Cell-free DNA as biomarker in Hodgkin lymphoma patients

Introduction Treatment failure or relapse occurs in approximately 10% of early stage Hodgkin lymphoma (HL) patients and 20% of patients who have advanced stage HL. Identifying this group of refractory and relapsed HL patients is crucial in order to plan for effective first line treatment. Considering that genomic aberrations can be readily detected in cfDNA, this method may allow for a non-invasive method of monitoring disease load and serve as a prognostic marker. We aimed to study the genomic profile of HL and correlate these findings to the levels of the circulating biomarker TARC and to other clinical characteristics.

Methods Cell-free DNA was isolated from 1-2 ml of plasma with the QIAamp Circulating Nucleic Acid kit according to manufacturer’s protocol. A targeted panel was designed including 46 genes commonly mutated in B cell lymphoma, genomic regions of immunoglobulin gene loci, MYC, BCL2 and BCL6 to detect chromosomal breaks and part of the EBV genome. Deep targeted sequencing (DTS) was performed with two hybrid capture-based next generation sequencing platforms, namely SureSelect^{XT HS} Target Enrichment System and Twist Custom Panel Multiplex Hybridization Kit. Prior to target enrichment, a small fraction of the indexed libraries was aliquoted for low-pass whole genome sequencing (LP-WGS) to check for copy number aberrations (CNAs). Variant calling was performed with SNPPET SNP caller using SureCall Software. The R package, CNAclinic was used for CNAs analysis.

Results The total yield of cfDNA ranged between 23-429 ng (median 100 ng). There was no significant correlation between cfDNA levels and TARC in this cohort. Using an input varying between 18 and 124 ng, we successfully generated libraries for NGS. Copy number aberrations were detected in four out of eleven samples. In vitro size selection of cfDNA fragments <150bp to enrich for tumor cell derived DNA, resulted in detection of CNAs in three additional samples. Variant calling with SNPPET SNP caller, revealed a median of 130 variants after filtering out common variants (allele frequency (AF) >1% in 1000 Genomes). Hotspot mutations in STAT6 were detected in two HL samples and in one of these two samples we also observed a XPO1 E571k hotspot mutation. These mutations were validated by ddPCR in diagnostic FFPE tissue samples of the two patients.

Conclusion Despite the low percentage of tumor cells in HL tissue samples, somatic mutations with AF as low as 0.5% were detected in cfDNA using a minimal input of 18ng.