An 8-gene machine learning model improves clinical prediction of severe dengue progression.
Biomarkers
Dengue
Gene signature
Host response
Machine learning
Prognostic
Severe dengue
Journal
Genome medicine
ISSN: 1756-994X
Titre abrégé: Genome Med
Pays: England
ID NLM: 101475844
Informations de publication
Date de publication:
29 03 2022
29 03 2022
Historique:
received:
23
10
2021
accepted:
24
02
2022
entrez:
29
3
2022
pubmed:
30
3
2022
medline:
5
4
2022
Statut:
epublish
Résumé
Each year 3-6 million people develop life-threatening severe dengue (SD). Clinical warning signs for SD manifest late in the disease course and are nonspecific, leading to missed cases and excess hospital burden. Better SD prognostics are urgently needed. We integrated 11 public datasets profiling the blood transcriptome of 365 dengue patients of all ages and from seven countries, encompassing biological, clinical, and technical heterogeneity. We performed an iterative multi-cohort analysis to identify differentially expressed genes (DEGs) between non-severe patients and SD progressors. Using only these DEGs, we trained an XGBoost machine learning model on public data to predict progression to SD. All model parameters were "locked" prior to validation in an independent, prospectively enrolled cohort of 377 dengue patients in Colombia. We measured expression of the DEGs in whole blood samples collected upon presentation, prior to SD progression. We then compared the accuracy of the locked XGBoost model and clinical warning signs in predicting SD. We identified eight SD-associated DEGs in the public datasets and built an 8-gene XGBoost model that accurately predicted SD progression in the independent validation cohort with 86.4% (95% CI 68.2-100) sensitivity and 79.7% (95% CI 75.5-83.9) specificity. Given the 5.8% proportion of SD cases in this cohort, the 8-gene model had a positive and negative predictive value (PPV and NPV) of 20.9% (95% CI 16.7-25.6) and 99.0% (95% CI 97.7-100.0), respectively. Compared to clinical warning signs at presentation, which had 77.3% (95% CI 58.3-94.1) sensitivity and 39.7% (95% CI 34.7-44.9) specificity, the 8-gene model led to an 80% reduction in the number needed to predict (NNP) from 25.4 to 5.0. Importantly, the 8-gene model accurately predicted subsequent SD in the first three days post-fever onset and up to three days prior to SD progression. The 8-gene XGBoost model, trained on heterogeneous public datasets, accurately predicted progression to SD in a large, independent, prospective cohort, including during the early febrile stage when SD prediction remains clinically difficult. The model has potential to be translated to a point-of-care prognostic assay to reduce dengue morbidity and mortality without overwhelming limited healthcare resources.
Sections du résumé
BACKGROUND
Each year 3-6 million people develop life-threatening severe dengue (SD). Clinical warning signs for SD manifest late in the disease course and are nonspecific, leading to missed cases and excess hospital burden. Better SD prognostics are urgently needed.
METHODS
We integrated 11 public datasets profiling the blood transcriptome of 365 dengue patients of all ages and from seven countries, encompassing biological, clinical, and technical heterogeneity. We performed an iterative multi-cohort analysis to identify differentially expressed genes (DEGs) between non-severe patients and SD progressors. Using only these DEGs, we trained an XGBoost machine learning model on public data to predict progression to SD. All model parameters were "locked" prior to validation in an independent, prospectively enrolled cohort of 377 dengue patients in Colombia. We measured expression of the DEGs in whole blood samples collected upon presentation, prior to SD progression. We then compared the accuracy of the locked XGBoost model and clinical warning signs in predicting SD.
RESULTS
We identified eight SD-associated DEGs in the public datasets and built an 8-gene XGBoost model that accurately predicted SD progression in the independent validation cohort with 86.4% (95% CI 68.2-100) sensitivity and 79.7% (95% CI 75.5-83.9) specificity. Given the 5.8% proportion of SD cases in this cohort, the 8-gene model had a positive and negative predictive value (PPV and NPV) of 20.9% (95% CI 16.7-25.6) and 99.0% (95% CI 97.7-100.0), respectively. Compared to clinical warning signs at presentation, which had 77.3% (95% CI 58.3-94.1) sensitivity and 39.7% (95% CI 34.7-44.9) specificity, the 8-gene model led to an 80% reduction in the number needed to predict (NNP) from 25.4 to 5.0. Importantly, the 8-gene model accurately predicted subsequent SD in the first three days post-fever onset and up to three days prior to SD progression.
CONCLUSIONS
The 8-gene XGBoost model, trained on heterogeneous public datasets, accurately predicted progression to SD in a large, independent, prospective cohort, including during the early febrile stage when SD prediction remains clinically difficult. The model has potential to be translated to a point-of-care prognostic assay to reduce dengue morbidity and mortality without overwhelming limited healthcare resources.
Identifiants
pubmed: 35346346
doi: 10.1186/s13073-022-01034-w
pii: 10.1186/s13073-022-01034-w
pmc: PMC8959795
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Research Support, U.S. Gov't, Non-P.H.S.
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
33Subventions
Organisme : NIAID NIH HHS
ID : R01 AI125197
Pays : United States
Organisme : NIAID NIH HHS
ID : U19 AI057229
Pays : United States
Organisme : NIAID NIH HHS
ID : U19 AI109662
Pays : United States
Organisme : National Institute of Allergy and Infectious Diseases
ID : U19AI057229
Organisme : NIAID NIH HHS
ID : U19 AI167903
Pays : United States
Informations de copyright
© 2022. The Author(s).
Références
Southeast Asian J Trop Med Public Health. 2008 Mar;39(2):252-9
pubmed: 18564710
Am J Respir Crit Care Med. 2016 Nov 1;194(9):1104-1115
pubmed: 27135599
Sci Rep. 2021 Mar 10;11(1):5554
pubmed: 33692368
J Clin Virol. 2007 Jun;39(2):76-81
pubmed: 17507286
Lancet Infect Dis. 2016 Aug;16(8):935-41
pubmed: 27091092
Sci Transl Med. 2015 May 13;7(287):287ra71
pubmed: 25972003
Epidemiol Perspect Innov. 2006 Oct 05;3:11
pubmed: 17022816
PLoS One. 2010 Jul 20;5(7):e11671
pubmed: 20652028
PLoS Negl Trop Dis. 2020 Dec 23;14(12):e0008960
pubmed: 33362244
Trop Med Int Health. 2011 Aug;16(8):936-48
pubmed: 21624014
Nat Commun. 2019 Jul 31;10(1):3422
pubmed: 31366921
Sci Rep. 2021 Apr 30;11(1):9387
pubmed: 33931684
Lancet Infect Dis. 2021 Jul;21(7):1014-1026
pubmed: 33640077
Biostatistics. 2007 Jan;8(1):118-27
pubmed: 16632515
PLoS Med. 2019 Apr 23;16(4):e1002786
pubmed: 31013272
J Infect Dis. 2018 May 5;217(11):1690-1698
pubmed: 29490079
Nucleic Acids Res. 2017 Jan 9;45(1):e1
pubmed: 27634930
Cancer. 1950 Jan;3(1):32-5
pubmed: 15405679
Trends Microbiol. 2020 Jun;28(6):478-486
pubmed: 31982232
J Virol. 2010 Dec;84(24):12982-94
pubmed: 20943967
Pac Symp Biocomput. 2016;22:144-153
pubmed: 27896970
BMJ. 1995 Feb 18;310(6977):452-4
pubmed: 7873954
IEEE Trans Biomed Eng. 2019 Oct;66(10):2861-2868
pubmed: 30716030
PLoS Negl Trop Dis. 2010 Jun 15;4(6):e710
pubmed: 20559541
Am J Trop Med Hyg. 2010 Sep;83(3):690-5
pubmed: 20810840
PLoS One. 2014 Apr 28;9(4):e96314
pubmed: 24777054
J Innate Immun. 2010;2(5):478-82
pubmed: 20571257
J Infect Dis. 2009 Feb 15;199(4):537-546
pubmed: 19138155
J Infect Dev Ctries. 2020 Jul 31;14(7):781-787
pubmed: 32794470
PLoS Negl Trop Dis. 2018 Feb 9;12(2):e0006258
pubmed: 29425194
BMC Infect Dis. 2013 Oct 24;13:498
pubmed: 24152678
Mol Syst Biol. 2018 Aug 27;14(8):e7862
pubmed: 30150281
Genes (Basel). 2019 Jan 28;10(2):
pubmed: 30696086
Int J Exp Pathol. 1999 Jun;80(3):143-9
pubmed: 10469270
Brief Bioinform. 2013 Jul;14(4):469-90
pubmed: 22851511
J Infect Dis. 2007 Apr 15;195(8):1097-107
pubmed: 17357045
Cell Host Microbe. 2014 Jul 9;16(1):115-27
pubmed: 24981333
Jpn J Infect Dis. 2018 May 24;71(3):239-243
pubmed: 29709965
PLoS Negl Trop Dis. 2020 Feb 14;14(2):e0007969
pubmed: 32059026
BMC Bioinformatics. 2011 Mar 17;12:77
pubmed: 21414208
Transl Res. 2017 Aug;186:62-78.e9
pubmed: 28683259
BMC Infect Dis. 2011 Apr 21;11:106
pubmed: 21510901
PLoS Negl Trop Dis. 2018 Oct 4;12(10):e0006497
pubmed: 30286085
Nat Commun. 2020 Mar 4;11(1):1177
pubmed: 32132525
Gigascience. 2020 Nov 3;9(11):
pubmed: 33140829
PLoS Negl Trop Dis. 2012;6(12):e1966
pubmed: 23285306
Med J Malaysia. 2020 Nov;75(6):635-641
pubmed: 33219170
Stat Med. 1997 Jul 15;16(13):1529-42
pubmed: 9249923
Lancet Infect Dis. 2016 Jun;16(6):712-723
pubmed: 26874619
PLoS Negl Trop Dis. 2013 Jul 11;7(7):e2298
pubmed: 23875036
PLoS Negl Trop Dis. 2020 Oct 13;14(10):e0008567
pubmed: 33048921
Pediatr Infect Dis J. 2002 Jan;21(1):34-8
pubmed: 11791096
Sci Transl Med. 2017 Aug 30;9(405):
pubmed: 28855396
Lancet Respir Med. 2016 Mar;4(3):213-24
pubmed: 26907218
J Med Assoc Thai. 2011 Aug;94 Suppl 3:S74-84
pubmed: 22043757
Immunity. 2021 Apr 13;54(4):753-768.e5
pubmed: 33765435
Cell Rep. 2019 Jan 29;26(5):1104-1111.e4
pubmed: 30699342
Am J Trop Med Hyg. 2012 Jan;86(1):166-70
pubmed: 22232468
Dement Geriatr Cogn Dis Extra. 2018 Sep 25;8(3):321-327
pubmed: 30386368
Proc Natl Acad Sci U S A. 2020 Sep 15;117(37):22727-22735
pubmed: 32868442
Cell Host Microbe. 2020 Jun 10;27(6):992-1000.e3
pubmed: 32320677
BMC Med Genomics. 2020 Feb 17;13(1):28
pubmed: 32066441
Biometrics. 1988 Sep;44(3):837-45
pubmed: 3203132
Trans R Soc Trop Med Hyg. 2009 Sep;103(9):871-7
pubmed: 19500813
Stat Methods Med Res. 2017 Aug;26(4):1936-1948
pubmed: 26152746
Immunity. 2015 Dec 15;43(6):1199-211
pubmed: 26682989
PLoS One. 2009 Nov 19;4(11):e7892
pubmed: 19936257