Adult patients with molecularly confirmed glioblastoma, IDH wildtype (GB-IDHwt), with radiological and clinical differential diagnoses including lymphoma in cases 4 and 5 (Table 1). CSF spinal tap cfDNA sequencing remained diagnostically inconclusive in each case. These patients were included to illustrate the limitation of our CSF analysis workflow, most likely due to insufficient cfDNA fractions released from GB-IDHwt cells.

Discrimination of CNS lymphoma from other neuroradiological mimics including small-cell cancers and non-neoplastic conditions such as encephalitis is of utmost clinical importance. At least to some extent, our approach can detect inflammatory signatures [9] and pinpoint neoplasia based on CNVs. Moreover, clearance of cf-tDNA from CSF may reflect a sustained tumour response in CNS-lymphoma patients [8], making our approach a potential monitoring tool, even though it is mainly qualitative.

Its universal applicability, low hands-on, and infrastructure requirements [9] may help to reduce the number of patients receiving steroids before lymphoma diagnosis. Such pretreatment often massively delays or even prevents timely CNS-lymphoma diagnosis. In cases with two positive confirmatory results (nanopore sequencing/ ddPCR), CNS biopsy may even be omitted bearing in mind that surgical complication rates of stereotactic brain biopsies of ~ 5%, mainly haemorrhages, negatively impact outcomes [3].

The sensitivity of our workflow under concurrent corticosteroid therapy remains to be elucidated. In urgent clinical settings, steroid therapy could theoretically be initiated as soon as an elevated CSF cfDNA content, suitable for further analysis, has been established. Importantly, CSF sampling and shipment at ambient temperature to a laboratory for sequencing can be completed within 24 h.

Whilst the small cohort size is a clear limitation, our observations still highlight the utility of nanopore CSF workup and call for democratising accessibility of high-quality DNAmeth/CNV-reference datasets to broaden applicability. Given the low rates at which CNS-lymphoma diagnoses can be obtained through conventional CSF analyses [2, 14], our findings could prove transformative for the clinical management of patients with suspected CNS lymphoma.