High karyotypic complexity is an independent prognostic factor in patients with CLL treated with venetoclax combinations

Chronic lymphocytic leukemia (CLL) is a malignancy of mature B cells and displays high clinical and genetic heterogeneity. Survival of patients with CLL has greatly improved over the last decade with the approval of targeted therapies, including inhibitors of Bruton tyrosine kinase and the B-cell lymphoma-2 (BCL-2) inhibitor, venetoclax.^1-14 With the increased use of these efficacious therapies, it is important to identify biomarkers that reliably predict the depth and duration of responses to guide clinical practice. Biomarkers such as TP53 aberrations, deletion of 11q [del(11q)], and immunoglobulin variable heavy chain (IGHV) mutational status that have prognostic significance in patients with CLL treated with chemoimmunotherapy (CIT) may have diminished significance in patients treated with novel therapies.^7,15-17 Hence, evaluation of current prognostic markers and discovery of novel prognostic markers for treatment with targeted therapies are of major importance.

Apart from the aforementioned genetic markers, the presence of >3 (complex karyotype; CKT) or >5 (high-CKT; hCKT) chromosomal aberrations (CAs) has been shown to be associated with poor prognosis in patients with CLL treated with CIT.^18-28 However, CKT has not yet been systematically incorporated into clinical decision-making, given that its definition and methodological aspects are still under debate. In addition to the numerical cut-off for CKT, uncertainty exists on the relevance of the presence of major structural chromosome abnormalities, such as translocations.^{18,26,27,29,30} Moreover, it is not yet established whether CKT affects patient prognosis independently of TP53 aberrations, as the presence of CKT and TP53 aberrations are strongly correlated, resulting in a heavily intertwined impact of those factors on patient outcome.^{19,20,22,23,25,31-33} 

For targeted therapies, studies evaluating the prognostic impact of CKT on patient outcome are more heterogeneous, not only in study size and methods but also in observed impact.^21,32,34-39 According to most reports evaluating patients with relapsed/refractory CLL, the efficacy of ibrutinib is influenced by the presence of CKT.^7,37,39 In contrast, none of the studies evaluating ibrutinib in treatment-naive patients found an impact of CKT. For venetoclax treatment, data regarding the impact of CKT on patient outcome is more limited. The presence of CKT was associated with inferior outcomes in 2 reports evaluating patients with relapsed/refractory CLL treated with venetoclax plus rituximab (RV), whereas no impact of CKT was demonstrated for treatment outcomes on ibrutinib + venetoclax.^35,36 In patients with untreated CLL, the efficacy of venetoclax + obinutuzumab (GV) treatment was not significantly affected by the presence of CKT.³⁴ These studies used different methods and definitions and did not uniformly examine the influence of TP53 aberrations and the impact of translocations. Hence, there is a need for more studies to fully establish the prognostic impact of karyotype complexity, its definition, and its impact independently of TP53 aberrations.

Here, we prospectively evaluate the impact of karyotypic complexity and translocations in patients with CLL without TP53 aberrations treated with venetoclax-based time-limited combinations within the international GAIA/CLL13 trial. Lastly, karyotype evolution at CLL progression was analyzed and compared between treatment modalities.

The phase 3 GAIA/CLL13 study randomly assigned treatment-naive patients with TP53 wild-type CLL to groups of standard CITs (6 cycles of fludarabine, cyclophosphamide, rituximab [FCR, ≤65 years] or bendamustine, rituximab [BR, >65 years]) and 12 cycles of RV, GV, or venetoclax-obinutuzumab-ibrutinib (GIV and ibrutinib until cycle 36 in case of failure to achieve undetectable minimal residual disease [uMRD]).⁴⁰ Additional information on study design can be found in the supplemental Data, available on the Blood website.

All patients provided written informed consent before enrollment. The study protocol was approved by the health authorities and institutional review boards of each participating site. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice Guidelines. The GAIA/CLL13 trial was registered (NCT02950051).

For the assessment of CKTs, peripheral blood samples at baseline and clinical progression of CLL were shipped to the Laboratory for Molecular Hematology, University Hospital of Cologne, Germany. Peripheral whole blood was cultured for 72 hours in the presence of 1 μM CpG-oligodeoxynucleotides (TIB MOLBIOL, Berlin, Germany) and 100 U/mL interleukin-2 (Biochrom, Berlin, Germany). Colcemid was added at 200 ng/mL for 1.5 hours for metaphase arrest. After hypotonic treatment, cell suspensions were fixed with 3:1 methanol and acetic acid and dropped for metaphase spreading on glass slides. After trypsinization and Giemsa staining, karyotypes were analyzed and described per the International System for Human Cytogenetic Nomenclature 2020. Noncomplex karyotype (nCKT) was defined as ≤2 CAs and CKT as ≥3 CAs. CKT was further subdivided into intermediate CKT (iCKT) with 3 or 4 CAs and hCKT with ≥5 CAs. If patients acquired at least 1 chromosomal aberration at the time point of clinical progression that was not detected in the baseline sample, the karyotype at progression was considered evolved.

Statistical methods are described in the supplemental Data.

Chromosome analysis was successfully performed at baseline in 895 of 926 patients (96.7%). Of these, 672 (RV, 231; GV, 218; and GIV, 223) were in the pooled venetoclax population and 223 in the CIT arm. CKT was present in 46 (20.6%), 44 (19.0%), 36 (16.5%), and 27 (12.1%) (iCKT was present in 30 [13.5%], 34 [14.7%], 25 [11.5%], and 21 [9.4%]; hCKT in: 16 [7.2%], 10 [4.3%], 11 [5.0%], and 6 [2.7%]) patients treated with CIT, RV, GV, and GIV, respectively). Male sex (P = .033), del(11q) (P < .001), del(6q) (P < .001), trisomy 12 (P < .001), and higher lactate dehydrogenase levels (P = .018) were associated with the presence of CKT (Table 1). All other patient and CLL characteristics were not significantly associated with the presence of CKT, in particular, no correlation was observed between IGHV status (P = .065) or CLL–international prognostic index (P = .16) and CKT.

There was no significant difference in the overall response rate at month 15 between patients with and without CKT, in both the CIT arm (76.1% vs 83.1%, P = .28) and the pooled venetoclax collective (89.7% vs 94.7%; P = .053; Figure 1A). uMRD rates in the peripheral blood at month 15 were lower in patients with CKTs compared with those with nCKTs (37.0% vs 57.1%; OR 0.44; 95% confidence interval [CI], 0.23-0.86; P = .016) in the CIT arm (Figure 1B). In the pooled venetoclax collective (73.8% vs 78.6%; P = .28) and in the separate venetoclax arms (RV, 52.3% vs 56.7% [P = .60]; GV, 86.1% vs 86.3% [P = .98]; and GIV, 92.6% vs 92.3% [P = .96]; data not shown), uMRD rates of patients with and without CKT were not significantly different. Patients with hCKT had a significantly lower uMRD rate at MO15 in the CIT arm (vs nCKT; 25.0% vs 57.1%; P = .021), whereas uMRD rates were not significantly different between patients with hCKT and CKT in the pooled venetoclax arms (70.4% vs 78.6%; P = .316).

The median observation time at the data cut used for this analysis was 38.8 months. In the CIT arm, median progression-free survival (PFS) was 38.4 months in the CKT group (3-year PFS rate, 51.9%) and not reached in the nCKT group (3-year PFS rate, 81.3%) (hazard ratio [HR], 2.58; 95% CI, 1.54-4.32; P < .001; Figure 2A). The overall survival (OS) was also significantly shorter in patients with CKT vs in nCKT (median, not reached for both subgroups; 3-year OS rate, 85.3% vs 97.2%; HR, 3.25; 95% CI, 1.03-10.26; P = .044; Figure 2C).

In the pooled venetoclax collective, patients with CKT also had shorter PFS than patients with nCKT (HR, 1.71; 95% CI, 1.11-2.62; P = .015; Figure 2B), but the OS was not different (Figure 2D). None of the individual venetoclax arms showed significant differences in PFS or OS between CKT and nCKT (supplemental Figure 1). To further investigate the PFS differences between CKT groups, we subdivided the CKT group into patients with iCKT and hCKT. In the CIT arm, those with hCKT and iCKT showed similar PFS (median, 31.8 and 38.4 months; HR, 1.10; 95% CI, 0.46-2.63; P = .828), which was significantly shorter than that in the nCKT group (not reached, hCKT vs nCKT: HR, 2.75; 95% CI, 1.29-5.85; P = .009; iCKT vs nCKT: HR, 2.49; 95% CI, 1.36-4.57; P = .003) (Figure 3A). In contrast, the presence of iCKT did not associate with inferior PFS in the venetoclax arms (iCKT vs nCKT: HR, 1.19; 95% CI, 0.69-2.06; P = .53), whereas hCKT did (hCKT vs nCKT: HR, 3.72; 95% CI, 2.03-6.82; P < .001; Figure 3B). This significant adverse impact of ≥5 vs <5 CA was observed across all individual study arms (CIT: HR, 2.34 [95% CI, 1.11-4.92; P = .025]; RV: HR, 2.78 [95% CI, 1.10-6.98; P = .030]; GV: HR, 3.33 [95% CI, 1.16-9.56; P = .025]; and GIV: HR, 7.90 [95% CI, 2.31-27.03; P = .001]; Figure 3C). In the subgroup of patients with unmutated IGHV, hCKT (vs nCKT: HR, 3.30; 95% CI, 1.78-6.14; P < .001), but not iCKT (vs nCKT: HR, 1.19; 95% CI, 0.61-2.32; P = .61), led to inferior PFS in the pooled venetoclax arms, whereas in the CIT arm, both hCKT (vs nCKT: HR, 2.28; 95% CI, 1.05-4.96; P = .038) and iCKT (vs nCKT: HR, 2.43; 95% CI, 1.24-4.75; P = .01) were associated with shorter PFS (supplemental Figure 2). In patients with mutated IGHV, neither iCKT nor hCKT negatively affected outcomes, both in the pooled venetoclax collective and the CIT arm (supplemental Figure 2). Patients with CKT that included trisomy 12 and additionally trisomy 18 and/or 19 appeared to have longer PFS than patients with other CKT and patients with nCKT (Figure 3D). However, absolute patient numbers in both subgroups (mutated IGHV: n = 53; trisomies 12, 18, or 19: n = 18) were small, so the validity of these results might be limited.

At baseline, translocations were observed in 52 patients (23.4%) in the CIT arm and in 160 patients (24.1%) in the venetoclax arms. Patients harboring translocations had shorter PFS than patients without translocations in all treatment arms (Figure 4A). Focusing on the venetoclax-treated population, this difference was mostly driven by unbalanced translocations (PFS of unbalanced translocations vs no translocations: HR, 3.83; 95% CI, 2.30-6.39; P < .001); however, balanced translocations still associated with inferior PFS compared with patients without any translocation (HR, 1.66; 95% CI, 1.04-2.65; P = .035; Figure 4B). In the presence of CKT, translocations still influenced PFS adversely: although patients with CKT but without any translocations did not have different PFS compared with patients with nCKT (HR, 0.53; 95% CI, 0.20-1.45; P = .217), patients with CKT plus translocations had significantly shorter PFS (vs nCKT: HR, 2.79; 95% CI, 1.76-4.42; P < .001; Figure 4C).

To further investigate possible biological mechanisms behind the adverse impact conferred by translocations, involved chromosomes were analyzed in more detail. The most frequently involved chromosomes were 13q (n = 71), 14q (n = 64), 18q (n = 41), and 8q (n = 31), and the most frequently observed translocations were t(14;18), t(14;19), t(1;13), t(2;14), and t(18;22) (supplemental Table 1). In the pooled venetoclax arms, PFS was not different in patients with or without translocations involving chromosomes 13 (HR, 0.96; 95% CI, 0.45-2.06; P = .91) or 14 (HR, 1.34; 95% CI, 0.65-2.76; P = .42; supplemental Figure 4), whereas translocations involving chromosomes 18 (HR, 2.48; 95% CI, 1.30-4.76; P = .006) or 8 (HR, 3.94; 95% CI, 2.11-7.34; P < .001) conferred shorter PFS (Figure 4D-E). In the CIT arm, patient numbers were too small for reliable correlative analyses.

The presence of chromosome deletions was associated with shorter PFS in both the CIT arm (HR, 2.33; 95% CI, 1.39-3.90; P = .001) and the pooled venetoclax collective (HR, 1.67; 95% CI, 1.15-2.42; P = .007; supplemental Figure 5). The impact of individual deletions varied strongly: the most frequent aberration, deletion 13q, was associated with longer PFS in the pooled venetoclax arms (HR, 0.67; 95% CI, 0.47-0.95; P = .026), whereas it did not influence PFS in the CIT arm (P = .538; supplemental Figure 5). Deletion 14q32 had no impact on PFS, and deletion 11q was only associated with shorter PFS in the CIT arm (HR, 2.68; 95% CI, 1.62-4.44; P < .001; supplemental Figure 5). Less prevalent deletions of chromosomes 4 (n = 12), 15 (n = 12), and 18 (n = 14) were associated with shorter PFS in the pooled venetoclax arms; however, these and multiple other deletions almost exclusively occurred in patients with CKT (supplemental Figure 5).

Separate multivariable analyses were performed for the CIT arm and the pooled venetoclax collective, including all relevant patient and CLL characteristics. In the CIT arm, age >65 years, unmutated IGHV and the presence of CKT were independently associated with shorter PFS (Table 2). In the CIT arm, patients’ age determined the treatment they received (FCR for patients ≤65 years, BR for patients >65 years), so it is not clear whether the observed impact on PFS can be attributed to age or the allocated treatment. In the venetoclax-treated collective, translocations and hCKTs were confirmed as independent prognostic factors for inferior PFS, along with β₂ microglobulin >3.5 mg/L and unmutated IGHV, whereas CKT was not an independent prognostic factor (Table 2).

At the time point of clinical progression of CLL, karyotyping data were available for 27 of 60 patients (45.0%) with disease progression in the CIT arm and 59 of 98 patients (60.2%) in the pooled venetoclax arms. Of these, 11 (40.7%) and 12 patients (20.3%), respectively, had disease progression that required a new treatment. Karyotype evolution was present in 48 of 86 patients (55.8%) (CIT, 17 of 27 [63.0%] and venetoclax, 31 of 59 [52.5%]) at the time of progression. Baseline karyotypes were associated with different frequencies of karyotype evolution. Patients with normal karyotypes at baseline showed the lowest incidence of karyotype evolution (20.0%; 4 of 20 patients). The presence of trisomy 12 was also associated with a comparatively low incidence of evolution (42.9%; 6 of 14 patients), whereas in patients with del13q (75%; 15 of 20 patients), del11q (76.2%; 16 of 21 patients), or del6q (100%; 8 of 8 patients), most patients showed karyotype evolution at progression (supplemental Table 2).

In the CIT arm, 7 of 20 patients (35.0%) who were categorized as having nCKT at baseline acquired CKT at progression, whereas only 6 of 45 (13.3%) patients with nCKT treated with venetoclax developed CKT at progression. The mean number of CAs increased from baseline to progression (2.0-3.4 CAs) in patients treated with CIT (Figure 5). In the venetoclax arms, the mean number of CAs remained stable at progression (2.0 CAs at both time points). The frequency of acquired 17p deletions was also higher in the CIT arm (14.8%; 4 of 27 patients) compared with the venetoclax arms (0 of 59 patients). Ideograms illustrating the acquisition of aberrations between baseline and clinical progression in the CIT arm and the pooled venetoclax population are depicted in supplemental Figure 6.

In this prospective study on the value of karyotype complexity in CLL, we identified the presence of hCKT and/or translocations as independent prognostic factors for PFS in patients treated with time-limited venetoclax-based combinations. Furthermore, we confirmed the adverse impact of CKT in the context of CIT. Dismal outcomes in patients harboring complex or highly complex karyotypes have also been observed in several other studies, both in patients with CLL treated with CIT and targeted therapy.^{7,8,19,21-24,32,34-39,41-44} However, in many studies, the impact of CKT was hard to distinguish from the impact of TP53 aberrations. The presence of CKT undoubtedly correlates with TP53 aberrations, and especially in patients treated with targeted therapies, the impact of CKT on outcome is frequently influenced by TP53 aberrations.^20,25,33,36 Given the exclusion of patients with TP53 aberrations from this study, it can be concluded that in the absence of TP53 aberrations, CKT/hCKT itself confers an inferior prognosis.

In CLL, CKT is currently used as a dichotomous variable (CKT vs nCKT) with a cut-off at 3 unrelated cytogenetic abnormalities, despite several reports addressing adverse outcomes with increasing karyotypic complexity beyond the threshold of 3 aberrations.^18,19,36,37 Our study supports an expansion of this definition, at least in the context of targeted therapy, as in patients treated with time-limited venetoclax combination regimens, the presence of hCKT and not CKT was identified as an independent adverse prognostic factor. Although patients with CKT only had a moderately increased risk of progression compared with nCKT, hCKT led to a more than threefold increase in the risk of progression or death compared with nCKT. This negative impact was observed in all treatment arms, so neither treatment with venetoclax nor the addition of ibrutinib to GV was able to abrogate the adverse prognostic impact of hCKT.

Our data on hCKT is partly supported by an analysis of the MURANO trial, which evaluated treatment with RV in relapsed/refractory CLL. Here, an inferior PFS was observed in patients harboring a CKT, with particularly impaired outcomes in patients with hCKT.³⁶ In the VISION/HO141 trial of ibrutinib + venetoclax for relapsed/refractory CLL, no impact of CKT or hCKT on outcome could be demonstrated.³⁵ The CLL14 study that evaluated a time-limited treatment with GV in previously untreated CLL did not observe any significant impact of CKT or hCKT on PFS.³⁴ However, the impact of CKT was analyzed at a rather early time point with limited PFS events and in a smaller patient population compared with the GAIA/CLL13 trial.

In the CIT arm, we did not find an increased impact on PFS by increasing complexity above the threshold of 3, whereas other studies evaluating patients treated with CIT did observe differences in the impact of hCKT compared with nCKT.^19,42 This difference might be caused by the distinct composition of the CKT patient group compared with other studies because of the exclusion of patients harboring TP53 aberrations in our cohort. Baliakas et al showed that, especially for these patients, each additional aberration has an impact on outcome.¹⁹ This distinct composition of the CKT group is also reflected by the lack of enrichment for unmutated IGHV genes in CKT patients in our study, which is observed in other studies.^{19,20,23,24,29,37,42,45} 

Despite being the first study, to our knowledge, to systematically demonstrate inferior outcomes for patients with CKT/hCKT in the context of first-line venetoclax-based treatment, our analysis still provides a strong rationale for the use of these regimens as opposed to CIT in patients with CKT. Firstly, the increase in the risk of disease progression associated with the presence of CKT was much smaller in the venetoclax-treated collective (HR 1.71) than with CIT (HR 2.58). Furthermore, in the CIT arm, an impaired OS in patients with CKT was already observed at this rather early time point, whereas no OS differences were seen in the venetoclax arms. Moreover, an additional advantage of the use of venetoclax-based therapies over CIT is the prevention of the development of a more complex karyotype and thus a harder-to-treat disease upon clinical progression. Karyotype complexity and incidence of del(17p) remained stable after venetoclax treatment, whereas they increased after CIT.

Our data indicate that for defining CKT in a proper manner, not only the number of cytogenetic abnormalities is important but also other factors such as the presence of (certain) translocations, coexisting trisomies, and the IGHV status. The presence of translocations has been suggested to be a prognostic marker for CLL in multiple studies; however, most of these analyses did not confirm translocations as an independent prognostic factor.^{20,27,29,30,37,45} This discrepancy might also be explained by the lack of TP53 aberrations in our patient group; in other studies, a substantial part of the patients with CKT had TP53 aberrations, inducing an inferior outcome irrespective of translocations.

The significance of specific translocations in CLL is largely unexplored. In this analysis, we found an inferior PFS upon venetoclax treatment when a translocation involving 8q (with 8q24 harboring c-MYC) or 18q (with 18q21 harboring BCL2) was present. The t(14;18)(q32;q21) associated with overexpression of BCL-2 in follicular lymphoma did, however, not lead to shorter PFS in this study.^46,47 Unfortunately, for CIT, the number of patients harboring these abnormalities was too low to draw conclusions.

The presence of chromosome deletions was associated with shorter PFS on univariate comparison; this analysis does not, however, support the use of this variable as a simple prognostic marker (any deletion, yes vs no) in the context of venetoclax-based combinations. The impact of individual deletions ranged from very favorable with deletion 13q to highly adverse with deletions of chromosomes 4, 15, and 18 that mostly co-occurred with multiple other deletions and hCKT, making it difficult to assess the impact of the individual deletions.

Subgroup analyses within the group of patients with CKT also included selection for IGHV status and trisomy 12 combined with trisomy 18 and/or 19. In patients with mutated IGHV, the presence of a CKT, both iCKT and hCKT, did not appear to affect PFS in our study, neither in patients treated with venetoclax or CIT. This partly confirms the findings of Baliakas et al, who also observed a reduced impact of iCKT on outcome in patients with mutated IGHV; however, patients with hCKTs still showed a strongly inferior outcome.¹⁹ Ramos-Campoy et al however observed the opposite effect.⁴⁵ The coexisting presence of trisomies of chromosomes 12 and 19 has previously been uniformly described as a very favorable characteristic in patients with CLL with CKT.^19,48,49 Our study suggests similar results, with no PFS events recorded in this specific subgroup of patients with a CKT including trisomy 12 and an additional trisomy 18 and/or 19 irrespective of the treatment arm. Given the previous findings that these favorable complex karyotypes almost exclusively occur in patients with mutated IGHV,^48,49 we speculate that in this data set without TP53 aberrations, the positive impact of CKT with multiple trisomies largely abrogates the adverse impact of CKT in patients with mutated IGHV. Absolute numbers are, however, too small (CKT with trisomy 12 + trisomy 18 and/or 19, n = 18 and CKT with mutated IGHV, n = 53) to draw reliable conclusions.

In general, this karyotype analysis is limited by its rather small proportion of patients with hCKT and unbalanced translocations; therefore, we caution against comparing HRs between individual treatment arms in these subgroups. Because of a prespecified test hierarchy in the study protocol, we are not allowed to formally compare the PFS between the individual venetoclax arms at this time point. We decided to pool the venetoclax arms for certain analyses. Although we made sure to only pool the venetoclax arms for analyses in which the individual treatment arms showed the same trend/direction, the pooled data set might potentially be confounded by trends shown in a single arm; therefore, we decided against presenting absolute measures of PFS (eg, 3-year PFS) in the pooled collective. It is also important to acknowledge that this analysis does not include patients with TP53 aberrations, and its findings might, therefore, not apply to patients with TP53 mutations and/or 17p deletions.

The findings of this study support the incorporation of karyotyping into the standard diagnostic workup of CLL, whereas guidelines of the International Workshop on CLL (iwCLL), the European Society for Medical Oncology (ESMO), and the European Research Initiative on CLL (ERIC) currently recommend only performing fluorescence in situ hybridization analysis with additional TP53 mutational analysis in routine diagnostics and do not yet recommend karyotyping to guide treatment decisions.^2,50,51 Compared with genomic microarrays, karyotyping has certain advantages: karyotyping enables the detection of balanced rearrangements and multiple clones, and thus the study of the impact of (specific) translocations on CLL prognosis. In contrast, the use of microarrays enables the detection of abnormalities with high resolution and without the need for dividing cells.^45,50 In a study in which karyotyping and microarrays were directly compared, no significant differences were found in the percentage of patients assigned to each CKT group or the impact CKT had on prognosis, although differences were observed when focusing on individual cases.⁴⁵ 

In conclusion, in patients with CLL lacking TP53 aberrations as detected by fluorescence in situ hybridization and Sanger sequencing, hCKT, but not CKT, was an independent prognostic factor for shorter PFS after treatment with time-limited venetoclax combinations. In patients treated with CIT, we confirmed CKT as an independent adverse prognostic factor. Moreover, the presence of translocations was associated with poor PFS in all treatment arms. Upon clinical progression of CLL, karyotyping revealed increased genomic complexity after CIT treatment as opposed to venetoclax-based treatments. These data support the application of (h)CKT and the presence of translocations as biomarkers in clinical decision-making. Further research is needed to determine if karyotype complexity and the presence of translocations are prognostic factors in other patient groups than treatment-naive patients with TP53-intact CLL.

The authors thank all patients and their families, physicians, and study teams at the sites for their participation and great contributions to this study. They especially thank Emily Holmes, Kerstin Löschke, Alana Hönig, Johanna Wesselmann, Lisa Engel, Verena Bonigut, Marina Stockem, Dana Engelhardt, Katharina Löhers, and Anne Westermann, who worked as project managers; Florian Drey, Ronald D'Brot, Dilara Celik, Jörn Harrandt, Viktoria Monar, Uyen Nguyen, Anna Schloßnagel, Johannes Pöppinghaus, Michael Verhülsdonk, Annette Niederhausen, Irene Preißler-Stodden, Olga Korf, Jan-Erik Mittler, Henrik Gerwin, Alwina Flock, Sally Huggins, Anna Wetzel, and Giovanna Gelmi, who worked as data managers; and Sabine Frohs, Tanja Annolleck, Berit Falkowski and Christina Rhein, who worked as safety managers in the GAIA trial.

The trial was sponsored by the German CLL Study Group with financial support and study drug provision from Roche, AbbVie, and Janssen. This analysis was partly supported by a Dutch Cancer Foundation grant (Coherent: 13650/2021-Infra).

Contribution: M.F., Y.J.T., B.E., K.-A.K., and A.P.K. designed the study; M.F., J.v.T., F.S., N.D.S., A.-M.F., K.F., and B.E. were responsible for the trial management. M.F., M.G., P.T., P.B.S., T.T., M.-D.L., C.d.C.-B., C.S., C.B.P., T.I., B.S., A.J., I.C., T.N., M. Baumann, H.H., T.G., J.C.R., E.C.D., V.L., G.J., A.W., J.G., N.G., F.S., C.-M.W., M.R., E.T., S.S., C.U.N., M.H., B.E., and A.P.K. enrolled patients, contributed data, and interpreted data; M.P. and K.-A.K. performed the karyotyping and M.F., Y.J.T., C.H.M.M., A.-M.v.d.K.-K., J.D., M.P., K.-A.K., and A.P.K. analyzed the raw data; M.F., Y.J.T., S.R., B.E., and A.P.K. designed and performed statistical analyses; M.S., S.S., C.U.N., M.H., B.E., K.-A.K., and A.P.K. supervised the project; M.F., Y.J.T., K.-A.K., and A.P.K. drafted the first version of the manuscript; and all authors revised and approved the final version of the manuscript.

Conflict-of-interest disclosure: J.v.T. was part of the speakers’ bureaus, has received honoraria for lectures/presentations, and has participated at advisory boards and received travel grants from AbbVie, AstraZeneca, Janssen, and Roche. M.G. was supported for attending meetings by AbbVie, AstraZeneca, Janssen, and Roche. C.S. has received honoraria for presentations by AbbVie and AstraZeneca and was supported for attending meetings from BeiGene. A.J. has received speaker fees from AbbVie, Amgen, AstraZeneca, BeiGene, Janssen, Incyte, Novartis, Sobi, Sanofi-Genzyme, and Takeda and participated at consultancy-ad boards for AbbVie, Amgen, AstraZeneca, BeiGene, Janssen, Incyte, Novartis, Sobi, Sanofi-Genzyme, Roche, and Takeda. In addition, she received travel grants from Amgen, AbbVie, and Celgene. T.N. has received honoraria for lectures/presentations and has participated at advisory boards from AbbVie, Roche, AstraZeneca, Gilead, BeiGene, and Janssen. M. Baumann has participated on a data safety monitoring/advisory board by AbbVie, AstraZeneca, Janssen, Roche, and Pfizer and was supported for attending meetings by AbbVie, Celgene/BMS, and Janssen. H.H. was supported for attending meetings from AbbVie, Roche and Janssen. J.G. has participated at advisory boards from AbbVie, Janssen, and Roche. F.S. was supported for travel attending meetings from Gilead. A.-M.F. was supported for attending meetings/travel from AbbVie and received payments for consulting and presentations from AstraZeneca. K.F. was part of the speakers’ bureaus, has received payments/honoraria for lectures/presentations from AbbVie and Roche, and has received consulting fees from AbbVie, AstraZeneca, and Roche. She was supported for attending meetings/travel by Roche and has participated on a Data Safety Monitoring/Advisory Board from AstraZeneca. C.-M.W. received research and travel grants from Hoffman-La Roche, Janssen-Cilag, and AbbVie and participated at advisory boards from Hoffman-La Roche, Janssen-Cilag, and AbbVie. M.R. has received payments/honoraria for lectures/presentations from Janssen, AstraZeneca and Roche and consulting fees from AbbVie, AstraZeneca, Janssen and Roche. He was supported for attending meetings/travel by AstraZeneca and has received grants/contracts from AbbVie, Hoffman-La Roche. M. Brüggemann received grants and personal fees from Amgen (advisory board, speakers bureau, travel support), and personal fees from Becton Dickinson, Janssen, Pfizer (speakers bureau), and Jazz (travel support), all outside the submitted work. E.T. reports travel support from Abbvie, Janssen, and BeiGene and honoraria for presentations and advisory board participation for AbbVie, AstraZeneca, BeiGene, Janssen, and Roche. S.S. was part advisory boards/received honoraria and research grants by AbbVie, Acerta, Amgen, AstraZeneca, BeiGene, BMS, Gilead, Hoffman-La Roche, Janssen, Novartis, and Sunesis. C.U.N. has received consulting fees from AbbVie, AstraZeneca, BeiGene, Eli Lilly, CSL Behring, Octapharma, Takeda, Genmab, and Janssen, honoraria for lectures/presentations from AbbVie and travel support from AstraZeneca. M.H. received honoraria, consulting fees and grant from Abbvie, Roche, Janssen, AstraZeneca, Gilead, and BeiGene. B.E. was part of the speakers’ bureaus and participated at advisory boards from AbbVie, Janssen, and Roche. A.P.K. received honoraria for consultancy and research funding from AstraZeneca, Janssen, Roche/Genentech, AbbVie, BMS, and LAVA. The remaining authors declare no competing financial interests.

Correspondence: Moritz Fürstenau, Department I of Internal Medicine, Faculty of Medicine and University Hospital Cologne, Center for Integrated Oncology Aachen Bonn Cologne Duesseldorf, German CLL Study Group, University of Cologne, Gleueler Str 176-178, 50935 Köln, Germany; e-mail: moritz.fuerstenau@uk-koeln.de; and Arnon P. Kater, Department of Hematology, Cancer Center Amsterdam, Amsterdam University Medical Centers, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands; e-mail: a.p.kater@amsterdamumc.nl.