Monitoring of Hematologic, Cardiac, and Hepatic Function in Post-Menopausal Women with HR+/ HER22 Metastatic Breast Cancer

ABSTRACT

Introduction: In the treatment of metastatic breast cancer (mBC), regular monitoring is key in helping physicians to make informed clinical decisions, managing treatment side effects, and maintaining patients’ quality of life. Therefore, we investigated the monitoring frequency in post-menopausal women with HR+/HER2- mBC stratified by first-line regimen.

Methods: Treatment monitoring was assessed using two complementary data sources: a medical chart review (chart review analysis) and a commercial claims database (claims analysis). Women with post-menopausal HR+/HER2- mBC who initiated first-line therapy for mBC were selected and classified under three cohorts, based on treatment received: cyclin-dependent kinase 4/6 (CDK4/6) inhibitor (i.e., palboci- clib—the only CDK4/6 approved at the time of the study), endocrine therapy (ET), and chemotherapy. Frequency of monitoring [com- plete blood count (CBC), electrocardiogram (EKG), and liver function test (LFT)] and labo- ratory abnormalities detected during the first line of therapy were analyzed.

Results: In the chart review analysis, 64 US oncologists abstracted medical information on 401 eligible patients, including 210 CDK4/6 users, 121 ET users, 51 chemotherapy users; 19 patients used other regimens. All patients had ≥ 1 CBC; between 8.3% (ET users) and 39.5% (CDK4/6 users) had ≥ 1 EKG; and over 98% of patients had ≥ 1 LFT across all three cohorts. Among monitored patients, 64.6% had a CBC abnormality, with anemia (39.9%), leukopenia (27.4%), and neutropenia (26.7%) being the most common. Abnormal EKG readings were detected in 8.4, 0.0%, and 7.7% of CDK4/6, ET, and chemotherapy users, respectively. LFT abnormalities were detected in 14.1–26.0% of CDK4/6 and chemotherapy users, respectively. Similar frequency of monitoring was observed in the claims analysis, with the exception of EKG monitoring, for which the proportion of patients tested was higher.

Conclusion: Post-menopausal women with HR+/HER2- mBC receiving first-line therapy with CDK4/6, ET, or chemotherapy were regu- larly monitored regardless of the first-line regi- men received.

Keywords: CDK4/6 inhibitor; HR+/HER2-; Metastatic breast cancer; Monitoring; Oncology; Post-menopausal

INTRODUCTION

With more than 252,000 newly diagnosed cases expected in the United States (US) in 2017, breast cancer (BC) is the cancer with the highest incidence among women [1]. The most com- mon BC histological subtype is hormone receptor-positive (HR+)/human epidermal growth factor receptor two negative (HER2-); a substantial proportion of patients with this subtype will develop metastatic BC (mBC) [2, 3]. The primary goal of mBC treatment is to improve patients’ quality of life and to prolong survival. Endocrine therapy (ET) has long been the standard of care of first-line treatment for HR+/HER2- mBC [4, 5], with fulvestrant 500 mg shown to impart a survival benefit over anastrozole 1 mg [6]. Following disease pro- gression, patients may receive an alternative single-agent endocrine therapy followed by chemotherapy. Chemotherapy is typically rec- ommended when there is clear evidence of resistance to endocrine therapy or when there is a need for rapid disease control [7]. Even though most patients with HR+ mBC initially respond to ET [8], approximately 30% will eventually show primary resistance towards ET, and others will develop secondary resistance [3, 9]. Recently, the treatment landscape in post- menopausal mBC has evolved significantly with the emergence of cyclin-dependent kinase 4/6 (CDK4/6) inhibitors, including palbociclib, ribociclib, and abemaciclib, which have been shown to significantly improve survival in patients with HR+/HER2- advanced or mBC [10–13]. CDK4/6 inhibitors are indicated for first-line treatment of post-menopausal HR+/ HER2- advanced or mBC (palbociclib and ribociclib), and in the second line for patients whose disease progressed following ET (palbo- ciclib and abemaciclib) [14–16].

Considering the importance of palliative care in the treatment of mBC, regular moni- toring is key in maintaining patients’ quality of life, managing side effects associated with therapies, and helping physicians make informed clinical decisions [7, 17]. The 2017 National Comprehensive Cancer Network (NCCN) guidelines recommend performing regular physical examinations, laboratory tests [i.e., complete blood counts (CBCs) and liver function tests (LFTs)], assessments of symptoms and biomarkers, and diagnostic imaging at intervals of 1–3 months for patients receiving ET and before beginning each chemotherapy cycle for patients receiving chemotherapy [7]. However, the NCCN does not yet make any specific recommendation on the monitoring of patients receiving targeted therapies such as CDK4/6 inhibitors. Nonetheless, product label recommendations stipulate that CBCs and LFTs should be performed regularly during the first 4–6 months of therapy for ribociclib and abe- maciclib, while only CBCs are recommended for palbociclib [14–16]. CBCs are recommended to detect hematologic toxicity (such as neutrope- nia) [14–16, 18, 19] and LFTs are recommended to detect hepatobiliary or hematologic toxicity [14–16]. There are no specific recommendations on cardiac monitoring in the NCCN guidelines. However, the product label of ribociclib rec- ommends that electrocardiograms (EKGs) be performed prior to treatment initiation with ribociclib, 2 weeks after initiating the first 28-day cycle, and at the beginning of the sec- ond cycle [15].

How monitoring recommendations are implemented by physicians in the real world remains largely unknown and may be influ- enced by many factors, such as physicians’ preferences, patients’ comorbidity profile, or performance status. Considering that post- menopausal women are older and have signifi- cant comorbidities, it cannot be denied that periodic monitoring would be tantamount to good patient care. Therefore, this retrospective study was conducted to assess the frequency of disease monitoring among post-menopausal women diagnosed with HR+/HER2- mBC in the US.

METHODS
Data Source

Two complementary data sources were used: a patient medical chart review (hereafter referred to as the chart review analysis) and a commer- cial claims database (hereafter referred to as the claims analysis).

The retrospective chart review was con- ducted between February and June 2017. US oncologists were invited by email to participate in this study. Physicians were asked to provide clinical information on up to ten eligible patients and to fill out a survey about their practice and treatment preferences. A chart ex- traction form was developed to collect infor- mation on patient characteristics and monitoring outcomes. Data collected from physicians did not include any patient-identi- fying information and the study was exempted from full review by the New England institu- tional review board.

The claims analysis was conducted using data from the Truven Health Analytics Mar- ketScan Commercial databases (January 2006–December 2015). The databases covered all census regions and included information on health plan enrollment history, demographics, diagnoses, and pharmacy and medical care ser- vices. Data were de-identified and complied with the confidentiality requirements of the Health Insurance Portability and Accountability Act.

Study Population and Cohorts

In the chart review analysis, eligible patients were required to have the following character- istics: (1) women aged ≥ 18 years, (2) diagnosed with de novo or recurrent HR+/HER2- or HR+/ HER2 equivocal mBC between March 2015 and January 2016 (i.e., allowing for a potential fol- low-up of 12 months), (3) post-menopausal at the initiation of first of line therapy for mBC, (4) had complete BC-related care information available since mBC diagnosis, and (5) never enrolled in a clinical trial for the treatment of BC.

In the claims analysis, eligible patients were required to have the following characteristics: (1) women with ≥ 2 claims with a diagnosis for BC, (2) ≥ 2 claims with a diagnosis for a sec- ondary neoplasm, (3) initiated a therapy for mBC in 2013 or later [since significant changes to Current Procedural Terminology (CPT) codes were effective as of January 2013, the first line of therapy has been defined based on a previously published algorithm [20]; the initiation date of the first line of therapy following secondary neoplasm diagnosis was defined as the index date], (4) ≥ 18 years old as of the index date, (5) continuous health plan enrollment for ≥ 6 months before (i.e., the baseline period) and ≥ 1 month after the index date, (6) HR+/ HER2- BC tumor, proxied by ≥ 1 prescription fill for (or administration of) an ET, everolimus, or palbociclib and no claims for trastuzumab, lapatinib, pertuzumab, ado-trastuzumab, or afatinib, and (7) post-menopausal, proxied by either (1) aged ≥ 60 years old as of the index date, (2) ≥ 1 claim for fulvestrant or an aro- matase inhibitor without gonadotropin-releas- ing hormone (GnRH) medication use, (3) bilateral oophorectomy before the index date, or (4) a diagnosis or procedure code related to post-menopausal status before the index date.

In both analyses, three cohorts were formed based on the treatment received in the first line of therapy: one cohort comprising patients who received a CDK4/6 inhibitor-based regimen (only palbociclib was approved at the time of data collection; hereafter referred to as the CDK4/6 cohort), one cohort comprising patients who received endocrine monotherapy (hereafter referred to as the ET cohort), and one cohort comprising patients who received a chemotherapy-based regimen (hereafter refer- red to as the chemotherapy cohort).

Study Measures

In the chart review analysis, information on patient demographics, including age, race, and insurance type, and on clinical characteristics, including comorbid conditions, metastatic sites present at first-line treatment initiation, first- line treatment regimen, and the frequency of monitoring received during the first line of therapy (i.e., CBCs, EKGs, and LFTs, and whe- ther or not an abnormality was detected during monitoring) was collected. The definition of abnormality was based on a non-protocol-dri- ven evaluation setting by physicians. Labora- tory abnormalities included anemia, leukopenia, neutropenia, thrombocytopenia, abnormal EKG readings, and abnormal liver functions such as increased alkaline phos- phatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and bilirubin levels. Finally, the chart extraction form also collected information on hospitalizations and emergency room (ER) admissions related to CBC, EKG, and LFT abnormalities.

In the claims analysis, demographics (i.e., age at mBC diagnosis and region) and clinical characteristics [i.e., Quan-Charlson comorbidity index (Quan-CCI) [21], comorbidities, and metastatic sites] were analyzed using informa- tion from the patients’ enrollment file and diagnoses and procedures recorded in their medical claims during the baseline period. CBCs, EKGs, and LFTs (including ALP, ALT, AST, and bilirubin tests) were defined based on CPT procedure codes recorded in medical claims during the first line of therapy.

Statistical Analyses

In both analyses, categorical variables were summarized using frequencies and percentages, and continuous variables were summarized using means, medians, and standard deviations. The proportions of patients who had at least one CBC, one LFT, and one EKG (hereafter referred to as monitored patients) during the first line of therapy were measured overall and separately for each cohort. In addition, for each type of monitoring, results were summarized using the number of tests conducted per patient per month (PPPM) during the line of therapy. In the chart review analysis, the proportions of monitored patients for which laboratory abnormalities were detected and the propor- tions of monitored patients with a hospitaliza- tion or ER admission associated with a laboratory abnormality were summarized using the aforementioned descriptive statistics.

Compliance with Ethics Guidelines

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For the com- mercial claims analysis, the data were de-iden- tified and fully compliant with the patient confidentiality requirements of the Health Insurance Portability and Accountability Act; therefore, no ethical review by an institutional review board was required. For the retrospective chart review, we obtained institutional review board exemption from the New England Insti- tutional Review Board.

RESULTS
Chart Review Analysis

A total of 64 US oncologists participated in the chart review, with a majority working in com- munity-based practice (95.3%). Most physicians (54.7%) practiced in small/intermediate size practice settings (i.e., 2–9 oncologists or sub- specialists), and 40.6% practiced in larger prac- tice settings (i.e., 10 or more oncologists or subspecialists). All four US census regions were represented by the sample of participating physicians.
The medical charts of 401 post-menopausal women with HR+/HER2- or HR+/HER2 equivocal mBC were collected (Tables 1, 2). A total of 210 (52.4%) patients were classified in the CDK4/6 cohort, 121 (30.2%) in the ET cohort, 51 (12.7%) in the chemotherapy cohort, and 19 (4.7%) were not classified as they received other treatment regimens (i.e., ever- olimus- or bevacizumab-based regimens). Overall, patients’ mean age was 67.0 years, 70.6% were Caucasian, 55.1% were Medicare- insured, and 41.1% were commercially insured. The most common metastatic sites present at first-line therapy initiation were bones (72.8%), lungs (39.2%), lymph nodes (28.7%), and liver (22.9%; Table 1).

First line of therapy was observed to have an average duration of 15.4, 14.6, and 6.8 months for the CDK4/6, ET, and chemotherapy cohorts, respectively. During this time, all patients received ≥ 1 CBC. The mean number of CBCs performed PPPM was 1.3, 0.7, and 2.6 for patients in the CDK4/6, ET, and chemotherapy cohorts, respectively. Among monitored patients, CBC abnormalities were detected in 172 (81.9%) patients of the CDK4/6 cohort, 37 (30.6%) patients of the ET cohort, and 40 (78.4%) patients of the chemotherapy cohort. The most commonly detected CBC abnormali- ties were neutropenia (CDK4/6 = 43.8%; ET = 2.5%; chemotherapy = 21.6%), leukopenia (CDK4/6 = 42.4%; ET = 2.5%; chemother- apy = 35.3%), and anemia (CDK4/6 = 40.0%; ET = 29.8%; chemotherapy = 60.8%). Among patients with CBC abnormalities, 4 (2.3%) CDK4/6 users, 0 (0.0%) ET users, and 2 (5.0%) chemotherapy users had a hospitalization or ER admission related to this abnormality (Table 2). EKG monitoring was performed in 83 (39.5%) CDK4/6 users, 10 (8.3%) ET users, and 13 (25.5%) chemotherapy users. Among monitored patients, abnormal EKG readings were detected in 7 (8.4%) CDK4/6 users, 0 (0.0%) ET users, and 1 (7.7%) chemotherapy user. A prolonged QT interval was detected in 2 (2.4%) monitored patients in the CDK4/6 cohort. Among monitored patients in the CDK4/6 cohort with an abnormal EKG, 1 (14.3%) patient had a hospitalization or ER admission related to an abnormal EKG reading. Other events related to abnormal EKG readings included non-specific arrhythmia, atrial fibrilla- tion, T-wave abnormalities, and ST-T changes. One patient had a hospitalization or ER admis- sion related to an EKG abnormality (Table 2).

Almost all patients received ≥ 1 LFT [CDK4/ 6 = 206 (98.1%); ET = 119 (98.3%); chemotherapy = 50 (98.0%)] during first-line mBC treat- ment. On average, a total of 0.9, 0.6, and 1.8 LFTs were performed PPPM in the CDK4/6, ET, and chemotherapy cohorts, respectively. Among LFT- monitored patients, abnormalities were detected in 29 (14.1%) CDK4/6 users, 27 (22.7%) ET users, and 13 (26.0%) chemotherapy users. Among patients with an abnormal LFT, 1 (3.7%) patient from the ET cohort had a hospitalization or ER admission related to an LFT abnormality, and no such event was observed in the CDK4/6 and chemotherapy cohorts (Table 2).

Claims Analysis

The claims analysis comprised 7468 post- menopausal women with HR+/HER2- mBC. In total, 214 (2.9%) patients were classified in the CDK4/6 cohort, 4725 (63.3%) in the ET cohort, 2323 (31.3%) in the chemotherapy cohort, and 206 (2.8%) were not classified due to the use of other treatment regimens. Overall, mean age at mBC diagnosis was 64.1 years. The most com- mon metastatic sites present at first line of therapy initiation were the lymph nodes 0.4 in the ET cohort, and 1.5 in the chemotherapy cohort (Fig. 1).

DISCUSSION

In this real-world observational study, the monitoring patterns of post-menopausal women with HR+/HER2- mBC were evaluated separately among patients receiving a CDK4/6- based regimen, ET monotherapy, and chemotherapy-based regimens using both a patient medical chart review and a claims-based approach. Results from both approaches high- light that monitoring tests, including CBCs, EKGs, and LFTs, are performed regularly among these patients regardless of the type of treat- ment used in the first line of therapy, suggesting that monitoring is part of comprehensive patient care practice.
The patient medical chart review and claims database analyses generally presented consis- tent findings in terms of the frequency of monitoring with CBCs and LFTs, which in turn appeared in line with the most recent NCCN treatment guidelines. In fact, for patients receiving chemotherapy, the NCCN recom- mends that CBCs and LFTs be performed before each cycle of chemotherapy, which typically lasts 14 or 28 days [7]. Based on these recom- mendations, between 1 and 3 tests PPPM were expected, which appears consistent with the average of 2.5 CBCs PPPM (2.4 PPPM in the claims analysis) and 1.8 LFTs PPPM (1.5 PPPM in the claims analysis) observed in the chart re- view analysis. Similarly, the NCCN recom- mends that CBCs and LFTs be performed every was approved at the time the study was con- ducted, recommends that CBC be performed at the beginning of each 28-day cycle, and every 2 weeks during the first two cycles, for a total of five CBCs over the first 2 months. In the chart review analysis, over an average follow-up period of 15.4 months, an average of 1.3 CBCs PPPM was observed; in the claims analysis, over an average follow-up period of 4.3 months, an average of 1.4 CBCs PPPM was observed. Despite the absence of specific recommendations in the product label, an average of 0.9 LFTs PPPM (1.0 in the claims analysis) was observed in the CDK4/6 cohort.

The frequencies of CBCs and LFTs were generally similar between the patient medical chart review and the claims data analyses, but tended to be slightly lower in the claims data- base analysis. This observation was quite unex- pected given that claims data usually capture a broader spectrum of medical services (i.e., all medical services reimbursed by payers for the claims analysis vs. only services recorded in the patient medical chart and available to the oncologist for the chart review analysis). A possible explanation is that medical services fully reimbursed by Medicare do not appear in
commercial claims databases. In support of this hypothesis, a sensitivity analysis conducted among patients under 65 years of age yielded frequencies of CBC and LFT that were closer to those observed in the chart review analysis [i.e., CBCs (chart review vs. claims sensitivity): 1.3 vs. 1.3 PPPM; LFTs (chart review vs. claims sensi- tivity): 0.9 vs. 1.0 PPPM]. In addition, the fre- quency of EKGs tended to be higher in the claims analysis relative to the chart review analysis, which is presumably because claims reflect medical services offered by any provider, while medical charts only include services offered by the treating oncologist. Intriguingly, LFTs tended to be performed less frequently than CBCs across all treatments cohorts regardless of the data source (i.e., in the chart review and claims). Given both LFTs and CBCs are conducted by venipuncture, further analyses are warranted to understand why LFTs and CBCs may not be performed simultaneously.

The results from the chart review analysis revealed that between 8.3% (over an average follow-up period of 14.6 months in the ET cohort) and 39.5% (over an average follow-up period of 15.4 months in the CDK4/6 cohort) of patients received EKG monitoring. In the claims analysis, a similar proportion of patients in the CDK4/6 cohort received EKG monitoring; 20.6% over an average follow-up period of 4.3 months; but a 12-month Kaplan–Meier rate of 36.3% (results not presented). Interestingly, conflicting results were obtained for the ET cohort: in the chart review analysis, the ET cohort had the lowest rate of EKG monitoring, while the converse was observed in the claims analysis. This observation suggests that EKGs are commonly performed among patients with mBC, but are not necessarily prescribed by the treating oncologist. In fact, EKG monitoring can be conducted as part of routine care for the treatment of comorbidities such as hyperten- sion and other cardiovascular disease. These comorbidities, which had a relatively high prevalence in the current study, are associated with an increased risk of QT prolongation and may warrant more EKG monitoring in affected patients.

Among monitored patients, the CBC moni- toring detected hematologic abnormalities in about two-thirds of patients; the proportion ran- ged between 30.6% in the ET cohort and 81.9% in the CDK4/6 cohort. Abnormal liver function was detected in 18.3% of patients, ranging from 14.1% in the CDK4/6 cohort to 26.0% in the chemotherapy cohort. Abnormal EKG reading was detected in 6.9% of the patients; no patients in the ET cohort, 1 patient (7.7%) in the chemotherapy cohort, and 7 patients (8.4%) in the CDK4/6 cohort. QT prolongation event accounted for only two of these events; other events included non-specific arrythmia, atrial fibrilation, T-wave abnormalities, and ST-T chan- ges; one patient was hospitalized or adminited to an emergency department related to this abnor- mal cardiac function. Any comparisons to clinical trial results should be taken with caution due to the different durations of follow-up, the non- protocol driven assessment of abnormalities, and most importantly, the inclusion of a mix of treatment in study cohorts. However, the pro- portions of abnormal monitoring tests in the current study appeared generally consistent with those observed in clinical trials, for example, the PALOMA-2 clinical trial (palbociclib + letrozole vs. letrozole)—considering the two largest cohorts in the current study, palbociclib-based regimens and ET monotherapy [10]. However, the rates of neutropenia in the CDK4/6 and ET cohorts were about half those observed in the PALOMA-2 trial (43.8% vs. 79.5%). Such a discrepancy may be the consequence of physicians’ subjective assessment of abnormalities: before asking about the specific type of CBC abnormality detected, physicians were asked whether an abnormality was found. Since low grade neutropenia is very common among patients receiving CDK4/6 inhibitors and highly expected from physicians, it is possible that many physicians responded ‘no’’ to this question. On the other hand, the rates of anemia observed for the CDK4/6 (40.0%) and ET (29.8%) cohorts were much higher than those observed in the palbociclib plus letrozole (24.1%) and letro- zole (9.0%) arms of the PALOMA-2 trial. Since anemia is often associated with chemotherapy [22], the higher rate of anemia observed in the CDK4/6 cohort may be partially attributed to use of chemothearpy in combination with a CDK4/6 inhibitor. However, a similar explanation appears unlikely to account for this discrepancy with respect to the ET cohort as the rates of adverse hematologic events appeared much lower in previous trials with letrozole [23] and other ETs such as anastrozole [24] and fulvestrant [25]. Further studies may be warranted to better understand why such a high rate of anemia has been reported.

Limitations

The present study is subject to a few limitations. First, in the chart review analysis, although participating physicians were instructed to provide information on randomly selected patients among those who met the inclusion criteria, a selection bias may exist (e.g., for patients with a better/worse prognosis, for patients treated more recently, etc.). Second, as in all chart review analyses, the data may not include monitoring tests ordered by specialists other than the surveyed physician. Third, lab- oratory abnormalities were defined based on the responding physicians’ subjective assessment rather than a formal definition. Fourth, in the claims analysis, only diagnostic and procedure codes that were recorded for reimbursement purposes were available. Fifth, this study did not assess the factors associated with the frequency of monitoring; further studies would be war- ranted to better understand such factors. Finally, the chart review and claims data are subject to coding errors, data entry errors, and data omissions.

CONCLUSION

In this retrospective observational study, the monitoring patterns of post-menopausal women with HR+/HER2- mBC were evaluated and reported separately for patients treated with CDK4/6 inhibitors, ET, and chemotherapy. The results revealed that CBCs, LFTs,PD-0332991 and EKGs are performed regularly regardless of the treatment used in first line.