The combination of supervised and unsupervised learning based risk stratification and phenotyping in pulmonary arterial hypertension-a long-term retrospective multicenter trial.
Atypical pulmonary arterial hypertension
Biomarkers
Mortality
Pulmonary arterial hypertension
Right-heart failure
Risk assessment
Journal
BMC pulmonary medicine
ISSN: 1471-2466
Titre abrégé: BMC Pulm Med
Pays: England
ID NLM: 100968563
Informations de publication
Date de publication:
25 Apr 2023
25 Apr 2023
Historique:
received:
11
10
2022
accepted:
06
04
2023
medline:
27
4
2023
pubmed:
26
4
2023
entrez:
25
4
2023
Statut:
epublish
Résumé
Accurate risk stratification in pulmonary arterial hypertension (PAH), a devastating cardiopulmonary disease, is essential to guide successful therapy. Machine learning may improve risk management and harness clinical variability in PAH. We conducted a long-term retrospective observational study (median follow-up: 67 months) including 183 PAH patients from three Austrian PAH expert centers. Clinical, cardiopulmonary function, laboratory, imaging, and hemodynamic parameters were assessed. Cox proportional hazard Elastic Net and partitioning around medoid clustering were applied to establish a multi-parameter PAH mortality risk signature and investigate PAH phenotypes. Seven parameters identified by Elastic Net modeling, namely age, six-minute walking distance, red blood cell distribution width, cardiac index, pulmonary vascular resistance, N-terminal pro-brain natriuretic peptide and right atrial area, constituted a highly predictive mortality risk signature (training cohort: concordance index = 0.82 [95%CI: 0.75 - 0.89], test cohort: 0.77 [0.66 - 0.88]). The Elastic Net signature demonstrated superior prognostic accuracy as compared with five established risk scores. The signature factors defined two clusters of PAH patients with distinct risk profiles. The high-risk/poor prognosis cluster was characterized by advanced age at diagnosis, poor cardiac output, increased red cell distribution width, higher pulmonary vascular resistance, and a poor six-minute walking test performance. Supervised and unsupervised learning algorithms such as Elastic Net regression and medoid clustering are powerful tools for automated mortality risk prediction and clinical phenotyping in PAH.
Sections du résumé
BACKGROUND
BACKGROUND
Accurate risk stratification in pulmonary arterial hypertension (PAH), a devastating cardiopulmonary disease, is essential to guide successful therapy. Machine learning may improve risk management and harness clinical variability in PAH.
METHODS
METHODS
We conducted a long-term retrospective observational study (median follow-up: 67 months) including 183 PAH patients from three Austrian PAH expert centers. Clinical, cardiopulmonary function, laboratory, imaging, and hemodynamic parameters were assessed. Cox proportional hazard Elastic Net and partitioning around medoid clustering were applied to establish a multi-parameter PAH mortality risk signature and investigate PAH phenotypes.
RESULTS
RESULTS
Seven parameters identified by Elastic Net modeling, namely age, six-minute walking distance, red blood cell distribution width, cardiac index, pulmonary vascular resistance, N-terminal pro-brain natriuretic peptide and right atrial area, constituted a highly predictive mortality risk signature (training cohort: concordance index = 0.82 [95%CI: 0.75 - 0.89], test cohort: 0.77 [0.66 - 0.88]). The Elastic Net signature demonstrated superior prognostic accuracy as compared with five established risk scores. The signature factors defined two clusters of PAH patients with distinct risk profiles. The high-risk/poor prognosis cluster was characterized by advanced age at diagnosis, poor cardiac output, increased red cell distribution width, higher pulmonary vascular resistance, and a poor six-minute walking test performance.
CONCLUSION
CONCLUSIONS
Supervised and unsupervised learning algorithms such as Elastic Net regression and medoid clustering are powerful tools for automated mortality risk prediction and clinical phenotyping in PAH.
Identifiants
pubmed: 37098543
doi: 10.1186/s12890-023-02427-2
pii: 10.1186/s12890-023-02427-2
pmc: PMC10131314
doi:
Types de publication
Observational Study
Multicenter Study
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
143Subventions
Organisme : Oesterreichische Nationalbank
ID : Project 17271
Informations de copyright
© 2023. The Author(s).
Références
J Stat Softw. 2010;33(1):1-22
pubmed: 20808728
BMC Med Res Methodol. 2013 Mar 06;13:33
pubmed: 23496923
Eur Respir J. 2019 Jan 24;53(1):
pubmed: 30545968
Heart. 2011 Jul;97(13):1054-60
pubmed: 21558476
Eur Heart J. 2018 Dec 14;39(47):4175-4181
pubmed: 28575277
Eur Respir J. 2010 May;35(5):1079-87
pubmed: 20032020
Stat Med. 1996 Feb 28;15(4):361-87
pubmed: 8668867
Ann Intern Med. 1991 Sep 1;115(5):343-9
pubmed: 1863023
Eur Respir J. 2023 Jan 6;61(1):
pubmed: 36028254
Am J Respir Crit Care Med. 2022 May 1;205(9):1102-1111
pubmed: 35081018
Eur Respir J. 2019 Jan 24;53(1):
pubmed: 30545971
Crit Rev Clin Lab Sci. 2015;52(2):86-105
pubmed: 25535770
J Stat Softw. 2011 Mar;39(5):1-13
pubmed: 27065756
Eur Respir J. 2017 Aug 3;50(2):
pubmed: 28775050
Eur Respir J. 2010 Sep;36(3):549-55
pubmed: 20562126
Eur Respir J. 2021 Apr 8;57(4):
pubmed: 33334941
Am Heart J. 2019 Feb;208:91-99
pubmed: 30580131
Eur Respir J. 2017 Aug 3;50(2):
pubmed: 28775047
Brief Bioinform. 2011 May;12(3):203-14
pubmed: 21324971
J Heart Lung Transplant. 2020 Dec;39(12):1435-1444
pubmed: 33082079
Eur Respir J. 2020 Aug 27;56(2):
pubmed: 32366491
Neural Comput. 2004 Jun;16(6):1299-323
pubmed: 15130251
Stat Med. 1999 Sep 15-30;18(17-18):2529-45
pubmed: 10474158
PLoS One. 2018 Aug 30;13(8):e0203396
pubmed: 30161261
Int J Cardiol. 2018 Dec 1;272S:37-45
pubmed: 30190158
Chest. 2013 Jul;144(1):274-283
pubmed: 23880678
PLoS One. 2019 Oct 25;14(10):e0224453
pubmed: 31652290
Chest. 2012 Feb;141(2):354-362
pubmed: 21680644
Chest. 2015 Oct;148(4):1043-54
pubmed: 26066077
Intern Emerg Med. 2020 Jun;15(4):573-585
pubmed: 32040829
J Am Coll Cardiol. 2011 Jul 12;58(3):300-9
pubmed: 21737024
BMC Pulm Med. 2018 Oct 16;18(1):161
pubmed: 30326867
Eur Respir J. 2015 Jul;46(1):152-64
pubmed: 25837032
Lancet Respir Med. 2016 Apr;4(4):306-22
pubmed: 26975810
Eur Respir J. 2019 Jun 5;53(6):
pubmed: 30923187
Chest. 2019 Aug;156(2):323-337
pubmed: 30772387