Clonal Hematopoiesis in Patients with Chronic Lymphocytic Leukemia Treated with Fixed-Duration Venetoclax-Obinutuzumab or Chlorambucil-Obinutuzumab: Insights from the Randomized CLL14 Trial

BACKGROUND

Chronic lymphocytic leukemia (CLL) and clonal hematopoiesis (CH) are conditions with high prevalence in elderly people. Previous work has shown that continuous treatment of CLL with BCL2 inhibitors may result in a unique profile of CH in remission (Blombery et al, 2022). However, CH has so far not been studied in the context of fixed-duration targeted therapy.

METHODS

The study cohort was selected based on sample availability from the CLL14 trial, in which patients (pts) with previously untreated CLL were randomized to 1-year fixed-duration Venetoclax-Obinutuzumab (Ven-Obi) or Chlorambucil-Obinutuzumab (Clb-Obi). DNA extracted from peripheral blood mononuclear cell samples from cycle 7 day 1 (C7D1, n=344), 2 years (2-FU, n=212) and 4 years (4-FU, n=132) after end of treatment were subjected to mutational profiling via error-corrected targeted NGS, based on a panel of 45 CH-associated genes and also BAX, BAK1, BCL2, BCL2L1, allowing for variant calling of VAFs ≥0.5%. Variants were correlated with clinical and molecular CLL characteristics; all statistical analyses are descriptive.

RESULTS

A cohort of 344 pts (of 432 CLL14 participants) with a median follow-up of 77.4 months was assessed; median age was 72 (range 41-89) years, 30.5% were female, 50.6% were randomized to Ven-Obi. At least one CH variant was detected in 223 pts (65%) at C7D1, most commonly DNMT3A (38.1%), TET2 (17.4%) and ASXL1 (5.8%); as well as BAX in 13 pts (3.8%). The median number of CH mutations per patient was 2 (max: 8) with 118 pts harboring >1 mutation. Mean VAF was 4.9% (range: 0.5-47%). Tracking the identified mutations from C7D1 back in a subset of pts (n=97), 84% were detected without filtering at C1D1 (160/191). Below detection limit at C1D1 were 7 of 16 BAX variants identified at C7D1.

The only clinical feature differing between pts with and without CH was a higher median age (73 vs 69, p<0.001). Recurrent genetic alterations in CLL, including del(17p), del(11q), del(13q), trisomy 12 or mutations in TP53, NOTCH1 or SF3B1 were not associated with CH, however, complex karyotype (≥3 aberrations) tended to be more common in pts with CH (17.1% vs 9.4%;P = 0.056).

Presence of CH with any variant ≥2% did not impact PFS or OS when adjusted for IGHV status and del(17p) (HR 1.02 [0.77-1.36, p=0.88] and 1.07 [0.68-1.71, p=0.76], respectively). However, in the Clb-Obi arm, large clones with VAFs ≥10% (n=27) were associated with shorter PFS (HR 2.05 [1.22-3.44, p=0.007]) (OS HR 2.12 [0.92-4.86, p=0.076]), but not in the Ven-Obi arm (n=16; PFS HR 1.29 [0.58-2.88, p=0.53] and OS HR 0.92 [0.26-3.29, p=0.9]).

Mutations in BAX were exclusively found in the Ven-Obi arm (13 pts, q=0.007). Serial analysis of these BAX variants from C7D1 showed an increase in clone size until 2-FU (2.2-fold), with no further increase at 4-FU. At 2-FU, an association of U2AF1 mutations with Ven-Obi treatment (7 pts, q=0.08) was observed that was not seen at C7D1. Longitudinal analysis revealed a 15-fold increase in VAF of U2AF1 mutations from C7D1 to 2-FU and no further increase at 4-FU.

Pts with CH at C7D1 did not manifest more side effects, including cytopenia, which occurred in 85 pts (54 with and 31 without CH; OR 1.01 [0.61-1.69], p=0.96) and cardiovascular events, which occurred in 50 pts (36 with and 14 without CH; OR 1.59 [0.82, 3.09], p=0.17). This was also true for major CH variants with VAF ≥10%.

In total, 70 pts experienced at least one second primary malignancy (SPM), of which 45 had and 25 did not have CH (OR 1.06 [0.61-1.83, p=0.85]). Of three secondary hematological malignancies, one blastoid mantle cell lymphoma occurred in a patient with a PTPN11 mutation and myelodysplastic syndromes in two pts with a DNMT3A mutation. Six pts in the cohort experienced Richter Transformation (RT), all of them had CH, two with DNMT3A and one with PPM1D, ZRSR2, TET2 and PTPN11 variants, respectively.

CONCLUSIONS

In this analysis, over 60% of elderly pts with CLL had CH. Major CH variants with a VAF ≥10% were associated with a shorter PFS in the context of chemoimmunotherapy, but not Ven-based therapy. Presence of CH was not associated with cardiovascular events, cytopenia or SPMs, however, all pts with RT had CH. Acquisition of BAX and U2AF1 mutations was associated with Ven-Obi exposure. With fixed-duration therapy, clonal expansion stabilized after end of treatment, consistent with a strong but transient selection effect of Ven treatment on these clones.