Standardized MRD Flow and ASO IGH RQ-PCR for MRD Quantification in CLL Patients after Rituximab-Containing Immunochemotherapy – a Comparative Analysis in 574 Samples from the Randomized GCLLSG CLL8 Trial

Combinations of purine analogues with the therapeutic anti-CD20 antibody Rituximab induce a high number of complete clinical remissions in CLL patients. Therefore, the detection of minimal residual disease (MRD) appears particularly warranted after such regimens. However, Rituximab prevents the assessment of CD20, one mainstay of flow cytometric MRD detection (MRD flow) in CLL. As no comparative analysis of MRD flow to ASO primer IGH real-time quantitative PCR (RQ-PCR) after Rituximab has been performed so far, doubts remained regarding the diagnostic utility of MRD flow in this clinical setting. We compared results obtained by RQ-PCR to 4-color 3-tube MRD flow in 574 samples from 69 patients randomized to receive fludarabine and cyclophosphamide (FC) or FC plus Rituximab (FCR). Thus we were able to investigate how treatment with FC and FCR impacts on the performance of DNA-based RQ-PCR and of antibody-dependent flow cytometric MRD detection. With 58.4 % positive and 27.0 % negative samples by both assays, we found a qualitative concordance between MRD flow and RQ-PCR of 85.4 % (490/574 samples). Only very few samples were exactly quantifiable with one method and negative by the other method (0.7% MRD flow pos/RQ-PCR neg and 1.4% MRD flow neg/RQ-PCR pos samples). Most discordant samples (12.5% of all samples) were MRD negative by flow but positive by RQ-PCR at low levels outside the quantitative range of MRD detection. We calculated a median quantitative range of 10⁻⁴ and a median sensitivity of 2×10⁻⁵ in those samples, demonstrating a higher sensitivity of PCR to detect low-level MRD. Considering a threshold of 10⁻⁴ for MRD positivity, 94% of all samples showed concordant results by both methods. The cases that were discordantly classified as to this threshold usually comprised samples with residual CLL cells close to 10⁻⁴ (3.7% RQ-PCR pos/MRD flow neg and 2.3% RQ-PCR neg/MRD flow pos samples). Quantitative MRD levels determined by both methods were closely correlated irrespective of therapy (Spearman r = 0.95 in FCR, r = 0.96 in FC). We next classified all samples according to the PCR result and compared MRD flow detection rates between the treatment arms. Regarding PCR positive samples within the quantitative range, MRD flow was positive in 96.8% and 98.0% of samples from the FCR and FC arms, respectively. Positive MRD flow results were recorded in 33.7% (FCR) and 40.0% (FC) of samples that concomitantly yielded PCR positive results outside the quantitative range. Within the PCR negative samples, MRD flow detected residual CLL cells in comparable proportions of samples (FCR 2.3% vs. FC 3.4%). We compared the MRD results obtained using combinations that include CD20 (CD20/CD5/CD19/CD43, CD79b/CD20/CD19/CD5) to the MRD results from the combination CD81/CD22/CD19/CD5 in individual samples. We found that CD20 is undetectable on residual CLL cells and therefore not informative for MRD flow in patients from the FCR arm up to 180 days after the last Rituximab infusion. For the same period of time benign B cells are significantly less frequent in patients after FCR than after FC (p<0.0001) and often account for less than 1 cell in 10,000 leukocytes after this therapy regimen. Thus the efficient reduction of background non-CLL B-cells by Rituximab likely explains the high specificity and sensitivity of MRD flow after FCR treatment. In summary, MRD assessment by flow and PCR are equally effective for MRD quantification in Rituximab treated CLL patients within the sensitivity range of 10⁻⁴, while PCR is more sensitive for detecting MRD below that level. The use of Rituximab does not influence MRD detection neither by flow nor by RQ-PCR.