Language function following preterm birth: prediction using machine learning

Evi Valavani; Manuel Blesa; Paola Galdi; Gemma Sullivan; Bethan Dean; Hilary Cruickshank; Magdalena Sitko-Rudnicka; Mark E Bastin; Richard Chin; Donald MacIntyre; Sue Fletcher-Watson; James P Boardman; Thanasis Tsanas

doi:https://doi.org/10.1038/s41390-021-01779-x

Language function following preterm birth: prediction using machine learning

Evi Valavani^*, Manuel Blesa, Paola Galdi, Gemma Sullivan, Bethan Dean, Hilary Cruickshank, Magdalena Sitko-Rudnicka, Mark E Bastin, Richard Chin, Donald MacIntyre, Sue Fletcher-Watson, James P Boardman, Thanasis Tsanas

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract / Description of output

BACKGROUND: Preterm birth can lead to impaired language development. This study aimed to predict language outcomes at
2 years corrected gestational age (CGA) for children born preterm.
METHODS: We analysed data from 89 preterm neonates (median GA 29 weeks) who underwent diffusion MRI (dMRI) at termequivalent age and language assessment at 2 years CGA using the Bayley-III. Feature selection and a random forests classifier were
used to differentiate typical versus delayed (Bayley-III language composite score <85) language development.
RESULTS: The model achieved balanced accuracy: 91%, sensitivity: 86%, and specificity: 96%. The probability of language delay at 2
years CGA is increased with: increasing values of peak width of skeletonized fractional anisotropy (PSFA), radial diffusivity (PSRD),
and axial diffusivity (PSAD) derived from dMRI; among twins; and after an incomplete course of, or no exposure to, antenatal
corticosteroids. Female sex and breastfeeding during the neonatal period reduced the risk of language delay.
CONCLUSIONS: The combination of perinatal clinical information and MRI features leads to accurate prediction of preterm infants
who are likely to develop language deficits in early childhood. This model could potentially enable stratification of preterm children
at risk of language dysfunction who may benefit from targeted early interventions.

Original language	English
Journal	Pediatric Research
DOIs	https://doi.org/10.1038/s41390-021-01779-x
Publication status	Published - 11 Oct 2021

Access to Document

https://doi.org/10.1038/s41390-021-01779-xLicence: Creative Commons: Attribution (CC-BY)

Pediatric_research_open_accessFinal published version, 1.26 MBLicence: Creative Commons: Attribution (CC-BY)

Cite this

@article{104e62c7c0184aabb551bf63c7eb8233,

title = "Language function following preterm birth: prediction using machine learning",

abstract = "BACKGROUND: Preterm birth can lead to impaired language development. This study aimed to predict language outcomes at2 years corrected gestational age (CGA) for children born preterm.METHODS: We analysed data from 89 preterm neonates (median GA 29 weeks) who underwent diffusion MRI (dMRI) at termequivalent age and language assessment at 2 years CGA using the Bayley-III. Feature selection and a random forests classifier wereused to differentiate typical versus delayed (Bayley-III language composite score <85) language development.RESULTS: The model achieved balanced accuracy: 91%, sensitivity: 86%, and specificity: 96%. The probability of language delay at 2years CGA is increased with: increasing values of peak width of skeletonized fractional anisotropy (PSFA), radial diffusivity (PSRD),and axial diffusivity (PSAD) derived from dMRI; among twins; and after an incomplete course of, or no exposure to, antenatalcorticosteroids. Female sex and breastfeeding during the neonatal period reduced the risk of language delay.CONCLUSIONS: The combination of perinatal clinical information and MRI features leads to accurate prediction of preterm infantswho are likely to develop language deficits in early childhood. This model could potentially enable stratification of preterm childrenat risk of language dysfunction who may benefit from targeted early interventions.",

author = "Evi Valavani and Manuel Blesa and Paola Galdi and Gemma Sullivan and Bethan Dean and Hilary Cruickshank and Magdalena Sitko-Rudnicka and Bastin, {Mark E} and Richard Chin and Donald MacIntyre and Sue Fletcher-Watson and Boardman, {James P} and Thanasis Tsanas",

year = "2021",

month = oct,

day = "11",

doi = "https://doi.org/10.1038/s41390-021-01779-x",

language = "English",

journal = "Pediatric Research",

issn = "0031-3998",

publisher = "INT PEDIATRIC RESEARCH FOUNDATION, INC",

}

TY - JOUR

T1 - Language function following preterm birth: prediction using machine learning

AU - Valavani, Evi

AU - Blesa, Manuel

AU - Galdi, Paola

AU - Sullivan, Gemma

AU - Dean, Bethan

AU - Cruickshank, Hilary

AU - Sitko-Rudnicka, Magdalena

AU - Bastin, Mark E

AU - Chin, Richard

AU - MacIntyre, Donald

AU - Fletcher-Watson, Sue

AU - Boardman, James P

AU - Tsanas, Thanasis

PY - 2021/10/11

Y1 - 2021/10/11

N2 - BACKGROUND: Preterm birth can lead to impaired language development. This study aimed to predict language outcomes at2 years corrected gestational age (CGA) for children born preterm.METHODS: We analysed data from 89 preterm neonates (median GA 29 weeks) who underwent diffusion MRI (dMRI) at termequivalent age and language assessment at 2 years CGA using the Bayley-III. Feature selection and a random forests classifier wereused to differentiate typical versus delayed (Bayley-III language composite score <85) language development.RESULTS: The model achieved balanced accuracy: 91%, sensitivity: 86%, and specificity: 96%. The probability of language delay at 2years CGA is increased with: increasing values of peak width of skeletonized fractional anisotropy (PSFA), radial diffusivity (PSRD),and axial diffusivity (PSAD) derived from dMRI; among twins; and after an incomplete course of, or no exposure to, antenatalcorticosteroids. Female sex and breastfeeding during the neonatal period reduced the risk of language delay.CONCLUSIONS: The combination of perinatal clinical information and MRI features leads to accurate prediction of preterm infantswho are likely to develop language deficits in early childhood. This model could potentially enable stratification of preterm childrenat risk of language dysfunction who may benefit from targeted early interventions.

AB - BACKGROUND: Preterm birth can lead to impaired language development. This study aimed to predict language outcomes at2 years corrected gestational age (CGA) for children born preterm.METHODS: We analysed data from 89 preterm neonates (median GA 29 weeks) who underwent diffusion MRI (dMRI) at termequivalent age and language assessment at 2 years CGA using the Bayley-III. Feature selection and a random forests classifier wereused to differentiate typical versus delayed (Bayley-III language composite score <85) language development.RESULTS: The model achieved balanced accuracy: 91%, sensitivity: 86%, and specificity: 96%. The probability of language delay at 2years CGA is increased with: increasing values of peak width of skeletonized fractional anisotropy (PSFA), radial diffusivity (PSRD),and axial diffusivity (PSAD) derived from dMRI; among twins; and after an incomplete course of, or no exposure to, antenatalcorticosteroids. Female sex and breastfeeding during the neonatal period reduced the risk of language delay.CONCLUSIONS: The combination of perinatal clinical information and MRI features leads to accurate prediction of preterm infantswho are likely to develop language deficits in early childhood. This model could potentially enable stratification of preterm childrenat risk of language dysfunction who may benefit from targeted early interventions.

U2 - https://doi.org/10.1038/s41390-021-01779-x

DO - https://doi.org/10.1038/s41390-021-01779-x

M3 - Article

SN - 0031-3998

JO - Pediatric Research

JF - Pediatric Research

ER -

Language function following preterm birth: prediction using machine learning

Abstract / Description of output

Access to Document

Fingerprint

Cite this