MRD growth decelerated with venetoclax combined with either rituximab or obinutuzumab in patients with R/R CLL

The CLL14 trial (NCT02242942) has previously demonstrated that the BCL2 inhibitor venetoclax (Ven) plus CD20 antibody obinutuzumab (Obi) (Ven-Obi), compared with chlorambucil-obinutuzumab (Clb-Obi), can induce deep remissions in patients with chronic lymphocytic leukemia (CLL) with coexisting conditions. Similarly, the MURANO phase III trial (NCT02005471) has also demonstrated improved and sustained progression-free survival (PFS) and overall survival (OS) with Ven + rituximab (VenR) compared with bendamustine + rituximab (BR) in patients with CLL. Both CLL14 and MURANO trials have been previously reported by the Lymphoma Hub.

The level of minimal residual disease (MRD) at the end of treatment (EoT) is a prognostic indicator for PFS in patients with CLL; however, progression of CLL is highly dynamic process, and EoT-MRD only provides a snapshot. Capturing and measuring MRD dynamics by regular and more sensitive methods may yield valuable information on disease-specific kinetics. During the 63rd American Society of Hematology (ASH) Annual Meeting and Exposition, results from the MRD kinetics model of the MURANO trial were presented by Arnon Kater.¹ In addition, Al-Sawaf et al.² recently published the extended follow-up analyses from the CLL14 trial in the Journal of Clinical Oncology. Key findings from the MRD analyses of the MURANO and CLL14 trials are summarized below.

Methods

The study design of both the MURANO and CLL14 trials have been previously reported. The methods described here only relate to the longitudinal analyses of MRD growth in both studies. Both studies employed a population-based logistic growth model with a non-linear mixed effects approach. The model was developed to assess the inter-individuality variability around the growth parameters, and to identify covariates with a significant impact on the appearance of detectable MRD.

The MRD model of the open-labelled, phase III randomized MURANO trial included patients who had completed treatment without prior progressive disease (PD) and had ≥ 2 measurable MRD samples post EoT by next generation sequencing (NGS).

The outcome of interest was median MRD in peripheral blood (PB) at cycle 4, 2-3 months after end of combination therapy (EoCT) and every 3-6 months thereafter measured by allele-specific oligonucleotide polymerase chain reaction (ASO-PCR) and/or flow cytometry.

• The undetected MRD (uMRD) threshold was < 10^-4.
• Prognostic markers and patient demographics were screened as covariates for impact on key model parameters.

The extended analyses of the open-labelled, phase III, randomized CLL14 trial included patients who had completed and been off treatment for a minimum of 3 years and had at least two measurable MRD samples by NGS on and after follow-up Month 3.

The outcome of interest was uMRD rates with the cutoff of 10^-4, 10^-5, and 10^-6 by NGS, and analyses using allele-specific oligonucleotide PCR.

• MRD in bone marrow (BM) was assessed in patients with treatment response at Cycle 9 and 2 months after EoT.
• MRD in PB was assessed at baseline, cycle 7, 9, and 12, and every 3 months after treatment completion until follow-up Month 18, and then every 6 months until 9 years from last patient enrollment.

Results

Baseline characteristics

In the MURANO trial, a total of 91 and 120 patients were included in the VenR and BR arms, while 153 and 154 patients were included in the Ven-Obi and Clb-Obi arms of the CLL14 study, respectively. Selected baseline variables used in covariate testing of the MRD models are shown in Table 1.

Table 1. Baseline variables for covariate testing*

aCGH, array comparative genomic hybridization; del(11q), chromosome 11q deletion; del(13q), chromosome 13q deletion; del(17p), chromosome 17p deletion; FISH, fluorescence in situ hybridization; GC, genomic complexity; IGHV, immunoglobulin heavy-chain variable-region; PCR, polymerase chain reaction; TLS, tumor lysis syndrome; WES; whole exome sequencing. *Adapted from Kater, et al.1 and Al-Sawaf, et al.2 †Variables excluded from the covariate screening in the MURANO study, either because the lowest patient number in a category for a binary outcome was <10% of study population or there was ≥ 20% missing data.
	MURANO		CLL14
Variables, n	VenR (n = 91)	BR (n = 120)	Ven-Obi (n = 153)	Clb-Obi (n = 154)
Age <65 years	47	45	—	—
Gender, female	27	28	54	50
Mutated IGHV (by PCR)	22	33	68	75
Disrupted del(17p)/TP53†	22	24	17	16
Mutations (by WES)
ATM†	14	22	17	17
BICR3†	2	5	6	4
NFKBIE†	5	7	3	8
NOTCH†	7	24	21	25
SF3B1†	9	14	20	26
XPO1†	9	13	6	14
TP53	18	22	14	11
Abnormal cytogenetics (by FISH)
del(17p)	16	18	13	4
del(11q)	28	39	29	23
Trisomy 12	12	23	—	—
del(13q)	16	24	—	—
Response	2	26	—	—
TLS high-risk	26	34	37	35
Prior lines of therapy ≥2	35	40	—	—

MURANO - MRD level at EoT

The median level of MRD at EoT was significantly lower in the VenR arm compared with the BR arm (p = 5.2 × 10^-8).

• There was no difference in median MRD level at EoT by immunoglobulin heavy-chain variable-region (IGHV) or TP53 mutation status in VenR arm (p = 0.79 and p = 0.48, respectively), whereas in the BR arm patients with TP53 mutation had significantly higher median MRD levels at EoT compared to those with wild type TP53 (p = 0.002).   

Median MRD doubling time was significantly longer in VenR arm compared to the BR arm (93 vs 53 days, respectively).

• Longer MRD doubling time was predictive of longer PFS, with median time to progression of 41.3 vs 16.1 months in patients with MRD doubling time ≥ 62 days compared with patients whose MRD doubling time was <62.6 days, respectively.

Covariates including treatment type, age, IGHV and TP53 mutation status, and tumor burden at study initiation were significantly associated with MRD growth rate.

CLL14 – MRD, PFS and OS

The MRD levels of patients 3 months after completion of therapy are shown in Table 2.

• More than half (55.8%) of patients in the Ven-Obi arm who had uMRD levels < 10^-6 at follow-up Month 3 already had uMRD levels < 10^-6 at Cycle 7 (after completion of therapy).

• 25.5% of patients in the Ven-Obi arm showed deepened MRD responses after continuing six cycles of Ven monotherapy.
• Of the 14 patients in the Ven-Obi arm with detectable MRD (≥ 10^-4) at follow-up Month 3, seven (50%) had decreasing MRD values from Cycle 7 to follow-up Month 3 and seven (50%) had increasing MRD values until EoT.

Table 2. MRD levels at follow-up month 3*

Clb-Obi, chlorambucil-obinutuzumab; MRD, minimal residual disease; NE, not evaluable; PD, progressive disease; Ven-Obi, venetoclax-obinutuzumab *Adapted from Al-Sawaf, et al.2
MRD levels by NGS, %	Ven-Obi	Clb-Obi
MRD <10-6	40	7
MRD ≥10-6 and <10-5	26	13
MRD ≥10-5 and <10-4	8	14
MRD ≥10-4 and <10-2	5	21
MRD ≥10-2, PD, Death, or NE	20	46

 

In the Ven-Obi arm, covariates including CLL-international prognostic index, disease burden, response to treatment, IGHV status, and complex karyotype had significant impact on the appearance of detectable MRD (p<0.05).

• Treatment arm, disease burden, CLL-international prognostic index, and BM MRD status at EoT were independent prognostic factors for MRD conversion in multivariate analysis.

The median MRD doubling time was longer for the Ven-Obi arm compared with the Clb-Obi arm (80 vs 69 days; p=0.0039).

• The median time from follow-up Month 3 to MRD level of 10^-2 was significantly longer in Ven-Obi arm compared with the Clb-Obi arm (1,259 vs 233 days; p<0.0001).
• Patients with detectable MRD in BM and uMRD in PB had shorter MRD doubling times compared with patients who had uMRD in both PB and BM (60 vs 80 days in the Ven-Obi arm and 42 vs 52 days in the Clb-Obi arm, respectively).
• Median time to MRD conversion was 21 vs 6 months in Ven-Obi and Clb-Obi arms, respectively.

The estimated 4-year PFS rate was higher in the Ven-Obi compared with the Clb-Obi arm (74% vs 35.4%).

• Sustained PFS was observed in Ven-Obi compared to Clb-Obi arm (median PFS not reached vs 36.4 months, respectively; p<0.001).
• PFS was also significantly longer in patients with TP53 mutations in the Ven-Obi arm compared to the Clb-Obi arm (49 vs 20.8 months; p=0.03).
• In both arms, PFS was longer for patients with mutated IGHV compared with unmutated IGHV (Ven-Obi arm, hazard ratio [HR], 0.47; p=0.02; Clb-Obi arm, HR, 0.33; p<0.0001).

No difference was observed in OS in both arms (HR, 0.85; p=0.49).

• The number of patients that died (34 vs 41) and the estimated 4-year OS rates (85.4% vs 83.1%) were similar in the Ven-Obi and Clb-Obi arms.

Conclusion

The MRD analyses from MURANO and CLL14 studies have demonstrated that post EoT MRD growth is decelerated in patients with relapsed/refractory CLL treated with the combination of either VenR or Ven-Obi. Taken together, these findings suggest that fixed duration Ven plus anti-CD20 therapy may contribute to delayed progression of CLL by slowing the growth of the residual disease clones. The studies are however limited by the absence of validation cohorts and absence of some data. Clonal growth may be influenced by biological and treatment related parameters; further studies are needed to explore this association.