Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Targeted gene expression profiling predicts meningioma outcomes and radiotherapy responses

Nature Medicine volume 29, pages 3067–3076 (2023)Cite this article

Subjects

Abstract

Surgery is the mainstay of treatment for meningioma, the most common primary intracranial tumor, but improvements in meningioma risk stratification are needed and indications for postoperative radiotherapy are controversial. Here we develop a targeted gene expression biomarker that predicts meningioma outcomes and radiotherapy responses. Using a discovery cohort of 173 meningiomas, we developed a 34-gene expression risk score and performed clinical and analytical validation of this biomarker on independent meningiomas from 12 institutions across 3 continents (N = 1,856), including 103 meningiomas from a prospective clinical trial. The gene expression biomarker improved discrimination of outcomes compared with all other systems tested (N = 9) in the clinical validation cohort for local recurrence (5-year area under the curve (AUC) 0.81) and overall survival (5-year AUC 0.80). The increase in AUC compared with the standard of care, World Health Organization 2021 grade, was 0.11 for local recurrence (95% confidence interval 0.07 to 0.17, P < 0.001). The gene expression biomarker identified meningiomas benefiting from postoperative radiotherapy (hazard ratio 0.54, 95% confidence interval 0.37 to 0.78, P = 0.0001) and suggested postoperative management could be refined for 29.8% of patients. In sum, our results identify a targeted gene expression biomarker that improves discrimination of meningioma outcomes, including prediction of postoperative radiotherapy responses.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Study design and gene expression biomarker characteristics.
Fig. 2: Gene expression biomarker discrimination of meningioma outcomes.
Fig. 3: Gene expression biomarker comparisons with other meningioma classification systems.
Fig. 4: Gene expression biomarker nomograms for meningioma outcomes.
Fig. 5: Gene expression biomarker prediction of meningioma radiotherapy responses and prognostic validation in samples from a prospective clinical trial.

Similar content being viewed by others

Data availability

Raw data from targeted gene expression panels are deposited in the NCBI Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE222054. Raw DNA methylation data from the UCSF WHO grade 2 or grade 3 validation cohort and the RTOG 0539 validation cohort are available under accession number GSE221029. Raw amplicon and targeted exome sequencing data from discovery and validation cohort meningiomas are deposited in the NCBI Sequencing Reads Archive (https://www.ncbi.nlm.nih.gov/sra) under project numbers PRJNA916225 and PRJNA916253. Matrices containing TPM data from RNA sequencing cohorts used for analytical validation are deposited along with code on GitHub (https://github.com/william-c-chen/Meningioma_GE_Biomarker). Accession numbers and publications containing previously reported data are available in Supplementary Table 4. The publicly available GRCh38 (hg38), CRCh37.p13 (hg19) and Kallisto index v10 datasets were used in this study. Source data are provided with this paper.

Code availability

Raw data from targeted gene expression panels are deposited in the NCBI Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE222054. Raw DNA methylation data from the UCSF WHO grade 2 or grade 3 validation cohort and the RTOG 0539 validation cohort are available under accession number GSE221029. Raw amplicon and targeted exome sequencing data from discovery and validation cohort meningiomas are deposited in the NCBI Sequencing Reads Archive (https://www.ncbi.nlm.nih.gov/sra) under project numbers PRJNA916225 and PRJNA916253. Matrices containing TPM data from RNA sequencing cohorts used for analytical validation are deposited along with code on github (https://github.com/william-c-chen/Meningioma_GE_Biomarker). Accession numbers and publications containing previously reported data are available in Supplementary Table 4. The publicly available GRCh38 (hg38), CRCh37.p13 (hg19), and Kallisto index v10 datasets were used in this study.

References

  1. Ostrom, Q. T. et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2015–2019. Neuro Oncol. 24, v1–v95 (2022).

    CAS  PubMed  PubMed Central  Google Scholar 

  2. Reihl, S. J. et al. A population study of clinical trial accrual for women and minorities in neuro-oncology following the NIH Revitalization Act. Neuro Oncol. 24, 1341–1349 (2022).

    PubMed  PubMed Central  Google Scholar 

  3. Brastianos, P. K. et al. Advances in multidisciplinary therapy for meningiomas. Neuro Oncol. 21, I18–I31 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  4. Goldbrunner, R. et al. EANO guideline on the diagnosis and management of meningiomas. Neuro Oncol. 23, 1821–1834 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. Louis, D. N. et al. The 2016 World Health Organization classification of tumors of the central nervous system: a summary. Acta Neuropathol. 131, 803–820 (2016).

  6. Chen, W. C. et al. Radiotherapy for meningiomas. J. Neuro Oncol. 160, 505–515 (2022).

    Google Scholar 

  7. Rogers, C. L. et al. Low-risk meningioma: initial outcomes from NRG Oncology/RTOG 0539. Neuro Oncol. 25, 137–145 (2023).

    CAS  PubMed  Google Scholar 

  8. Rogers, C. L. et al. High-risk meningioma: initial outcomes from NRG Oncology/RTOG 0539. Int. J. Radiat. Oncol. Biol. Phys. https://doi.org/10.1016/j.ijrobp.2019.11.028 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  9. Rogers, L. et al. Intermediate-risk meningioma: Initial outcomes from NRG Oncology RTOG 0539. J. Neurosurg. 129, 35–47 (2018).

    PubMed  Google Scholar 

  10. Weber, D. C. et al. Adjuvant postoperative high-dose radiotherapy for atypical and malignant meningioma: a phase-II parallel non-randomized and observation study (EORTC 22042-26042). Radiother. Oncol. https://doi.org/10.1016/j.radonc.2018.06.018 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  11. Louis, D. N. et al. The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol. 23, 1231–1251 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  12. Youngblood, M. W. et al. Correlations between genomic subgroup and clinical features in a cohort of more than 3000 meningiomas. J. Neurosurg. 133, 1345–1354 (2020).

    CAS  Google Scholar 

  13. Clark, V. E. et al. Genomic analysis of non-NF2 meningiomas reveals mutations in TRAF7, KLF4, AKT1, and SMO. Science 339, 1077–1080 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  14. Sahm, F. et al. TERT promoter mutations and risk of recurrence in meningioma. J. Natl Cancer Inst. 108, djv377 (2016).

    PubMed  Google Scholar 

  15. Guyot, A. et al. Analysis of CDKN2A gene alterations in recurrent and non-recurrent meningioma. J. Neuro Oncol. https://doi.org/10.1007/s11060-019-03333-6 (2019).

    Article  Google Scholar 

  16. Driver, J. et al. A molecularly integrated grade for meningioma. Neuro Oncol. 24, 796–808 (2022).

    CAS  PubMed  Google Scholar 

  17. Maas, S. L. N. et al. Integrated molecular-morphologic meningioma classification: a multicenter retrospective analysis, retrospectively and prospectively validated. J. Clin. Oncol. 39, 3839–3852 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  18. Nassiri, F. et al. A clinically applicable integrative molecular classification of meningiomas. Nature 597, 119–125 (2021).

    CAS  PubMed  Google Scholar 

  19. Patel, A. J. et al. Molecular profiling predicts meningioma recurrence and reveals loss of DREAM complex repression in aggressive tumors. Proc. Natl Acad. Sci. USA 116, 21715–21726 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  20. Vasudevan, H. N. et al. Comprehensive molecular profiling identifies FOXM1 as a key transcription factor for meningioma proliferation. Cell Rep. 22, 3672–3683 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  21. Choudhury, A. et al. Hypermitotic meningiomas harbor DNA methylation subgroups with distinct biological and clinical features. Neuro Oncol. 25, 520–530 (2023).

    CAS  PubMed  Google Scholar 

  22. Choudhury, A. et al. Meningioma DNA methylation groups identify biological drivers and therapeutic vulnerabilities. Nat. Genet. 54, 649–659 (2022).

    CAS  PubMed  PubMed Central  Google Scholar 

  23. Olar, A. et al. Global epigenetic profiling identifies methylation subgroups associated with recurrence-free survival in meningioma. Acta Neuropathol. 133, 431–444 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  24. Sahm, F. et al. DNA methylation-based classification and grading system for meningioma: a multicentre, retrospective analysis. Lancet Oncol. 18, 682–694 (2017).

    CAS  PubMed  Google Scholar 

  25. Sparano, J. A. et al. Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer. N. Engl. J. Med. 379, 111–121 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  26. Hu, J. C. et al. Clinical utility of gene expression classifiers in men with newly diagnosed prostate cancer. JCO Precis. Oncol. https://doi.org/10.1200/po.18.00163 (2018).

  27. Spratt, D. E. et al. Development and validation of a novel integrated clinical-genomic risk group classification for localized prostate cancer. J. Clin. Oncol. https://doi.org/10.1200/JCO.2017.74.2940 (2018).

    Article  PubMed  Google Scholar 

  28. Smith, J. C. & Sheltzer, J. M. Genome-wide identification and analysis of prognostic features in human cancers. Cell Rep. 38, 110569 (2022).

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Schaeffer, E. M. et al. NCCN Clinical Practice Guidelines in Oncology: Prostate Cancer Version 2.2022 (National Comprehensive Cancer Network, 2021).

  30. Gradishar, W. J. et al. Breast Cancer, Version 3.2022, NCCN Clinical Practice Guidelines in Oncology. J. Natl Compr. Cancer Netw. 20, 691–722 (2022).

    Google Scholar 

  31. Chen, W. C. et al. A prognostic gene-expression signature and risk score for meningioma recurrence after resection. Neurosurgery 88, 202–210 (2021).

    Google Scholar 

  32. Vuong, Q. H. Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica 57, 307 (1989).

    Google Scholar 

  33. Nabors, L. B. et al. Central nervous system cancers, version 3.2020. J. Natl Compr. Cancer Netw. 18, 1537–1570 (2020).

    Google Scholar 

  34. Nassiri, F. et al. DNA methylation profiling to predict recurrence risk in meningioma: development and validation of a nomogram to optimize clinical management. Neuro. Oncol. https://doi.org/10.1093/neuonc/noz061 (2019).

  35. Näslund, O, et al. Meningioma classification by immunohistochemistry: a replicability study. Brain and Spine https://doi.org/10.1016/j.bas.2022.101711 (2023).

  36. Bray, D. P. et al. Adjuvant radiotherapy in Grade II, atypical meningioma of the skull base. J. Neurol. Surg. B 83, 418–422 (2022).

    Google Scholar 

  37. Bray, D. P. et al. Adjuvant radiotherapy versus watchful waiting for world health organization grade II atypical meningioma: a single-institution experience. Neurosurgery 88, E435–E442 (2021).

    PubMed  Google Scholar 

  38. Byun, H. K. et al. Adjuvant radiotherapy versus surveillance for Grade 2 intracranial meningiomas: a multi-institutional propensity score-matched study. Front. Oncol. 12, 877244 (2022).

    CAS  PubMed  PubMed Central  Google Scholar 

  39. Chen, W. C. et al. Histopathological features predictive of local control of atypical meningioma after surgery and adjuvant radiotherapy. J. Neurosurg. https://doi.org/10.3171/2017.9.JNS171609 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  40. Dohm, A. et al. Early or late radiotherapy following gross or subtotal resection for atypical meningiomas: clinical outcomes and local control. J. Clin. Neurosci. 46, 90–98 (2017).

    PubMed  PubMed Central  Google Scholar 

  41. Kent, C. L. et al. Long-term outcomes for patients with atypical or malignant meningiomas treated with or without radiation therapy: a 25-year retrospective analysis of a single-institution experience. Adv. Radiat. Oncol. 7, 100878 (2022).

    PubMed  Google Scholar 

  42. Pan, P. C. et al. Outcomes following upfront radiation versus monitoring in atypical meningiomas: 16-year experience at a tertiary medical center. Neuro Oncol. Adv. 3, vdab094 (2021).

    Google Scholar 

  43. Rydzewski, N. R. et al. Gross total resection and adjuvant radiotherapy most significant predictors of improved survival in patients with atypical meningioma. Cancer 124, 734–742 (2018).

    PubMed  Google Scholar 

  44. Barrett, O. C. et al. Pathologic predictors of local recurrence in atypical meningiomas following gross total resection. Int. J. Radiat. Oncol. Biol. Phys. 103, 453–459 (2019).

    PubMed  Google Scholar 

  45. Graffeo, C. S. et al. Revisiting adjuvant radiotherapy after gross total resection of world health organization Grade II meningioma. World Neurosurg. 103, 655–663 (2017).

    PubMed  Google Scholar 

  46. Keric, N. et al. Impact of postoperative radiotherapy on recurrence of primary intracranial atypical meningiomas. J. Neuro Oncol. 146, 347–355 (2020).

    CAS  Google Scholar 

  47. Momin, A. A. et al. Adjuvant radiation versus observation with salvage radiation after gross-total resection of WHO grade II meningiomas: a propensity score-adjusted analysis. J. Neurosurg. https://doi.org/10.3171/2021.4.JNS21559 (2021).

  48. Rebchuk, A. D. et al. Survival and recurrence outcomes following adjuvant radiotherapy for grade 2 intracranial meningiomas: 13-year experience in a tertiary-care center. World Neurosurg. 161, e748–e756 (2022).

    PubMed  Google Scholar 

  49. Sun, S. Q. et al. Management of atypical cranial meningiomas, part 1: predictors of recurrence and the role of adjuvant radiation after gross total resection. Neurosurgery 75, 347–54 (2014). discussion 354-5; quiz 355.

    PubMed  Google Scholar 

  50. Wang, F. et al. Risk factors associated with postoperative recurrence in atypical intracranial meningioma: analysis of 263 cases at a single neurosurgical centre. Acta Neurochir. 161, 2563–2570 (2019).

    PubMed  Google Scholar 

  51. Zeng, Q. et al. Effectiveness of postoperative radiotherapy in patients with atypical meningiomas after gross total resection: analysis of 260 cases. World Neurosurg. 162, e580–e586 (2022).

    PubMed  Google Scholar 

  52. Preusser, M. et al. Trabectedin for recurrent WHO grade 2 or 3 meningioma: a randomized phase II study of the EORTC Brain Tumor Group (EORTC-1320-BTG). Neuro Oncol. 24, 755–767 (2022).

    CAS  PubMed  Google Scholar 

  53. Capper, D. et al. DNA methylation-based classification of central nervous system tumours. Nature 555, 469–474 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  54. Morin, O. et al. Integrated models incorporating radiologic and radiomic features predict meningioma grade, local failure, and overall survival. Neuro Oncol. Adv. 1, vdz011 (2019).

    Google Scholar 

  55. Magill, S. T. et al. Multiplatform genomic profiling and magnetic resonance imaging identify mechanisms underlying intratumor heterogeneity in meningioma. Nat. Commun. 11, 4803 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  56. Rogers, C. L. et al. Pathology concordance levels for meningioma classification and grading in NRG Oncology RTOG Trial 0539. Neuro Oncol. 18, 565–574 (2016).

    PubMed  Google Scholar 

  57. Geiss, G. K. et al. Direct multiplexed measurement of gene expression with color-coded probe pairs. Nat. Biotechnol. 26, 317–25 (2008).

    CAS  PubMed  Google Scholar 

  58. Breiman, L. Bagging predictors. Mach. Learn. https://doi.org/10.1007/BF00058655 (1996).

  59. Baik, B., Yoon, S. & Nam, D. Benchmarking RNA-seq differential expression analysis methods using spike-in and simulation data. PLoS ONE 15, e0232271 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  60. Bi, W. L. et al. Genomic landscape of high-grade meningiomas. NPJ Genom. Med. https://doi.org/10.1038/s41525-017-0014-7 (2017).

  61. Von Spreckelsen, N. et al. KLF4 K409Q-mutated meningiomas show enhanced hypoxia signaling and respond to mTORC1 inhibitor treatment. Acta Neuropathol. Commun. 8, 41 (2020).

    Google Scholar 

  62. Leek, J. T., Johnson, W. E., Parker, H. S., Jaffe, A. E. & Storey, J. D. The SVA package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28, 882–883 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

The authors thank J. de Groot, S. Hervey-Jumper, C. Park, S. Chang and N. Willmarth for providing comments, and F. Jiang for providing biostatistical feedback during the conception of this study. This study was supported by NIH grants P50 CA097257 (W.C.C., D.R.R. and M.S.B.), P50 CA221747 (K.M., J.P.C., C.M.H. and S.T.M.), U01 CA180868 and U10 CA180822 (M.Y.C.P., M.W., S.L.P. and M.P.M.), R01 CA262311 (D.R.R.), F32 CA213944 (S.T.M.), F30 CA246808 and T32 GM007618 (A.C.), a UCSF Catalyst Program Award (W.C.C. and D.R.R.), the Northwestern Medicine Malnati Brain Tumor Institute of the Lurie Cancer Center (K.M., J.P.C., C.M.H. and S.T.M.), the UCSF Wolfe Meningioma Program Project (D.R.R. and M.W.M.) and the Trenchard Family Charitable Fund (D.R.R.). This work was also supported by the Northwestern University NUSeq Core Facility and a Conquer Cancer Herman H. Freckman, MD, Endowed Young Investigator Award from the American Society of Clinical Oncology (W.C.C.), and by the Helen Diller Family Cancer Center Physician Scientist Program in Clinical Oncology program (W.C.C., 5K12 CA260225-03). W.C.C. is a Chan Zuckerberg Biohub San Francisco Physician-Scientist Fellow. Any opinions, findings and conclusions expressed in this material are those of the author(s) and do not necessarily reflect those of Conquer Cancer or the American Society of Clinical Oncology.

Author information

Author notes

  1. These authors contributed equally: Abrar Choudhury, Mark W. Youngblood, Mei-Yin C. Polley.

  2. These authors jointly supervised this work: Stephen T. Magill, David R. Raleigh.

Authors and Affiliations

  1. Department of Radiation Oncology, University of California San Francisco, San Francisco, CA, USA

    William C. Chen, Abrar Choudhury, Kanish Mirchia, Minh P. Nguyen, Harish N. Vasudevan, Jason W. Chan, Jessica Chew, Jacob S. Young, Lauren Boreta, Steve E. Braunstein, Penny K. Sneed & David R. Raleigh

  2. Department of Neurological Surgery, University of California San Francisco, San Francisco, CA, USA

    William C. Chen, Abrar Choudhury, Kanish Mirchia, Abigail K. Suwala, Minh P. Nguyen, Harish N. Vasudevan, Jacob S. Young, Nicholas Butowski, Nancy Ann Oberheim Bush, Mitchel S. Berger, Arie Perry & David R. Raleigh

  3. Department of Pathology, University of California San Francisco, San Francisco, CA, USA

    William C. Chen, Abrar Choudhury, Kanish Mirchia, Minh P. Nguyen, Jacob S. Young, David A. Solomon, Arie Perry & David R. Raleigh

  4. Medical Scientist Training Program, University of California San Francisco, San Francisco, CA, USA

    Abrar Choudhury

  5. Department of Neurological Surgery, Northwestern University, Chicago, IL, USA

    Mark W. Youngblood, Kathleen McCortney, James P. Chandler, Craig M. Horbinski & Stephen T. Magill

  6. NRG Statistics and Data Management Center, NRG Oncology, Philadelphia, PA, USA

    Mei-Yin C. Polley, Minhee Won, C. Leland Rogers, Stephanie L. Pugh & Minesh P. Mehta

  7. Department of Pathology, Johns Hopkins University, Baltimore, MD, USA

    Calixto-Hope G. Lucas

  8. Department of Pathology, Leiden University Medical Center, Leiden, the Netherlands

    Sybren L. N. Maas

  9. Department of Pathology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, the Netherlands

    Sybren L. N. Maas

  10. Department of Neuropathology, University Hospital Heidelberg and CCU Neuropathology, German Consortium for Translational Cancer Research, German Cancer Research Center, Heidelberg, Germany

    Abigail K. Suwala & Felix Sahm

  11. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, TX, USA

    James C. Bayley & Akash J. Patel

  12. Department of Neurosurgery, Baylor College of Medicine, Houston, TX, USA

    Akdes S. Harmanci

  13. Center for Secure Artificial Intelligence for Healthcare, Center for Precision Health, School of Biomedical Informatics, The University of Texas Health Science Center, Houston, TX, USA

    Arif O. Harmanci

  14. Department of Molecular and Human Genetics, Baylor College of Medicine, and Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, TX, USA

    Tiemo J. Klisch

  15. Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, USA

    Theresa J. Yu & Javier E. Villanueva-Meyer

  16. Department of Neurosurgery, Brigham and Women’s Hospital, and Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA

    Varun Bhave & Wenya Linda Bi

  17. Department of Clinical Oncology, The University of Hong Kong, Pokfulam, China

    Tai-Chung Lam

  18. Division of Neurosurgery, Department of Surgery, The University of Hong Kong, Pokfulam, China

    Jenny Kan-Suen Pu, Lai-Fung Li & Gilberto Ka-Kit Leung

  19. Department of Radiation Oncology, Ohio State University, Columbus, OH, USA

    Haley K. Perlow & Joshua D. Palmer

  20. Division of Neuropathology and Neurochemistry, Department of Neurology, Medical University of Vienna, Vienna, Austria

    Christine Haberler

  21. Division of Oncology, Department of Medicine, Medical University of Vienna, Vienna, Austria

    Anna S. Berghoff & Matthias Preusser

  22. Caris Life Sciences, Phoenix, AZ, USA

    Theodore P. Nicolaides & David Spetzler

  23. Department of Neuropathology, University of Magdeburg, Magdeburg, Germany

    Christian Mawrin

  24. Department of Neurological Surgery, University of Pittsburgh, Pittsburgh, PA, USA

    Sameer Agnihotri

  25. Department of Neurological Surgery, Children’s Hospital of Philadelphia, Philadelphia, PA, USA

    Adam Resnick

  26. Brain Tumor Institute, Children’s National Hospital, Washington, DC, USA

    Brian R. Rood

  27. Neurosciences Department, University of California San Diego, La Jolla, CA, USA

    Jessica Schulte

  28. Department of Neurology, University of California San Francisco, San Francisco, CA, USA

    Nicholas Butowski

  29. Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA

    Sandro Santagata

  30. Miami Neuroscience Institute, Baptist Health, Miami, FL, USA

    Minesh P. Mehta & Michael W. McDermott

  31. Department of Pathology, Northwestern University, Chicago, IL, USA

    Craig M. Horbinski

Authors
  1. William C. Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Abrar Choudhury
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Mark W. Youngblood
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Mei-Yin C. Polley
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Calixto-Hope G. Lucas
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Kanish Mirchia
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Sybren L. N. Maas
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Abigail K. Suwala
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Minhee Won
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. James C. Bayley
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Akdes S. Harmanci
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Arif O. Harmanci
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Tiemo J. Klisch
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Minh P. Nguyen
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Harish N. Vasudevan
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. Kathleen McCortney
    View author publications

    You can also search for this author in PubMed Google Scholar

  17. Theresa J. Yu
    View author publications

    You can also search for this author in PubMed Google Scholar

  18. Varun Bhave
    View author publications

    You can also search for this author in PubMed Google Scholar

  19. Tai-Chung Lam
    View author publications

    You can also search for this author in PubMed Google Scholar

  20. Jenny Kan-Suen Pu
    View author publications

    You can also search for this author in PubMed Google Scholar

  21. Lai-Fung Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  22. Gilberto Ka-Kit Leung
    View author publications

    You can also search for this author in PubMed Google Scholar

  23. Jason W. Chan
    View author publications

    You can also search for this author in PubMed Google Scholar

  24. Haley K. Perlow
    View author publications

    You can also search for this author in PubMed Google Scholar

  25. Joshua D. Palmer
    View author publications

    You can also search for this author in PubMed Google Scholar

  26. Christine Haberler
    View author publications

    You can also search for this author in PubMed Google Scholar

  27. Anna S. Berghoff
    View author publications

    You can also search for this author in PubMed Google Scholar

  28. Matthias Preusser
    View author publications

    You can also search for this author in PubMed Google Scholar

  29. Theodore P. Nicolaides
    View author publications

    You can also search for this author in PubMed Google Scholar

  30. Christian Mawrin
    View author publications

    You can also search for this author in PubMed Google Scholar

  31. Sameer Agnihotri
    View author publications

    You can also search for this author in PubMed Google Scholar

  32. Adam Resnick
    View author publications

    You can also search for this author in PubMed Google Scholar

  33. Brian R. Rood
    View author publications

    You can also search for this author in PubMed Google Scholar

  34. Jessica Chew
    View author publications

    You can also search for this author in PubMed Google Scholar

  35. Jacob S. Young
    View author publications

    You can also search for this author in PubMed Google Scholar

  36. Lauren Boreta
    View author publications

    You can also search for this author in PubMed Google Scholar

  37. Steve E. Braunstein
    View author publications

    You can also search for this author in PubMed Google Scholar

  38. Jessica Schulte
    View author publications

    You can also search for this author in PubMed Google Scholar

  39. Nicholas Butowski
    View author publications

    You can also search for this author in PubMed Google Scholar

  40. Sandro Santagata
    View author publications

    You can also search for this author in PubMed Google Scholar

  41. David Spetzler
    View author publications

    You can also search for this author in PubMed Google Scholar

  42. Nancy Ann Oberheim Bush
    View author publications

    You can also search for this author in PubMed Google Scholar

  43. Javier E. Villanueva-Meyer
    View author publications

    You can also search for this author in PubMed Google Scholar

  44. James P. Chandler
    View author publications

    You can also search for this author in PubMed Google Scholar

  45. David A. Solomon
    View author publications

    You can also search for this author in PubMed Google Scholar

  46. C. Leland Rogers
    View author publications

    You can also search for this author in PubMed Google Scholar

  47. Stephanie L. Pugh
    View author publications

    You can also search for this author in PubMed Google Scholar

  48. Minesh P. Mehta
    View author publications

    You can also search for this author in PubMed Google Scholar

  49. Penny K. Sneed
    View author publications

    You can also search for this author in PubMed Google Scholar

  50. Mitchel S. Berger
    View author publications

    You can also search for this author in PubMed Google Scholar

  51. Craig M. Horbinski
    View author publications

    You can also search for this author in PubMed Google Scholar

  52. Michael W. McDermott
    View author publications

    You can also search for this author in PubMed Google Scholar

  53. Arie Perry
    View author publications

    You can also search for this author in PubMed Google Scholar

  54. Wenya Linda Bi
    View author publications

    You can also search for this author in PubMed Google Scholar

  55. Akash J. Patel
    View author publications

    You can also search for this author in PubMed Google Scholar

  56. Felix Sahm
    View author publications

    You can also search for this author in PubMed Google Scholar

  57. Stephen T. Magill
    View author publications

    You can also search for this author in PubMed Google Scholar

  58. David R. Raleigh
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

A.C., M.W.Y. and M.Y.C.P. contributed equally as second authors and S.T.M. and D.R.R. contributed equally as last authors. All authors contributed to the acquisition, analysis or interpretation of data. W.C.C., M.Y.C.P., M.W., C.L.R., S.L.P., M.P.M. and D.R.R. performed or supervised statistical analyses. W.C.C., A.C., S.L.N.M., A.K.S., J.C.B., A.S.H., A.H., T.K., H.N.V., V.B., W.L.B., A.J.P., F.S. and D.R.R. performed or supervised bioinformatic analyses. W.C.C., C.H.G.L., K.M. and S.T.M. processed tumor samples and extracted DNA and RNA. W.C.C. and D.R.R. drafted the manuscript. All authors critically edited the manuscript and provided administrative, technical or material support. W.C.C. and D.R.R. conceived and designed the study. S.T.M. and D.R.R. supervised the study and contributed equally.

Corresponding authors

Correspondence to William C. Chen, Stephen T. Magill or David R. Raleigh.

Ethics declarations

Competing interests

M.P. has received honoraria for lectures, consultation or advisory board participation from the following for-profit companies: Bayer, Bristol Myers Squibb, Novartis, Gerson Lehrman Group (GLG), CMC Contrast, GlaxoSmithKline, Mundipharma, Roche, BMJ Journals, MedMedia, Astra Zeneca, AbbVie, Lilly, Medahead, Daiichi Sankyo, Sanofi, Merck Sharp & Dome, Tocagen, Adastra, and Gan & Lee Pharmaceuticals. W.C.C. and D.R.R. are the inventors on patent PCT/US 21/70288 describing the use of targeted gene expression profiling to predict meningioma outcomes and radiotherapy responses. The remaining authors declare no competing interests.

Peer review

Peer review information

Nature Medicine thanks Ingo Mellinghoff, Fenghai Duan, Adelheid Woehrer and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary handling editor: Ulrike Harjes, in collaboration with the Nature Medicine team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Prognostic and predictive gene expression biomarker development.

a, C-index for LFFR models based on meningioma Nanostring targeted gene expression profiling plotted against the natural-log of the lambda parameter during algorithm training. Center and error bars shows the mean C-index +/− estimated SEM. Candidate model performance was estimated using tenfold cross validation in the UCSF discovery cohort (N = 173). An optimal gene set (N = 34 genes, dotted lines, Supplementary Table 3) was selected within 1 standard error of the model achieving maximal c-index to reduce over-fitting. The number of genes in each model is displayed at the top of the graph. In order to further reduce over-fitting, improve calibration and stability, and facilitate risk score calculations using FFPE meningiomas or gene expression quantification from RNA sequencing or microarrays (Extended Data Fig. 4), bootstrap aggregation58 was used to train 500 ridge-regression sub-models using the normalized and log-transformed gene counts as inputs and discovery cohort risk scores between 0 and 1 as target variables. In brief, this procedure nominates a bootstrap aggregated risk score defined as the arithmetic mean across sub-model risk scores. b, Log2 values for Nanostring counts or RNA sequencing data (transcripts per million, TPM) for the 34 genes comprising the gene expression risk score in the frozen meningiomas from the UCSF discovery cohort (N = 173), revealing high concordance with R2 = 0.81 (two sided F-test P < 2.2 × 10−16). Similarly, a non-regularized Cox model for LFFR using RNA sequencing TPMs for the same 34 genes also achieved excellent performance in the UCSF discovery cohort (N = 173 meningiomas, LFFR c-index 0.89 ± 0.02, OS c-index 0.84 ± 0.02), and outperformed 10,000 randomly sampled sets of 34 genes (one-sided bootstrap P < 0.0001). c, Limitations to using RNA sequencing for targeted gene expression biomarker discovery, in comparison to using RNA sequencing for targeted gene expression biomarker validation or implementation (as described for b and as shown in Extended Data Fig. 4). The distribution of univariate LFFR Cox model two-sided unadjusted Wald test P-values for all RNA sequencing genes in the UCSF discovery cohort are shown (N = 58,830 genes, N = 173 meningiomas). A background uniform distribution is evident, with a peak towards lower P-values. Between P = 0.0 and 0.2, at least 6904 of 17437 P-values (40%) could be expected to be false positives rather than related to true biologicalfunction. Spike-in experiments and simulations report false discovery rates across bioinformatic methods and experimental conditions for observational whole transcriptomic approaches between 10-75%, depending on the ground-truth prevalence59. d, Log10 \(\beta\) coefficient magnitudes (red for positive coefficients, blue for negative coefficients) versus log2 transformed P-values from individual Cox models from the discovery cohort RNA sequencing are shown, demonstrating challenges with feature selection using sparse observations in high dimensional space. Despite the limitations in using RNA sequencing to discover clinical biomarkers, we show RNA sequencing can be used to validate and implement the 34-gene expression biomarker that was developed using Nanostring targeted gene expression profiling (Extended Data Fig. 4).

Source data

Extended Data Fig. 2 Gene expression biomarker characteristics in the discovery cohort.

a, Gaussian fits (left) to the distribution of gene expression risk scores in the UCSF discovery cohort, stratified by cases with local recurrence (red) or without local recurrence (blue) on clinical follow up. The gene expression risk score was also plotted against the time to censorship or local recurrence (right), and higher risk score correlated with greater risk of local recurrence and shorter time to recurrence. b, Kaplan Meier curves for LFFR or OS in the UCSF discovery cohort stratified by the gene expression risk score. c, Gene expression risk score distributions stratified by clinical characteristics in the UCSF discovery cohort. Mean +/- standard error measurements are shown for gene expression risk scores stratified by tumor location (skull base, N = 60; falx, N = 13; parasagittal, N = 34; convexity, N = 54; other, N = 10), extent of resection (GTR, N = 110; STR, N = 63), setting (primary, N = 144; recurrent, N = 29), WHO 2016 grade (grade 1, N = 83; grade 2, N = 66; grade 3, N = 24), and WHO 2021 grade (grade 1, N = 81; grade 2, N = 61; grade 3, N = 29). There was no significant difference across meningioma locations (ANOVA, two-sided unadjusted P = 0.22), but gene expression risk scores were higher among subtotally resected meningiomas (Student’s t-test, two-sided P = 0.02), recurrent meningiomas (Student’s t-test, two-sided P < 0.0001), and were stratified by WHO 2016 or 2021 grade (P < 0.0001). Convexity meningiomas arise adjacent to the cerebral convexity underlying the calvarium, while parasagittal meningiomas abut or involve the parasagittal sinus along the calvarial midline, falx meningiomas involve the falx without extending superiorly to the parasagittal sinus, and skull base meningiomas arise adjacent to the bones of the skull base. d, UCSF discovery cohort gene expression risk score scatter plots across clinical or molecular variables associated with meningioma biology or outcomes (blue, low risk; purple, intermediate risk; red, high risk). There was no clear association between patient age and gene expression risk score, but risk score was loosely correlated with MIB1 labeling index39, genomic instability as defined by the proportion of non-centromeric, non-acrocentric chromosomes affected by copy number gain or loss60, and DNA methylation of the CDKN2A locus21. P values shown are from a two-sided, unadjusted F-test. Thus, the gene expression biomarker correlated with surrogate markers of aggressive meningiomas. e, Disease specific survival among patients in the UCSF discovery cohort stratified by gene expression risk score.

Source data

Extended Data Fig. 3 Gene expression biomarker across somatic short variants in the discovery cohort.

Targeted DNA sequencing of recurrent somatic short variants was performed on 171 meningiomas from the UCSF discovery cohort (98.8%). a, Oncoplot distribution of identified pathogenic short somatic variants with variant allele frequency (VAF) of at least 5.0% (N = 98 variants, median VAF 38.0%, interquartile range [IQR] 29-43%, median sequencing depth 551.5, IQR 354-856). Consistent with prior reports, variants in NF2 were most common (N = 67, 39.2%), followed by TRAF7 (N = 10, 5.8%) and AKT1 (N = 8, 4.7%). A minority of meningiomas (N = 16, 9.4%) were identified without alteration of NF2 or loss of chromosome 22q, but with a characteristic pathogenic variant in one of the following genes: TRAF7, AKT1, PIK3CA, SMARCB1, SMARCE1, SMO, SUFU, KLF4, or POLR2A. The majority of these were WHO 2021 grade 1 meningiomas (N = 9, 60.0%), and were associated with favorable histologic characteristics and outcomes (median MIB1 labeling index 2.0%, range 0.5-4.0, 5-year LFFR 90.9%). TERT promoter C228T and C250T hotspot mutations were not identified in the discovery cohort. BAP1 mutations were rare (N = 5, 2.9%) and correlated with high histological grade and poor outcomes (N = 3 [60.0%] WHO 2016 grade 2 or 3, 5-year LFFR 40.0%). Homozygous CDKN2A/B loss, derived from meningioma DNA CNVs (Supplementary Methods), was identified in 10 meningiomas from the UCSF discovery cohort (5.8%, 80.0% WHO 2016 grade 3, 20.0% WHO 2016 grade 2) and was associated with poor outcomes (5-year LFFR 14.2%). These findings were supported by targeted DNA sequencing of recurrent somatic short variants in 35 consecutive clinical validation cohort meningiomas from The University of Hong Kong using the same approach. b, Same oncoplot from the UCSF discovery cohort as in a, but ordered by VAF instead of gene expression risk score.

Source data

Extended Data Fig. 4 Gene expression biomarker characteristics in the analytical validation cohort.

Gene expression risk score concordance across multiple conditions and replicates. Test-retest conditions (combined N = 44, R = 0.94, P < 0.0001) were comprised of varying probe batches (N = 10, R = 0.98, P < 0.0001), within probe batch testing of technical replicates (N = 12, R = 0.98, P < 0.0001), and test-retest conditions for meningiomas with serial RNA extraction on the same FFPE block or frozen tumor chunk at least 4 weeks apart (N = 22, R = 0.94, P < 0.001). Gene expression risk scores on paired frozen/FFPE meningiomas also demonstrated high concordance (N = 90, R = 0.88, P < 0.001), and FFPE gene expression risk scores provided excellent discrimination of outcomes across FFPE clinical validation datasets, including a prospective clinical trial (Figs. 2, 3, 5 and Extended Data Figs. 5, 6, 8). b, Principal component analysis on FFPE gene expression risk scores from meningiomas processed at multiple laboratories spanning academic institutions (Northwestern University, San Francisco Veterans Administration) or Clinical Laboratory Improvement Amendments (CLIA) certified private industry (Canopy Biosciences), demonstrating no laboratory batch effects. c, Publicly available microarray and clinical data were used to test the performance of the gene expression risk score on a non-Nanostring platform (N = 33 of 34 genes available). No paired microarray/Nanostring data was available to train a calibration model, which precluded direct comparison. Thus, the RNA sequencing calibration model described below (and described in further detail in the Supplementary Methods) was adapted to microarray data as an exploratory analysis, yielding prognostic risk groups as shown in the Kaplan Meier plot (P = 0.0014, Log-rank test). d, Concordance of gene expression risk scores derived from RNA sequencing or Nanostring targeted gene expression profiling on the same meningiomas (N = 469 meningiomas, R = 0.89, F-test two-sided unadjusted P < 0.0001). e, Distribution of gene expression risk scores (mean +/- SEM is shown) derived from RNA sequencing of cohorts overlapping (UCSF, The University of Hong Kong, N = 502) or non-overlapping (Caris Life Sciences, Heidelberg University, Brigham and Women’s Hospital, University Hospital Magdeburg, Children’s Brain Tumor Network, Baylor College of Medicine, N = 640) with the discovery or clinical validation cohorts, comprising 1142 unique meningiomas. Gene expression risk scores remained well distributed across all datasets, including RNA sequencing of pediatric meningiomas (Children’s Brain Tumor Network, N = 29), meningiomas with KLF4 or AKT1 somatic short variants61 (University Hospital Magdeburg, N = 31), or FFPE (N = 428) or frozen (N = 718) meningiomas, and demonstrated similar stratification by 2016 WHO histological grade as with Nanostring targeted gene expression profiling analyses (Extended Data Fig. 2c). f, Principal component analysis of gene expression risk scores across RNA sequencing cohorts after correction for batch effects using the COMBAT62 pipeline in the sva package in R. g, LFFR or OS stratified by gene expression risk scores from RNA sequencing of cohorts with available clinical data (UCSF discovery, The University of Hong Kong, and Baylor College of Medicine).

Source data

Extended Data Fig. 5 Gene expression biomarker characteristics in the clinical validation cohort.

A, Kaplan Meier curves for LFFR stratified by gene expression risk score (blue lines, low risk; purple lines, intermediate risk; red lines, high risk) for individual clinical validation cohorts, including The University of Hong Kong (Frozen N = 339; c-index=0.80; low risk N = 122, 5-year LFFR 95.1%; intermediate risk N = 151, 5-year LFFR 73.6%; high risk N = 66, 5-year LFFR 19.5%), Northwestern University (Frozen and FFPE N = 180; c-index=0.74; low risk N = 42, 5-year LFFR 90.0%; intermediate risk N = 98, 5-year LFFR 76.0%; high risk N = 42, 5-year LFFR 21.4%), UCSF WHO grade 2 or grade 3 (FFPE N = 158; c-index=0.78; low risk N = 24, 5-year LFFR 87.4%; intermediate risk N = 69, 5-year LFFR 77.5%; high risk N = 65, 5-year LFFR 22.0%), Baylor College of Medicine (Frozen N = 116; c-index=0.77; low risk N = 35, 5-year LFFR 90.0%; intermediate risk N = 61, 5-year LFFR 63.0%; high risk N = 20, 5-year LFFR 0.0%), and Heidelberg University plus the Medical University of Vienna (FFPE N = 61 with LFFR data; c-index=0.76; low risk N = 24, 5-year LFFR 80.4%; intermediate risk N = 23, 5-year LFFR 48.1%; high risk N = 14, 5-year LFFR 19.3%). The gene expression risk score remained well calibrated across multiple independent clinical validation cohorts comprising both frozen and FFPE meningiomas. When assessed separately within each independent retrospective cohort site, the gene expression risk score remained independently prognostic in multivariate analysis combining the risk score with WHO 2016 grade (P < 0.001 in all cases, two-sided unadjusted Wald test P-value). B, Kaplan Meier curves for LFFR in clinical validation cohort meningiomas stratified by gene expression risk score within WHO 2021 grades, demonstrating that the gene expression biomarker remained discriminatory across WHO 2021 grade 1 (low risk N = 114, intermediate risk N = 127, high risk N = 26), WHO 2021 grade 2 (low risk N = 7, intermediate risk N = 17, high risk N = 26), and WHO 2021 grade 3 meningiomas (low risk N = 2, intermediate risk N = 46, high risk N = 98). Shown are two-sided unadjusted Log-rank P-values. C, Forest plots of hazard ratios (HR) with 95% confidence intervals (CI) for local recurrence (left) or death (right) for each 0.1 increase in gene expression risk score are shown (center and error bars denote the hazard ratio and 95% confidence interval). The gene expression biomarker was prognostic across all molecular classification systems tested for both recurrence and survival. HRs according to gene expression risk score across meningioma settings, extent of resection (EOR), and WHO grades from Fig. 2b are re-presented for ease of comparison to HRs in molecular classification systems. P values shown are from two-sided unadjusted Wald’s tests. Overall, in the retrospective clinical validation cohort, LFFR Harrel’s c-index/Uno’s c-index for the gene expression biomarker was 0.78/0.77 (N = 854), while LFFR c-index was 0.68/0.66 for WHO 2016 grade (N = 854), 0.72/0.71 for WHO 2021 grade (N = 462), 0.72/0.73 for integrated score (N = 398), 0.73/0.73 for integrated grade (N = 460), 0.68/0.69 for DNA methylation groups (N = 460), 0.69/0.70 for DNA methylation subgroups (N = 460), 0.74/0.73 for DNA methylation probes (N = 455), and 0.70/0.71 for gene expression type (N = 389). OS Harrel’s c-index/Uno’s c-index for the gene expression biomarker was 0.78/0.78 (N = 863), while OS c-index was 0.72/0.72 for WHO 2016 grade (N = 863), 0.74/0.73 for WHO 2021 grade (N = 463), 0.73/0.72 for integrated score (N = 410), 0.75/0.75 for integrated grade (N = 460), 0.66/0.66 for DNA methylation groups (N = 460), 0.68/0.68 for DNA methylation subgroups (N = 460), 0.73/0.74 for DNA methylation probes (N = 455), and 0.70/0.67 for gene expression type (N = 386).

Source data

Extended Data Fig. 6 Molecular classification comparisons in the clinical validation cohort.

Kaplan Meier curves are shown for LFFR (shown are two-sided unadjusted Log-rank P-values) in clinical validation cohort meningiomas stratified by molecular risk groups (blue lines, low risk; purple lines, intermediate risk; red lines, high risk) using the gene expression biomarker in a, or 2 contemporary supervised meningioma classification systems based on combined molecular and clinical features: integrated grade16 based on CNVs and mitoses in b, or integrated score17 based on CNVs, DNA methylation families24, and WHO 2016 grade in c. In a, the gene expression biomarker remained robustly discriminatory across integrated grade or integrated score risk groups, concordant with the independent prognostic value of the gene expression risk score on multivariate analyses (Supplementary Tables 10, 11) and within groups from the 6 other molecular and/or histological classification systems tested (Fig. 3a). The converse was examined in b and c, where integrated grade was unable to discriminate outcomes across gene expression risk score groups, and integrated score had limited discriminatory power for intermediate and high gene expression risk score groups and was not discriminatory for low gene expression low risk groups.

Source data

Extended Data Fig. 7 Gene expression biomarker nomograms for meningioma outcomes.

a, Nomograms are shown for prediction of 5-year LFFR or OS based on gene expression risk score, extent of resection, setting, and WHO 2016 histologic grade. To use the nomograms, use a straight-edge to draw a vertical line between the variable of interest and the points scale at the top of the nomogram to determine the contribution in points to the total score for each variable. Add up the points from each variable, and then draw a vertical line from the total points scale at the bottom of the nomogram to the 5-year outcome scale to determine the estimated outcome. b, Calibration curves are shown for the models corresponding to the nomograms in Fig. 4a for LFFR (top) and OS (bottom) using the gene expression risk score, extent of resection, primary vs recurrent status, and WHO 2021 grade (and the addition of age for OS). The calibrate function from the rms package in R was used, with B = 1000 iterations and N = 75 samples per group. Center and error bars denote the predicted 5-year LFFR versus the observed 5-year LFFR calculated via the Kaplan Meier method, with a 95% confidence interval. c, Calibration curves corresponding to the nomograms in a, for LFFR (top) and OS (bottom) using 150 samples per group. Center and error bars denote the predicted 5-year LFFR versus the observed 5-year LFFR calculated via the Kaplan Meier method, with a 95% confidence interval. d, Time dependent AUC is shown for LFFR and OS for the retrospective clinical validation cohort (N = 866) as a function of time.

Source data

Extended Data Fig. 8 Molecular classification systems and response to treatment.

Kaplan Meier curves are shown for LFFR or OS (shown are two-sided unadjusted Log-rank P-values) in retrospective clinical validation cohort meningiomas or prospective RTOG 0539 meningiomas, stratified by gene expression risk score (blue lines, low risk; purple, intermediate risk lines; red lines, high risk), extent of resection, postoperative observation (Obs), or postoperative radiotherapy (RT). a, Primary retrospective clinical validation cohort meningiomas receiving postoperative external beam radiotherapy (N = 89) stratified by gene expression risk score (low risk N = 14, intermediate risk N = 45, high risk N = 30), revealing the gene expression biomarker remained prognostic among patients receiving postoperative radiotherapy. b, Primary WHO 2016 grade 2 meningiomas with GTR from the retrospective clinical validation cohort stratified by gene expression risk score (N = 21 low risk, N = 63 intermediate risk, N = 18 high risk). The gene expression risk score remained prognostic among gross totally resected primary WHO grade 2 meningiomas (N = 102, HR for local recurrence of 1.75 per 0.1 increase, 95% CI 1.18-2.59, P = 0.0057). c, Primary WHO grade 2 meningiomas with GTR from the retrospective clinical validation cohort stratified by postoperative radiotherapy (N = 28) or observation (N = 74). Patients with meningiomas meeting these criteria were eligible for 2 Phase III randomized multi-institutional trials (NRG BN003 and ROAM-EORTC 1308) examining clinical outcomes with postoperative radiotherapy versus observation. Postoperative radiotherapy did not offer a benefit to patients with meningiomas meeting these criteria in the retrospective clinical validation cohort. d, Retrospective clinical validation cohort meningiomas stratified by gene expression risk score across RTOG 0539 clinical risk groups (low clinical risk, primary WHO grade 1 meningiomas; intermediate clinical risk, recurrent WHO grade 1 meningiomas or primary WHO grade 2 meningiomas status post GTR; high clinical risk, recurrent or STR WHO grade 2 meningiomas or WHO grade 3 meningiomas after any resection). The gene expression biomarker remained prognostic across RTOG 0539 low clinical risk (gene expression risk score low risk N = 173, intermediate risk N = 224, high risk N = 27), RTOG 0539 intermediate clinical risk (gene expression risk score low risk N = 32, intermediate risk N = 80, high risk N = 38), and RTOG 0539 high clinical risk groups (gene expression risk score low risk N = 16, intermediate risk N = 75, high risk N = 128). e, Primary WHO grade 1 meningiomas from the retrospective clinical validation cohort (equivalent to RTOG 0539 low clinical risk meningiomas) stratified by gene expression risk score (N = 173 low risk, N = 224 intermediate risk, N = 27 high risk, 5-year LFFR 92.7%, 77.3%, and 43.0% for low, intermediate, or high risk meningiomas, respectively). f and g, Prospective validation cohort meningiomas from RTOG 0539 identified as low risk by the gene expression biomarker stratified by postoperative radiotherapy (N = 12 WHO 2016 grade 2 or 3 or recurrent WHO 2016 grade 1 meningiomas) or observation (N = 27 primary WHO 2016 grade 1 meningiomas). These analyses showed favorable outcomes for prospectively collected meningiomas with low gene expression risk scores across clinical risk strata, consistent with findings from retrospective clinical validation cohort meningiomas. More broadly, these data support the hypothesis that the gene expression biomarker may be useful for identifying meningiomas where postoperative radiotherapy could be safely omitted, even in the setting of conventionally high risk clinical features. h, Meningiomas treated with surgical monotherapy from the retrospective clinical validation cohort stratified by integrated score17 (the only contemporary molecular classification system potentially providing additional prognostic information for LFFR within gene expression biomarker strata, Extended Data Fig. 6c) and extent of resection. Favorable (light blue) and unfavorable (yellow) groups were identified using the same criteria for identification of biomarker/surgical strata (Fig. 5a). I, Favorable and unfavorable strata based on integrated score were unable to identify meningiomas benefitting from postoperative radiotherapy even after propensity matching on integrated score, extent of resection, and WHO 2016 grade. j, OS in the same meningiomas as Fig. 5c (that is propensity matched favorable and unfavorable retrospective clinical validation cohort meningiomas based on biomarker/surgical strata), demonstrating a trend towards benefit with postoperative radiotherapy for unfavorable meningiomas.

Source data

Extended Data Fig. 9 Biomarker risk score gene distribution in tumor versus microenvironment cell types from meningioma single-cell RNA sequencing.

a, Single-cell RNA sequencing uniform manifold approximation and projection (UMAP) of 57,114 transcriptomes from 8 human meningioma samples and 2 human dura samples shaded by cell clusters that were defined using cell signature gene sets, cell cycle analysis, and differentially expressed cluster marker genes, as previously reported22. Image is reproduced with permission. b, Feature plots showing normalized biomarker risk score gene expression across reduced dimensionality clusters of meningioma and tumor microenvironment cells. 33 of 34 biomarker genes were available for analysis in meningioma single-cell RNA sequencing data from a, although several were sparsely captured in single-cell RNA data (a known limitation that can be overcome using bulk RNA sequencing or Nanostring hybridization targeted gene expression profiling).

Source data

Supplementary information

Supplementary Information

Supplementary Methods, Tables 1–15 (plus legend for Supplementary Table 16, which is provide in .xlsx format separately as described below) and References.

Reporting Summary

Supplementary Table 1

Correlations between 100 meningioma genes in the discovery cohort. Data provided in excel format are Pearson correlations between normalized gene expression levels of the 100 meningioma genes from Supplementary Table 1.

Source data

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, W.C., Choudhury, A., Youngblood, M.W. et al. Targeted gene expression profiling predicts meningioma outcomes and radiotherapy responses. Nat Med 29, 3067–3076 (2023). https://doi.org/10.1038/s41591-023-02586-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41591-023-02586-z

This article is cited by

Search

Advanced search

Quick links

Nature Briefing: Cancer

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

Get what matters in cancer research, free to your inbox weekly. Sign up for Nature Briefing: Cancer