PCNSL: biomarker better than biopsy?

In this issue of Blood, Baraniskin and colleagues report on microRNAs (miRNAs) as a possible biomarker for the diagnosis of primary central nervous system lymphoma (PCNSL).¹  Levels of miR-21, miR-19, and miR-92a were significantly increased in cerebrospinal fluid (CSF) samples from PCNSL patients compared with controls with inflammatory CNS disease or other neurologic disorders.

The diagnosis of PCNSL is most commonly achieved via stereotactic brain biopsy. Contemporary imaging methods (CT, MRI, PET) fail to reliably differentiate inflammatory processes, solid-tumor metastases, and primary or secondary CNSL. A misinterpretation of findings can lead to a delay in initiating therapy on the one hand, or to unnecessary resection of PCNSL on the other, in conjunction with related morbidity. Both of these situations can affect the patient's outcome dramatically. The success of stereotactic biopsy, while the histologic gold standard, depends on accessible lesions, and it is sometimes unfeasible when lesions lie close to or within critical brain structures. Although most brain biopsy procedures are safely performed there is up to a 7% risk of hemorrhage and up to a 35% risk of failure to achieve a definitive histologic diagnosis.²  CSF examination can only provide definitive evidence of PCNSL in the presence of leptomeningeal dissemination and is < 50% sensitive for the diagnosis of PCNSL in this setting.³  Antithrombin III,⁴  soluble CD-27,⁵  and free immunoglobulin light chains⁶  are CSF biomarkers that have not yet achieved general acceptance in clinical practice. miRNAs are short RNA molecules that bind the 3′-untranslated regions of mRNA transcripts (see figure panel A). They inhibit gene expression at a posttranscriptional level by interfering with the translational initiation or degradation of mRNA.⁷ 

Baraniskin et al were the first to define the role of miRNAs in the CSF of lymphoma patients. Taking a “candidate approach” and quantifying miRNA via qRT-PCR, the authors identified significant levels of miRNAs in the CSF of PCNSL patients. In particular, miR-21, miR-19, and miR-92a had diagnostic value in distinguishing PCNSL from inflammatory CNS diseases and other neurologic disorders. Despite their small patient sample (n = 23), using combined miRNA analyses they demonstrated that these candidate miRNAs have high sensitivity (95.7%) and specificity (96.7%) for PCNSL diagnosis (see figure panel B). The authors also demonstrated that miRNAs in the CSF exhibited remarkable stability with resistance to exogenous RNase, repeated freeze-thaw cycles, and long-term storage of CSF specimens. Possible explanations for this phenomenon include miRNA protection in exosomes or association of miRNA with other molecules (eg, CSF proteins). These features may enhance the practical utility of CSF miRNA for diagnostic purposes.

The article by Baraniskin et al advances the field of diagnostic markers in CNSL. Perhaps the analysis of miRNAs in the CSF of patients with suspected PCNSL will expand the diagnostic tools at our disposal, especially in patients in whom biopsy appears too risky or when histologic findings are equivocal. However, as these data were generated from a small number of patients, it will be up to future studies to validate the diagnostic utility of miRNAs in the PCNSL patient population. Finally, there is the intriguing possibility that miRNAs derived from primary brain tumors like PCNSL may also circulate in blood,⁸  which could offer a readily accessible source of tumor-derived RNA for future study.