Racial and Ethnic Characteristics and Outcomes of Patients Diagnosed with CLL/SLL in the USA Open Access

Chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL, hereafter CLL) is a cancer that is most commonly diagnosed among patients of White race in the USA. According to Surveillance, Epidemiology and End Results (SEER) data, the incidence rate of CLL in 2020 was 4.3 per 100,000 among individuals of White race and only 2.7 per 100,000 among those who are Black [1]. Although historically, Black patients with CLL have had significantly shorter overall survival (OS) when compared to White patients, the relative infrequency of diagnosis, compounded with the lower rates of enrollment to clinical trials, has limited the information available about patient outcomes for all other races in the current era of targeted therapies [2, 3].

BTK is a member of the TEC (tyrosine kinase expressed in hepatocellular carcinoma) family of nonreceptor tyrosine kinases and a key component of the B-cell receptor signaling complex. The B-cell receptor is required for normal and malignant B-cell maturation, proliferation, and survival [4, 5]. In 2014, the first covalent BTKi (cBTKi) was approved in the USA for the care of patients with CLL. Since then, there have been a total of three cBTKi agents approved by the US Food and Drug Administration (ibrutinib, zanubrutinib, and acalabrutinib). Pirtobrutinib, a noncovalent BTKi, has been approved for patients with CLL who have received at least two lines of therapy, including a cBTKi and a B-cell lymphoma 2 inhibitor (BCL2i) [6, 7]. From 2013 to 2016, the annual mortality rates for CLL decreased by 5.5%, generally corresponding to the introduction and adoption of these therapies [1]. In 2019, the BCL2i, venetoclax, was also approved for the care of patients with CLL, adding to the targeted therapies [8]. These agents have changed the landscape for patient care since the introduction of cBTKi therapy in 2014.

A study of outcomes in CLL patients by race conducted from 2004 to 2018 found that Black patients were younger, more likely to be uninsured, and to have more comorbidities [3]. While Black patients were more likely to receive immediate therapy than White patients, those who were Black had shorter OS than other races (median 7.0 years for Black patients vs. 9.0 for those who were White) [3]. Another study of patient outcomes early in the BTKi era (2015–2019) similarly found that there were differences in baseline characteristics such as younger age, worse Eastern Cooperative Oncology Group (ECOG) performance status, and more advanced stage at diagnosis among those who were Black; however, survival disparities were not observed by race, suggesting that these novel agents may have had a role in reducing healthcare disparities [9].

Given these intriguing data from the late 2010s suggesting the role of novel therapies in the reduction of survival disparities, this study was designed and conducted exclusively in the current time period to evaluate the hypotheses raised regarding race and survival outcomes in CLL after introduction and adoption of targeted therapy. Specifically, this study evaluated patient demographic and clinical characteristics, treatment patterns, and outcomes by both race and ethnicity, with a focus on the utilization of cBTKi therapies since their regulatory approval.

This study used the CLL Analytic Cohort from the nationwide Flatiron Health electronic health record-derived de-identified longitudinal database, comprising de-identified patient-level structured and unstructured data, curated via technology-enabled abstraction [10, 11]. De-identified data are not considered human subjects in accordance with US 45 CFR 46 [12]. Flatiron Health de-identified data originate from approximately 280 cancer clinics (∼800 sites of care) in the USA. Patients in the CLL Analytic Cohort are those who have CLL as identified by ICD-9: 204.1x or ICD-10: C91.1x or C83.0x codes, at least two documented clinical visits on different days occurring on or after January 1, 2011, and evidence of receiving antineoplastic therapy for CLL. Patients were included in the CLL Analytic Cohort if they received the initial CLL diagnosis on or after November 13, 2013 (the first approval date for cBTKi therapy) and had at least 121 days of follow-up in the database to allow for the observation of study outcomes.

Within the Flatiron Health database, race variables are collected as White, Black, Asian, other, or missing. Race in this study was evaluated for White, Black, and all non-White patients (including Black, Asian, and other race categories). Ethnicity is recorded as Hispanic or Latino, not Hispanic or Latino, or missing. Patients with missing race or ethnicity were excluded from these analyses and no imputation was made as to missing variables. However, those with missing race or ethnicity variables were retained in the study cohort and reported descriptively. Duration of cBTKi therapy was recorded in two ways. First, the duration of the initial cBTKi exposure was defined from the first use of cBTKi to the last use of cBTKi within a single line of therapy. Rechallenge of cBTKi in later lines of therapy was not included in this calculation. Second, the cumulative duration of cBTKi exposure was defined from the first use of cBTKi to the last use of cBTKi observed, excluding gap periods of 60 days or longer, to estimate the lifetime cumulative cBTKi exposure. Duration of therapy was estimated using Kaplan-Meier method; patients who were still receiving cBTKi therapy at the end of the database were censored at the patient’s last observation in the data. Lines of therapy reported in the Flatiron Health database were oncologist-defined and rule-based.

Outcomes evaluated included use of cBTKi therapy, time to next treatment or death (TTNT-D), and OS. Patients were considered to have received cBTKi therapy if they received ibrutinib, zanubrutinib, or acalabrutinib at any time on or after diagnosis; utilization was summarized descriptively by race and ethnicity. TTNT-D was defined as the time from the start of the line of therapy until the start of the immediate next line of therapy or death, whichever occurred first. If the time gap between the last follow-up and death date was less than or equal to 90 days, then that patient was considered to have an event for TTNT-D at the death date. Otherwise, patients were censored at the last follow-up date. OS was evaluated from the start of first-line therapy as well as from initial diagnosis, until time of death. Unadjusted time-to-event analyses were conducted using the unadjusted Kaplan-Meier method. Analyses were also conducted using a Cox proportional hazards regression model, adjusting for baseline covariates: year of CLL diagnosis, age at diagnosis, sex, race, ethnicity, site of care (community vs. academic practice), socioeconomic status, practice size (total number of patients with CLL treated at the site of care during the year of the patient’s index diagnosis), insurance type, and del(17p) status. Two interactions were also explored in the Cox proportional hazards regression model for OS. The first interaction was between ethnicity and site of care and the second interaction was between race and site of care while retaining all other covariates in the model. Patients without events were censored at the last activity date in the database. All statistical tests were conducted at a 5% significance level. All analyses were conducted using SAS version 9.4.

The statistical analysis plan included both unadjusted and adjusted analyses for the three primary study outcomes: use of cBTKi therapy, TTNT-D, and OS. Baseline characteristics were presented descriptively and compared using χ² or t tests for categorical or continuous variables. Comparisons were conducted for White versus Black race, White versus all non-White races (including Black, Asian, and other race categories), and Hispanic or Latino versus not Hispanic or Latino. As mentioned previously, patients with missing race were excluded from the comparative analyses for race, and patients with missing ethnicity data were excluded for the comparisons by Hispanic or Latino ethnicity.

A total of 6,372 patients with CLL met eligibility criteria and were included in this study (shown in Fig. 1). White patients represented 71.8% of the cohort (n = 4,572), Black represented 8.8% (n = 558) of the cohort, and 18.5% were non-White (n = 1,180). Race was missing for 620 (9.7%). A small proportion of patients were Hispanic or Latino (n = 270; 4.2%), while most patients were not Hispanic or Latino (n = 4,897; 76.9%). Ethnicity was missing for 1,205 (18.9%). A summary of patient characteristics by race and ethnicity, including all covariates used in the comparative analyses, is included in Tables 1 and 2. Patients who were Black were significantly younger, more likely to be female, of lower socioeconomic status, and more likely to have unmutated IGHV gene and del(11q) than White patients. Patients who were Black also had a shorter time from diagnosis to initiation of first-line therapy. White patients were more likely to harbor the prognostically more favorable del(13q) molecular subtype. Patients who were Hispanic or Latino were significantly younger and had lower SES, but no significant differences were observed for sex or molecular subtypes, including IGHV mutation status. Both patients who were Black and those who were Hispanic or Latino were significantly more likely to be treated in centers with a lower total patient volume and less likely to receive care in an academic practice setting.

The increasing utilization of cBTKi therapy by the year of initiation (Fig. 2) was not significantly different by the racial group (Table 2). As shown in Table 3, patterns of cBTKi utilization were significantly different by race: 67.4% of Black versus 55.8% of White patients received cBTKi therapy (p < 0.0001). Non-White patients also had a significantly higher use of cBTKi therapy than White patients (60.8 vs. 55.8%, p = 0.002). No significant differences were observed by ethnicity (53.3 Hispanic/Latino vs. 57.2% not Hispanic/Latino, p = 0.21). There was less time from diagnosis to cBTKi therapy for Black and non-White patients versus White as well as for Hispanic or Latino patients versus those who were not Hispanic or Latino. There were significantly higher rates of cBTKi use in the first-line setting for Black patients than for those who were White (49.2 vs. 40.3%, p = 0.0001), as well as for those who were non-White versus those who were White (44.5 vs. 40.3% p = 0.01). Both patients who were Black as well as those who were not Hispanic or Latino had a longer duration of cumulative cBTKi therapy and a longer duration of first-line cBTKi therapy than those who were White or who were Hispanic or Latino, respectively (all p < 0.05). There were no differences by race for the year of the first observed cBTKi, nor were there differences by race for BCL2i or PI3Ki utilization (Table 2).

There were no significant differences in TTNT-D for either race or ethnicity in either adjusted or unadjusted analyses (Table 4). Median time from the start of cBTKi therapy to the start of the next treatment or death was 36.9 months (95% CI: 32.6–41.4) for Black, 35.0 months (95% CI: 33.3–36.9) for White, and 34.1 (95% CI: 29.1–37.8) for non-White patients (all p > 0.05). TTNT-D for those who were Hispanic or Latino was also not significantly different from those who were not Hispanic or Latino (33.9 months, 95% CI: 23.2–49.0 vs. 35.3 months, 95% CI: 33.5–37.1).

Survival outcomes by race were not significantly different in either unadjusted or adjusted analyses (Table 5). Median OS from the start of therapy was 79.4 months (95% CI: 70.1 – not evaluable) for Black, 84.1 months (95% CI: 81.1–91.3) for White, 81.7 months (95% CI: 72.3-not evaluable) for non-White patients. Median OS was 89.6 months (95% CI: 70.8–not evaluable) for those who were Hispanic or Latino and 84.1 months (95% CI: 80.5–91.3) for those not Hispanic or Latino. Findings were consistent when evaluating OS from the date of diagnosis, accounting for the time from diagnosis to initiation of treatment in the calculation of total survival time. Survival curves are shown in Figure 3. Factors significantly associated with OS from the start of first-line therapy by race included del 17p, age, sex, year of start of first-line therapy, and site of care (Table 6). Factors not significantly associated with OS by race included socioeconomic status, practice volume, insurance type, and ethnicity. Interactions between ethnicity and site of care and also between race and site of care were not statistically significant.

CLL (inclusive of SLL) is a heterogeneous hematological malignancy, characterized by different molecular subtypes that have historically been associated with worse survival outcomes. Unmutated IGHV status, for instance, is associated with shorter treatment-free interval as well as inferior survival, irrespective of stage at the time of diagnosis [13, 14]. Del(11q) has been associated with a more symptomatic presentation at the time of diagnosis, including B symptoms and extensive lymphadenopathy, possibly contributing to earlier treatment initiation [15]. On the other hand, del(13q) alterations are associated with a much more indolent biological behavior and longer survival [15].

Black patients have been described as younger and portending more symptoms upon initial diagnosis and, in some analysis, twice more likely to die from CLL in the first 5 years of diagnosis than their White counterparts [16, 17]. A retrospective analysis of CLL cases seen at 2 different academic institutions demonstrated higher rates of molecular alterations associated with worse outcomes such as del(17p) and del(11q) and unmutated IGVH gene among individuals of Black race, as well as worse OS than the non-Black population [18]. Similarly, this study demonstrates differences in molecular subtypes of CLL that may explain aggressive biological behavior of this disease in Black patients that may have contributed to potentially worse outcomes observed in these earlier studies.

While data have shown racial disparities for patients with CLL in the USA over the past 2 decades, these studies do not assess the treatments received or other factors that could be related to survival outcomes after the introduction of targeted therapies. A retrospective analysis of 47,867 cases diagnosed from 2006 to 2016 from the SEER 18 database demonstrated Black race to be associated with worse OS (HR = 1.5, 95% CI: 1.4–1.6), remaining the strongest independent factor for OS after controlling for SES [19]. An analysis of the National Cancer Database (NCDB) of 97,804 CLL cases diagnosed between 2004 and 2018 also demonstrated a relationship with Black race and shorter OS compared to their White counterparts (HR = 1.5, 95% CI: 1.5–1.6) [3]. Neither study describes treatment patterns nor exposure to cBTKi therapy as the majority of the time period studied was in the pre-cBTKi era. In the current study of 6,372 patients diagnosed with CLL from 2013 to 2022, patients of Black race displayed higher rates of unmutated IGHV status and del(11q), presented at younger age and at more advanced stage than White patients. They were, however, more likely to be treated with a cBTKi than White patients, despite lower SES. Similar to the current study, an analysis of 1,459 patients captured by the Inform CLL registry treated with cBTKi from 2015 to 2019 showed that there were no racial disparities in survival despite Black patients presenting at younger age, with a higher Rai stage and a worse ECOG performance status, underscoring the potential of access to effective therapy to overcome unfavorable outcomes [9]. The findings of the Inform CLL registry, consistent with the current study, suggest that the racial disparity in survival may have been reduced in recent years. Furthermore, when including the time period of diagnosis to treatment in the duration of OS, the results remain unchanged, suggesting that there are not delays in time to treatment that could have impacted the OS findings observed. Importantly, this study is limited to the treated population, and OS findings cannot be extrapolated to patients with CLL who do not receive systemic therapy.

The introduction of cBTKi therapy has revolutionized the treatment of CLL, especially patients without TP53 mutations/del(17p). Multiple trials have demonstrated the efficacy of cBTKi therapy versus cytotoxic chemotherapy [20‒22]. However, access to oral agents may be restricted to certain patients due to lack of insurance coverage, potentially generating disparities in outcomes. These data support the emerging evidence, suggesting that the outcome disparities seen before the cBTKi era may no longer be present and leads to hypotheses about the causal nature of this relationship that should be further investigated. Causal inference regarding cBTKi therapy or the timing of when this therapy is received cannot be made in the current study. The introduction of BCL2i therapy and its use in sequence with both covalent and noncovalent BTKi therapy could also be evaluated in future research, once sufficient follow-up is available for patients with access to all these agents. Prospective trials or retrospective analyses incorporating time-varying confounders would be needed to infer causality. Therefore, this study, while demonstrating the association of these variables, cannot conclude that the use of a targeted or other novel therapy is directly responsible for the outcomes observed.

While the results from this study demonstrate similar survival outcomes by race, additional data are needed to confirm this conclusion. Patients who are Black continue to present with clinical characteristics associated with poor prognosis (e.g., younger age, unmutated IGHV, more advanced stage of disease) as in the pre-cBTKi era, yet the current data show that these patients now receive faster treatment and are more likely to receive BTKi-based therapy than those who are White. This study did not specifically look at time-to-treatment outcomes, nor was there access to data from the pre-diagnosis period, to identify reasons for the later stage at diagnosis. The potential gaps in care leading to a later-stage diagnosis as well as rates of treatment are also important topics of future investigation, as patients who did not receive systemic therapy are not recorded in this dataset. These promising data demonstrate the potential impact of reducing barriers to novel treatments and emphasize the need to continue to reduce barriers for all patients. However, in adjusted analyses, there were differences in OS related to ethnicity; the reasons for this finding remain unclear, as sample sizes were small and the follow-up was limited. This outcome deserves more attention in a larger cohort with a longer follow-up. Though this study specifically evaluated race and ethnicity, these data are limited with regard to other races. For example, there are no data recorded for Native American or Hawaiian populations, and the Asian population sample size was too small to report without risking patient identification. These issues demonstrate the gap in understanding about other racial groups that are diagnosed with CLL. Underrepresented populations in clinical research supporting these novel agents are not limited to racial or ethnic groups. The care and outcomes of patients treated in settings without access to clinical trials, patients who are uninsured or underinsured, as well those who reside in rural regions for example, remain largely unknown. Additional work and expansion of data that include more robust social determinants of health data are needed to ensure equity in the care of all patients with CLL.

Patients who are Black tend to present with characteristics suggesting more aggressive disease at diagnosis of CLL as compared to White patients. Although these findings have been consistent through time and warrant further study to understand and address root causes, recent analyses have shown that fortunately progress has been made in closing the survival gap. Studies such as this one conducted in the post-cBTKi era have demonstrated that Black patients do not have significantly different survival outcomes as those who are White. Furthermore, and interestingly, among all treated patients, those who are Black are more likely to initiate therapy earlier and more likely to receive cBTKi therapy for CLL than White patients.

De-identified data are not considered human subjects in accordance with US 45 CFR 46.

Drs. Hess, Winfree, and Escalon and Ms. Li are employees of Eli Lilly and Company.

This was an unfunded research project supported by employee time and database access provided by Eli Lilly and Company.

Debora Bruno: conceptualization; investigation; and writing – review and editing. Manoj Khanal and Xiaohong Ivy Li: formal analysis, validation, data curation, and writing – review and editing. Mercer Escalon: conceptualization and writing – review and editing. Katherine Winfree: investigation and writing – review and editing. Lisa Hess: conceptualization, methodology, validation, investigation, writing – original draft, and visualization.

The data that support the findings of this study were originated by and are the property of Flatiron Health, Inc., which has restrictions prohibiting the authors from making the dataset publicly available. Requests for data sharing by license or by permission for the specific purpose of replicating results in this manuscript can be submitted to [email protected].