Background: CNS involvement in aggressive lymphomas (including diffuse large B-cell lymphoma, DLBCL) is associated with poor survival, requiring aggressive CNS-directed therapy to avert dismal outcomes. Early detection of clinically occult CNS invasion is thus critical. Clinical risk factors have low specificity to select patients for CNS prophylaxis, and routine evaluations of the CSF by cytology or flow cytometry have low sensitivity (Doolittle et al., Blood 2008). Moreover, most CNS recurrences of DLBCL occur in the brain parenchyma and remain occult on CSF examination. There is a high unmet need for an accurate method to identify subclinical CNS invasion and help non-invasively diagnose parenchymal disease. Clonotype-specific cell-free DNA (cfDNA) can be detected in the plasma of patients with DLBCL using high-throughput sequencing (HTS) combined with multiplex PCR, and cfDNA reappearance after therapy precedes systemic recurrence (Roschewski et al., Lancet Oncol. 2015). However, clonotypic cfDNA has not been systematically evaluated in the CSF compartment. Because of the blood-brain barrier that isolates CSF from plasma, clonotypic cfDNA might hold promise as a sensitive method of CNS lymphoma detection. Therefore, we sought to evaluate the ability of HTS-based cfDNA assay in the CSF to detect CNS invasion.
Methods: We collected CSF samples from consenting patients with aggressive lymphomas who either had clinically overt CNS disease, or who were newly diagnosed and without evident CNS disease, but at high clinical risk. Genomic DNA from primary tumors (formalin-fixed paraffin-embedded [FFPE] tissue; Fig. A) was analyzed for tumor-specific clonotype using a quantitative HTS of rearranged IGK, IGH (VJ or DJ), or IGL loci, as previously described (Carlson CS et al., Nat Commun. 2013). Clonotypes in each primary tumor were selected for subsequent tracking in the CSF based on frequency relative to other B cells and all nucleated cells, as well as distribution. HTS of B-cell receptor (BCR) was performed on 5 mL CSF centrifuged at 1000x for 5 min at 4°C and separated into acellular portion for cfDNA, and a cell pellet suspension. Clonotype copy numbers are expressed per mL for acellular fluid, and per 106 cells for cell pellets.
Results: The HTS assay identified median 3 (range, 2-6) dominant immunoglobulin sequences in each primary FFPE lymphoma sample (N=9), with median dominant clonotype frequency 39.0% (range, 22.0-97.9%) and median copy count 753,618 per 106 cells (range 47,455-1,407,748). In the CSF, the HTS/ASO-PCR assay detected the dominant clonotype in 5 out of 9 samples (Table), including all cases (N=3) with overt CNS disease, and 2 without known CNS invasion (sensitivity=100% and specificity=67% for clinically overt disease). Of note, the cfDNA assay detected tumor-specific clonotype in the CSF of a patient with primary CNS lymphoma (PCNSL) who had negative CSF evaluations by flow cytometry and IGH PCR. In the cell-free CSF, median detectable cfDNA clonotype copy count was 1,649/mL (range, 2-5,620), and in the CSF cell suspension it was 389,558 per 106 cells (range, 10,373-1,181,171). Copy counts were significantly higher (P=.004 for cfDNA in CSF fluid, P=.004 for cell suspension) in patients with overt leptomeningeal involvement (positive CSF cytology) than those with parenchymal or clinically occult disease (Fig. B).
The 2 cases with positive CSF cfDNA but negative clinical staging and CSF cytology were notable for high risk of CNS dissemination due to ruptured epidural tumor with spinal cord compression, or double-hit histology (MYC-IGH and BCL2-IGH rearrangements) with extensive extranodal invasion of kidneys, pleura, and liver. Both patients were recommended CNS-directed prophylaxis based on clinical risk.
Conclusions: HTS of BCR loci in the CSF of patients with aggressive lymphomas can identify intraparenchymal CNS or leptomeningeal disease with high sensitivity. More importantly, it can detect intraparenchymal involvement not identifiable using traditional means of CSF analysis (flow cytometry, standard IGH PCR). Detection of cfDNA in CSF of some high-risk patients without overt CNS disease will be further explored in a larger sample. It suggests that pre-treatment HTS of BCR loci could be prospectively tested to predict the risk of CNS recurrence, with a future potential to enable more sensitive and specific selection of patients for CNS prophylaxis therapy.
Olszewski:Spectrum Pharmaceuticals: Research Funding; TG Therapeutics: Research Funding; Adaptive Biotechnologies: Research Funding; Genentech: Research Funding. Reagan:Incyte: Membership on an entity's Board of Directors or advisory committees; Pfizer: Research Funding. Mullins:Adaptive Biotechnologies: Employment.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal