Identifying drivers of first-line HR+/HER2− metastatic breast cancer treatment choices

Abstract

Aim: Assess factors associated with first-line (1L) treatment for HR+/HER2- metastatic breast cancer.

Materials & methods: A cross-sectional survey of 250 US oncologists was conducted. Correlations were calculated between treatment class and demographics, treatment perceptions and other clinical/nonclinical characteristics.

Results: Efficacy and safety/tolerability were critical in oncologists’ 1L decision-making. CDK4/6i use positively correlated with proportion of Medicare and postmenopausal patients (r = 0.54–0.67). Chemotherapy use demonstrated positive correlations with perimenopausal and premenopausal patients and symptom burden (r = 0.31–0.42). Aromatase inhibitor (AI) monotherapy correlated positively with anticipated treatment compliance (r = 0.42).

Conclusion: Efficacy and safety/tolerability were most important to 1L decision-making. Clinical characteristics corresponded with CDK4/6i and chemotherapy use. Anticipated compliance was associated with AI monotherapy use.

Plain Language Summary

Patients in the USA with a certain type of metastatic breast cancer (mBC, i.e., HR+/HER2−) might get chemotherapy or hormone therapy alone instead of new and potentially better medicines called cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) as their first treatment.

Researchers wanted to understand how US cancer specialists decided the first treatment for this type of mBC. In a survey of 250 cancer specialists, researchers looked at different factors that might influence decision-making, including patient characteristics, doctors’ opinions about the treatments and other medical and non-medical features. This study also examined the connections between these factors and the cancer specialists’ choice of first treatment.

Researchers found that cancer specialists care most about how well a treatment works and how safe it is when choosing the first treatment for HR+/HER2− mBC. They are more likely to use CDK4/6i if their patients have Medicare coverage or are older (i.e., women who have been through menopause). Chemotherapy is chosen if their patients are younger (i.e., women who are near and before menopause) or have more symptoms. Cancer specialists tend to choose first treatment with hormone therapy alone if they think their patients have a hard time following their treatment plan. The results showed that patient characteristics, doctors’ opinions of treatments and other medical and non-medical factors play a role in choosing treatment for HR+/HER2− mBC. By understanding these factors, researchers can work toward improving treatment choices for patients with this type of mBC.

Tweetable Abstract

US #oncologists consider efficacy and safety as most important when selecting first-line treatment for #HR+/HER2− #metastaticbreastcancer. Different patient characteristics correlated with greater selection of #aromataseinhibitors, #chemotherapy and #CDK4/6inhibitors.

Article highlights.

• Breast cancer (BC) is the most prevalent cancer globally, with 6% of women in the USA diagnosed with de novo metastatic disease.

• HR+/HER2− is the most common BC subtype, accounting for 87.4 new cases per 100,000 women.

• Endocrine therapy (ET) was traditionally preferred for HR+/HER2− metastatic disease (mBC), until recent therapeutic advances introduced cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) in combination with ET.

• Current guidelines recommend CDK4/6i+ET as the preferred first-line (1L) treatment, with approximately 60–63% adoption in routine US clinical practice.

• Findings from this study highlight that US oncologists prioritize anticipated treatment efficacy and safety/tolerability in 1L treatment decisions. Patient demographics, clinical characteristics and nonclinical factors are also considerations in treatment decisions.

• Factors positively correlated with higher CDK4/6i use include a higher proportion of postmenopausal patients, Medicare coverage and emphasis on treatment efficacy and performance status.

• Increased chemotherapy use in the 1L setting is correlated with higher proportions of premenopausal and perimenopausal patients, along with an emphasis on symptom burden.

• Factors associated with higher aromatase inhibitor (AI) monotherapy use include greater consideration of expected lack of treatment compliance, a higher proportion of perimenopausal patients and reliance on health insurance marketplace/state exchange coverage.

• Oncologists perceive CDK4/6i+ET as the appropriate 1L treatment, while chemotherapy is considered for patients with a high symptom burden and AI monotherapy is chosen when tolerability is a concern.

1. Background

Breast cancer (BC) is the most commonly diagnosed cancer worldwide, with 6% of women in the USA having metastatic disease (mBC) at initial diagnosis [1]. The USA represents 11.8% of global BC cases [2], of which HR+/HER2− is the most common subtype, with an age-adjusted rate of 87.4 new cases per 100,000 women [3].

Previously, endocrine therapy (ET; including aromatase inhibitors [AI], selective estrogen receptor degraders [SERDs] and selective estrogen receptor modulators) was the preferred frontline treatment approach for HR+/HER2− mBC until patients developed endocrine resistance [4–6]. Recent therapeutic advancement has been the incorporation of targeted therapy, including cyclin-dependent kinase 4/6 inhibitors (CDK4/6i; i.e., palbociclib, ribociclib and abemaciclib), in the management of HR+/HER2− mBC [5]. The combination of CDK4/6i+ET has demonstrated survival benefits over ET alone in patients with HR+/HER2− mBC in both clinical trials and real-world settings [7–14]. This has led the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines^®) to list CDK4/6i+AI as the preferred first-line (1L) treatment option for post-menopausal or premenopausal women receiving ovarian ablation or suppression with HR+/HER2− recurrent unresectable (local or regional) or stage IV (M1) disease [15]. According to a recent analysis of treatment patterns in routine US clinical practice, between 2020 and 2022, approximately 60–63% of women with HR+/HER2− mBC received a CDK4/6i+ET regimen as 1L treatment [16]. Chemotherapy is still the recommended regimen if ET options exhibit no clinical benefit or in case of rapid progression or visceral crisis [17,18].

Since the introduction of CDK4/6i in 2015, utilization of ET monotherapy and chemotherapy has been decreasing, although a substantial proportion of patients continue to receive these treatments in the 1L setting [16]. Prior to the COVID-19 pandemic, in 2019, 23.9% received ET monotherapy and 15.7% received chemotherapy; however, 1L treatment patterns varied during the pandemic. For example, chemotherapy utilization dropped to 11.7% in 2021 but returned to pre-pandemic distributions (16.7%) in 2022 [16].

US oncologists may consider treatment expectations, patient clinical characteristics and nonclinical factors when selecting 1L mBC treatments. Prior qualitative research revealed clear differences in the types of factors that influence treatment selection among US oncologists who frequently use CDK4/6i+ET, ET monotherapy, or chemotherapy in the 1L setting for patients with HR+/HER2− mBC [19]. In that study, oncologists who highly utilized CDK4/6i+ET considered this regimen to be appropriate for almost all 1L patients, rarely relying on nonclinical factors to guide treatment. Those with high ET monotherapy utilization valued the perceived balance of efficacy and tolerability achieved with that regimen, particularly for elderly patients. Furthermore, their patients encountered affordability challenges with CDK4/6i+ET. High chemotherapy users valued its fast response and high compliance rate; moreover, they indicated patient comorbidities influenced regimen selection [19]. These insights demonstrate clear differences in oncologists’ perceptions of the added value and applicability of the various 1L treatment options for patients with HR+/HER2− mBC. However, the extent to which these concerns are observed across oncologists could not be evaluated with the small sample.

Ultimately, understanding the decision-making process in 1L mBC treatment selection among a broad sample of oncologists is a critical step toward ensuring equitable care for patients with HR+/HER2− mBC. Accordingly, this study aimed to describe the clinical and nonclinical factors considered by US oncologists in the selection of 1L treatment for HR+/HER2− mBC.

2. Methods

2.1. Study design & data source

This was a self-reported, cross-sectional, online quantitative survey that collected data from US oncologists across community and academic settings from August to October 2021; participants who completed the survey received an honorarium. The survey instrument for this study was developed using insights from prior qualitative research with oncologists describing key factors influencing their decision-making for 1L treatment of patients with HR+/HER2− mBC [19]. The quantitative survey was preceded by cognitive interviews with three oncologists who met eligibility criteria to ensure the survey was clear and easily understood. Key components of the survey were reviewed to assess that interpretations are consistent across respondents, adding explanations for key terms where necessary and ensuring that the survey was concise and manageable by respondents to minimize respondent burden. Interviews were conducted via one-on-one, semi-structured telephone calls using Glance, a secure desktop-sharing platform, for sharing the survey instrument. The interviews were audio-recorded for analysis. Each cognitive interview, lasting approximately 45 min, allowed time for additional questions, understanding of respondents’ rationale and addressing any queries. Participating oncologists received fair market incentives for their time. Insights from cognitive interviews were used to refine the final survey instrument.

Oncologists were recruited using the Kantar Profiles All Global Circle physician panel. Panel members are recruited using multiple sources, including the American Medical Association Physician Masterfile, social media campaigns, referral campaigns and organic sources. Additional panelist recruitment methodologies include opt-in emails, co-registration, e-newsletter campaigns, traditional banner placements and internal and external affiliate networks. The protocol, informed consent statement and study materials were reviewed and granted an exemption from expedited or full ethical review by the Sterling Institutional Review Board (Atlanta, Georgia) per US Department of Health and Human Services Policy 45 CFR §46.104(d).

2.2. Study sample

Oncologists were included in the study if they were board certified, in practice between 2 and 30 years, spent ≥50% of time dedicated to direct patient care and have managed ≥5 1L patients with HR+/HER2− mBC in the past 3 months. Participants who were unable or unwilling to provide informed consent or employed by a pharmaceutical or biotechnology company, government health agency, or market research or advertising agency were excluded.

2.3. Variables & measures

The study assessed oncologist demographic (age and gender) and practice (primary specialty, USA region and location of practice, practice and residency setting, the status of onsite multidisciplinary breast/infusion clinic, years in practice, time spent in direct patient care and oncologist perceptions of the primary decision-maker for 1L treatment selection) characteristics, patient load characteristics (patient age, menopausal status, total patients with HR+/HER2− mBC treated in the past 3 months who received 1L therapy, change in newly diagnosed patients with de novo mBC from the previous year, type of insurance and 1L treatment distribution) and treatment approach for IL recommendation (oncologists’ treatment perceptions, the importance of each type of anticipated treatment expectation, efficacy data, patient clinical and nonclinical characteristics in determining 1L HR+/HER2− mBC treatment, financial barriers and factors limiting the ability to recommend CDK4/6i+ET [on a scale from 1=does not limit at all to 7=extremely limiting]). To measure the perceived importance of different factors in determining 1L treatment selection, oncologists were asked to allocate 100 points across all factors for each of four categories (general factors, efficacy data, clinical characteristics and nonclinical characteristics), with a greater number of points allocated to those factors in each category perceived as more important.

2.4. Statistical analysis

Descriptive analysis was conducted for all study variables for the aggregate sample. Frequencies and percentages were reported for categorical variables; means, medians and standard deviations (SDs) were reported for continuous or discrete variables. Pearson and point-biserial correlations were conducted to evaluate the strength and direction of associations between oncologists’ self-reported 1L utilization rate (any CDK4/6i+ET, AI monotherapy and chemotherapy) and key oncologist demographic and practice characteristics, treatment expectations and clinical and nonclinical patient characteristics. Bootstrapped 95% confidence intervals (CIs) and standard errors (SEs) were computed and two-tailed p-values were reported. Correlation coefficients with p-values <0.05, two-tailed, were considered statistically significant. Correlation coefficients (r) with absolute value ≥0.30 represented at least a medium effect size [20]. All statistical analyses were conducted using SPSS^® version 28.0 (IBM Corp., NY, USA).

3. Results

A total of 250 oncologists were included in the analyses. Oncologist demographic, practice and patient load characteristics are shown in Tables 1 and 2.

Table 1.

Oncologist demographic and practice characteristics.

	Overall sample (N = 250)
Age, mean (SD)	48.7 (9.7)
Gender, n (%)
Male	179 (71.6)
Female	52 (20.8)
Other	1 (0.4)
Prefer not to answer	18 (7.2)
Primary medical specialty, n (%)
Medical oncology	113 (45.2)
Hematology oncology	137 (54.8)
Professional designation, n (%)
MD	241 (96.4)
DO	9 (3.6)
USA region of practice, n (%)
Northeast	62 (24.8)
Midwest	44 (17.6)
South	94 (37.6)
West	50 (20.0)
Academic or community practice setting, n (%)
Academic	82 (32.8)
Community	168 (67.2)
Years in practice, mean (SD)	15.5 (6.7)
Proportion of professional time spent in direct patient care, mean % (SD)	91.7 (7.4)
Residency setting, n (%)
Office- or clinic-based private practice	78 (31.2)
Non-teaching/community hospital-based practice	46 (18.4)
Teaching/university hospital-based practice	59 (23.6)
Academic cancer specialist center	58 (23.2)
Non-academic cancer specialist center	9 (3.6)
VA/DoD hospital	0 (0.0)
Other	0 (0.0)
Community of practice, n (%)
Major metropolitan area, population >500,000	83 (33.2)
Urban area, population between 100,000 and 500,000	76 (30.4)
Suburb of a large city, population >100,000	70 (28.0)
Small city, population between 30,000 and 100,000	14 (5.6)
Rural or small town, population <30,000	7 (2.8)
Onsite multidisciplinary breast clinic, n (%)
Yes	131 (52.4)
No	119 (47.6)
Onsite infusion clinic, n (%)
Yes	228 (91.2)
No	22 (8.8)
Primary decision maker for 1L treatment in patients with HR+/HER2− mBC in past 3 months†, n (%)
1 (My patients are the primary decision maker when it comes to their mBC treatment)	0 (0.0)
2	1 (0.4)
3	3 (1.2)
4 (My patients and I make mBC treatment decisions together)	51 (20.4)
5	69 (27.6)
6	89 (35.6)
7 (I am the primary decision maker when it comes to my patients’ mBC treatment)	37 (14.8)

^†In the survey, oncologists were asked to select a number on the scale that most closely matches the primary decision maker(s) for first-line mBC treatment.

1L: First-line; DO: Doctor of osteopathic medicine; mBC: Advanced/metastatic breast cancer; MD: Medical doctor; SD: Standard deviation; VA/DoD: Department of Veterans Affairs/Department of Defense.

3.1. Self-reported oncologists’ demographics & practice characteristics

Overall, most oncologists were male (71.6%), with a mean age of 48.7 ± 9.7 years; 31.2% completed their residency in an office- or clinic-based private practice. Most oncologists practiced in a community setting (67.2%) located either in a major metropolitan or urban area (63.6%). A majority reported hematology oncology as their primary specialty (54.8%), and 37.6% were in practice in the USA South. Overall, oncologists had been in practice for an average of 15.5 ± 6.7 years. Moreover, 52.4% reported having a multidisciplinary breast clinic onsite, and nearly all had an onsite infusion clinic (91.2%). Roughly one in five (20.4%) oncologists reported that they make mBC treatment decisions together with their patients (Table 1).

3.2. Patient load characteristics from treatment experience

Oncologists reported that among their patients with HR+/HER2− mBC receiving 1L treatment in the past 3 months, on average, 35.0% (±11.4) were aged 51–65 years, and 50.1% (±21.1) were postmenopausal. A majority (57.2%) reported having about the same amount of newly diagnosed patients with de novo mBC as compared with the prior year; 38.4 and 4.4% reported having higher and lower amounts of these patients, respectively. Oncologists reported seeing a median of 80 (interquartile range [IQR]: 48–120) patients with BC in the past 3 months, of which, a median of 40 (IQR: 20–75) had HR+/HER2− mBC. Also, a median of 20 (IQR: 12–50) of these HR+/HER2− mBC patients were treated with 1L therapy in the past 3 months, of which, the highest mean percentage of patients were prescribed any CDK4/6i+ET (53.3 ± 33.7). Furthermore, on average, oncologists also reported that the greatest percentage of their patients with HR+/HER2− mBC had commercial insurance (37.7 ± 16.9) or Medicare (34.4 ± 14.7; Table 2).

Table 2.

Patient characteristics and prescribing experience.

	Overall sample (N = 250)
Percentage of patients treated by age group†, mean (SD)
18–50 years	22.1 (11.4)
51–65 years	35.0 (11.4)
66–75 years	27.8 (9.7)
>75 years	15.1 (9.9)
Percentage of patients treated by menopausal status†, mean (SD)
Premenopausal	26.4 (12.7)
Perimenopausal	23.5 (14.2)
Postmenopausal	50.1 (21.1)
Number of patients treated in past 3 months‡, median (IQR)
BC	80 (48.0–120.0)
Head and neck cancer	25 (10.0–50.0)
NSCLC	50 (30.0–88.0)
Ovarian cancer	25 (10.0–50.0)
Hematological malignancies	56.5 (22.0–100.0)
Number of patients with HR+/HER2− mBC treated in past 3 months, median (IQR)	40 (20.0–75.0)
Newly diagnosed patients presenting with de novo mBC compared with previous years, n (%)
A higher percentage of newly diagnosed patients presented with metastatic disease in the past 12 months	96 (38.4)
A lower percentage of newly diagnosed patients presented with metastatic disease in the past 12 months	11 (4.4)
About the same amount of newly diagnosed patients presented with metastatic disease in the past 12 months	143 (57.2)
Number of patients with HR+/HER2− mBC treated with 1L therapy in past 3 months, median (IQR)	20 (12.0–50.0)
Insurance type among patients with HR+/HER2− mBC treated in past 3 months, mean (SD)
Commercial HMO/PPO/POS	37.7 (16.9)
Medicare	34.4 (14.7)
Medicaid	14.1 (11.8)
Health insurance marketplace or state exchanges	8.8 (12.1)
Cash/self-pay	4.7 (7.2)
Other	0.3 (1.6)
Percentage of each 1L mBC therapy prescribed in past 3 months, mean (SD)
Single-agent/combination chemotherapy	11.4 (13.2)
Any ET monotherapy	34.9 (27.6)
AI monotherapy	16.0 (16.4)
SERM	10.5 (11.5)
SERD monotherapy	8.4 (10.1)
Any CDK4/6i+ET	53.3 (33.7)
CDK4/6i+AI	37.5 (31.6)
CDK4/6i+SERD	15.9 (16.0)
Other	0.4 (2.2)

^†Percentage of patients treated by age group and menopausal status refer to those with HR+/HER2− mBC.

^‡Number of patients treated in past 3 months refers to overall patients.

1L: First-line; AI: Aromatase inhibitor; BC: Breast cancer; CDK4/6i: Cyclin-dependent kinases 4/6 inhibitor; ET: Endocrine therapy; HMO: Health maintenance organization; IQR: Interquartile range; mBC: Advanced/metastatic breast cancer; NSCLC: Non-small-cell lung cancer; POS: Point of service; PPO: Preferred provider organization; SD: Standard deviation; SERD: Selective estrogen receptor down-regulator; SERM: Selective estrogen receptor modulator.

3.3. Treatment approach

The highest rated (i.e., assigned greatest mean number of points) general factor influencing 1L treatment selection was anticipated treatment efficacy (26.9 ± 19.3), followed by anticipated safety/tolerability (19.5 ± 11.7; Figure 1A). Figure 1 describes the average number of points oncologists allocated for each type of factor within each category in determining 1L therapy. Among the various types of efficacy data available to oncologists, OS was rated highest in determining 1L treatment selection (33.3 ± 20.8; Figure 1B). Performance status (20.7 ± 17.0; Figure 1C) and cost to patient/insurance coverage (24.9 ± 20.5; Figure 1D) were the highest-rated patient clinical and nonclinical characteristics, respectively, by oncologists in determining their 1L treatment selection.

Figure 1.

Importance of (A) general factors and specific types of (B) efficacy data, (C) patient clinical characteristics and (D) nonclinical characteristics in determining 1L HR+/HER2− mBC treatment selection.

Note: Oncologists were asked to assign a total of 100 points across all factors shown under each category, with a greater number of points assigned to those factors perceived as being more important to their 1L treatment selection.

1L: First-line; mBC: Advanced/metastatic breast cancer; OS: Overall survival; PFS: Progression-free survival; QoL: Quality of life; SD: Standard deviation.

Furthermore, patient insurance coverage/cost to the patient was the factor that most limited the ability to prescribe CDK4/6i as 1L treatment (4.9 ± 1.5). One in four oncologists (26.3%) reported that cost is often/always a barrier for patients with HR+/HER2− mBC after consultation with a financial counselor (Table 3). Oncologists generally perceived that a CDK4/6i+ET regimen was most appropriate for 1L treatment, relative to other available therapies, especially for patients with Eastern Cooperative Oncology Group (ECOG) grades 0 or 1 (84.8%), postmenopausal patients (84.0%) and patients with a high expression (≥50%) of hormone receptors (84.0%). Chemotherapy was most often perceived as an appropriate option for patients with a high symptom burden (59.2%), and ET monotherapy was considered the most appropriate option for patients when oncologists have tolerability concerns (58.4%; Table 4).

Table 3.

Limitations and barriers related to treatment approach.

	Overall sample (N = 250)
Factors limiting the ability to recommend CDK4/6i†, mean (SD)
Presence of psychiatric or cognitive disorders	4.3 (1.5)
Presence of respiratory comorbidities	4.3 (1.5)
Presence of cardiac comorbidities	4.6 (1.5)
Presence of infection-related comorbidities	4.4 (1.5)
Presence of GI comorbidities	4.4 (1.4)
Presence of hematological comorbidities	4.7 (1.4)
Patient's insurance coverage/cost to the patient	4.9 (1.5)
Clinical pathways/formulary	4.0 (1.8)
Patient preference	4.3 (1.6)
Expected AEs	4.5 (1.5)
AE management	4.4 (1.4)
Monitoring requirements	4.3 (1.6)
Expected lack of patient compliance	4.5 (1.5)
Inconvenience to patient	4.3 (1.5)
Cost still barrier for HR+/HER2− mBC patients after financial counselor consultation, n (%)
Always	11 (4.6)
Often	52 (21.7)
Sometimes	121 (50.4)
Rarely	45 (18.8)
Never	8 (3.3)
Patients have not ever consulted with a financial counselor	3 (1.3)

^†Rating scale for this ranged from 1 (does not limit at all) to 7 (extremely limiting).

AE: Adverse event; CDK4/6i: Cyclin-dependent kinases 4/6 inhibitor; GI: Gastrointestinal; mBC: Advanced/metastatic breast cancer; SD: Standard deviation.

Table 4.

Prescriber treatment perceptions.

Perceptions of treatment for HR+/HER2− mBC, n (%)	Overall sample (N = 250)
	CDK4/6i+ET	Chemotherapy	ET monotherapy
Current label limits use	99 (39.6)	30 (12.0)	29 (11.6)
Limitations of GPO or formulary listing	105 (42.0)	32 (12.8)	24 (9.6)
Limitations of clinical guidelines or pathways	90 (36.0)	33 (13.2)	25 (10.0)
Concern about drug coverage	146 (58.4)	32 (12.8)	23 (9.2)
High cost to the patient	176 (70.4)	46 (18.4)	25 (10.0)
High cost to the system	156 (62.4)	43 (17.2)	28 (11.2)
Requires extensive AE management	138 (55.2)	110 (44.0)	25 (10.0)
Requires extensive monitoring	140 (56.0)	105 (42.0)	32 (12.8)
Is preferred by patients	131 (52.4)	31 (12.4)	84 (33.6)
Appropriate for patients with extensive metastatic disease	186 (74.4)	146 (58.4)	57 (22.8)
Appropriate for patients with visceral crisis	140 (56.0)	142 (56.8)	40 (16.0)
Appropriate for patients 70 years old or older	193 (77.2)	83 (33.2)	129 (51.6)
Appropriate for patients younger than 70 years old	203 (81.2)	132 (52.8)	108 (43.2)
Appropriate for premenopausal patients	193 (77.2)	112 (44.8)	97 (38.8)
Appropriate for postmenopausal patients	210 (84.0)	105 (42.0)	109 (43.6)
Appropriate for patients with ECOG grades 0 or 1	212 (84.8)	132 (52.8)	102 (40.8)
Appropriate for patients with ECOG grades 2+	148 (59.2)	49 (19.6)	140 (56.0)
Appropriate for frail patients	142 (56.8)	40 (16.0)	139 (55.6)
Appropriate for patients with many comorbidities	161 (64.4)	34 (13.6)	132 (52.8)
Appropriate for patients with few comorbidities	196 (78.4)	117 (46.8)	95 (38.0)
Appropriate for patients with a high symptom burden	176 (70.4)	148 (59.2)	44 (17.6)
Appropriate for patients with a low symptom burden	187 (74.8)	60 (24.0)	122 (48.8)
Appropriate for patients with a high tumor burden	174 (69.6)	143 (57.2)	43 (17.2)
Appropriate for patients with a low tumor burden	189 (75.6)	55 (22.0)	127 (50.8)
Appropriate for patients with only bone metastasis	193 (77.2)	58 (23.2)	102 (40.8)
Appropriate for patients with a disease-free interval greater than 12 months after adjuvant ET	202 (80.8)	73 (29.2)	88 (35.2)
Appropriate for patients with disease-free interval less than or equal to 12 months after adjuvant ET	203 (81.2)	106 (42.4)	45 (18.0)
Appropriate for patients for whom compliance is a concern	117 (46.8)	96 (38.4)	84 (33.6)
Appropriate for patients for whom tolerability is a concern	155 (62.0)	43 (17.2)	146 (58.4)
Appropriate for patients with a low expression (<50%) of hormone receptors	189 (75.6)	119 (47.6)	71 (28.4)
Appropriate for patients with a high expression (≥50%) of hormone receptors	210 (84.0)	66 (26.4)	115 (46.0)

1L: First-line; AE: Adverse event; CDK4/6i: Cyclin-dependent kinases 4/6 inhibitor; ECOG: Eastern Cooperative Oncology Group; ET: Endocrine therapy; GPO: Group purchasing organization; mBC: Advanced/metastatic breast cancer.

3.4. Correlation analysis

Correlation analysis results are represented in Supplementary Figure S1. The strongest positive correlations with 1L CDK4/6i+ET use was observed with the proportion of postmenopausal patients (r = 0.67), Medicare patients (r = 0.54), emphasis on treatment efficacy (r = 0.44) and performance status (r = 0.38), proportion of patients aged 66–75 years (r = 0.37) and oncologist's practice located in the Northeast region (r = 0.35; all p < 0.05). No oncologist characteristics demonstrated reportable correlations (r ≥ 0.30) with CDK4/6i+ET use in the 1L treatment setting. The proportion of perimenopausal (r = 0.42) and premenopausal patients (r = 0.31) in the oncologist's practice and consideration for a patient's symptom burden (r = 0.31; all p < 0.05) had strong positive correlations to chemotherapy use. Furthermore, the strongest positive correlations to AI monotherapy use were observed with greater consideration of the expected lack of treatment compliance (r = 0.42), the proportion of perimenopausal patients (r = 0.36) and the proportion of patients with health insurance marketplace/state exchange insurance coverage (r = 0.32; all p < 0.05). Additionally, AI prescribing showed a positive correlation with the practice's cost/logical challenges (r = 0.41; p = 0.007).

4. Discussion

This study assessed the clinical and nonclinical factors that drive the selection of 1L regimens for HR+/HER2− mBC from the perspective of US oncologists. Anticipated efficacy and safety/tolerability were found to be the most important factors in determining 1L treatment selection. Clinical characteristics corresponded with greater 1L CDK4/6i and chemotherapy utilization, with symptom burden providing additional critical information when selecting chemotherapy. Compliance challenges were associated with greater 1L AI monotherapy utilization.

In the current study, oncologists reported that 53.3% of their patients with HR+/HER2− mBC received CDK4/6i+ET as 1L therapy. These results are consistent with a previous analysis of electronic medical records data that estimated 56.7% of US adults with HR+/HER2− mBC initiated 1L treatment with CDK4/6i+ET in 2019 [16]. That study also estimated 1L chemotherapy and ET monotherapy utilization in 2019 as 15.7 and 23.9%, respectively. Conversely, in the current study, oncologists reported using chemotherapy in the 1L setting with only 11.4% of patients, with ET monotherapy used for 34.9% of patients.

In the present study, >80% of oncologists perceived CDK4/6i+ET to be an appropriate 1L treatment for postmenopausal patients with ECOG grades 0 or 1, high expression (≥50%) of hormone receptors, patients with disease-free interval ≤12 months after adjuvant ET and <70 years old. A majority perceived chemotherapy was appropriate for patients with a high symptom burden, whereas ET monotherapy was perceived to be appropriate mainly when tolerability was a concern. Overall, these results suggest that oncologists perceived a CDK4/6i+ET regimen to be appropriate for 1L HR+/HER- mBC treatment, relative to other available therapies. This is likewise in accord with a prior survey of French oncologists’ knowledge, perceptions and practice characteristics in which 98.3% concurred that 1L treatment for advanced BC should be a CDK4/6i+ET regimen, assuming visceral crisis is not present [21]. Yet, the current study's results suggest that US oncologists may be reluctant to prescribe 1L CDK4/6i+ET more widely due to higher cost, increased adverse event risk, or concerns about patient compliance, relative to ET monotherapy. Moreover, the misperception that CDK4/6i+ET is associated with a low tumor response rate in the 1L setting, compared with chemotherapy, is a potential barrier in utilizing CDK4/6i+ET as a 1L treatment for HR+/HER2− mBC [22]. Furthermore, considerations about more severe or aggressive disease appear to drive the perceived appropriateness of using 1L chemotherapy.

Notably, the current study reported anticipated treatment efficacy and safety/tolerability as the most important factors considered by the oncologists in 1L treatment selection. Additionally, oncologists perceived OS and PFS as the most important efficacy data when selecting 1L therapy. Furthermore, for 1L therapy selection, the patient clinical and nonclinical characteristics rated highest by oncologists were performance status and cost to patient/insurance coverage, respectively. These findings are aligned with results from a prior USA-based physician survey that reported efficacy data as being of foremost importance to oncologists when determining their 1L therapy selection [23]. Additionally, our study showed that financial toxicity and patient performance status are also key considerations.

The current study revealed that certain nonclinical characteristics (percentage of patients on Medicare, age 66–75 and postmenopausal and Northeast region of practice) and clinical characteristics (emphasis on treatment efficacy and performance status) were positively correlated with higher 1L CDK4/6i+ET utilization. Moreover, having a higher proportion of postmenopausal patients was most strongly correlated (r = 0.67) with 1L CDK4/6i use, consistent with the FDA-approved indications for CDK4/6i [24–26]. These findings are likewise in accord with pivotal clinical trials in the 1L metastatic setting that demonstrated PFS benefits for the CDK4/6i+ET treatment arm over the control arm with palbociclib (PALOMA-2; hazard ratio [HR]=0.58, 95% confidence interval [CI]: 0.46–0.72) [27], abemaciclib (MONARCH 3; HR = 0.54, 95% CI: 0.41–0.72) [28] and ribociclib (MONALEESA-2; HR = 0.56; 95% CI: 0.43–0.72) [29], thereby supporting the use of a CDK4/6i+ET regimen as 1L treatment for postmenopausal women with HR+/HER2− mBC. The proportion of patients with Medicare also evidenced a large effect size association with 1L CDK4/6i+ET utilization and may reflect the typical age profile of patients with HR+/HER2− mBC [30]. The positive correlations of 1L CDK4/6i+ET use with greater emphasis on performance status may be consistent with evidence-based prescribing habits and clinical trial data showing that CDK4/6i+ET regimens demonstrated significantly improved PFS, relative to ET monotherapy, among patients with HR+/HER2− mBC [31,32]. Furthermore, the European Society for Medical Oncology (ESMO) guidelines recommend CDK4/6i+ET as the standard-of-care for HR+/HER2− mBC, with improved survival and a good toxicity profile [17,33].

Chemotherapy use in the 1L setting was positively correlated with patient clinical (greater emphasis on symptom burden) and nonclinical (higher percentage of premenopausal and perimenopausal patients) characteristics. The emphasis on symptom burden concurs with national treatment guidelines, which recommend that patients with visceral disease should be considered for chemotherapy [15]. Additionally, this is consistent with the European School of Oncology (ESO)-ESMO international consensus guidelines that recommend chemotherapy for patients with life-threatening visceral metastases or the need for rapid symptom and/or disease control [17]. Furthermore, the correlation of 1L chemotherapy use with a higher percentage of premenopausal and perimenopausal patients is aligned with previous findings that suggested younger patients may have more aggressive disease [30].

Lastly, AI monotherapy was found to be positively correlated with patient characteristics (higher percentage of patients with health insurance marketplace/state exchange and higher percentage of perimenopausal patients) and nonclinical characteristics (greater emphasis on anticipated lack of patient compliance with CDK4/6i+ET and cost/logical challenges). These characteristics may reflect patients who face more economic barriers to treatment. Ultimately, ET monotherapy is fairly inexpensive, with monthly costs <US$50 [34], and may therefore be more easily afforded by patients with financial challenges. Moreover, correlations of AI monotherapy with expected lack of patient compliance with CDK4/6i+ET may be due to the simpler schedule associated with AI monotherapy, relative to the greater complexity of some CDK4/6i+ET regimens (i.e., greater pill burden/on-off schedule). Furthermore, the positive correlations of 1L AI monotherapy use with perimenopausal patients are consistent with the ESO-ESMO international consensus guidelines recommended indications for ET monotherapy [17].

In addition to the reasons described above, it is possible that those oncologists who are choosing chemotherapy or AI monotherapy in the 1L setting might be considering other clinical evidence in their treatment decision-making. For instance, in the PEARL trial, PFS with palbociclib+ET was not superior to capecitabine monotherapy [35]. Importantly, however, palbociclib+ET was found to be better tolerated than capecitabine, which was further supported by significantly delayed decline in global health status and improvements in physical, role, cognitive and social functioning and in symptoms, including fatigue, nausea/vomiting and appetite loss, for palbociclib+ET versus capecitabine [36]. In addition, the SONIA trial found no statistically significant or clinically meaningful improvement in PFS for use of a 1L CDK4/6i+ET regimen, compared with second-line (2L) use; furthermore, time on a CDK4/6i was longer when used in 1L than 2L, which the researchers surmised would raise toxicity risks and treatment costs [37]. Nonetheless, SONIA was an open-label trial comparing 1L CDK4/6i+AI, followed by 2L fulvestrant and 1L AI monotherapy, followed by 2L CDK4/6i+fulvestrant; as fulvestrant monotherapy is typically not a preferred 2L treatment option after using a CDK4/6i in 1L, the inferences that can be drawn from SONIA's findings are thus limited. It is important to note that results from YOUNG PEARL, which was a phase II trial conducted in premenopausal women, demonstrated benefits, both in terms of PFS and quality of life, for palbociclib+ET, compared with capecitabine [38,39].

Over 60% of oncologists in this study reported observing either no change or a decrease in the proportion of de novo mBC patients, relative to the prior year. Given data were collected in 2021, this finding may reflect downstream effects from restricted healthcare services access during the COVID-19 pandemic. Specifically, patients with BC have reported disruptions and delays in treatment, clinical care and mental health services during the pandemic [40,41]. It is thus plausible that these delays could have also impacted patients’ ability to schedule and receive appropriate diagnostic testing.

Overall, this study provided insights into US oncologists’ perspectives on clinical and nonclinical factors that drive their 1L therapy selection. Their perceptions of the appropriateness of 1L regimens varied with each treatment regimen. Altogether, the study findings detail the decision-making process involved in 1L mBC treatment selection among a broad sample of oncologists. However, given that the study sample is reflective of US oncologists and their real-world practice, any interpretations of study implications for international clinical practice need to be made with caution, recognizing differences in practice environments. Moreover, by highlighting the different factors associated with the utilization of 1L HR+/HER2− mBC treatment regimens, findings can inform strategies for mitigating disparities in care. For example, more generous subsidies could be provided to reduce financial barriers to treatment, and medication management support services could be implemented to improve patient compliance. Ultimately, by describing the clinical and nonclinical factors associated with selecting different 1L treatment regimens for patients with HR+/HER2− mBC, the current study provides a greater understanding of the decision-making process in 1L mBC treatment selection in the real-world US oncology practice setting.

4.1. Strength & limitations

All data, including oncologist practice characteristics, the clinical characteristics of their patients, and 1L utilization patterns, were self-reported and therefore could not be confirmed for accuracy and may be subject to reporting bias. However, any potential reporting bias was anticipated to be constant across oncologists. Additionally, self-reported data collection is a standard approach for assessing respondent perceptions and other subjective outcomes. Of note, the current study was conducted in the USA, where financial considerations for patients may be a more salient concern when selecting treatment, relative to countries with universal healthcare systems, where there may be lesser financial impact on patients due to the greater coverage provided.

The large sample of 250 oncologists was a key strength of the current study. Nevertheless, the convenience sample included in this study may not fully represent the overall population of US oncologists who treat patients with HR+/HER2– mBC, which may limit generalizability, although recruitment was monitored during fielding to ensure broader geographic representation. The demographic distribution of our sample was generally consistent with the US oncology workforce in terms of age (current study: mean age of 49 years vs national data: 64% aged 41–63 years), although the study sample had a smaller proportion of females (current study: 21% vs national data: 36% female) and oncologists in practice in rural communities (current study: 3% vs national data: 11%) [42]. We cannot exclude the possibility that the results were influenced, at least to some extent, by variations between the study sample and the broader population of US oncologists in demographics, practice setting, specialty area, or other observed characteristics; future research studies will need to discern the direction and magnitude of any such associations. Results reflect associations of oncologist demographic and practice characteristics, treatment expectations and clinical and nonclinical characteristics with 1L utilization patterns, but causality cannot be inferred. Other physician, patient and institutional factors not measured in the survey may affect treatment selection choices among patients with HR+/HER2– mBC. Therefore, these results are not comprehensive, and extrapolation of study findings to oncology practices and physicians treating a different mix of patients may be unreliable.

5. Conclusion

A variety of patient clinical and nonclinical factors contributed to US oncologists’ selection of 1L treatment for HR+/HER2− mBC. Findings indicated that anticipated efficacy and safety/tolerability are the most important factors considered by the oncologists in 1L treatment decisions. A CDK4/6i+ET regimen was perceived to be the most appropriate 1L treatment for patients with HR+/HER2− mBC by oncologists. Patient demographic and clinical characteristics, including postmenopausal status, Medicare coverage, emphasis on treatment efficacy and performance status, age 66–75 years and Northeast region of practice corresponded with greater 1L CDK4/6i+ET use. Emphasis on symptom burden and perimenopausal and premenopausal status were associated with greater 1L chemotherapy use, whereas cost and compliance challenges were associated with greater 1L AI monotherapy use. By acknowledging the limitations and highlighting the implications for clinical practice, this study aimed to stimulate further research within the medical community, thereby promoting improved patient outcomes through evidence-based and guideline-aligned practices.

Funding Statement

The study was sponsored by Pfizer Inc. A Brufsky has received research grants from Agendia and AstraZeneca, and has received consulting fees or honoraria from AstraZeneca, Pfizer, Novartis, Lilly, Genentech/Roche, Seagen, Daiichi Sankyo, Merck, Agendia, Sanofi and Puma. AS Cha-Silva, L Shahied Arruda, W Heck and SK Kurosky are employees and stockholders of Pfizer Inc. MC Maculaitis and L Kopenhafer are employees of Oracle Life Sciences, which received funding from Pfizer to conduct and report on the study; P Olsen was an employee of Oracle Life Sciences at the time the study was conducted.

Supplementary material

Supplementary data for this article can be accessed at https://doi.org/10.1080/14796694.2024.2350294

Author contributions

MC Maculaitis was responsible for the data analysis and the initial draft of the article. All authors contributed to the conception and design of the study, interpretation of the data, and critically revising the article for important intellectual content. All authors approved of the final version to be submitted.

Financial disclosure

The study was sponsored by Pfizer Inc. A Brufsky has received research grants from Agendia and AstraZeneca, and has received consulting fees or honoraria from AstraZeneca, Pfizer, Novartis, Lilly, Genentech/Roche, Seagen, Daiichi Sankyo, Merck, Agendia, Sanofi and Puma. AS Cha-Silva, L Shahied Arruda, W Heck and SK Kurosky are employees and stockholders of Pfizer Inc. MC Maculaitis and L Kopenhafer are employees of Oracle Life Sciences, which received funding from Pfizer to conduct and report on the study; P Olsen was an employee of Oracle Life Sciences at the time the study was conducted. 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.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The protocol, interview screener and study questionnaire containing the informed consent statement were submitted to Sterling IRB (Atlanta, GA) and granted an exemption determination per the US Department of Health and Human Services Policy for Protection of Human Research Subjects at 45 CFR §46.104(d) (IRB ID#: 9060-MMaculaitis). All respondents provided informed consent electronically prior to participating.

Data availability statement

The data that support the findings of this study are not publicly available due to the data collection only being granted exemption determination from an IRB for this specific protocol. The data can be made available for non-commercial use from the authors upon reasonable request.