Patient samples and timepoints
All patient material used in this study was obtained from the Dutch HOVON-139/GIVE trial (NTR6043), operated by the HOVON consortium. The HOVON-139/GIVE trial is a prospective, open-label, multicenter randomized phase-II trial intended to evaluate the efficacy and safety of pre-induction with two cycles of obinutuzumab, followed by induction with six cycles of obinutuzumab and venetoclax and six cycles of venetoclax alone, followed by either 12 cycles of venetoclax maintenance or MRD-conditional treatment cessation(7). Inclusion was restricted to previously untreated CLL patients, unfit for treatment with fludarabine, cyclophosphamide and rituximab-like regimens. More detailed information on in- and exclusion criteria and the trial protocol is available in the primary publication(7). MRD was measured by NGS at four timepoints: during the fourth week of venetoclax ramp up, at the end of induction treatment (after cycle 12), six months after randomization, and one year after randomization (i.e., at the end of optional consolidation treatment).
IGHV-leader NGS
MRD was quantified using the previously described IGHV-leader NGS assay(8). Briefly, IGHV libraries were amplified from genomic DNA, using the IGHV leader primer set developed by the EuroClonality-NGS working group(9). Per sample, we aimed for 6μg of DNA input, in PCR triplicates of 2 μg. After PCR cycling, libraries were checked using the D1000 DNA ScreenTape assay on a 4200 Agilent TapeStation (Agilent, Santa Clara, CA, USA) and purified using the AMPure XP kit (Beckman Coulter, Brae, CA, USA) with a volume ratio of 1.8X. PCR products were diluted to 4nM and equimolarly pooled in a single library-pool. Multiplexed libraries were denatured and sequenced at a final loading concentration of 8-20pM, with 5-20% of PhiX (Illumina, San Diego, CA, USA). Paired-end sequencing (2x300 cycles) was performed on a MiSeq system using a 600-cycle v3 kit (Illumina). NGS output was uploaded to the ARResT/Interrogate immunoprofiler(10). Reads were converted to cell-equivalents using the EuroClonality central in-tube quality control(11). The clonotypic IGH sequence, identified in a pre-treatment sample, was used to define malignant cell-equivalents. MRD depth was calculated by dividing the number of malignant cell-equivalents by the total number of cells equivalent to DNA input.
Limit of detection and quantification
As previously determined, the limit of detection and quantification (LoD/LoQ) of the IGHV-leader NGS assay are 3.4 and 3.8 malignant cell-equivalents per assay, respectively. Consequently, 6μg DNA input allows detection and quantification down to MRD 4,1*10-6. However, in some samples, final PCR input was either higher or lower (range 2-12μg, see Supplementary Table 1). To allow for uniform analysis of all samples, a LoD/LoQ cut-off was established at MRD 10-5. All samples assessable to at least MRD 10-5, where MRD levels were not detectable or <10-5, were designated as undetectable MRD (uMRD). Samples without detectable MRD that were not assessable to MRD 10-5 were excluded from the analysis.
IGHV repertoire diversity
The workflow of the IGHV-leader NGS assay involves PCR-based amplification and sequencing of all IGHV rearrangements present in a sample. Not only does this capture information on the presence or absence of residual leukemic cells, but it also allows characterization of the healthy background IGHV repertoire. The sequence diversity of the IGHV repertoire, reflective of the size of the polyclonal healthy B cell pool, can be quantified using Shannon’s diversity index (H), which has 0 as minimal value (i.e., only one specie exists in the population), but has no theoretical upper limit (14). IGHV repertoire analysis was restricted to samples with at least 6μg DNA input. Samples from sequencing runs with an above-average error-rate were excluded from the analysis. Annotated IGHV repertoires were downloaded from ARResT/Interrogate and loaded to R, version 4.1.0(15). Shannon’s diversity index was calculated using the vegan package(16). Healthy donor IGHV repertoires had previously been obtained from a cohort of older individuals without signs or symptoms of any lymphoproliferative disease(17).
Multicolor Flow-cytometry
MRD measurements had been previously performed by multicolor flow cytometry (MFC) at the end of induction treatment, six months after randomization and one year after randomization, with a LoD of MRD 10-4 (12,13). In this approach, MRD is expressed as the number of residual leukemic cells as a fraction of the total leukocyte population, including granulocytes. To facilitate comparison to NGS measurements, MFC measurements were corrected using the concurrently measured granulocyte percentage.
Statistical analysis
Data analysis was performed in R. For continuous data, statistical significance was assessed using an unpaired Welch’s t-test. For categorical data, statistical significance was assessed using a χ2 test, or using Fisher’s exact test if these were fewer than 5 observations per condition. Pearson’s r was calculated using log10-transformed data. Kaplan-Meier survival analysis was performed using the survival package(18). Differences in survival time were evaluated using a log-rank test. Due to insufficient power, the trial protocol precludes any statistical comparison between the two randomization arms of the trial, regarding treatment efficacy and prognosis.
Data availability statement
IGHV sequencing data has been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO series accession number GSE225043(19). All other original data can be obtained upon reasonable request.