
==== Front
Acta Neuropathol
Acta Neuropathol
Acta Neuropathologica
0001-6322
1432-0533
Springer Berlin Heidelberg Berlin/Heidelberg

39225884
2793
10.1007/s00401-024-02793-z
Correspondence
Rapid brain lymphoma diagnostics through nanopore sequencing of cytology-negative cerebrospinal fluid
http://orcid.org/0000-0002-9884-1240
Hench J. jurgen.hench@gmail.com

1
Hultschig C. 1
Bratic Hench I. 1
Sadasivan H. 2
Yaldizli Ö 3
Hutter G. 4
Dirnhofer S. 1
Tzankov A. 1
Frank S. 1
1 grid.410567.1 0000 0001 1882 505X Institute for Pathology and Medical Genetics, University Hospital Basel, Basel, Switzerland
2 grid.34477.33 0000000122986657 AI Group, AMD and Paul G Allen School of CSE, University of Washington, Seattle, USA
3 grid.410567.1 0000 0001 1882 505X Dept. of Neurology, University Hospital Basel, Basel, Switzerland
4 grid.410567.1 0000 0001 1882 505X Dept. of Neurosurgery, University Hospital Basel, Basel, Switzerland
3 9 2024
3 9 2024
2024
148 1 364 7 2024
5 8 2024
19 8 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Keywords

Liquid biopsy
CNS lymphoma
Rapid testing
Nanopore sequencing
DNA methylation
Tumor classification
Non-invasive diagnostics
Unsupervised machine learning
Epigenetics
Point-of-care testing
Stiftung für krebskranke Kinder, Regio Basiliensis2023-F006/007 Hench J. University of BaselOpen access funding provided by University of Basel

issue-copyright-statement© Springer-Verlag GmbH Germany, part of Springer Nature 2024
==== Body
pmcHaematological neoplasms affecting the CNS, most prominently aggressive B-cell lymphomas, require rapid diagnosis, typically by stereotactic biopsy, to initiate treatment, prompting non-invasive modalities [6]. In challenging cases, DNA methylation (DNAmeth) and copy number variant (CNV) profiling of cerebrospinal fluid (CSF) may meet this demand. Healthy and neoplastic cells, including CNS lymphomas [11], may shed DNA fragments into the bloodstream and CSF as cell-free DNA (cfDNA). In addition, cellular debris can contribute to sediment DNA (seDNA). In several paediatric high-grade brain tumours, CSF contains sufficient amounts of tumour-derived cfDNA (cf-tDNA) for methylation- and CNV-based tumour classification through ligation-based nanopore sequencing [1], and for disease monitoring by various approaches [15]. Currently, these methods require laborious sample processing and expensive infrastructure.

Here, we have CSF-adapted our fast-track unsupervised machine learning (ML) approach [9] for cases with a differential diagnosis of lymphoma and other malignant brain tumours including metastases. We demonstrate its clinical application in two CNS-lymphoma cases. Comparison of nanopore sequencing-derived methylation patterns to pan-cancer epigenomic and CNV data allowed next-day diagnosis [9] and treatment initiation. Our point-of-care protocol can significantly reduce neurological impairment through timely and non-invasive testing. In short, we have adapted CSF preservation (Cell-Free DNA BCT CE, Streck, USA) and cfDNA extraction protocols (MagMAX™, Cat: A29319; KingFisher™ Cat: 5400110, Thermo Fisher, USA) to comply with a wide range of clinical scenarios (Fig. 1a); CSF seDNA was extracted as described (DNeasy kit ID: 69504; QiaCube, Qiagen, Germany) [9]. Sequencers were controlled, SQK-RBK004 kit (ONT, UK) was used, and data were analysed with NanoDiP as previously described [9] (Fig. 1a, open source code: see material availability). Patient characteristics are summarised in Table 1.Fig. 1 CSF analysis workflow and resulting CNV plots. a CSF drawn into a cfDNA preservation tube is centrifuged to separate sediment from supernatant. cfDNA and seDNA are extracted separately using commercial systems for cfDNA (supernatant) or tissue/cells (sediment). Photometric absorption measurement quantifies DNA. The fraction with the higher amount of DNA is subjected to NanoDiP analysis. Photograph demonstrates the small footprint of the NanoDiP device, consisting of a Jetson (Nvidia, USA) AGX Xavier computer (black box, top) to which an Mk1B (silver box, bottom) sequencer (ONT, UK) is connected through a USB3 cable. Sequencer control and all downstream data analysis occurs offline on the Jetson computer. Sequencing and reference methylation data are contained on the solid state drive attached to the Jetson computer (PCIe module on the left edge of the black box). b, c CNV plots, case 1. a: cfDNA, b: seDNA. Note the amplitude difference in both CNV profiles relative to the chromosome X vs. baseline difference. While cfDNA mainly consists of cf-tDNA, as confirmed independently by ddPCR, the seDNA does not contain significant amounts of tumour DNA. The seDNA likely stems from admixed leukocytes. Respective methylation profile had similarities to reactive lesions (not shown). d CNV plot from seDNA of case 2. Note the copy number alterations that indicate a clonal, neoplastic cell population. The amplitude of the aberrations is high relative to the chromosome X vs. baseline amplitude, suggesting a high contribution of neoplastic cell DNA to the seDNA

Table 1 Case summary

Case	Sex	Age [y]	Clin. differential	cfDNA [ng/µl]	seDNA [ng/µl]	Fraction analysed	Run [h]	Integrated diagnosis	Ancillary findings	Cytology	
1	m	68	Multifocal CNS lymphoma or glioma	 ~ 10	 ~ 4	cfDNA	17.5	DLBCL	MYD88: p.L265P, AF 95%	Negative	
2	m	48	NK/TCL recurrence in CNS	 ~ 4	 ~ 9	seDNA	16.5	NK/TCL	Clonal TCR rearr.	Sparse unclassifiable lymphoid cells, FACS inconclusive	
3	f	8	AT/RT recurrence	 ~ 4	n.d	cfDNA	17.6	AT/RT (recurrent)	-	2 × negative	
4	m	71	Unclear intracerebral mass lesion; lymphoma, glioma, other?	 ~ 3	 < 1	cfDNA	19.5	Inconclusive	Biopsy: glioblastoma, IDHwt	n.d	
5	w	72	Lymphoma vs. glioma	 ~ 2	 ~ 1	cfDNA	14.8	Inconclusive	Biopsy: glioblastoma, IDHwt	n.d	
6	m	67	Glioblastoma IDHwt 1.5y prior; recurrence vs. radionecrosis?	 ~ 6	 ~ 2	cfDNA	42.8	Inconclusive	Biopsy: glioblastoma, IDHwt	n.d	

Case 1

Biopsy of a small periventricular lesion, suspicious for CNS lymphoma, revealed reactive changes. Following negative CSF cytology, a second CSF spinal tap sample submitted as a “liquid biopsy” revealed elevated cfDNA in the supernatant. CNV (Fig. 1b) and DNAmeth profiles (100% match) of cfDNA completed the next day suggested diffuse large B-cell lymphoma, enabling initiation of immediate treatment. Droplet Digital PCR (ddPCR) for MYD88:p.L265P [6, 10] (Cat# WT:10042967/ MUT:10042964, BioRad, USA, manufacturer-supplied protocol) was positive (same cfDNA extract and day), independently supporting the diagnosis. Post hoc nanopore sequencing and ddPCR of seDNA (research purpose) showed a flat CNV, unspecific DNAmeth profiles (Fig. 1c), and negative MYD88-ddPCR.

Case 2

Suspected leptomeningeal relapse of an extranodal EBV-positive NK/T-cell lymphoma (NK/TCL) diagnosed 10 months ago. CSF cytology was negative. A subsequent CSF spinal tap “liquid biopsy” had higher seDNA than cfDNA; seDNA was sequenced. Due to a lack of T-cell lymphoma references in our pan-cancer database, the DNAmeth profile remained unclassifiable, whilst the CNV plot revealed alterations recurrently found in NK/TCL [12], warranting the integrated diagnosis of NK/TCL relapse (Fig. 1d). Blood analysis by a parallel sequencing panel revealed clonal T-cell receptor gamma rearrangement 9 days later (Cat: A51562, Thermo Fisher, USA).

Case 3

Child with atypical teratoid/rhabdoid tumour (AT/RT; CNS WHO grade 4) diagnosed 5.5 months ago. CSF monitoring cytology was negative twice. A third “liquid biopsy” CSF spinal tap displayed elevated cfDNA and nanopore sequencing revealed the DNAmeth profile of ATRT_TYR [4] as well as CNVs already established from the primary tumour tissue specimen. This case demonstrates the general applicability of our diagnostic CSF analysis workflow.

Cases 4, 5, 6

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.

Acknowledgements

We thank Dr Hasindu Gamaarachchi for his continued support concerning the methylation caller f5c: https://github.com/hasindu2008/f5c/releases/tag/v1.2. We acknowledge the excellent technical support by Bruno Grilli and Julian Pollinger.

Funding

Open access funding provided by University of Basel. Stiftung für krebskranke Kinder,Regio Basiliensis,2023-F006/007,Juergen Hench

Material availability

Due to changes in nanopore sequencing chemistry, we adapted methylation calling algorithms [9] to the current “RBK-114” rapid sequencing kit on MinION and PromethION (Oxford Nanopore Technologies, UK). Analysis tools were adapted to the ORIN Developer Kit (Nvidia, USA) [5, 7, 13]. Source: https://github.com/neuropathbasel/nanodip_dev

Data availability

Source code is available at: https://github.com/neuropathbasel/nanodip_dev

Declarations

Ethical approval

All patients agreed to the scientific workup of their cases (local ethics committee EKNZ Req-2019–00553). Methylation data was used according to EKNZ Req-2019–0123.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References

1. Afflerbach AK Rohrandt C Brändl B Sönksen M Hench J Frank S Classification of brain tumors by nanopore sequencing of cell-free DNA from cerebrospinal fluid Clin Chem 2023 10.1093/clinchem/hvad115
Afflerbach AK, Rohrandt C, Brändl B, Sönksen M, Hench J, Frank S et al (2023) Classification of brain tumors by nanopore sequencing of cell-free DNA from cerebrospinal fluid. Clin Chem. 10.1093/clinchem/hvad11510.1093/clinchem/hvad115
2. Baraniskin A Schroers R Liquid biopsy and other non-invasive diagnostic measures in PCNSL Cancers 2021 13 2665 10.3390/cancers13112665 34071407
Baraniskin A, Schroers R (2021) Liquid biopsy and other non-invasive diagnostic measures in PCNSL. Cancers 13:2665. 10.3390/cancers1311266534071407 10.3390/cancers13112665
3. Barkley AS Sullivan LT Gibson AW Camacho D Barber JK Ko AL Stereotactic brain biopsy hemorrhage risk factors and implications for postoperative care at a single institution: an argument for postoperative imaging World Neurosurg 2020 144 e807 e812 10.1016/j.wneu.2020.09.084 32956884
Barkley AS, Sullivan LT, Gibson AW, Camacho D, Barber JK, Ko AL et al (2020) Stereotactic brain biopsy hemorrhage risk factors and implications for postoperative care at a single institution: an argument for postoperative imaging. World Neurosurg 144:e807–e812. 10.1016/j.wneu.2020.09.08432956884 10.1016/j.wneu.2020.09.084
4. Capper D Jones DTW Sill M Hovestadt V Schrimpf D Sturm D DNA methylation-based classification of central nervous system tumours Nature 2018 555 469 474 10.1038/nature26000 29539639
Capper D, Jones DTW, Sill M, Hovestadt V, Schrimpf D, Sturm D et al (2018) DNA methylation-based classification of central nervous system tumours. Nature 555:469–474. 10.1038/nature2600029539639 10.1038/nature26000
5. Dong J Liu X Sadasivan H Sitaraman S Narayanasamy S mm2-gb: GPU accelerated minimap2 for long read dna mapping bioRxiv 2024 10.1101/2024.03.23.586366 39211171
Dong J, Liu X, Sadasivan H, Sitaraman S, Narayanasamy S (2024) mm2-gb: GPU accelerated minimap2 for long read dna mapping. bioRxiv. 10.1101/2024.03.23.58636639211171 10.1101/2024.03.23.586366
6. Ferreri AJM Calimeri T Cwynarski K Dietrich J Grommes C Hoang-Xuan K Primary central nervous system lymphoma Nat Rev Dis Primer 2023 9 29 10.1038/s41572-023-00439-0
Ferreri AJM, Calimeri T, Cwynarski K, Dietrich J, Grommes C, Hoang-Xuan K et al (2023) Primary central nervous system lymphoma. Nat Rev Dis Primer 9:29. 10.1038/s41572-023-00439-010.1038/s41572-023-00439-0
7. Gamaarachchi H Lam CW Jayatilaka G Samarakoon H Simpson JT Smith MA GPU accelerated adaptive banded event alignment for rapid comparative nanopore signal analysis BMC Bioinform 2020 10.1186/s12859-020-03697-x
Gamaarachchi H, Lam CW, Jayatilaka G, Samarakoon H, Simpson JT, Smith MA et al (2020) GPU accelerated adaptive banded event alignment for rapid comparative nanopore signal analysis. BMC Bioinform. 10.1186/s12859-020-03697-x10.1186/s12859-020-03697-x
8. Grommes C Tang SS Wolfe J Kaley TJ Daras M Pentsova EI Phase 1b trial of an ibrutinib-based combination therapy in recurrent/refractory CNS lymphoma Blood 2019 133 436 445 10.1182/blood-2018-09-875732 30567753
Grommes C, Tang SS, Wolfe J, Kaley TJ, Daras M, Pentsova EI et al (2019) Phase 1b trial of an ibrutinib-based combination therapy in recurrent/refractory CNS lymphoma. Blood 133:436–445. 10.1182/blood-2018-09-87573230567753 10.1182/blood-2018-09-875732
9. Hench J Hultschig C Brugger J Mariani L Guzman R Soleman J EpiDiP/NanoDiP: a versatile unsupervised machine learning edge computing platform for epigenomic tumour diagnostics Acta Neuropathol Commun 2024 12 51 10.1186/s40478-024-01759-2 38576030
Hench J, Hultschig C, Brugger J, Mariani L, Guzman R, Soleman J et al (2024) EpiDiP/NanoDiP: a versatile unsupervised machine learning edge computing platform for epigenomic tumour diagnostics. Acta Neuropathol Commun 12:51. 10.1186/s40478-024-01759-238576030 10.1186/s40478-024-01759-2
10. Montesinos-Rongen M Godlewska E Brunn A Wiestler OD Siebert R Deckert M Activating L265P mutations of the MYD88 gene are common in primary central nervous system lymphoma Acta Neuropathol (Berl) 2011 122 791 792 10.1007/s00401-011-0891-2 22020631
Montesinos-Rongen M, Godlewska E, Brunn A, Wiestler OD, Siebert R, Deckert M (2011) Activating L265P mutations of the MYD88 gene are common in primary central nervous system lymphoma. Acta Neuropathol (Berl) 122:791–792. 10.1007/s00401-011-0891-222020631 10.1007/s00401-011-0891-2
11. Mutter JA Alig SK Esfahani MS Lauer EM Mitschke J Kurtz DM Circulating tumor DNA profiling for detection, risk stratification, and classification of brain lymphomas J Clin Oncol 2023 41 1684 1694 10.1200/JCO.22.00826 36542815
Mutter JA, Alig SK, Esfahani MS, Lauer EM, Mitschke J, Kurtz DM et al (2023) Circulating tumor DNA profiling for detection, risk stratification, and classification of brain lymphomas. J Clin Oncol 41:1684–1694. 10.1200/JCO.22.0082636542815 10.1200/JCO.22.00826
12. Ng SB Chung TH Kato S Nakamura S Takahashi E Ko YH Epstein-Barr virus-associated primary nodal T/NK-cell lymphoma shows a distinct molecular signature and copy number changes Haematologica 2018 103 278 287 10.3324/haematol.2017.180430 29097495
Ng SB, Chung TH, Kato S, Nakamura S, Takahashi E, Ko YH et al (2018) Epstein-Barr virus-associated primary nodal T/NK-cell lymphoma shows a distinct molecular signature and copy number changes. Haematologica 103:278–287. 10.3324/haematol.2017.18043029097495 10.3324/haematol.2017.180430
13. Sadasivan H Maric M Dawson E Iyer V Israeli J Narayanasamy S Accelerating minimap2 for accurate long read alignment on GPUs J Biotechnol Biomed 2023 6 13 23 10.26502/jbb.2642-91280067 36937168
Sadasivan H, Maric M, Dawson E, Iyer V, Israeli J, Narayanasamy S (2023) Accelerating minimap2 for accurate long read alignment on GPUs. J Biotechnol Biomed 6:13–23. 10.26502/jbb.2642-9128006736937168 10.26502/jbb.2642-91280067
14. Schroers R Baraniskin A Heute C Vorgerd M Brunn A Kuhnhenn J Diagnosis of leptomeningeal disease in diffuse large B-cell lymphomas of the central nervous system by flow cytometry and cytopathology Eur J Haematol 2010 85 520 528 10.1111/j.1600-0609.2010.01516.x 20727005
Schroers R, Baraniskin A, Heute C, Vorgerd M, Brunn A, Kuhnhenn J et al (2010) Diagnosis of leptomeningeal disease in diffuse large B-cell lymphomas of the central nervous system by flow cytometry and cytopathology. Eur J Haematol 85:520–528. 10.1111/j.1600-0609.2010.01516.x20727005 10.1111/j.1600-0609.2010.01516.x
15. Wadden J Ravi K John V Babila CM Koschmann C Cell-free tumor DNA (cf-tDNA) liquid biopsy: current methods and use in brain tumor immunotherapy Front Immunol 2022 13 882452 10.3389/fimmu.2022.882452 35464472
Wadden J, Ravi K, John V, Babila CM, Koschmann C (2022) Cell-free tumor DNA (cf-tDNA) liquid biopsy: current methods and use in brain tumor immunotherapy. Front Immunol 13:882452. 10.3389/fimmu.2022.88245235464472 10.3389/fimmu.2022.882452
