A Feasibility Study Evaluating the Use of Cell-free DNA Analysis in Laboratory Brain Cancer Investigations

Abstract

Abstract Aims Circulating tumour DNA (ctDNA), shed from solid cancers in to the plasma, represents an exciting analyte for diagnosis and monitoring of disease in cancer patients. However, its use in glioma brain cancer patients represents a challenge, due to reduced permeability of the blood brain barrier. This pilot study sought to investigate the practical aspects and clinical utility of using cell-free DNA (cfDNA) in glioma tests in a NHS diagnostic laboratory. Firstly, we investigated the potential of ctDNA as a proxy for the brain cancer biopsy; where cfDNA analysis was compared to the paired FFPE brain specimen for relevant glioma genetic biomarkers. Secondly, ctDNA constitutes a portion of the overall cfDNA and there is evidence cfDNA metrics per se may also be of value as prognostic tools and surrogates of tumour burden. Additionally, we investigated a potential role for cfDNA metrics in prognostic impact; linking cfDNA concentrations to clinical outcome measures. Method 10ml peripheral blood was collected in specialist preservative tubes and cfDNA isolated using an extraction kit (Qiagen MinElute ccfDNA kit). cfDNA concentration and purity was assessed using chip-based automated electrophoresis. Where relevant (12/39 cases), cfDNA samples were run though laboratory tests of IDH variant detection, 1p19q co-deletion assessment and MGMT promoter methylation analysis. Results were compared with ‘standard of care’ brain biopsy tests. A potential correlate of cfDNA concentration and clinical outcomes data were assessed in a sub-cohort of glioblastoma patients (n=32). The cohort was divided in to 2 groups – high cfDNA vs. low cfDNA - based on whether a subject’s extracted sample cfDNA concentration fell above or below the mean. Comparison of overall survival in months between subjects was checked for normal distribution using the Shapiro-Wilk t-test. The test of equity of survival distributions for the high cfDNA vs. low cfDNA was then analysed as a Kaplan-Meier curve. Results The protocol delivered cfDNA of high purity, averaging 91%, within the plasma nucleic acid fraction, however the cfDNA concentrations (mean ≈1ng µl-1) fell below the conventional limit of detection of the laboratory tests. In spite of the low concentration, cfDNA samples did generate test PCR amplicon; however results reflected the germline DNA profile rather than the new somatic changes of the tumour. The cfDNA analysis did not pick up the tumour biomarkers seen in the paired tumour biopsy sample. In a second part of the study, cfDNA concentrations for the glioblastoma cohort were assessed in the context of their clinical outcomes data. The data showed a correlate where high cfDNA concentration in the extracted sample was independently associated with inferior outcome in terms of overall survival, with Log Rank significance p=0.014 (Figure 1). Conclusion The cfDNA yields from a 10ml blood sample were consistently too low to meet the limit of detection requirements of the standard laboratory neuropathology genetic tests and glioma tumour profile could not be picked up against the germline background. Thus, in spite of the considerable advantages to glioma plasma molecular testing, using cfDNA as a proxy for a brain biopsy would currently not be possible in our routine diagnostic environment. However, within the limitations of the pilot project testing strategy, the data showed an interesting correlate where high cfDNA concentration was independently associated with inferior outcome in terms of overall survival for glioblastoma patients. Given the simplicity of obtaining this quantifiable metric, there are grounds for further investigations as to its utility; not only with survival outcomes, but also potential correlation with the clinical assessment of tumour burden, blood brain barrier integrity and disease pseudoprogression.