Molecular Profiling and Disease Tracking of Post-Transplant Lymphoproliferative Diseases Using Cell-Free DNA

Introduction.

Post-transplant lymphoproliferative disorder (PTLD) comprises a spectrum of hematological malignancies that develop after solid organ transplantation (SOT). Unlike diffuse large B-cell lymphoma (DLBCL), initial treatment of PTLD patients starts with rituximab monotherapy. Identifying patients who do not respond to rituximab is essential to prevent delay in administering chemotherapy. We previously showed that cell-free tumor DNA (ctDNA) isolated from plasma represents a promising approach to profile PTLD at diagnosis [PID 37705050]. The use of ctDNA as biomarker has shown feasible in DLBCL, but has not yet been established for PTLD. We aimed to explore the feasibility of ctDNA for genomic profiling and response assessment in PTLD.

Methods.

Plasma samples of 28 PTLD patients were collected in this prospective observational multi-center cohort study in the Netherlands (National trial registry 7402). Low-coverage whole genome sequencing (lcWGS) was performed for detection of copy number variants (CNVs). Ultra deep (median ~15.000x target coverage) targeted next generation sequencing (NGS) was performed for the detection of single nucleotide variants (SNVs). The targeted panel contained 117 genes (comprising 391 kb) known to be recurrently mutated in B-cell lymphomas. SNVs were called using an in-house pipeline, optimized for detecting small somatic variants with low variant allele frequencies (VAFs) [PID 37705050]. For baseline samples, the VAF cutoff was set to >0.5%. For follow-up timepoints, this cutoff was removed to further increase sensitivity. ctDNA concentrations based on the VAF of SNVs were expressed as haploid genomic equivalent per milliliter (hGE/mL) of plasma. IchorCNA was used to call CNVs.

Results.

Median age of patients was 57 years (range 29-74), with 68% presenting with Ann Arbor stage III - IV. Median time between SOT and PTLD was 68 months (range 3-379). SOT consisted of kidney (n=17, 61%), liver (n=7, 25%), lung (n=3, 11%) and hematopoietic stem cell transplant (n=1, 4%). Pathology review showed DLBCL (19/24), Burkitt like (1/24), polymorphic PTLD (1/24) and undetermined B-cell lymphomas (3/24). EBV was negative for 10 (45%), positive for 12 (55%) and unknown for 2 cases. Out of 28 patients, 21 (75%) started initial treatment with rituximab, 6 (21%) initiated R-CHOP therapy and one patient (4%) did not receive treatment.

In total 92 plasma samples were analyzed. At diagnosis, a total of 285 SNVs were detected in cfDNA samples of 27 out of 28 (96%) patients. The median number of SNVs per sample was 5 (range 0-35). KMT2C was the most frequently mutated gene (12/28, 43%), followed by ARID1B (8/28, 29%), P2RY8 (7/28, 25%), KMT2D, IGLL5 and TP53 (5/28, 18%). CNVs were detected in 12 out of 21 baseline samples (57%). The most frequent CNVs were gains of 11q and 18q.

Out of the 28 PTLD patients, 26 exhibited sufficient baseline mutations to facilitate response assessment, while six cases lacked sufficient number of mutations to reliably calculate and track the hGE/ml. Recurring SNVs were detected in the remaining 20 patients (71%). Of these 20 cases, 16 patients initiated treatment with rituximab monotherapy, whereas 4 patients started directly with R-CHOP. Within the subset receiving rituximab monotherapy, 7 patients continued with rituximab alone, while 9 patients escalated to R-CHOP. Among the 7 patients who continued rituximab monotherapy, a rapid decline in ctDNA levels was observed in 6 patients as early as four weeks post-treatment initiation. Of the 9 patients who escalated to R-CHOP, plasma samples from interim timepoints revealed that 7 patients tested positive for minimal residual disease (MRD), while 2 patients tested negative. For the 4 patients who initiated treatment with R-CHOP, follow-up assessments revealed that 2 tested positive for MRD, while the remaining 2 tested negative. Overall, we were able to detect MRD with a sensitivity of as low as 0.2% VAF or 1 hGE/mL of plasma in follow-up timepoints. Clinical outcomes were available for most patients (n=17), while this is pending for the other patients (n=11).

Conclusion.

Our study demonstrates that ctDNA analyses are highly suitable for detecting CNVs and SNVs in diagnostic and follow-up plasma samples from PTLD patients. This approach can be used to molecularly profile PTLD and track disease activity over time. Correlation of ctDNA dynamics with FDG-PET scan results and clinical outcome is ongoing.

van der Poel:Kite, a Gilead company: Honoraria; Takeda: Honoraria. Diepstra:Takeda: Research Funding.