Skip to content
Licensed Unlicensed Requires Authentication Published by De Gruyter February 2, 2023

Accurate stratification between VEXAS syndrome and differential diagnoses by deep learning analysis of peripheral blood smears

  • Floris Chabrun , Valentin Lacombe , Xavier Dieu , Franck Geneviève and Geoffrey Urbanski ORCID logo EMAIL logo

Abstract

Objectives

VEXAS syndrome is a newly described autoinflammatory disease associated with UBA1 somatic mutations and vacuolization of myeloid precursors. This disease possesses an increasingly broad spectrum, leading to an increase in the number of suspected cases. Its diagnosis via bone-marrow aspiration and UBA1-gene sequencing is time-consuming and expensive. This study aimed at analyzing peripheral leukocytes using deep learning approaches to predict VEXAS syndrome in comparison to differential diagnoses.

Methods

We compared leukocyte images from blood smears of three groups: participants with VEXAS syndrome (identified UBA1 mutation) (VEXAS); participants with features strongly suggestive of VEXAS syndrome but without UBA1 mutation (UBA1-WT); participants with a myelodysplastic syndrome and without clinical suspicion of VEXAS syndrome (MDS). To compare images of circulating leukocytes, we applied a two-step procedure. First, we used self-supervised contrastive learning to train convolutional neural networks to translate leukocyte images into lower-dimensional encodings. Then, we employed support vector machine to predict patients’ condition based on those leukocyte encodings.

Results

The VEXAS, UBA1-WT, and MDS groups included 3, 3, and 6 patients respectively. Analysis of 33,757 images of neutrophils and monocytes enabled us to distinguish VEXAS patients from both UBA1-WT and MDS patients, with mean ROC-AUCs ranging from 0.87 to 0.95.

Conclusions

Image analysis of blood smears via deep learning accurately distinguished neutrophils and monocytes drawn from patients with VEXAS syndrome from those of patients with similar clinical and/or biological features but without UBA1 mutation. Our findings offer a promising pathway to better screening for this disease.


Corresponding author: Geoffrey Urbanski, MD, MPH, PhD, University of Angers, Inserm 1083, CNRS 6015, Mitolab Team, MITOVASC Institute, SFR ICAT, F-49000 Angers, France; Service de Médecine Interne et Immunologie Clinique, Centre Hospitalier Universitaire Angers, Angers, France; and Department of Orofacial Sciences, School of Dentistry, University of California, San Francisco, CA, USA, Phone: +33 2 41 35 40 03, +1-415/476-5976, E-mail:
Floris Chabrun and Valentin Lacombe contributed equally to this work.

Acknowledgments

This work was granted access to the HPC resources of IDRIS under the allocation 2022-AD011011303R2 made by GENCI. We are thankful to Dr. Marc Ferré for providing those resources, and to Mr. Samuel Ross Gilbert for proofreading the English version of this manuscript.

  1. Research funding: None declared.

  2. Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.

  3. Competing interests: Authors state no conflict of interest.

  4. Informed consent: Informed consent was obtained from all individuals included in this study.

  5. Ethical approval: This study was approved by the ethics committee of Angers University Hospital (#2022–094) and was conducted in compliance with the Declaration of Helsinki. All participants gave non-opposition informed consent. We applied the STARD (Standards for Reporting Diagnostic accuracy studies) recommendations.

  6. Data availability: Processed data (encodings) along with deep learning models are publicly available online at “https://github.com/fchabrun/VEXAS-BloodSmear”.

References

1. Beck, DB, Ferrada, MA, Sikora, KA, Ombrello, AK, Collins, JC, Pei, W, et al.. Somatic mutations in UBA1 and severe adult-onset autoinflammatory disease. N Engl J Med 2020;383:2628–38. https://doi.org/10.1056/nejmoa2026834.Search in Google Scholar

2. van der Made, CI, Potjewijd, J, Hoogstins, A, Willems, HPJ, Kwakernaak, AJ, de Sevaux, RGL, et al.. Adult-onset autoinflammation caused by somatic mutations in UBA1: a Dutch case series of patients with VEXAS. J Allergy Clin Immunol 2022;149:432–9.e4. https://doi.org/10.1016/j.jaci.2021.05.014.Search in Google Scholar PubMed

3. Georgin-Lavialle, S, Terrier, B, Guedon, AF, Heiblig, M, Comont, T, Lazaro, E, et al.. Further characterization of clinical and laboratory features in VEXAS syndrome: large-scale analysis of a multicentre case series of 116 French patients. Br J Dermatol 2022;186:564–74. https://doi.org/10.1111/bjd.20805.Search in Google Scholar PubMed

4. Lacombe, V, Kosmider, O, Prévost, M, Lavigne, C, Urbanski, G. Severe joint involvement in VEXAS syndrome: a case report. Ann Intern Med 2021;174:1025–7. https://doi.org/10.7326/l21-0023.Search in Google Scholar

5. Beaumesnil, S, Boucher, S, Lavigne, C, Urbanski, G, Lacombe, V. Ear, nose, throat, and bronchial involvements in VEXAS syndrome: specifying the spectrum of clinical features. JAMA Otolaryngol Head Neck Surg 2022;148:284. https://doi.org/10.1001/jamaoto.2021.4092.Search in Google Scholar PubMed

6. Koster, MJ, Kourelis, T, Reichard, KK, Kermani, TA, Beck, DB, Cardona, DO, et al.. Clinical heterogeneity of the VEXAS syndrome. Mayo Clin Proc 2021;96:2653–9. https://doi.org/10.1016/j.mayocp.2021.06.006.Search in Google Scholar PubMed

7. Poulter, JA, Collins, JC, Cargo, C, De Tute, RM, Evans, P, Ospina Cardona, D, et al.. Novel somatic mutations in UBA1 as a cause of VEXAS syndrome. Blood 2021;137:3676–81. https://doi.org/10.1182/blood.2020010286.Search in Google Scholar PubMed PubMed Central

8. Templé, M, Duroyon, E, Croizier, C, Rossignol, J, Huet, T, Friedrich, C, et al.. Atypical splice-site mutations causing VEXAS syndrome. Rheumatology 2021;60:e435–7. https://doi.org/10.1093/rheumatology/keab524.Search in Google Scholar PubMed

9. Lacombe, V, Prevost, M, Bouvier, A, Thépot, S, Chabrun, F, Kosmider, O, et al.. Vacuoles in neutrophil precursors in VEXAS syndrome: diagnostic performances and threshold. Br J Haematol 2021;195:286–9. https://doi.org/10.1111/bjh.17679.Search in Google Scholar PubMed

10. Moen, E, Bannon, D, Kudo, T, Graf, W, Covert, M, Van Valen, D. Deep learning for cellular image analysis. Nat Methods 2019;16:1233–46. https://doi.org/10.1038/s41592-019-0403-1.Search in Google Scholar PubMed PubMed Central

11. Dwibedi, D, Aytar, Y, Tompson, J, Sermanet, P, Zisserman, A. With a little help from my friends: nearest-neighbor contrastive learning of visual representations. In: 2021 IEEE/CVF international conference on computer vision (ICCV); 2021:9568–77 pp.10.1109/ICCV48922.2021.00945Search in Google Scholar

12. Hadsell, R, Chopra, S, LeCun, Y. Dimensionality reduction by learning an invariant mapping. In: 2006 IEEE computer society conference on computer vision and pattern recognition (CVPR’06); 2006:1735–42 pp.Search in Google Scholar

13. McInnes, L, Healy, J, Saul, N, Grossberger, L. UMAP: Uniform Manifold approximation and projection. J Open Source Software 2018;3:861. https://doi.org/10.21105/joss.00861.Search in Google Scholar

14. Gulati, G, Song, J, Florea, AD, Gong, J. Purpose and criteria for blood smear scan, blood smear examination, and blood smear review. Ann Lab Med 2013;33:1–7. https://doi.org/10.3343/alm.2013.33.1.1.Search in Google Scholar PubMed PubMed Central

15. D’Angelo, G. Hematopoietic cells vacuolation, not always a reactive event. The VEXAS syndrome. Int J Lab Hematol 2023;45:e15–6. https://doi.org/10.1111/ijlh.13955.Search in Google Scholar PubMed


Supplementary Material

This article contains supplementary material (https://doi.org/10.1515/cclm-2022-1283).


Received: 2022-12-18
Accepted: 2023-01-17
Published Online: 2023-02-02
Published in Print: 2023-06-27

© 2023 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 5.3.2024 from https://www.degruyter.com/document/doi/10.1515/cclm-2022-1283/html
Scroll to top button