Abstract
Background Currently available clinical and biological prognostic factors do not adequately identify first-line DLBCL patients at risk for treatment failure. In DLBCL, circulating tumor DNA (ctDNA) can be utilized as a for molecular disease classification, detect minimal residual disease (MRD) or predict disease progression. We hypothesized that ctDNA detected by the EuroClonality immunoglobulin-based next generation sequencing (IG-NGS) assay correlates with treatment response and outcome.
Patients and methods Overall, 451 samples (41 diagnostic fresh frozen paraffin embedded (FFPE) biopsies, 64 peripheral blood mononuclear cells (PBMC) and 346 plasma samples) from patients from a Phase Ib study in de novo DLBCL treated with tafasitamab and R-CHOP or tafasitamab and R-CHOP in combination with Lenalidomide (First-MIND; MOR208C107) were analyzed. DNA from FFPE or diagnostic PBMC was sequenced by using a 2-step PCR approach with the Euroclonality NGS IGH-VJ, IGH-DJ and IGK primer sets (euroclonality.org) (Brüggemann, Leukemia, 2019) with a number of primers adapted for usage in short fragment cfDNA. IG markers were identified at abundance level ≥5% of annotated IG reads and a >1 log higher abundance to the next most frequent clonotype. Disease related clonotypes were traced in plasma during (C2D1 and C4D1 n=34) and at end of treatment (EOT n=32) by a 1-step PCR approach and sequencing of at least 5000 human genome equivalents (hGE) of cell free (cf)DNA. ctDNA levels were determined as the number of specific clonotype molecules per input genome equivalents normalized by a reference standard DNA spiked into each sample as cIT-QC (Knecht, Leukemia, 2019) and reported per plasma volume. Data were analysed by ARResT/Interrogate (Bystry, Bioinformatics, 2017). Patients were reported MRD positive if ≥1 clonal IG rearrangement was detected.
Results From 41 patients with available diagnostic FFPE and sufficient DNA quality for marker screening, 34 (89%) had at least 1 detectable clonal marker (IGH-VJ, IGH-DJ or IGK). One clonal IG marker was identified in 20/34 (59%) patients, whereas two or three markers were identified in 9 (27%) and 5 (14%) patients, respectively. Disease specific clonotypes identified in FFPE were recovered in 11/34 PBMC samples (32%) with a median abundance of 0,29% (range 0.006-21.9%) demonstrating a peripheral blood involvement. In 32/33 (97%) pretreatment plasma samples, ctDNA was detected with a median copy number of 172.5/ml plasma. cfDNA and ctDNA levels prior to treatment correlated with IPI and LDH (Fig1). ctDNA dynamics during treatment demonstrated rapid treatment response at C2D1. 23/34 samples were either MRD neg (n=19) or showed a ctDNA reduction ≥ 2 log levels (Fig2). At C4D1 7/34 (20%) samples had low-level detectable MRD and at EOT 29/32 (91%) patients achieved MRD negativity after treatment with tafasitamab and R-CHOP +/- lenalidomide (LOD 2 x 10 -4 for follow-up samples). Early ctDNA dynamics predicted metabolic response when applying a threshold of ctDNA reduction to <1% after the first treatment cycle. 25/28 (89%) patients completing treatment and achieving PET-response could be identified early during treatment by ctDNA assessment.
Conclusion Baseline ctDNA levels determined by the EuroClonality IG-NGS assay and absolute cfDNA amounts correlated with pre-treatment risk factors. ctDNA dynamics after the 1 st treatment cycle demonstrated rapid treatment response to tafasitamab + RCHOP +/- lenalidomide and correlated with metabolic response. Assessment of ctDNA dynamics allows early response assessment and might be useful for risk stratification in patients with DLBCL.
Kuffer: Morphosys AG: Current Employment. Blair: Morphosys AG: Current Employment.
