A Combined Radiomics and Machine Learning Approach to Overcome the Clinicoradiologic Paradox in Multiple Sclerosis

Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.BACKGROUND AND PURPOSEC...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:American journal of neuroradiology : AJNR Jg. 42; H. 11; S. 1927
Hauptverfasser: Pontillo, G, Tommasin, S, Cuocolo, R, Petracca, M, Petsas, N, Ugga, L, Carotenuto, A, Pozzilli, C, Iodice, R, Lanzillo, R, Quarantelli, M, Brescia Morra, V, Tedeschi, E, Pantano, P, Cocozza, S
Format: Journal Article
Sprache:Englisch
Veröffentlicht: 01.11.2021
ISSN:1936-959X, 1936-959X
Online-Zugang:Weitere Angaben
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.BACKGROUND AND PURPOSEConventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.In this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training (n = 400) and test (n = 100) sets, while the site 2 cohort (n = 104) constituted the external test set. After feature selection of clinicodemographic and MR imaging-derived variables, different machine learning algorithms predicting disability as measured with the Expanded Disability Status Scale were trained and cross-validated on the training cohort and evaluated on the test sets. The effect of different algorithms on model performance was tested using the 1-way repeated-measures ANOVA.MATERIALS AND METHODSIn this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training (n = 400) and test (n = 100) sets, while the site 2 cohort (n = 104) constituted the external test set. After feature selection of clinicodemographic and MR imaging-derived variables, different machine learning algorithms predicting disability as measured with the Expanded Disability Status Scale were trained and cross-validated on the training cohort and evaluated on the test sets. The effect of different algorithms on model performance was tested using the 1-way repeated-measures ANOVA.The selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy (r approaching 0.80) and excellent intra- and intersite generalizability (r ≥ 0.73). The machine learning algorithm had no relevant effect on the performance.RESULTSThe selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy (r approaching 0.80) and excellent intra- and intersite generalizability (r ≥ 0.73). The machine learning algorithm had no relevant effect on the performance.The multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.CONCLUSIONSThe multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.
AbstractList Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.BACKGROUND AND PURPOSEConventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.In this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training (n = 400) and test (n = 100) sets, while the site 2 cohort (n = 104) constituted the external test set. After feature selection of clinicodemographic and MR imaging-derived variables, different machine learning algorithms predicting disability as measured with the Expanded Disability Status Scale were trained and cross-validated on the training cohort and evaluated on the test sets. The effect of different algorithms on model performance was tested using the 1-way repeated-measures ANOVA.MATERIALS AND METHODSIn this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training (n = 400) and test (n = 100) sets, while the site 2 cohort (n = 104) constituted the external test set. After feature selection of clinicodemographic and MR imaging-derived variables, different machine learning algorithms predicting disability as measured with the Expanded Disability Status Scale were trained and cross-validated on the training cohort and evaluated on the test sets. The effect of different algorithms on model performance was tested using the 1-way repeated-measures ANOVA.The selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy (r approaching 0.80) and excellent intra- and intersite generalizability (r ≥ 0.73). The machine learning algorithm had no relevant effect on the performance.RESULTSThe selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy (r approaching 0.80) and excellent intra- and intersite generalizability (r ≥ 0.73). The machine learning algorithm had no relevant effect on the performance.The multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.CONCLUSIONSThe multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.
Author Petsas, N
Ugga, L
Petracca, M
Pozzilli, C
Cocozza, S
Cuocolo, R
Pantano, P
Carotenuto, A
Iodice, R
Tedeschi, E
Pontillo, G
Tommasin, S
Lanzillo, R
Brescia Morra, V
Quarantelli, M
Author_xml – sequence: 1
  givenname: G
  surname: Pontillo
  fullname: Pontillo, G
– sequence: 2
  givenname: S
  surname: Tommasin
  fullname: Tommasin, S
– sequence: 3
  givenname: R
  surname: Cuocolo
  fullname: Cuocolo, R
– sequence: 4
  givenname: M
  surname: Petracca
  fullname: Petracca, M
– sequence: 5
  givenname: N
  surname: Petsas
  fullname: Petsas, N
– sequence: 6
  givenname: L
  surname: Ugga
  fullname: Ugga, L
– sequence: 7
  givenname: A
  surname: Carotenuto
  fullname: Carotenuto, A
– sequence: 8
  givenname: C
  surname: Pozzilli
  fullname: Pozzilli, C
– sequence: 9
  givenname: R
  surname: Iodice
  fullname: Iodice, R
– sequence: 10
  givenname: R
  surname: Lanzillo
  fullname: Lanzillo, R
– sequence: 11
  givenname: M
  surname: Quarantelli
  fullname: Quarantelli, M
– sequence: 12
  givenname: V
  surname: Brescia Morra
  fullname: Brescia Morra, V
– sequence: 13
  givenname: E
  surname: Tedeschi
  fullname: Tedeschi, E
– sequence: 14
  givenname: P
  surname: Pantano
  fullname: Pantano, P
– sequence: 15
  givenname: S
  surname: Cocozza
  fullname: Cocozza, S
BookMark eNpNjktLAzEcxIMo2FYvfoIcvWzN5rHZHJfFF7RUfIC3kib_bVOySU224sd3RQ-eZuYHM8wUnYYYAKGrksxZKfmN3oc0bySV_ARNSsWqQgn1fvrPn6NpzntCiFCSTlBqcBv7jQtg8bO2LvbOZKyDxUttdiPGC9ApuLDFzeGQ4gjxEPHqE5KJPeBhB7j1LjgT00_dx60z-EmPIX5hF_Dy6Ad38IBfjIcUs8sX6KzTPsPln87Q293ta_tQLFb3j22zKAwTYijK2ihrOSuZtQwIVTXtlDAV3yjSSWZqXdcgCCdABIAk0NXAtNwILTQtjaYzdP27O97-OEIe1r3LBrzXAeIxr6mQnJOqEoR-A4HOYS0
CitedBy_id crossref_primary_10_1007_s00415_024_12651_3
crossref_primary_10_61882_nl_4_2_107
crossref_primary_10_1016_j_msard_2023_104750
crossref_primary_10_1007_s11547_024_01904_w
crossref_primary_10_1016_j_cmpb_2025_108771
crossref_primary_10_1016_j_msard_2023_104869
crossref_primary_10_1111_jon_70069
crossref_primary_10_3390_ijms231810651
crossref_primary_10_1007_s11910_024_01354_x
crossref_primary_10_1111_jon_70089
crossref_primary_10_1080_01616412_2025_2532038
crossref_primary_10_3390_jpm15050167
crossref_primary_10_1016_j_crad_2025_106921
crossref_primary_10_1098_rsos_241052
crossref_primary_10_1016_j_metrad_2024_100068
crossref_primary_10_1109_JBHI_2022_3151741
crossref_primary_10_3389_fnins_2024_1294574
crossref_primary_10_1007_s00415_025_13253_3
crossref_primary_10_3389_fnins_2023_1167612
crossref_primary_10_1007_s10278_024_01031_y
ContentType Journal Article
Copyright 2021 by American Journal of Neuroradiology.
Copyright_xml – notice: 2021 by American Journal of Neuroradiology.
DBID 7X8
DOI 10.3174/ajnr.A7274
DatabaseName MEDLINE - Academic
DatabaseTitle MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
Database_xml – sequence: 1
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Medicine
EISSN 1936-959X
GroupedDBID ---
.55
23M
2WC
53G
5GY
5RE
5VS
6J9
7X8
ACGFO
ACIWK
ACPRK
ADBBV
AENEX
AFHIN
AFRAH
ALMA_UNASSIGNED_HOLDINGS
BAWUL
BTFSW
CS3
E3Z
EBS
EJD
EMOBN
F5P
F9R
GX1
H13
INIJC
KQ8
MV1
N9A
OK1
P2P
P6G
R0Z
RHI
RPM
TNE
TR2
W8F
WOQ
WOW
X7M
ZCG
ID FETCH-LOGICAL-c355t-18c9dd4313dd3e02982f95c64b90f73c8a88e5040e05ee70ef8e3a7b5a5a21ca2
IEDL.DBID 7X8
ISICitedReferencesCount 22
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000697385500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1936-959X
IngestDate Fri Jul 11 08:57:26 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 11
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c355t-18c9dd4313dd3e02982f95c64b90f73c8a88e5040e05ee70ef8e3a7b5a5a21ca2
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
OpenAccessLink http://www.ajnr.org/content/ajnr/42/11/1927.full.pdf
PQID 2574406650
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2574406650
PublicationCentury 2000
PublicationDate 20211101
PublicationDateYYYYMMDD 2021-11-01
PublicationDate_xml – month: 11
  year: 2021
  text: 20211101
  day: 01
PublicationDecade 2020
PublicationTitle American journal of neuroradiology : AJNR
PublicationYear 2021
SSID ssj0005972
Score 2.4712033
Snippet Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for...
SourceID proquest
SourceType Aggregation Database
StartPage 1927
Title A Combined Radiomics and Machine Learning Approach to Overcome the Clinicoradiologic Paradox in Multiple Sclerosis
URI https://www.proquest.com/docview/2574406650
Volume 42
WOSCitedRecordID wos000697385500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LS8QwEA7qinjxLb6J4LXupmna5CRFXLzsuviA3pY0mUo9tLpdxZ_vpM2yBy-Cx0JLhvDNzNfM5BtCrqwSSkieB0wnSRDlYYw-p0wgwBSCGcvww3bYRDIeyyxTE3_g1vi2ykVMbAO1rY07I-8jtKLI1QkGN-8fgZsa5aqrfoTGKulxpDKupSvJlmrhSJbDrqocB2hT1smTYsaM-vqtml2nmLyjX0G4zSzD7f_atEO2PKekaQeCXbIC1R7ZGPmq-T6ZpRTdHn-BwdJHbUt3EbmhurJ01LZSAvUqq6809RLjdF7TB0Q52gEUOSLt9EMRL7bswiWdaHyov2lZ0ZHvSqRPuD6m3bI5IC_Du-fb-8DPWggMMo55wKRR1iKb4NZycLrsYaGEiaNcDYqEG6mlBIEeDwMBkAygkMB1kgstdMiMDg_JWlVXcESodBJo3HIOUEQFY0oDcC4skg8W54Idk8vFbk4Ry65AoSuoP5vpcj9P_vDOKdkMXYNJezHwjPQK9Fc4J-vma142s4sWCj-Hfr_K
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=A+Combined+Radiomics+and+Machine+Learning+Approach+to+Overcome+the+Clinicoradiologic+Paradox+in+Multiple+Sclerosis&rft.jtitle=American+journal+of+neuroradiology+%3A+AJNR&rft.au=Pontillo%2C+G&rft.au=Tommasin%2C+S&rft.au=Cuocolo%2C+R&rft.au=Petracca%2C+M&rft.date=2021-11-01&rft.issn=1936-959X&rft.eissn=1936-959X&rft.volume=42&rft.issue=11&rft.spage=1927&rft_id=info:doi/10.3174%2Fajnr.A7274&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1936-959X&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1936-959X&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1936-959X&client=summon