Clinical Usefulness of Sequential Circulating-Tumor DNA Analysis in Patients with Advanced Stage Diffuse Large B-Cell Lymphoma Treated with R-CHOP Immunochemotherapy

Background

Circulating-tumor DNA (ctDNA) has emerged as an innovative approach to analyze the genetic information in various malignancies. Previous studies reported that pretreatment ctDNA level and molecular response during treatment were prognostic factors for diffuse large B-cell lymphoma (DLBCL). In this prospective study, we aimed to evaluate the usefulness of plasma ctDNA to determine the mutational profile, predict clinical response in newly diagnosed advanced stage DLBCL patients who were treated with R-CHOP.

Methods

Patients who were newly diagnosed with DLBCL between 2022 March and 2023 April, with advanced stage (Ann Arbor stage 3 or 4) and stage 2 with bulky disease (≥ 8cm), and scheduled for 6 cycles of R-CHOP treatment were included in this study. Plasma samples were collected sequentially at diagnosis (baseline), after 3^rd cycle (interim) and after 6^th cycle of R-CHOP (final). Paired formalin-fixed paraffin-embedded (FFPE) tissue samples were collected from available patients. Treatment responses were assessed according to the Lugano classification criteria using ¹⁸F-FDG PET/CT assessment.

Genomic DNA from peripheral blood mononuclear cells (PBMCs) and FFPE was extracted using the QIAamp DNA Mini Kit (QIAGEN) and QIAamp DNA FFPE Advanced UNG Kit (QIAGEN). Library preparation was performed using the DxSeq Lymphoma & Myeloma ctDNA, which includes 112 lymphoma-related genes for the Illumina Platform Kit. Paired-end sequencing was performed on the NovaSeq 6000 System (Illumina) with a >28,000× average sequencing depth. Variants in both cfDNA and PBMC were determined as germline variants or variants of clonal hematopoiesis of indeterminate potential (CHIP).

Results

Fifty-two patients were enrolled in this study. Median age of the patients was 70 years (rage 28-84), and 58.5% of the patients were male. Four patients (7.5%) were assessed as stage 2 with bulky disease and 48 patients were stage 3-4. CtDNA could be obtained from all 52 patients, and FFPE tissue samples were obtained in 49 patients. A total of 148 and 225 somatic mutations were identified in the 49 tumor tissues and 52 plasma ctDNA samples at baseline. The overall sensitivity of plasma ctDNA in discovering biopsy-confirmed mutations was 100.0% with a specificity of 99.5%. The concordance rate between the mutations assessed by ctDNA and tumor tissues was 74.7%. In addition, 39 patients were successfully classified by genetic subtype according to baseline ctDNA analyses (MCD, 28.9%; EZB, 23.1%; BN2, 15.4%; ST2, 7.7%; Unclassifiable, 25.0%).

After 6 cycles of R-CHOP, 38 patients were assessed as complete response (CR), 2 patients were partial response (PR), 1 patient was stable disease (SD) and 10 patients were progressive disease (PD). The median values of maxVAF, meanVAF at interim analysis were higher in non-responder group (SD/PD) than responder group (CR/PR) with statistical significance (maxVAF, 0.3 vs. 1.0, p=0.021; meanVAF, 0.2 vs. 0.6, p=0.017). Among the non-responder group (n=11), one or more key driver mutations at the interim analysis were identified in 6 patients. However, 4 out of 6 patients achieved a partial or complete metabolic response based on interim PET assessment. The median values of interim maxVAF, meanVAF in 2 patients who had less than PR at interim PET/CT were significantly higher than the others (maxVAF, 0.3 vs. 29.3, p=0.002; meanVAF, 0.2 vs. 16.1, p=0.001). In responder group (n=40), key driver mutations at baseline was disappeared in all patients at interim analysis.

Conclusion

Pre-treatment ctDNA reflected the mutational profile of tumor tissues with a high concordance rate. Interim ctDNA analysis was significantly associated with final clinical response of R-CHOP and could augment the interim response assessment for predicting treatment failure.

No relevant conflicts of interest to declare.