Key Point
-
•
TTE outcomes are worse in DR CLL vs DE and DE-only CLL, indicating reduced response durability after cBTKi/BCL2i failure.
Visual Abstract
Abstract
Patients with chronic lymphocytic leukemia (CLL) who are double-exposed (DE) or double-refractory (DR) to covalent Bruton tyrosine kinase inhibitors (cBTKi) and B-cell lymphoma 2 inhibitors (BCL2i) have limited treatment options and poor outcomes. We evaluated real-world treatment patterns, response, and time-to-event (TTE) outcomes in patients with DE CLL (N = 264), and among the DE-only (ie, DE but not DR; n =157) and DR (n = 107) subcohorts, using the Flatiron Health Research Database. The DE cohort included patients who started first-line therapy for CLL on/after 1 January 2011, had a line of therapy (LOT) for CLL after being DE and on/before 30 November 2022, and met other inclusion criteria. Real-world response rate (rwRR) and TTE outcomes (duration of response [rwDOR], progression-free survival [rwPFS], time to next treatment [rwTTNT], and overall survival [rwOS]), were evaluated. Most patients were retreated with a cBTKi- and/or a BCL2i-based therapy (DE, 69%; DE only, 75%; DR, 54%) as their first index LOT. rwRR was similar between groups (DE, 39%; DE only, 41%; DR, 36%), but median TTE outcomes were shorter in the DR subcohort (DR vs DE vs DE only: rwDOR, 11.5 vs 20.3 vs 26.2 months; rwPFS, 5.0 vs 8.8 vs 11.3 months; rwTTNT, 8.0 vs 13.8 vs 22.1 months; rwOS, 14.7 vs 25.2 vs 38.2 months) and shorter with each subsequent index LOT. Our results suggest an unmet need among patients with DR CLL. Future studies should continue to evaluate response durability with different treatments in DE and DR CLL populations.
Introduction
Chronic lymphocytic leukemia (CLL) is a malignant disorder of B lymphocytes and one of the most common leukemias in adults. In the United States, nearly 24 000 new cases of CLL and 4460 deaths from the disease are estimated for 2025.1,2 Historically, CLL treatment consisted primarily of chemoimmunotherapy but has shifted with the advent of targeted agents such as covalent Bruton tyrosine kinase inhibitors (cBTKi; ibrutinib, acalabrutinib, zanubrutinib) and the B-cell lymphoma 2 inhibitor (BCL2i) venetoclax, which have demonstrated high efficacy in frontline and relapsed/refractory settings.3, 4, 5 Although most patients achieve durable responses to cBTKi- and BCL2i-based therapies, many eventually develop disease progression. Thus, CLL is considered largely incurable and additional treatment options are needed to improve disease control and overall survival (OS).6, 7, 8, 9, 10 In particular, a high unmet need exists for patients who require subsequent therapy after being double exposed (DE) to cBTKi and BCL2i, because treatment options are limited and outcomes are poor after subsequent treatment.10, 11, 12, 13
Recent retrospective studies suggest that patients with CLL with double-refractory (DR) disease to cBTKi and BCL2i may have worse survival outcomes than those with DE-only CLL (ie, DE but not DR), with median real-world progression-free survival (rwPFS) ranging from 6.8 to 9.9 months for DR (vs 11.4-15.4 months for DE only) and median real-world OS (rwOS) from 13.1 to 26.4 months for DR (vs 16.3 months to not reached [NR] for DE only).14, 15, 16 Previous real-world studies have evaluated combined DE cohorts (ie, DE with/without DR), regardless of reason for discontinuation, and lack specific focus on DR disease. To our knowledge, only a few studies have evaluated the DR population exclusively,17, 18, 19, 20 and only 2 of these studies have reported responses to subsequent therapy.19,20 In 1 small study (N = 17), 4 patients achieved a response (3 complete, 1 partial),19 and the other study20 reported an objective response rate (RR) of 40%. These and other studies reported rwOS ranging from 3.6 to 33.0 months,17, 18, 19, 20 which may have varied by treatment regimens and differences in defining the index date (eg, at the start of second-line targeted therapy vs subsequent treatment line after DR disease). In addition, most patients in previous studies of DE, DE-only, and DR CLL received care at academic centers, which limits our understanding of the outcomes and unmet needs of patients receiving care in community-based settings. To address these gaps, this real-world observational study evaluated treatment patterns, clinical outcomes, and response among patients with DE, DE-only, and DR CLL in a predominantly community-based US population.
Methods
Study design
We conducted a retrospective observational cohort study using data from the US-based electronic health record–derived deidentified Flatiron Health Research Database (FHRD; https://flatiron.com/database-characterization),21 described herein according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.22 Patients were included (Figure 1) if they were diagnosed with CLL (International Classification of Diseases nineth revision, 204.1X; or International Classification of Diseases tenth revision C91.1x C83.0X), with a known date of diagnosis and physician documentation of CLL/treatment for CLL; if they were at least 18 years of age at diagnosis; and if they had at least 2 documented clinical visits on different days, at least 1 order for an antineoplastic therapy, and sequential or combination treatment with cBTKi (acalabrutinib, ibrutinib, zanubrutinib) and BCL2i (venetoclax) in any line of therapy (LOT), on or after 1 January 2011, and a subsequent LOT (ie, index LOT) on or before 30 November 2022, to allow for at least 6 months of potential follow-up time before the data cutoff date of 31 May 2023. These patients comprised the DE cohort. Subsets of the DE cohort were analyzed separately; these included patients with DE but not DR CLL (DE-only subcohort) and patients with DR CLL (DR subcohort). DR was defined as relapsed/refractory to a cBTKi and BCL2i, in any LOT, in the same or separate LOT, regardless of LOT sequence. The first index LOT was a nonmaintenance LOT initiated after the earliest cBTKi- and BCL2i-containing LOTs and could not include a clinical study drug or be preceded by disease transformation (Richter transformation). The second index LOT was the second LOT after cohort entry (ie, next LOT after first index LOT), and the third index LOT was the third LOT after cohort entry (ie, next LOT after second LOT; supplemental Figure 1). See supplemental Table 1 and supplemental Methods for more information. Patients were excluded if relevant unstructured data were unavailable in the FHRD. This study did not require ethics approval or patient informed consent because it used deidentified data and did not constitute human patient research according to the Common Rule.
Figure 1.
Attrition diagram. ICD, International Classification of Diseases; SLL, small lymphocytic lymphoma.
End points
The results of this study focused on analyses of confirmed real-world RR (rwRR) with select evidence (clinician-cited imaging, laboratory, or pathologic evidence of rwRR) because clinician use of an objective measure for evaluating response was expected to yield results more comparable with those from a clinical trial.23, 24, 25, 26 See the supplemental Methods for sensitivity definitions of rwRR and source evidence and detailed definitions and censoring methods for other outcomes.
Statistical analysis
This study evaluated a cohort of all eligible patients in the FHRD who met the inclusion criteria. Demographic and clinical characteristics were described for patients at initiation of the first index LOT and up to 2 subsequent index LOTs. For categorical variables, frequencies and percentages were generated. For continuous variables, descriptive statistics included medians and interquartile ranges.
Confirmed real-world duration of response (rwDOR) was estimated among patients with an rwRR during the LOT, using the first real-world complete response or real-world partial response (rwPR) as the index date. Time-to-event (TTE) outcomes were estimated from the start date of the index LOT to the date of event or death, as described in the supplemental Methods. Kaplan-Meier curves, median estimates, and probabilities at key milestones (12 and 24 months) were generated for rwDOR, rwPFS, real-world time to next treatment (rwTTNT), and rwOS.
Analyses were conducted for rwRR, rwDOR, rwPFS, rwTTNT, and rwOS in the first index LOT and second and third subsequent index LOTs. Analyses for the first index LOT were stratified by number of previous LOTs, history of chemotherapy exposure before the index LOT, and therapy class. For stratified analyses using the Kaplan-Meier method, P values were calculated using the log-rank test to compare survival distributions across strata (with the exception of sequence of index LOT strata, as those strata were not mutually exclusive). Sensitivity analyses included restricting to patients who received therapies with full approval (defined as therapeutic regimens with regular approval [not accelerated approval] to treat CLL as per the US Food and Drug Administration [FDA] product labels as of April 2024), with at least 12 months of potential follow-up, without evidence of previous clinical study drug exposure, without line 0 (line 0 flags patients with CLL treatment [as per unstructured data] that began >30 days before the start of structured activity indicating missing frontline therapy data), or without additional malignancy in the 3 years before index date. The association between rwPFS and rwOS was evaluated with criteria adapted from the Institute for Quality and Efficiency in Health Care27 (supplemental Methods).
Results
Patient characteristics
This study included 264 patients in the DE cohort, 157 in the DE-only subcohort, and 107 in the DR subcohort (see Figure 1 for attrition diagram). Table 1 describes the characteristics of the patients in the DE cohort and DE-only and DR subcohorts, stratified by the sequence of the index LOT. Characteristics between the cohorts were largely similar across index LOTs. Patients had a median age of ∼60 years at CLL diagnosis and 70 years at first index LOT and were predominantly male, White, and treated in the community setting, with an Eastern Cooperative Oncology Group performance status score of 0 to 2. Before the first index LOT, nearly all patients received ibrutinib as the most immediate cBTKi, and almost half had a known history of chemotherapy exposure (DE, n = 122 [46%]; DE only, n = 70 [45%]; DR, n = 52 [49%]). The median number (range) of LOTs before the first index LOT was 3 (1-12) for the DE cohort, 3 (1-12) for the DE-only subcohort, and 3 (2-12) for the DR subcohort. The median time from first index LOT to last confirmed activity (defined as the most recent record of vital information, a medication administration, a laboratory test/result being reported, or abstracted oral episode) was 1.1 years for DE, 1.2 years for DE only, and 0.9 years for DR.
Table 1.
Patient demographic and clinical characteristics in the DE cohort, DE-only subcohort, and DR subcohort, stratified by the sequence of the index line
| DE cohort (N = 264) |
DE-only subcohort (n = 157) |
DR subcohort (n = 107) |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| First index LOT (N = 264) | Second index LOT (n = 90) | Third index LOT (n = 35) | First index LOT (n = 157) | Second index LOT (n = 41) | Third index LOT (n = 16) | First index LOT (n = 107) | Second index LOT (n = 41) | Third index LOT (n = 12) | |
| Age at CLL diagnosis, Median (IQR), y | 60.0 (52.8-68.0) | 57.0 (51.0-65.8) | 55.0 (49.5-65.5) | 62.0 (52.0-69.0) | 56.0 (51.0-65.0) | 60.0 (51.8-68.5) | 58.0 (53.0-66.5) | 58.0 (50.0-66.0) | 52.5 (47.8-58.3) |
| Age at given index line, Median (IQR), y | 71.0 (63.0-78.0) | 69.5 (65.0-78.0) | 69.0 (65.5-78.0) | 73.0 (65.0-79.0) | 69.0 (63.0-76.0) | 69.5 (65.8-81.5) | 69.0 (62.5-77.0) | 70.0 (66.0-78.0) | 68.5 (61.0-73.0) |
| Birth sex, n (%) | |||||||||
| Male | 170 (64.4) | 64 (71.1) | 26 (74.3) | 98 (62.4) | 30 (73.2) | 12 (75.0) | 72 (67.3) | 29 (70.7) | 9 (75.0) |
| Female | 94 (35.6) | 26 (28.9) | 9 (25.7) | 59 (37.6) | 11 (26.8) | 4 (25.0) | 35 (32.7) | 12 (29.3) | 3 (25.0) |
| Race, n (%) | |||||||||
| White | 197 (74.6) | 63 (70.0) | 27 (77.1) | 128 (81.5) | 31 (75.6) | 13 (81.3) | 69 (64.5) | 25 (61.0) | 8 (66.7) |
| Black or African American | 36 (13.6) | 18 (20.0) | 6 (17.1) | 14 (8.9) | 7 (17.1) | 2 (12.5) | 22 (20.6) | 10 (24.4) | 3 (25.0) |
| Asian | 3 (1.1) | 0 (0.0) | 0 (0.0) | 2 (1.3) | 0 (0.0) | 0 (0.0) | 1 (0.9) | 0 (0.0) | 0 (0.0) |
| Other race | 18 (6.8) | 7 (7.8) | 2 (5.7) | 8 (5.1) | 2 (4.9) | 1 (6.3) | 10 (9.3) | 5 (12.2) | 1 (8.3) |
| Unknown | 10 (3.8) | 2 (2.2) | 0 (0.0) | 5 (3.2) | 1 (2.4) | 0 (0.0) | 5 (4.7) | 1 (2.4) | 0 (0.0) |
| Practice type, n (%) | |||||||||
| Academic | 72 (27.3) | 28 (31.1) | 9 (25.7) | 44 (28.0) | 13 (31.7) | 3 (18.8) | 28 (26.2) | 11 (26.8) | 3 (25.0) |
| Community | 187 (70.8) | 59 (65.6) | 24 (68.6) | 109 (69.4) | 26 (63.4) | 12 (75.0) | 78 (72.9) | 29 (70.7) | 8 (66.7) |
| Both | 5 (1.9) | 3 (3.3) | 2 (5.7) | 4 (2.5) | 2 (4.9) | 1 (6.3) | 1 (0.9) | 1 (2.4) | 1 (8.3) |
| Disease subtype, n (%) | |||||||||
| CLL | 215 (81.4) | 80 (88.9) | 29 (82.9) | 128 (81.5) | 37 (90.2) | 13 (81.3) | 87 (81.3) | 37 (90.2) | 10 (83.3) |
| CLL/SLL∗ | 33 (12.5) | 7 (7.8) | 5 (14.3) | 21 (13.4) | 3 (7.3) | 3 (18.8) | 12 (11.2) | 2 (4.9) | 1 (8.3) |
| SLL | 16 (6.1) | 3 (3.3) | 1 (2.9) | 8 (5.1) | 1 (2.4) | 0 (0.0) | 8 (7.5) | 2 (4.9) | 1 (8.3) |
| ECOG PS, n (%)† | |||||||||
| 0 | 85 (32.2) | 27 (30.0) | 10 (28.6) | 54 (34.4) | 16 (39.0) | 6 (37.5) | 29 (27.1) | 11 (26.8) | 3 (25.0) |
| 1 | 92 (34.8) | 38 (42.2) | 10 (28.6) | 50 (31.8) | 13 (31.7) | 5 (31.3) | 45 (42.1) | 19 (46.3) | 4 (33.3) |
| 2 | 26 (9.8) | 13 (14.4) | 9 (25.7) | 16 (10.2) | 5 (12.2) | 3 (18.8) | 9 (8.4) | 8 (19.5) | 3 (25.0) |
| 3-4 | 52 (19.7) | 10 (11.1) | 4 (11.4) | 29 (18.5) | 6 (14.6) | 2 (12.5) | 21 (19.6) | 2 (4.9) | 2 (16.7) |
| Unknown | 9 (3.4) | 2 (2.2) | 2 (5.7) | 8 (5.1) | 1 (2.4) | 0 (0.0) | 3 (2.8) | 1 (2.4) | 0 (0.0) |
| Rai stage at diagnosis, n (%) | |||||||||
| 0 | 21 (8.0) | 6 (6.7) | 3 (8.6) | 15 (9.6) | 3 (7.3) | 3 (18.8) | 6 (5.6) | 3 (7.3) | 0 (0.0) |
| I | 36 (13.6) | 11 (12.2) | 5 (14.3) | 21 (13.4) | 4 (9.8) | 2 (12.5) | 15 (14.0) | 6 (14.6) | 1 (8.3) |
| II | 26 (9.8) | 11 (12.2) | 3 (8.6) | 17 (10.8) | 8 (19.5) | 2 (12.5) | 9 (8.4) | 3 (7.3) | 1 (8.3) |
| III | 35 (13.3) | 12 (13.3) | 4 (11.4) | 18 (11.5) | 5 (12.2) | 1 (6.3) | 17 (15.9) | 6 (14.6) | 2 (16.7) |
| IV | 65 (24.6) | 19 (21.1) | 7 (20.0) | 33 (21.0) | 6 (14.6) | 2 (12.5) | 32 (29.9) | 10 (24.4) | 3 (25.0) |
| Unknown | 81 (30.7) | 31 (34.4) | 13 (37.1) | 53 (33.8) | 15 (36.6) | 6 (37.5) | 28 (26.2) | 13 (31.7) | 5 (41.7) |
| Deletion 17p status, n (%)‡ | |||||||||
| Present | 93 (35.2) | 31 (34.4) | 13 (37.1) | 48 (30.6) | 14 (34.1) | 5 (31.3) | 45 (42.1) | 14 (34.1) | 4 (33.3) |
| Absent | 131 (49.6) | 52 (57.8) | 20 (57.1) | 84 (53.5) | 24 (58.5) | 10 (62.5) | 47 (43.9) | 24 (58.5) | 8 (66.7) |
| Unknown/not documented | 14 (5.3) | 1 (1.1) | 0 (0.0) | 8 (5.1) | 0 (0.0) | 0 (0.0) | 6 (5.6) | 1 (2.4) | 0 (0.0) |
| No evidence of testing | 26 (9.8) | 6 (6.7) | 2 (5.7) | 17 (10.8) | 3 (7.3) | 1 (6.3) | 9 (8.4) | 2 (4.9) | 0 (0.0) |
| TP53 mutation status, n (%)§ | |||||||||
| Present | 20 (7.6) | 6 (6.7) | 2 (5.7) | 12 (7.6) | 2 (4.9) | 1 (6.3) | 8 (7.5) | 3 (7.3) | 1 (8.3) |
| Absent | 37 (14.0) | 16 (17.8) | 4 (11.4) | 24 (15.3) | 7 (17.1) | 3 (18.8) | 13 (12.1) | 8 (19.5) | 1 (8.3) |
| Unknown/not documented | 18 (6.8) | 5 (5.6) | 1 (2.9) | 12 (7.6) | 3 (7.3) | 1 (6.3) | 6 (5.6) | 1 (2.4) | 0 (0.0) |
| No evidence of testing | 189 (71.6) | 63 (70.0) | 28 (80.0) | 109 (69.4) | 29 (70.7) | 11 (68.8) | 80 (74.8) | 29 (70.7) | 10 (83.3) |
| IgHV status, n (%)¶ | |||||||||
| Mutated | 32 (12.1) | 7 (7.8) | 1 (2.9) | 23 (14.6) | 4 (9.8) | 0 (0.0) | 9 (8.4) | 3 (7.3) | 1 (8.3) |
| Unmutated | 105 (39.8) | 37 (41.1) | 15 (42.9) | 65 (41.4) | 18 (43.9) | 7 (43.8) | 40 (37.4) | 19 (46.3) | 6 (50.0) |
| Unknown/not documented | 9 (3.4) | 6 (6.7) | 3 (8.6) | 3 (1.9) | 3 (7.3) | 2 (12.5) | 6 (5.6) | 1 (2.4) | 0 (0.0) |
| No evidence of testing | 118 (44.7) | 40 (44.4) | 16 (45.7) | 66 (42.0) | 16 (39.0) | 7 (43.8) | 52 (48.6) | 18 (43.9) | 5 (41.7) |
| Line number of given index line, n (%)|| | |||||||||
| 2-3 | 95 (36.0) | 1 (1.1) | 0 (0.0) | 55 (35.0) | 1 (2.4) | 0 (0.0) | 34 (31.8) | 0 (0.0) | 0 (0.0) |
| 4-5 | 114 (43.2) | 53 (58.9) | 10 (28.6) | 72 (45.9) | 23 (56.1) | 3 (18.8) | 47 (43.9) | 24 (58.5) | 5 (41.7) |
| ≥6 | 55 (20.8) | 36 (40.0) | 25 (71.4) | 30 (19.1) | 17 (41.5) | 13 (81.3) | 26 (24.3) | 17 (41.5) | 7 (58.3) |
| Year of given index line start, n (%)|| | |||||||||
| 2016-2018 | 38 (14.4) | 5 (5.6) | 3 (8.6) | 24 (15.3) | 3 (7.3) | 1 (6.3) | 11 (10.3) | 2 (4.9) | 1 (8.3) |
| 2019-2020 | 120 (45.5) | 26 (28.9) | 9 (25.7) | 67 (42.7) | 14 (34.1) | 5 (31.3) | 49 (45.8) | 9 (22.0) | 2 (16.7) |
| 2021-2022 | 106 (40.2) | 59 (65.6) | 23 (65.7) | 66 (42.0) | 24 (58.5) | 10 (62.5) | 47 (43.9) | 30 (73.2) | 9 (75.0) |
| Presence of line 0, n (%)# | |||||||||
| Yes | 96 (36.4) | 42 (46.7) | 19 (54.3) | 49 (31.2) | 15 (36.6) | 7 (43.8) | 47 (43.9) | 23 (56.1) | 9 (75.0) |
| No | 168 (63.6) | 48 (53.3) | 16 (45.7) | 108 (68.8) | 26 (63.4) | 9 (56.3) | 60 (56.1) | 18 (43.9) | 3 (25.0) |
| Last cBTKi received before given index line, n (%)|| | |||||||||
| Acalabrutinib only | 20 (7.6) | 18 (20.0) | 8 (22.9) | 16 (10.2) | 6 (14.6) | 3 (18.8) | 6 (5.6) | 12 (29.3) | 3 (25.0) |
| Ibrutinib only | 243 (92.0) | 71 (78.9) | 27 (77.1) | 140 (89.2) | 34 (82.9) | 13 (81.3) | 101 (94.4) | 29 (70.7) | 9 (75.0) |
| Zanubrutinib only | 0 (0.0) | 1 (1.1) | 0 (0.0) | 0 (0.0) | 1 (2.4) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Combination with >1 cBTKi | 1 (0.4) | 0 (0.0) | 0 (0.0) | 1 (0.6) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Evidence of a clinical study drug in LOTs before given index line, n (%),∗∗|| | |||||||||
| Yes | 30 (11.4) | 16 (17.8) | 4 (11.4) | 20 (12.7) | 7 (17.1) | 1 (6.3) | 10 (9.3) | 8 (19.5) | 2 (16.7) |
| No | 234 (88.6) | 74 (82.2) | 31 (88.6) | 137 (87.3) | 34 (82.9) | 15 (93.8) | 97 (90.7) | 33 (80.5) | 10 (83.3) |
| Time from given index date to last confirmed activity, median (IQR),†† y | 1.1 (0.5-2.3) | 0.9 (0.5-1.5) | 0.5 (0.3-1.3) | 1.2 (0.6-2.5) | 0.9 (0.5-1.9) | 0.5 (0.4-1.0) | 0.9 (0.4-1.7) | 0.7 (0.3-1.5) | 0.5 (0.3-1.5) |
ECOG PS, Eastern Cooperative Oncology Group performance status; IgHV, immunoglobulin heavy chain variable region gene; IQR, interquartile range; SLL, small lymphocytic lymphoma; TP53, tumor protein p53.
As documented by the health care provider.
Extracted from unstructured documents using natural language processing when ECOG PS information from structured data is unavailable in the ECOG table. ECOG PS values represent the score closest to the index date within the eligible window (30 days before 7 days after the given index line). ECOG PS values of 5 are transformed to “unknown” for deidentification purposes.
Status at any time before or up to 30 days after first index line start date. “Unknown/not documented” represents patients who had fluorescence in situ hybridization and/or classical cytogenetic testing in the window of interest but did not have documentation of results or had only unsuccessful/indeterminate results. “No evidence of testing” represents patients without evidence of testing in the window of interest. Tests with unknown dates are considered to be outside the window of interest. If there are multiple tests in the window of interest, the following hierarchy of test results was used: (1) present, (2) absent, (3) unknown/not documented.
Status at any time before or up to 30 days after first index line start date. “Unknown/not documented” represents patients who had molecular cytogenetic testing in the window of interest but did not have documentation of results or had only unsuccessful/indeterminate results. “No evidence of testing” represents patients without evidence of testing in the window of interest. Tests with unknown dates are considered to be outside the window of interest. If there are multiple tests in the window of interest, the following hierarchy of test results was used: (1) present, (2) absent, (3) unknown/not documented.
The first index LOT was a nonmaintenance LOT initiated after the earliest cBTKi- and BCL2i-containing LOTs, on or before 30 November 2022, and could not include a clinical study drug or be preceded by disease transformation to a more aggressive lymphoma (Richter transformation). The second index LOT was the second LOT after cohort entry (ie, next LOT after first index LOT), and the third index LOT was the third LOT after cohort entry (ie, next LOT after second LOT).
Status at any time before or up to 30 days after first index line start date. “Unknown/not documented” represents patients who had testing in the window of interest but did not have documentation of results or had only unsuccessful/indeterminate results. “No evidence of testing” represents patients without evidence of testing in the window of interest. Tests with unknown dates are considered to be outside the window of interest. If there are multiple tests in the window of interest, the following hierarchy of test results was used: (1) mutated; (2) unmutated; (3) unknown/not documented.
Presence of line 0 indicates that patients' CLL treatment start (captured through unstructured data) is >30 days before the start of structured activity, suggesting an increased possibility of missing frontline data.
Note that clinical study drugs are masked and drug names are not known.
Last confirmed activity was defined as most recent record of vital information, a medication administration, a laboratory test/result being reported, or abstracted oral episode.
Treatment patterns
The first index LOT occurred on, or after, 2019 for most patients in the study (DE, n = 226 [86%]; DE only, n = 133 [85%]; DR, n = 96 [90%]; Table 1), and included a CLL therapy regimen with full approval by the FDA in many cases (DE, 51%; DE only, 61%; DR, 43%; supplemental Tables 2-5). Figure 2 shows treatment patterns across index LOTs in the DE cohort, DE-only subcohort, and DR subcohort. A cBTKi- and/or a BCL2i-based therapy was used most commonly as retreatment in the first index LOT (DE, 69%; DE only, 75%; DR, 54%) and in approximately one-third to two-thirds of patients in the second (DE, 53%; DE only, 61%; DR, 41%) and third index LOTs (DE, 43%; DE only, 63%; DR, 33%). No patient in the DR subcohort received a cBTKi in the third index LOT. Use of chemotherapy-based regimens (with or without anti-CD20 therapy) increased with each index LOT in the DE cohort and DR subcohort (DE:, 9.5%, 24.5%, and 31.4% in the first, second, and third index LOTs, respectively; DR, 11.2%, 26.9%, and 50%). Use in the DE-only subcohort increased similarly from the first to second index LOTs; however, a further increase was not observed in the third index LOT (8.9%, 26.8%, and 12.5%). Patients without a line 0 (ie, potentially missing frontline therapy data) had the same top 4 treatment classes for the first index LOT (data not shown).
Figure 2.
Treatment pattern by index line and cohort/subcohort. Treatment patterns in the DE cohort (A), DE-only subcohort (B), and DR subcohort (C). Anti-CD20, anti-CD20 monoclonal antibody; chemo, chemotherapy; mono, monotherapy.
rwRR and rwDOR
rwRR was similar for the DE cohort regardless of LOT (first index LOT, 39% [102/264] vs second index LOT, 42% [38/90] vs third index LOT, 40% [14/35]). rwRR was also similar between the DE-only subcohort and DR subcohort for the first index LOT (41% [65/157] vs 36% [38/107]), and third index LOT (38% [6/16] vs 42% [5/12]); however, rwRR was higher for the DE-only subcohort than for the DR subcohort in the second index LOT (51% [21/41] vs 34% [14/41]).
Among patients with a real-world response, median rwDOR was longer in the DE cohort and DE-only subcohort than in the DR subcohort for the first index LOT (20.3 vs 26.2 vs 11.5 months; Table 2). Furthermore, median rwDOR indexed to the second (DE, 6.0 months; DE only, 12.2 months; DR, 6.0 months) and third (DE, 3.1 months; DE only, 3.5 months; DR, 2.9 months) index LOT was shorter than when indexed to the first index LOT (Table 2). Comparisons between the DE cohort, DE-only subcohort, and DR subcohort in the first index line remained similar when implementing alternative definitions of rwRR and using different types of clinician-cited source evidence (supplemental Table 6).
Table 2.
Real-world response and clinical outcomes by index LOT in the DE cohort, DE-only subcohort, and DR subcohort
| DE cohort |
DE-only subcohort |
DR subcohort |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| First index LOT (N = 264) | Second index LOT (n = 90) | Third index LOT (n = 35) | First index LOT (n = 157) | Second index LOT (n = 41) | Third index LOT (n = 16) | First index LOT (n = 107) | Second index LOT (n = 41) | Third index LOT (n = 12) | |
| rwRR, n (%; 95% CI)∗ | 102 (39; 33-45) | 38 (42; 32-53) | 14 (40; 24-58) | 65 (41; 34-50) | 21 (51; 35-67) | 6 (38; 16-64) | 38 (36; 27-45) | 14 (34; 21-51) | 5 (42; 16-71) |
| Best response for rwRR, n (%)† | |||||||||
| rwCR | 26 (10) | 3 (3) | 2 (6) | 20 (13) | 3 (7.3) | 1 (6.3) | 6 (6) | 0 (0) | 0 (0) |
| rwPR | 76 (29) | 35 (39) | 12 (34) | 45 (29) | 18 (44) | 5 (31) | 32 (30) | 14 (34) | 5 (42) |
| No confirmed response | 117 (44) | 39 (43) | 15 (43) | 64 (41) | 15 (37) | 6 (38) | 48 (45) | 23 (56) | 6 (50) |
| No rwA | 45 (17) | 13 (14) | 6 (17) | 28 (18) | 5 (12) | 4 (25) | 21 (20) | 4 (10) | 1 (8) |
| rwDOR | |||||||||
| Median (95% CI), mo | 20.3 (13.5-32.6) | 6.0 (5.5 to NR) | 3.1 (2.7 to NR) | 26.2 (11.9 to NR) | 12.2 (6.0 to NR) | 3.5 (1.8 to NR) | 11.5 (7.9 to 23.5) | 6.0 (3.6 to NR) | 2.9 (2.7 to NR) |
| 1-year probability, % (95% CI) | 61 (50-74) | 43 (27-66) | NE | 62 (49-79) | 55 (34-90) | NE | 50 (33-75) | 44 (22-89) | NE |
| 2-year probability, % (95% CI) | 43 (31-58) | NE | NE | 59 (46-76) | NE | NE | 13 (4-45) | NE | NE |
| rwPFS | |||||||||
| Median (95% CI), mo | 8.8 (7.5-10.6) | 5.1 (4.0-7.3) | 4.3 (2.9-6.3) | 11.3 (9.2-17.1) | 10.9 (4.3-15.6) | 5.7 (2.4 to NR) | 5.0 (4.0-7.3) | 3.6 (2.6-5.7) | 4.7 (2.7 to NR) |
| 1-year probability, % (95% CI) | 41 (35-48) | 25 (17-36) | 4 (1-28) | 50 (42-59) | 38 (25-58) | NE | 23 (16-33) | 17 (9-35) | NE |
| 2-year probability, % (95% CI) | 24 (18-30) | 8 (3-19) | NE | 36 (28-46) | 10 (4-30) | NE | 9 (5-18) | 10 (4-28) | NE |
| rwTTNT | |||||||||
| Median (95% CI), mo | 13.8 (10.3-18.3) | 7.9 (5.7-11.3) | 5.7 (4.2-14.6) | 22.1 (12.2-38.1) | 9.8 (5.6-17.2) | 9.0 (3.8 to NR) | 8.0 (6.2-11.5) | 5.9 (4.4-11.4) | 5.0 (4.2 to NR) |
| 1-year probability, % (95% CI) | 52 (46-58) | 35 (26-47) | 32 (18-54) | 58 (51-67) | 42 (29-61) | 37 (16-84) | 37 (29-48) | 28 (17-47) | 28 (11-72) |
| 2-year probability, % (95% CI) | 35 (29-42) | 21 (12-34) | 15 (5-44) | 47 (39-57) | 22 (11-44) | 25 (7.9-78) | 18 (11-29) | 19 (9-42) | NE |
| rwOS | |||||||||
| Median (95% CI), mo | 25.2 (22.1-37.1) | 15.2 (11.4-22.6) | 6.6 (5.7 to NR) | 38.2 (25.4 to NR) | 15.6 (11.5 to NR) | 9.6 (6.0 to NR) | 14.7 (12.3 to 22.6) | 11.4 (5.9 to NR) | 6.6 (5.0 to NR) |
| 1-year probability, % (95% CI) | 71 (66-77) | 56 (46-68) | 45 (30-66) | 74 (68-82) | 61 (47-79) | 47 (26-88) | 60 (51-71) | 45 (32-65) | 46 (24-87) |
| 2-year probability, % (95% CI) | 52 (46-60) | 35 (25-49) | 26 (13-51) | 61 (53-71) | 46 (32-67) | 36 (15-82) | 33 (24-45) | 29 (16-53) | 34 (15-81) |
CI, confidence interval; NE, not estimable; rwA, real-world assessment; rwCR, real-world complete response.
rwRR was calculated as confirmed rwRR. All responses were based on select evidence (ie, clinician-cited evidence limited to imaging, pathology, and/or laboratory report records).
Best response for rwRR was defined hierarchically as either rwCR or rwPR for patients with a confirmed response in the LOT. Patients with at least 1 rwA in the LOT, but with no confirmed response were categorized as having “no confirmed response.” All other patients with no rwA in the LOT were categorized as such. The proportion of patients in each of these mutually exclusive categories were calculated among patients who received the LOT of interest.
When rwRR was stratified by the number of previous LOTs before the index LOT, results were similar (1-3 vs ≥4 previous LOTs: DE, 41% vs 34%; DE only, 42% vs 40%; DR, 38% vs 32%). However, median rwDOR was approximately twice as long for patients with fewer previous LOTs in the DE cohort (1-3 vs ≥4 previous LOTs, 23.5 vs 10.9 months) and DR subcohort (2-3 vs ≥4 previous LOTs, 13.5 vs 7.9 months), but the differences were not statistically significant (P > .07; Table 3). In contrast, within the DE-only subcohort, median rwDOR was >3 times longer for patients with fewer previous LOTs (1-3 vs ≥4 previous LOTs, 32.6 vs 9.9 months), representing a statistically significant difference (P = .012; Table 3).
Table 3.
Real-world response, rwDOR, and rwTTE outcomes in the DE cohort, DE-only subcohort, and DR subcohort, stratified by number of previous LOT
| DE (N = 264) |
DE-only subcohort (n = 157) |
DR (n = 107) |
||||
|---|---|---|---|---|---|---|
| 1-3 previous lines (n = 175) | ≥4 previous lines (n = 89) | 1-3 previous lines (n = 104) | ≥4 previous lines (n = 53) | 2-3 previous lines (n = 69) | ≥4 previous lines (n = 38) | |
| rwRR, % (95% CI)∗ | 41 (34-49) | 34 (24-45) | 42 (33-52) | 40 (27-54) | 38 (27-50) | 32 (18-49) |
| Best response rwCR, n (%)✝ | 20 (11) | 6 (7) | 16 (15) | 4 (7.5) | 3 (4) | 3 (8) |
| Best response rwPR, n (%)✝ | 52 (30) | 24 (27) | 28 (27) | 17 (32) | 23 (33) | 9 (24) |
| rwDOR, median (95% CI), mo | 23.5 (14.3 to NR) | 10.9 (8.3 to NR) | 32.6 (25.3 to NR) | 9.9 (5.7 to NR) | 13.5 (9.2 to NR) | 7.9 (3.2 to NR) |
| rwPFS, median (95% CI), mo | 11.0 (8.1-14.3) | 7.0 (5.1-9.3) | 19.1 (12.5-36.7) | 7.6 (5.2-10.6) | 4.6 (3.7-7.4) | 5.1 (3.7-8.3) |
| rwTTNT, median (95% CI), mo | 18.4 (14.2-23.4) | 8.2 (5.7-11.5) | 38.1 (22.7 to NR) | 9.5 (6.7-13.8) | 8.9 (6.3-15.4) | 5.5 (4.6-11.5) |
| rwOS, median (95% CI), mo | 35.6 (23.1-40.7) | 18.3 (12.2-37.5) | NR (37.1 to NR) | 19.8 (10.7 to NR) | 15.4 (12.3-25.2) | 14.1 (11.1 to NR) |
CI, confidence interval; rwA, real-world assessment; rwCR, real-world complete response.
rwRR was calculated as confirmed rwRR. All responses were based on select evidence (ie, clinician-cited evidence limited to imaging, pathology, and/or laboratory report records).
Best response for rwRR was defined hierarchically as either rwCR or rwPR for patients with a confirmed response in the LOT. Patients with at least 1 rwA in the LOT, but with no confirmed response were categorized as having “no confirmed response.” All other patients with no rwA in the LOT were categorized as such. The proportion of patients in each of these mutually exclusive categories were calculated among patients who received the LOT of interest.
When analyses were stratified by therapy class, rwRR was higher in patients receiving BCL2i with anti-CD20–based therapy or cBTKi-based therapy compared with those receiving other major therapy classes (Table 4). The longest median rwDOR was observed in patients whose first index LOT was BCL2i with anti-CD20–based therapy in the DE cohort (26.2 months, n = 78), phosphoinositide 3-kinase inhibitors (PI3Ki) with or without anti-CD20 in the DE-only subcohort (NR, n = 8), and cBTKi-based therapy in the DR subcohort (20.3 months, n = 30; Table 4). Of note, the cBTKi-based therapy class included cBTKi and BCL2i combination–based therapy in addition to other cBTKi-based therapy. cBTKi-based therapies consisted of cBTKi and BCL2i combination–based therapy in 28% of the DE cohort and 18% of the DE-only subcohort, whereas this was slightly higher (43%) in the DR subcohort (data not shown). Results of other stratified and sensitivity analyses were generally similar to those from the main analysis (supplemental Table 7).
Table 4.
Real-world response, rwDOR, and rwTTE outcomes in the DE cohort, DE-only subcohort, and DR subcohort, stratified by therapy class of first index LOT
| DE cohort (N = 264) |
DE-only subcohort (n = 157) |
DR subcohort (n = 107) |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CIT, anti-CD20, or chemotherapy (n = 47) | PI3K ± anti-CD20 (n = 29) | cBTKi-based∗ (n = 64) | BCL2i + anti-CD20 (n = 78) | CIT, anti-CD20, or chemotherapy (n = 28) | PI3K ± anti-CD20 (n = 8) | cBTKi-based∗ (n = 34) | BCL2i + anti-CD20 (n = 50) | CIT, anti-CD20, or chemotherapy (n = 21) | PI3K ± anti-CD20 (n = 23) | cBTKi-based∗ (n = 30) | BCL2i + anti-CD20 (n = 27) | |
| rwRR, % (95% CI)✝ | 26.0 (14.0-41.0) | 24.0 (11.0-44.0) | 39.0 (27.0-52.0) | 58.0 (46.0-69.0) | 36 (19-56) | 25 (4.5-64) | 38 (23-56) | 62 (47-75) | 19.0 (6.0-43.0) | 26.0 (11.049.0) | 47.0 (29.0-65.0) | 48.0 (29.0-68.0) |
| Best response rwCR, n (%)‡ | 3 (6.0) | 0 (0.0) | 1 (2.0) | 18 (23.0) | 3 (11) | 0 (0) | 1 (2.9) | 13 (26) | 1 (5.0) | 0 (0.0) | 0 (0.0) | 4 (15.0) |
| Best response rwPR, n (%)‡ | 9 (19.0) | 7 (24.0) | 24 (38.0) | 27 (35.0) | 7 (25) | 2 (25) | 12 (35) | 18 (36) | 3 (14.0) | 6 (26.0) | 14 (47.0) | 9 (33.0) |
| rwDOR, median (95% CI), mo | 17.9 (9.2 to NR) | 23.5 (4.0 to NR) | 8.5 (5.7 to NR) | 26.2 (14.3 to NR) | 25.3 (10.5 to NR) | NR (NR to NR) | 8.4 (5.7 to NR) | 35.4 (26.2 to NR) | 9.2 (2.8 to NR) | 4.0 (4.0 to NR) | 20.3 (4.6 to NR) | 13.5 (7.9 to NR) |
| rwPFS, median (95% CI), mo | 7.0 (2.5-9.3) | 4.0 (2.4-6.4) | 8.0 (6.3-14.6) | 13.8 (10.0-26.7) | 8.1 (3.4 to NR) | 5.9 (1.1 to NR) | 11.3 (6.9 to NR) | 26.7 (10.3 to NR) | 2.0 (1.4 to 8.9) | 4.0 (2.5-5.1) | 6.6 (4.6-14.0) | 8.8 (6.5-14.3) |
| rwTTNT, median (95% CI), mo | 5.7 (3.0-10.5) | 5.3 (3.4-13.5) | 14.7 (7.0-25.2) | 26.6 (18.4-49.7) | 9.5 (3.2 to NR) | 4.4 (2.1 to NR) | 19.1 (12.2 to NR) | 44.4 (23.4 to NR) | 2.6 (1.6-8.4) | 5.3 (4.0-14.2) | 10.7 (6.2 to NR) | 15.4 (11.0 to NR) |
| rwOS, median (95% CI), mo | 19.9 (11.2 to NR) | 6.4 (4.7-14.2) | 32.0 (23.7 to NR) | 40.7 (27.6 to NR) | 37.5 (17.0 to NR) | 6.4 (5.8 to NR) | 53.5 (19.1 to NR) | NR (39.0 to NR) | 8.3 (3.6-19.9) | 6.4 (4.7-14.9) | 32.0 (14.1 to NR) | 22.6 (15.4 to NR) |
Anti-CD20, anti-CD20 monoclonal antibody; CI, confidence interval; CIT, chemoimmunotherapy; rwA, real-world assessment; rwCR, real-world complete response.
Includes cBTKi plus BCL2i combination therapies.
rwRR was calculated as confirmed rwRR. All responses were based on select evidence (ie, clinician-cited evidence limited to imaging, pathology, and/or laboratory report records).
Best response for rwRR was defined hierarchically as either rwCR or rwPR for patients with a confirmed response in the LOT. Patients with at least 1 rwA in the LOT but with no confirmed response were categorized as having “no confirmed response.” All other patients with no rwA in the LOT were categorized as such. The proportion of patients in each of these mutually exclusive categories were calculated among patients who received the LOT of interest.
rwPFS, rwTTNT, and rwOS
Median times to clinical and treatment-related outcomes were shorter for the DR subcohort than for the DE cohort and DE-only subcohort in the first index LOT (DR vs DE vs DE-only medians: rwPFS, 5.0 vs 8.8 vs 11.3 months; rwTTNT, 8.0 vs 13.8 vs 22.1 months; rwOS, 14.7 vs 25.2 vs 38.2 months), and were shorter in subsequent index LOTs for all groups (Table 2; supplemental Figure 2). When stratified by number of previous LOTs before the index LOT (1-3 vs ≥4 LOTs; Table 3), outcomes were generally similar across strata in the DR subcohort (P > .30 for all outcomes) but longer in patients with fewer vs more previous LOTs before index LOT in the DE cohort (P < .05 for all outcomes) and DE-only subcohort (P < .01 for all outcomes; supplemental Figure 3).
In each cohort/subcohort, median rwTTNT, rwPFS, and rwOS were longer for patients receiving a BCL2i with anti-CD20–based therapy or cBTKi-based therapy (including cBTKi and BCL2i combination– or other cBTKi-based therapy) in the first index LOT than for those receiving chemotherapy and/or anti-CD20–based therapy or PI3Ki with/without anti-CD20–based therapy (Table 4; Figure 3). In the DE cohort, DE-only subcohort, and DR subcohort, results of sensitivity analyses were generally similar to those for the main analysis (supplemental Table 8).
Figure 3.
Kaplan-Meier curves for TTE outcomes in the DE cohort, DE-only subcohort, and DR subcohort, stratified by therapy class of the first index LOT. (A) DE cohort, rwTTNT; (B) DE cohort, rwPFS; (C) DE cohort, rwOS; (D) DE-only subcohort, rwTTNT; (E) DE-only subcohort, rwPFS; (F) DE-only subcohort, rwOS; (G) DR subcohort, rwTTNT; (H) DR subcohort, rwPFS; (I) DR subcohort, rwOS. Anti-CD20, anti-CD20 monoclonal antibody; chemo, chemotherapy; CI, confidence interval; CIT, chemoimmunotherapy.
Discussion
This large, real-world study highlights the unmet need among patients with DR CLL, primarily in the community oncology setting and in the absence of investigational therapies. Although rwRR was similar between the DE cohort, DE-only subcohort, and DR subcohort (39% vs 41% vs 36%), median rwDOR, rwPFS, rwTTNT, and rwOS were shorter in the DR subcohort than in the DE cohort and DE-only subcohort, reflecting reduced response durability after cBTKi/BCL2i failure. The discordance between similar rwRRs and markedly shorter TTE outcomes suggests that early disease control may be achievable in DR CLL but is not sustained, potentially reflecting underlying genomic instability or acquired resistance mechanisms that are not captured in real-world data sets. Retreatment with cBTKi- and/or BCL2i-based regimens was common in our study population (DE, 69%; DE only, 75%; DR, 54%) and use of chemotherapy-based regimens increased with each subsequent index LOT, emphasizing a lack of novel treatment options for this patient population.
The proportion of patients retreated with cBTKi- and/or BCL2i-based regimens in our study was high, consistent with a previous real-world study,10 and our rwRR results align with previous studies reporting rwRRs of 34% to 52% for DE and 24% to 40% for DR.12,20,28 rwRR has not been studied in DE-only CLL, to our knowledge. In our DE cohort and DE-only and DR subcohorts, we observed higher rwRRs with BCL2i with anti-CD20–based therapy and cBTKi-based therapies than with PI3Ki with/without anti-CD20–based therapy and chemotherapy with/without anti-CD20–based therapy. Previous studies reported similar findings with the exception of PI3Ki with/without anti-CD20–based therapy, which had rwRRs similar to those with BCL2i with anti-CD20–based therapy and cBTKi-based therapies.20,29,30 Furthermore, 1 small cohort study demonstrated a very high rwRR (100%) in 11 patients with DR CLL who received combination therapy with cBTKi and BCL2i, a more intensive off-label regimen that may not be used as commonly in patients with DE-only CLL.18 A higher proportion of patients in our DR subcohort vs the DE cohort and DE-only subcohort received combination cBTKi and BCL2i therapy, which may, in part, account for the higher rwRR and longer rwDOR observed in cBTKi-based therapy class subgroup of the DR subcohort compared with the DE cohort and DE-only subcohort.
The shorter median TTE outcomes in the DR subcohort than in the DE cohort and DE-only subcohort is consistent with recent publications (DR vs DE vs DE only: median OS in months, 21.2 vs 36.6 vs NR; and median PFS in months, 6.8 vs 8.4 vs 15.414; DR vs DE only: median OS in months, 26.0 vs NR15), reflecting reduced response durability among patients with cBTKi/BCL2i failure. When stratifying by number of previous LOTs, similar to what has been demonstrated in another study, TTE outcomes were longer for patients with fewer (1-3) vs more (≥4) LOTs before the index LOT among the DE cohort and DE-only subcohort.10 However, outcomes did not appear to differ by number of previous LOTs in the DR subcohort, further suggesting poor prognosis regardless of treatment history once dual resistance develops. When stratified by select classes of therapy, compared with the DE cohort and DE-only subcohort, TTE outcomes were also shorter among patients in the DR subcohort who received chemotherapy and/or anti-CD20 therapy and BCL2i with anti-CD20 therapy. Patients treated with PI3Ki-based regimens had similarly poor rwPFS, rwTTNT, and rwOS outcomes among those in the DE and DR groups, likely due to the large overlap in populations (23 of 29 patients in the DE cohort receiving PI3Ki were also in the DR cohort). rwDOR, however, appeared more favorable in the DE cohort, whereas estimates in the DE-only subcohort were limited by small sample size (n = 8). Differences in outcomes between the DE cohort, DE-only subcohort, and DR subcohort among cBTKi-treated patients were found, with rwDOR longest in the DR subcohort, and rwOS highest in the DE-only subcohort. These findings support treating DE and DR CLL as biologically and clinically distinct populations in the design and interpretation of clinical trials, including consideration of DR-restricted enrollment or stratified analyses, similar to approaches used in other refractory hematologic malignancies. Standardized definitions of DE and DR may be required to facilitate the design of such studies.
Strengths and limitations
Our study has notable strengths in that it was a large, retrospective study, leveraging structured electronic health record–derived and unstructured (abstracted) patient-level data from routine clinical practice in community oncology and academic cancer centers across the United States, which have patient populations generally considered to be different from a clinical trial population. In addition, index lines that included an investigational therapy were excluded from analysis. Thus, our results more closely represent care in community-based settings in which clinicians use treatments readily available to them. Our CLL study population was derived from the FHRD, which consists of patients with generally similar demographic and geographic distribution to those in the Surveillance, Epidemiology, and End Results (SEER) program and National Program of Cancer Registries.31 Many subgroup and sensitivity analyses were conducted, including multiple definitions of rwRR, which allowed us to evaluate the impact of various approaches on the results.
Several limitations must be considered when interpreting the findings of this study. Despite study patients being generally similar to other US patients with CLL, the results may not be generalizable to global populations of patients with CLL due to variation in patient demographics and clinical characteristics and standards of care. In particular, data on resistance mutations (eg, BTK or BCL2 variants), complex karyotype, or clonal evolution were unavailable, precluding assessment of their contribution to response and TTE outcomes. In addition, there may have been some misclassification of LOTs because more than one-third of patients had a line 0, an indicator for missing frontline therapy data. Results from sensitivity analyses restricted to patients without a line 0, which were conducted to evaluate a subset of patients whose treatment histories were most complete and least likely to be misclassified, were similar to the main analysis results. Given that clinical study drugs are masked in the FHRD and therefore could not determine whether a patient had been exposed to a cBTKi and/or BCL2i as an investigational agent, additional sensitivity analyses were conducted in which patients with evidence of previous clinical study drug exposure were excluded to minimize misclassification. The DE-only subcohort comprised patients without documented refractory status, but absence of documentation does not necessarily confirm nonexistence, which could result in some misclassification of these patients. Our main analysis used a “select evidence” approach clinician-cited evidence from imaging, laboratory tests, or pathology that aligns with previous real-world studies and approximates trial methods. Although not identical to clinical trial protocols, this approach reflects the realities of routine practice, in which variability in response assessments and definitions can affect outcome estimates.12,24, 25, 26,32 Thus, response type and associated outcomes captured in real-world data may vary between subcohorts, treatments, and studies, and when compared with clinical trials.23, 24, 25, 26 For example, rwPR with lymphocytosis could not be distinguished as a separate response category from rwPR. In clinical trials, PR with lymphocytosis is not accepted by the FDA as part of the primary end point definition of overall RR. The inability to distinguish rwPR with lymphocytosis may have led to a higher rwRR in this study than if responses could have been restricted to patients who achieved a PR without lymphocytosis or better. This was mitigated by (1) the presence of the "other" (mixed) rwA category (which was not considered a response in rwRR analyses), (2) requiring a confirmatory rwRR, and (3) noninclusion of end points that could be more sensitive to this limitation, especially real-world time to first response. Accordingly, comparisons of response type across LOT should be interpreted cautiously, whereas TTE outcomes are more robust to this limitation.
Certain subanalyses, especially among patients with DR CLL and stratifications, were limited by small sample sizes and should be interpreted with caution. In addition, given the survival reported in similar previous research (median, rwOS of ∼25 months from the start of the first index LOT after double discontinuation11) and relatively short observed follow-up times in our study, mortality data may not have been sufficiently mature for certain rwOS analyses. However, when analyses were restricted to index LOTs with at least 12 months of potential follow-up from the index LOT start date, results were similar to those of the main analysis. Finally, the data cutoff date for our study was 31 May 2023, before the accelerated approvals of pirtobrutinib and lisocabtagene maraleucel (a chimeric antigen receptor T-cell [CAR-T] therapy) therapies for CLL. Thus, our results do not reflect the contribution of these therapies to patient outcomes in the real-world setting.
The findings of this study suggest a lack of standard of care and lack of effective options/durable therapy for patients with CLL who require subsequent treatment after double exposure to cBTKi and BCL2i therapies. Although rwRRs were similar between patients with DE, DE only, and DR CLL, TTE outcomes (rwDOR, rwPFS, rwTTNT, and rwOS) were shorter for patients with DR CLL and became shorter with each subsequent index LOT for all groups. Our results suggest a high unmet need in the DE CLL population, particularly in those with CLL that is DR to cBTKi and BCL2i therapies. Even with the recent approvals of new therapies, PFS (pirtobrutinib, 16.8 months33; CAR-T, 11.9 months34) may not improve substantially over the rwPFS observed in this study (8-11 months overall) when neither of these therapies were available, with the caveat that clinical trial PFS and rwPFS are not directly comparable. Novel therapies with different mechanisms of action (eg, bispecific antibodies, CAR-T therapies, BTK degraders) are being evaluated in DR CLL, highlighting the importance of future studies that continue to evaluate the durability of response with different treatments in CLL populations.
Conflict-of-interest disclosure: J.P. reports employment with Merck Sharp & Dohme LLC (a subsidiary of Merck & Co, Inc, Rahway, NJ); and stock ownership in Merck & Co, Inc, Rahway, NJ. C.P. reports employment with Pine Mountain Consulting, LLC; stock ownership in Roche; and consulting fees from Flatiron Health, Canopy Care, IQ Solutions, Medicus Economics, Omada Health, Outcomes4Me, Pomelo Care, and TTi Health Research and Economics. A.S. reports employment with Flatiron Health Inc, an independent member of the Roche Group; and reports stock ownership in Roche. E.D.N., J.Y., E.B., I.S., C.S., X.L., and M.Z.H.F. report employment with Merck Sharp & Dohme LLC (a subsidiary of Merck & Co, Inc, Rahway, NJ); and stock ownership in Merck & Co, Inc, Rahway, NJ. S.A.P. reports research funding provided to the institution from Janssen, AstraZeneca, Merck, and Genentech for clinical studies in which S.A.P. is a principal investigator; and reports honoraria provided to the institution from Pharmacyclics, Merck, AstraZeneca, Janssen, BeiGene, Genentech, Amgen, MingSight Pharmaceuticals, TG Therapeutics, Novalgen Limited, Kite Pharma, and AbbVie for S.A.P’s participation in consulting activities/advisory board meetings.
Acknowledgments
The authors thank Jennifer Swanson from Flatiron Health for medical writing and editing support, and Olga Tymejczyk, formerly from Flatiron Health, for conception and design.
This study was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co, Inc, Rahway, NJ.
Authorship
Contribution: C.S., J.P., E.D.N., M.Z.H.F., I.S., J.Y., and S.A.P. conceptualized the study; A.S., C.S., J.P., E.D.N., E.B., and X.L. developed the methodology; C.M.P. contributed to formal analysis, investigation, visualization, and project administration; C.M.P. and J.P. wrote the original draft; C.M.P., A.S., C.S., J.P., E.D.N., M.Z.H.F., I.S., E.B., J.Y., X.L., and S.A.P. wrote, reviewed, and edited the manuscript; and C.M.P., C.S., and J.P. supervised the study.
Footnotes
Presented at the 2025 annual meeting of the American Society of Clinical Oncology, Chicago, IL, 30 May to 3 June 2025.
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 data set publicly available. Requests for data sharing by license or by permission for the specific purpose of replicating results in this study can be submitted to PublicationsDataAccess@flatiron.com.
The full-text version of this article contains a data supplement.
Supplementary Material
References
- 1.Siegel RL, Kratzer TB, Giaquinto AN, Sung H, Jemal A. Cancer statistics, 2025. CA Cancer J Clin. 2025;75(1):10–45. doi: 10.3322/caac.21871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Surveillance, Epidemiology, and End Results. Chronic Lymphocytic Leukemia - Cancer Stat Facts. National Cancer Institute; 2025. http://seer.cancer.gov/statfacts/html/clyl.html [Google Scholar]
- 3.Eichhorst B, Robak T, Montserrat E, et al. Chronic lymphocytic leukaemia: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2015;26(suppl 5):v78–v84. doi: 10.1093/annonc/mdv303. [DOI] [PubMed] [Google Scholar]
- 4.Hallek M, Fischer K, Fingerle-Rowson G, et al. Addition of rituximab to fludarabine and cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet. 2010;376(9747):1164–1174. doi: 10.1016/S0140-6736(10)61381-5. [DOI] [PubMed] [Google Scholar]
- 5.Bewarder M, Stilgenbauer S, Thurner L, Kaddu-Mulindwa D. Current treatment options in CLL. Cancers. 2021;13(10):2468. doi: 10.3390/cancers13102468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Woyach JA, Ruppert AS, Guinn D, et al. BTKC481S-mediated resistance to ibrutinib in chronic lymphocytic leukemia. J Clin Oncol. 2017;35(13):1437–1443. doi: 10.1200/JCO.2016.70.2282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Barr PM, Owen C, Robak T, et al. Up to 8-year follow-up from RESONATE-2: first-line ibrutinib treatment for patients with chronic lymphocytic leukemia. Blood Adv. 2022;6(11):3440–3450. doi: 10.1182/bloodadvances.2021006434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Al-Sawaf O, Zhang C, Jin HY, et al. Transcriptomic profiles and 5-year results from the randomized CLL14 study of venetoclax plus obinutuzumab versus chlorambucil plus obinutuzumab in chronic lymphocytic leukemia. Nat Commun. 2023;14(1):2147. doi: 10.1038/s41467-023-37648-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Seymour JF, Kipps TJ, Eichhorst B, et al. Enduring undetectable MRD and updated outcomes in relapsed/refractory CLL after fixed-duration venetoclax-rituximab. Blood. 2022;140(8):839–850. doi: 10.1182/blood.2021015014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mato AR, Hess LM, Chen Y, et al. Outcomes for patients with chronic lymphocytic leukemia (CLL) previously treated with both a covalent BTK and BCL2 inhibitor in the United States: a real-world database study. Clin Lymphoma Myeloma Leuk. 2023;23(1):57–67. doi: 10.1016/j.clml.2022.09.007. [DOI] [PubMed] [Google Scholar]
- 11.Mato AR, Davids MS, Sharman J, et al. Recognizing unmet need in the era of targeted therapy for CLL/SLL: “what’s past is prologue” (Shakespeare) Clin Cancer Res. 2022;28(4):603–608. doi: 10.1158/1078-0432.CCR-21-1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Eyre TA, Hess LM, Sugihara T, et al. Clinical outcomes among patients with chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL) who received treatment with a covalent BTK and BCL2 inhibitor in the United States: a real-world database study. Leuk Lymphoma. 2023;64(5):1005–1016. doi: 10.1080/10428194.2023.2190436. [DOI] [PubMed] [Google Scholar]
- 13.Zuber M, Akkala S, Li N, Veettil SK, Tan CJ, Villa Zapata L. Efficacy and effectiveness outcomes of treatments for double-exposed chronic lymphocytic leukemia and small lymphocytic lymphoma patients: a systematic literature review. Cancer Med. 2024;13(18) doi: 10.1002/cam4.70258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Samples L, Ujjani CS, Khajaviyan S, et al. Clinical outcomes in patients treated with both covalent Btkis and venetoclax and the significance of “double-refractory” status in patients with CLL/SLL. Blood. 2023;142(suppl 1):4650. [Google Scholar]
- 15.Yoon JT, Zhou Y, Mikhaleva M, et al. Characteristics and outcomes of patients with double refractory and double exposed chronic lymphocytic leukemia. Blood Adv. 2025;9(11):2808–2817. doi: 10.1182/bloodadvances.2025016006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ysebaert L, Ferrant E, Dilhuydy MS, et al. Progression on covalent Bruton tyrosine kinase inhibitors predicts poor outcomes in cBRKi-exposed and cBTKi-venetoclax double-exposed populations: a retrospective FILO study in 154 R/R CLL patients. Paper presented at: European Hematology Association; EHA 2024 congress was held in Madrid, Spain from June 13-16, 2024. https://library.ehaweb.org/eha/2024/eha2024-congress/420754/
- 17.Lew TE, Lin VS, Cliff ERS, et al. Outcomes of patients with CLL sequentially resistant to both BCL2 and BTK inhibition. Blood Adv. 2021;5(20):4054–4058. doi: 10.1182/bloodadvances.2021005083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hampel PJ, Rabe KG, Call TG, et al. Combined ibrutinib and venetoclax for treatment of patients with ibrutinib-resistant or double-refractory chronic lymphocytic leukaemia. Br J Haematol. 2022;199(2):239–244. doi: 10.1111/bjh.18357. [DOI] [PubMed] [Google Scholar]
- 19.Din MAU, Knepper T, Chavez J, et al. Treatment strategies and outcomes for double refractory chronic lymphocytic leukemia; a single center experience. Paper presented at: International Workshop on CLL; Boston, Massachusetts from October 6-9, 2023. https://1fb872b80d3df585b41f-41f06c41ae393ec809a826abae176f86.ssl.cf1.rackcdn.com/2501130-1693932752.pdf
- 20.Chatzikonstantinou T, Scarfò L, Karakatsoulis G. Outcomes of patients with chronic lymphocytic leukemia that failed BTK inhibitor and venetoclax-based treatment, and international multicenter study conducted by ERIC. Presented at: European Hematology Association; EHA 2024 congress was held in Madrid, Spain from June 13-16, 2024. 14 May 2024. https://library.ehaweb.org/eha/2024/eha2024-congress/420738
- 21.Flatiron Health Database Characterization Guide. https://flatiron.com/database-characterization
- 22.STROBE STROBE - Strengthening the reporting of observational studies in epidemiology. https://www.strobe-statement.org/
- 23.Hallek M, Cheson BD, Catovsky D, et al. iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood. 2018;131(25):2745–2760. doi: 10.1182/blood-2017-09-806398. [DOI] [PubMed] [Google Scholar]
- 24.Therasse P. Measuring the clinical response. What does it mean? Eur J Cancer. 2002;38(14):1817–1823. doi: 10.1016/s0959-8049(02)00182-x. [DOI] [PubMed] [Google Scholar]
- 25.Feinberg BA, Bharmal M, Klink AJ, Nabhan C, Phatak H. Using response evaluation criteria in solid tumors in real-world evidence cancer research. Future Oncol. 2018;14(27):2841–2848. doi: 10.2217/fon-2018-0317. [DOI] [PubMed] [Google Scholar]
- 26.Vo TT, Espirito JL, Boyd M, et al. Real-world outcomes of systemic therapy in locally advanced, recurrent or metastatic cutaneous squamous cell carcinoma. Future Oncol. 2022;18(17):2087–2099. doi: 10.2217/fon-2021-1229. [DOI] [PubMed] [Google Scholar]
- 27.Validity of Surrogate Endpoints in Oncology: Executive Summary of Rapid Report A10-05, Version 1.1. Institute for Quality and Efficiency in Health Care; 2011. [PubMed] [Google Scholar]
- 28.Lin KH, Huynh L, Yang X, et al. Clinical characteristics, treatment patterns, and outcomes of patients with chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) refractory to covalent Bruton’s tyrosine kinase inhibitor (BTKi) and exposed to B-cell lymphoma 2 inhibitor (BCL2i) Blood. 2023;142(suppl 1):1910. [Google Scholar]
- 29.Mato AR, Roeker LE, Jacobs R, et al. Assessment of the efficacy of therapies following venetoclax discontinuation in CLL reveals BTK inhibition as an effective strategy. Clin Cancer Res. 2020;26(14):3589–3596. doi: 10.1158/1078-0432.CCR-19-3815. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Thompson MC, Roeker LE, Coombs CC, et al. Addressing a new challenge in chronic lymphocytic leukemia: outcomes of therapies after exposure to both a covalent Bruton’s tyrosine kinase inhibitor and venetoclax. Blood. 2021;138(suppl 1):2628. [Google Scholar]
- 31.Ma X, Long L, Moon S, Adamson BJS, Baxi SS. Comparison of population characteristics in real-world clinical oncology Databases in the US: Flatiron Health, SEER, and NPCR. medRxiv. Preprint posted online 16 March 2020 doi: 10.1101/2020.03.16.20037143. [DOI] [Google Scholar]
- 32.Ma X, Bellomo L, Magee K, et al. Characterization of a real-world response variable and comparison with RECIST-based response rates from clinical trials in advanced NSCLC. Adv Ther. 2021;38(4):1843–1859. doi: 10.1007/s12325-021-01659-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Mato AR, Woyach JA, Brown JR, et al. Pirtobrutinib after a covalent BTK inhibitor in chronic lymphocytic leukemia. N Engl J Med. 2023;389(1):33–44. doi: 10.1056/NEJMoa2300696. [DOI] [PubMed] [Google Scholar]
- 34.Siddiqi T, Gauthier J, Kenderian SS, et al. Lisocabtagene maraleucel (liso-cel) in patients (pts) with relapsed or refractory (R/R) chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL): updated follow-up of transcend CLL 004. Blood. 2024;144(suppl 1) 4633-4633. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.