Karyotype and outcome in CLL

In this issue of Blood, Fürstenau et al¹ describe the spectrum of conventional karyotypes and their prognostic impact in chronic lymphocytic leukemia (CLL) from analysis of 895 patients recruited to the GAIA/CLL13 study.

The importance of karyotype in CLL was first recognized by Catovsky,² Juliusson,³ and others, in the 1980s, when the ability to generate metaphases for chromosome banding analysis was much less efficient. Fluorescence in situ hybridization (FISH) in CLL, published by Döhner et al⁴ in 2000, described the impact of 4 genomic abnormalities, 13q−, +12, 11q−, and 17p−, and had a major influence lasting >20 years. FISH has been the only chromosome evaluation used in most clinical trials, and it is the only technique recommended in many CLL guidelines, including the International Workshop on CLL (iwCLL).⁵ Although these 4 gene targets are both common and important, it became clear with improvements in standard karyotyping (96.7% success in this article¹ using CpG + interleukin-2) as well as the advent of genomic microarrays (“molecular karyotyping”),⁶ that there is a wide range of chromosomal aberrations that occur commonly in CLL; and for many of these, we know little of their biological or prognostic relevance. Recent interest has focused on “complex karyotype” (CKT) with simple addition of the number of aberrations. Initially, ≥3 aberrations were used as a definition, but more recently, Baliakis et al⁷ showed ≥5 aberrations have much more significance. There has been little focus on the individual aberrations and to what extent each contributed to the prognosis. Many patients with “CKT” have a 17p deletion as one of these aberrations, and it has been difficult to discriminate a “CKT effect” from that of 17p deletion. This study excludes 17p deletion and TP53 mutation and clearly shows an impact of CKT separated from TP53 dysfunction. As with virtually all clinical trials, only patients with progressive CLL who required treatment were recruited.

The GAIA/CLL13 clinical trial compared chemoimmunotherapy (CIT) (2 regimens: fludarabine, cyclophosphamide, and rituximab [FCR] for patients ≤65 years old and bendamustine, and rituximab [BR] for >65 years old) with 3 venetoclax regimens: 2 “doublets” of venetoclax with antibody, rituximab or obinutuzumab (RV or GV, respectively), and the “triplet” obinutuzumab, ibrutinib, and venetoclax (GIV). Therapy duration differed: CIT for 6 months; RV and GV for 12 months; and for GIV, ibrutinib monotherapy was continued to 36 months in those who did not achieve undetectable minimal residual disease (uMRD). In this study, the median observation time was 38.8 months.

Fürstenau et al initially focus on CKT status using a stratification from “non-complex" (nCKT) with ≤2 aberrations to “complex” with ≥3 aberrations. The latter is subdivided into “intermediate CKT” (iCKT) with 3 to 4 aberrations and “high CKT” (hCKT) with ≥5 aberrations (see figure). hCKT proportions were low in all treatment arms, although slightly higher in CIT, with 7.2%, 4.3%, 5.0%, and 2.7% in CIT, RV, GV, and GIV, respectively. CKT did not influence overall response rate, whereas uMRD rates were lower with CKT when using CIT, but not with venetoclax regimens. Progression-free survival (PFS) with CIT was adverse with both iCKT and hCKT, whereas in the pooled venetoclax arms, only hCKT had worse outcome. For both CIT and venetoclax, this only occurred for those with unmutated immunoglobulin heavy chain variable region (IGHV) genes (unmutated CLL [U-CLL]); in mutated IGHV (mutated CLL [M-CLL]), CKT had no statistical effect.

A small group of ∼2.3% formally with CKT have an outstandingly good outcome: patients with “triple trisomy” (often +12, +18, and +19, or more precisely +12 with +18 and/or +19; and all M-CLL). This "CKT" group was identified some years ago to have a highly favorable prognosis with FCR,⁷ essentially 100% PFS, although this seems not widely known. It is pleasing to see the same excellent outcome with venetoclax, also 100% durable PFS. Despite the small proportion, this group deserves wider recognition to avoid more intensive and inappropriate therapies.

Translocations were present in ∼24% of patients and imparted an inferior PFS in all treatment arms. The most common were t(14;18), t(14;19), t(1;13), t(2;14), and t(18;22). In venetoclax arms, shorter PFS was mainly driven by unbalanced translocations and when translocations involved chromosomes 8q and 18q. Deletions, overall, had shorter PFS but varied substantially by individual chromosome, with 13q− favorable, whereas deletions of chromosomes 4, 15, and 18 were highly unfavorable, and often occurred with other aberrations and CKT. There are many similarities and some differences between CIT and venetoclax regimens. In both, unmutated IGHV, CKT (vs none), and β2M confer worse outcome, whereas 13q− is favorable. Other factors have interesting differences. Age has no impact on venetoclax regimens, whereas with CIT, it is more adverse at age >65 years as the CIT regimen changes from FCR to BR. The 11q− is well recognized to be adverse with CIT but had no significance with venetoclax. Many aberrations had numbers too small in the CIT arm for statistical analysis. This divergence of outcome with IGHV status with both CIT and venetoclax is consistent with the CLL14 study,⁸ whereas with ibrutinib, M-CLL and U-CLL have identical outcomes.⁹ It will be interesting to see the longer-term follow-up from the GIV arm.

CLL has a risk of clonal evolution with the development of additional and adverse karyotypes potentially producing more refractory disease that may be therapy driven.¹⁰ Thus far, only small numbers have progressed and fewer still have required next line therapy. Nevertheless, more patients with CIT acquired a higher CKT level and 17p−. This study excluded 17p− and TP53-mutated patients, and so in the real-world environment, this situation will be more complex.

This article nicely demonstrates the importance of karyotype in CLL, both for CIT and now also with venetoclax-based regimens. Readers of Blood with an interest in CLL will be rewarded by close study of the main article and its supplementary data. However, just under 900 patients is not enough to evaluate and understand less frequent karyotypes. Karyotypic evaluation must be incorporated into CLL clinical trials, especially those with larger numbers. Furthermore, karyotype can clearly help refine the prognosis and expectations for individual patients and hence should be included in the standard diagnostic workup for all patients with CLL.

Conflict-of-interest disclosure: The author declares no competing financial interests.