Toward Generalizable and Transdiagnostic Tools for Psychosis Prediction: An Independent Validation and Improvement of the NAPLS-2 Risk Calculator in the Multisite PRONIA Cohort.
Clinical high-risk states
First-episode depression
Machine learning
Psychosis prediction
Reciprocal external validation
Risk calculators
Journal
Biological psychiatry
ISSN: 1873-2402
Titre abrégé: Biol Psychiatry
Pays: United States
ID NLM: 0213264
Informations de publication
Date de publication:
01 11 2021
01 11 2021
Historique:
received:
05
02
2021
revised:
03
06
2021
accepted:
27
06
2021
pubmed:
7
9
2021
medline:
3
11
2021
entrez:
6
9
2021
Statut:
ppublish
Résumé
Transition to psychosis is among the most adverse outcomes of clinical high-risk (CHR) syndromes encompassing ultra-high risk (UHR) and basic symptom states. Clinical risk calculators may facilitate an early and individualized interception of psychosis, but their real-world implementation requires thorough validation across diverse risk populations, including young patients with depressive syndromes. We validated the previously described NAPLS-2 (North American Prodrome Longitudinal Study 2) calculator in 334 patients (26 with transition to psychosis) with CHR or recent-onset depression (ROD) drawn from the multisite European PRONIA (Personalised Prognostic Tools for Early Psychosis Management) study. Patients were categorized into three risk enrichment levels, ranging from UHR, over CHR, to a broad-risk population comprising patients with CHR or ROD (CHR|ROD). We assessed how risk enrichment and different predictive algorithms influenced prognostic performance using reciprocal external validation. After calibration, the NAPLS-2 model predicted psychosis with a balanced accuracy (BAC) (sensitivity, specificity) of 68% (73%, 63%) in the PRONIA-UHR cohort, 67% (74%, 60%) in the CHR cohort, and 70% (73%, 66%) in patients with CHR|ROD. Multiple model derivation in PRONIA-CHR|ROD and validation in NAPLS-2-UHR patients confirmed that broader risk definitions produced more accurate risk calculators (CHR|ROD-based vs. UHR-based performance: 67% [68%, 66%] vs. 58% [61%, 56%]). Support vector machines were superior in CHR|ROD (BAC = 71%), while ridge logistic regression and support vector machines performed similarly in CHR (BAC = 67%) and UHR cohorts (BAC = 65%). Attenuated psychotic symptoms predicted psychosis across risk levels, while younger age and reduced processing speed became increasingly relevant for broader risk cohorts. Clinical-neurocognitive machine learning models operating in young patients with affective and CHR syndromes facilitate a more precise and generalizable prediction of psychosis. Future studies should investigate their therapeutic utility in large-scale clinical trials.
Sections du résumé
BACKGROUND
Transition to psychosis is among the most adverse outcomes of clinical high-risk (CHR) syndromes encompassing ultra-high risk (UHR) and basic symptom states. Clinical risk calculators may facilitate an early and individualized interception of psychosis, but their real-world implementation requires thorough validation across diverse risk populations, including young patients with depressive syndromes.
METHODS
We validated the previously described NAPLS-2 (North American Prodrome Longitudinal Study 2) calculator in 334 patients (26 with transition to psychosis) with CHR or recent-onset depression (ROD) drawn from the multisite European PRONIA (Personalised Prognostic Tools for Early Psychosis Management) study. Patients were categorized into three risk enrichment levels, ranging from UHR, over CHR, to a broad-risk population comprising patients with CHR or ROD (CHR|ROD). We assessed how risk enrichment and different predictive algorithms influenced prognostic performance using reciprocal external validation.
RESULTS
After calibration, the NAPLS-2 model predicted psychosis with a balanced accuracy (BAC) (sensitivity, specificity) of 68% (73%, 63%) in the PRONIA-UHR cohort, 67% (74%, 60%) in the CHR cohort, and 70% (73%, 66%) in patients with CHR|ROD. Multiple model derivation in PRONIA-CHR|ROD and validation in NAPLS-2-UHR patients confirmed that broader risk definitions produced more accurate risk calculators (CHR|ROD-based vs. UHR-based performance: 67% [68%, 66%] vs. 58% [61%, 56%]). Support vector machines were superior in CHR|ROD (BAC = 71%), while ridge logistic regression and support vector machines performed similarly in CHR (BAC = 67%) and UHR cohorts (BAC = 65%). Attenuated psychotic symptoms predicted psychosis across risk levels, while younger age and reduced processing speed became increasingly relevant for broader risk cohorts.
CONCLUSIONS
Clinical-neurocognitive machine learning models operating in young patients with affective and CHR syndromes facilitate a more precise and generalizable prediction of psychosis. Future studies should investigate their therapeutic utility in large-scale clinical trials.
Identifiants
pubmed: 34482951
pii: S0006-3223(21)01433-5
doi: 10.1016/j.biopsych.2021.06.023
pmc: PMC8500930
mid: NIHMS1721825
pii:
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
632-642Subventions
Organisme : NIMH NIH HHS
ID : U01 MH082022
Pays : United States
Organisme : NCATS NIH HHS
ID : UL1 TR001863
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH081902
Pays : United States
Organisme : NIMH NIH HHS
ID : R01 MH076989
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH081988
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH066069
Pays : United States
Organisme : NIMH NIH HHS
ID : P50 MH066286
Pays : United States
Organisme : NIMH NIH HHS
ID : R01 MH119219
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH066134
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH081928
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH081857
Pays : United States
Organisme : NIMH NIH HHS
ID : U01 MH081944
Pays : United States
Commentaires et corrections
Type : CommentIn
Informations de copyright
Copyright © 2021 Society of Biological Psychiatry. Published by Elsevier Inc. All rights reserved.
Références
BMC Med Res Methodol. 2013 Mar 06;13:33
pubmed: 23496923
Brain. 2015 Jul;138(Pt 7):2059-73
pubmed: 25935725
PLoS Med. 2013;10(2):e1001381
pubmed: 23393430
JAMA Psychiatry. 2016 Dec 1;73(12):1260-1267
pubmed: 27784037
Am J Psychiatry. 2016 Oct 1;173(10):980-988
pubmed: 27363508
NPJ Schizophr. 2021 Feb 12;7(1):8
pubmed: 33580020
JAMA Psychiatry. 2018 Dec 1;75(12):1289-1297
pubmed: 30347013
Schizophr Res. 2019 Jun;208:41-43
pubmed: 30745066
JAMA Psychiatry. 2018 Nov 1;75(11):1156-1172
pubmed: 30267047
Schizophr Bull. 2010 Jan;36(1):182-91
pubmed: 18579555
Schizophr Bull. 2019 Apr 25;45(3):562-570
pubmed: 29897527
Schizophr Bull. 2007 May;33(3):688-702
pubmed: 17440198
JAMA Psychiatry. 2020 Jul 1;77(7):755-765
pubmed: 32159746
World Psychiatry. 2015 Jun;14(2):189-97
pubmed: 26043337
JAMA Psychiatry. 2021 Feb 1;78(2):195-209
pubmed: 33263726
Transl Psychiatry. 2019 Oct 17;9(1):259
pubmed: 31624229
Schizophr Res. 2005 Sep 1;77(1):11-24
pubmed: 16005381
JAMA Netw Open. 2020 Apr 1;3(4):e203221
pubmed: 32315069
Schizophr Res. 2017 Dec;190:90-95
pubmed: 28372906
Arch Gen Psychiatry. 2001 Feb;58(2):158-64
pubmed: 11177117
Am J Psychiatry. 2020 Feb 1;177(2):155-163
pubmed: 31711302
Arch Gen Psychiatry. 2008 Jan;65(1):28-37
pubmed: 18180426
Schizophr Res. 2004 Jun 1;68(2-3):283-97
pubmed: 15099610
Eur Psychiatry. 2017 Feb;40:65-75
pubmed: 27992836
Transl Psychiatry. 2021 May 24;11(1):312
pubmed: 34031362
Front Psychiatry. 2019 Oct 24;10:774
pubmed: 31708822
Biol Psychiatry. 2011 May 15;69(10):959-66
pubmed: 21167475
Schizophr Bull. 2015 Mar;41(2):419-28
pubmed: 25103207
Schizophr Res. 2014 Apr;154(1-3):100-6
pubmed: 24613572
Schizophr Bull. 2014 Jan;40(1):120-31
pubmed: 23180756
Schizophr Res. 2021 Jan;227:10-17
pubmed: 32402605
World Psychiatry. 2015 Oct;14(3):322-32
pubmed: 26407788
Schizophr Bull. 2021 Mar 16;47(2):284-297
pubmed: 32914178
Am J Psychiatry. 2016 Oct 1;173(10):989-996
pubmed: 27363511
Arch Gen Psychiatry. 2010 Mar;67(3):241-51
pubmed: 20194824
Schizophr Bull. 2012 Mar;38(2):247-57
pubmed: 22258882
Psychopathology. 2014;47(5):292-6
pubmed: 25011580
Schizophr Bull. 2021 Jan 23;47(1):249-258
pubmed: 32634220
JAMA Psychiatry. 2019 May 1;76(5):516-525
pubmed: 30698613
Front Psychiatry. 2019 Apr 26;10:253
pubmed: 31105603
Am J Psychiatry. 2018 Sep 1;175(9):906-908
pubmed: 30173545
Int J Epidemiol. 2015 Oct 8;45(2):408-416
pubmed: 26452388
Aust N Z J Psychiatry. 1996 Oct;30(5):587-99
pubmed: 8902166
Biol Psychiatry. 2020 Aug 15;88(4):349-360
pubmed: 32305218
Biol Psychiatry. 2017 Feb 15;81(4):325-335
pubmed: 27519822
Eur Heart J. 2019 Feb 14;40(7):621-631
pubmed: 30476079
Schizophr Res. 2021 Jan;227:95-100
pubmed: 33046334
Mol Psychiatry. 2021 Jun;26(6):2070-2081
pubmed: 32398722
Schizophr Bull. 2018 Feb 15;44(2):276-285
pubmed: 29438561
Psychol Med. 2019 Sep;49(12):1990-1998
pubmed: 30213278
Schizophr Bull. 2016 May;42(3):732-43
pubmed: 26591006
JAMA Psychiatry. 2017 May 1;74(5):493-500
pubmed: 28355424
Eur Psychiatry. 2019 Jan;55:102-108
pubmed: 30447416
Schizophr Bull. 2012 Nov;38(6):1200-15
pubmed: 21576280
Med J Aust. 2014 Feb 3;200(2):108-11
pubmed: 24484115
Eur Child Adolesc Psychiatry. 2018 Jun;27(6):725-737
pubmed: 29058115
BJPsych Open. 2019 Jun 13;5(4):e54
pubmed: 31530297
Schizophr Bull. 1996;22(2):353-70
pubmed: 8782291
World Psychiatry. 2018 Jun;17(2):133-142
pubmed: 29856558
Schizophr Res. 2012 Dec;142(1-3):77-82
pubmed: 23043872
Schizophr Bull. 2015 Mar;41(2):471-82
pubmed: 24914177
Schizophr Res. 2018 Dec;202:333-340
pubmed: 30539771
Psychiatry Res. 2019 Sep;279:9-14
pubmed: 31279247
Schizophr Res. 2021 Jan;227:52-60
pubmed: 32571619