events), as shown in Figure 4F. For patients with PD-L1-positive tumors treated with PD-(L)1i ± CTx in 1L, a median rwOS of 23.2 months (95% CI 17.5-28.2 months; 50.5% events) was observed, and for patients with PD-L1-negative tumors who received 1L CTx, a median rwOS of 16.0 months (95% CI 13.9-20.7 months; 67.3% events) was observed (Figure 4G and H, respectively). It should be noted that these are descriptive, unadjusted survival analyses. Therefore, outcomes of the different subgroups should not be compared directly, as patients may differ in further baseline characteristics besides PD-L1 status and treatment received.
In our study cohort, patients with endocrine-based treatment during metastatic setting (n = 17) were excluded (Figure 1); of these, six patients received this type of treatment in later lines of therapy, but not in the 1L setting. To check for immortal time bias, outcome analyses including these patients were carried out, which revealed comparable results for rwPFS and rwOS.

Discussion
In recent years, the treatment landscape for TNBC in the 1L setting has evolved, offering targeted therapies for patients with PD-L1-positive and germline BRCA1/2-mutated tumors. But how many patients can benefit from these treatment approaches in clinical routine, and what are the current clinical outcomes in real world? In this study, we present baseline characteristics, 1L treatment patterns, and outcomes of patients with mTNBC treated in clinical routine in Germany between January 2018 and August 2023, based on prospective RWD from the OPAL registry.
Taking a look at the total study cohort and the subgroups of patients with PD-L1-negative tumors and tumors with unknown PD-L1 status, CTx monotherapy, followed by CTx combination therapy, represent the most common 1L treatment strategies, which is in line with the current recommendations for PD-L1-negative and BRCA1/2 wild-type mTNBC.7,8
The PD-L1 status was unknown for 32.3% (n = 119/368) of patients in our cohort. In a previous analysis, we found that PD-L1 testing was rapidly implemented in clinical routine after the European Medicines Agency (EMA) approval of the first PD-(L)1i in August 2019. The PD-L1 testing rate for advanced TNBC at start of 1L treatment increased from 15% in 2018 to 72%-82% in 2020-2022.23 So, most of the patients whose tumors had unknown PD-L1 status started their 1L before approval of PD-(L)1i. After PD-(L)1i approval, a PD-L1 test result was documented for the majority of the patients (81.1%, n = 228/281).
Among the patients tested positive for PD-L1, 81.2% received immunotherapy with PD-(L)1i ± CTx, indicating that analogous to PD-L1 testing, the use of PD-(L)1i was also rapidly integrated into clinical routine. Atezolizumab was administered at a considerably higher frequency than pembrolizumab, possibly as a result of its earlier approval.
Within the PD-L1-positive subgroup, 18.8% (n = 22/117) received CTx without PD-(L)1i, which may be caused by several factors. Some of these patients had initiated their 1L therapy before approval of PD-(L)1i (n = 5). Moreover, this group comprises 10 patients who are ≥75 years of age. In prior analyses, it was observed that PD-L1 testing and treatment with checkpoint inhibitors were lower among patients aged ≥75 years compared with patients aged <75 years, assuming that novel therapies are often implemented with delay in elderly patients in clinical routine.24 In addition, physicians need to consider various relevant factors when selecting the most suitable therapy for an individual patient, including the performance status, comorbidities, potential toxicities, or prior therapies.
Despite the 2019 EMA approval of PARPi, and although current guidelines recommend testing, BRCA1/2 testing was documented only for 22.8% of patients, assuming that this test might not be frequently carried out before start of 1L therapy in clinical routine. In the total cohort, 4.3% of patients exhibited a BRCA1/2 mutation; PARPi were rarely administered in 1L, with only 0.3% of patients receiving this treatment. Similar observations were recently reported by a large, retrospective real-world study of patients with mTNBC in the United States, which found 1L treatments with PARPi ranging from 1% to 3% between 2019 and 2022, and a large proportion of patients with unknown BRCA1/2 status (possibly due to low testing rates, but also missing test results).25
At least 57.3% of all patients received a 2L treatment, and 25.8% died before start of 2L therapy. Moreover, 5.2% of patients defined as ‘lost to follow-up’ will probably not receive a 2L therapy. This is comparable to the observations from the study carried out in the United States, where 51% of patients received 2L treatment and 34% did not survive to receive 2L, highlighting the need for additional treatment approaches in 1L.25 Similar findings were also reported in an earlier study, which found that 2L treatment was administered to at least 60% of patients with advanced TNBC.26
The median rwOS from start of 1L therapy in the total cohort was 17.6 months (95% CI 15.6-19.7 months), which seemed to be slightly longer than the median OS observed in most other real-world studies on mTNBC with median OS of 11.3 months (95% CI 10.7-12.0 months),25 11.6 months (95% CI 9.9-17.3 months),27 14.5 months (95% CI 11.6-18.9 months),28 15.2 months (95% CI 14.5-16.2 months) for women aged 40-69 years,29 and 16.8 months (95% CI 11.5-22.0 months).26 The median rwPFS in our total cohort was 6.8 months (95% CI 5.9-7.6 months), which was in line with other real-world studies, observing a median PFS of 5.9 (95% CI 5.7-6.2 months; for women aged 40-69 years)29 and 6.5 months (95% CI 5.2-8.4 months).28 However, it should be considered that current RWD are limited and that these real-world studies vary in terms of recruitment periods and, consequently, in treatment regimens administered. In addition, differences in baseline characteristics across the real-world cohorts analyzed may impact clinical outcomes, given that various factors such as age at diagnosis, disease status (de novo versus recurrent disease), presence and number of metastatic sites, performance status as well as comorbidity score, and patient’s descent are known to be associated with prognosis in mTNBC.30, 31, 32, 33 With respect to the study recently carried out in the United States (recruitment from 2011 to 2022), our real-world cohort includes a smaller proportion of patients with brain metastases (6.8% versus 13%) and more patients with de novo mTNBC (37.8% ± 2.4% unknown versus 30% ± 7% unknown), which along with other characteristics might contribute to differences in outcomes.25 Furthermore, the use of different CTx regimens across countries may also impact clinical outcomes.
In the subgroup of patients with PD-L1-positive tumors treated with 1L PD-(L)1i, a median rwPFS of 7.2 months (95% CI 6.5-9.1 months) and a median rwOS of 23.2 months (95% CI 17.5-28.2 months) were determined. These results align with the real-world ANASTASE study, focusing on patients with PD-L1-positive mTNBC receiving atezolizumab plus nab-paclitaxel in 1L (median PFS 6.3 months, 95% CI 3.9-8.7 months),34 and the phase III trials, which reported similar outcomes for patients with PD-L1-positive tumors receiving atezolizumab or pembrolizumab: IMpassion130 (median PFS 7.5 months, 95% CI 6.7-9.2 months/median OS 25.0 months, 95% CI 19.6-30.7 months)9,35 and KEYNOTE-355 (median PFS 9.7 months/median OS 23.0 months, 95% CI 19.0-26.3 months).10,36
Despite novel advances in recent years, the treatment options for advanced TNBC are still limited and clinical outcomes remain poor compared with HER2-positive and hormone receptor-positive/HER2-negative BC subtypes.3,26,28 Consequently, a high unmet need for effective therapeutic approaches persists, especially among patients who do not benefit from immunotherapies or PARPi.
Several promising therapeutic approaches are currently under investigation, with the potential to improve future outcomes for patients diagnosed with mTNBC. Trophoblast cell-surface antigen 2 (TROP2)-directed ADCs have demonstrated encouraging results in phase III trials. In the 1L setting, sacituzumab govitecan in combination with pembrolizumab significantly improved PFS compared with chemotherapy plus pembrolizumab in PD-L1-positive advanced TNBC (median PFS 11.2 months, 95% CI 9.3-16.7 months versus 7.8 months, 95% CI 7.3-9.3 months),37 and for patients who were not candidates for treatment with PD-(L)1i, sacituzumab govitecan led to significantly longer PFS than chemotherapy (median PFS 9.7 months, 95% CI 8.1-11.1 months versus 6.9 months, 95% CI 5.6-8.2 months).38 Furthermore, 1L datopotamab deruxtecan demonstrated significant improvements compared with chemotherapy in both OS (median OS 23.7 months, 95% CI 19.8-25.6 months versus 18.7 months, 95% CI 16.0-21.8 months) and PFS (median PFS 10.8 months, 95% CI 8.6-13.0 months versus 5.6 months, 95% CI 5.0-7.0 months) for patients ineligible for immunotherapy.39 Multiple innovative drugs are currently being evaluated in pivotal phase III trials for 1L treatment, including novel ADCs, such as the bispecific ADC izalontamab brengitecan targeting epidermal growth factor receptor and HER3 (IZABRIGHT-Breast01),40 and the anti-TROP2-directed ADC sacituzumab tirumotecan ± pembrolizumab (TroFuse-011),41 as well as pumitamig, a bispecific antibody against PD-L1 and vascular endothelial growth factor A (ROSETTA Breast-01).42
The present study has some limitations. It should be noted that descriptive and unadjusted survival analyses were carried out; therefore, outcomes across different subgroups are not directly comparable, as patients may differ in baseline characteristics beyond PD-L1 status and treatment. Due to the observational design of the study, there are no specifications as to the timing, frequency, or criteria of tumor assessment; thus, registry rwPFS data should be considered as the best clinical approximation and might not be identical to PFS determined in clinical trials. Moreover, it must be kept in mind that further treatment lines also influence the 1L rwOS determined. Despite these limitations, our study demonstrates significant strengths through its prospective collection of data from multiple sites located across Germany, providing valuable RWD on patients with mTNBC.

Conclusion
Based on data from the large, prospective, multicenter OPAL registry, this real-world study provides valuable insights on treatment patterns and clinical outcomes of patients diagnosed with mTNBC who started their 1L treatment between January 2018 and August 2023. Although new treatment options have evolved, such as PD-(L)1i for patients with PD-L1-positive tumors, which have been rapidly integrated into clinical routine, many patients are not eligible to receive these novel treatment approaches. The clinical outcomes for TNBC in general remain poor. In our total cohort, at least one-quarter of patients did not survive to initiate 2L therapy and thus could not benefit from therapeutic options approved for subsequent treatment lines. Taken together, these findings emphasize a significant unmet clinical need in patients with mTNBC and the urgent requirement for effective 1L treatment options.

Declaration of Generative AI and AI-Assisted Technologies in the Writing Processs
During the preparation of this work, the authors used the AI platform Langdock (Langdock GmbH, Berlin, Germany) in order to improve the style of the text. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.