Borrowing strength from adults : Transferability of AI algorithms for paediatric brain and tumour segmentation
Copyright © 2022 Elsevier B.V. All rights reserved..
PURPOSE: AI brain tumour segmentation and brain extraction algorithms promise better diagnostic and follow-up of brain tumours in adults. The development of such tools for paediatric populations is restricted by limited training data but careful adaption of adult algorithms to paediatric population might be a solution. Here, we aim exploring the transferability of algorithms for brain (HD-BET) and tumour segmentation (HD-GLIOMA) in adults to paediatric imaging studies.
METHOD: In a retrospective cohort, we compared automated segmentation with expert masks. We used the dice coefficient for evaluating the similarity and multivariate regressions for the influence of covariates. We explored the feasibility of automatic tumor classification based on diffusion data.
RESULTS: In 42 patients (mean age 7 years, 9 below 2 years, 26 males), segmentation was excellent for brain extraction (mean dice 0.99, range 0.85-1), moderate for segmentation of contrast-enhancing tumours (mean dice 0.67, range 0-1), and weak for non-enhancing T2-signal abnormalities (mean dice 0.41). Precision was better for enhancing tumour parts (p < 0.001) and for malignant histology (p = 0.006 and p = 0.012) but independent from myelinisation as indicated by the age (p = 0.472). Automated tumour grading based on mean diffusivity (MD) values from automated masks was good (AUC = 0.86) but tended to be less accurate than MD values from expert masks (AUC = 1, p = 0.208).
CONCLUSION: HD-BET provides a reliable extraction of the paediatric brain. HD-GLIOMA works moderately for contrast-enhancing tumours parts. Without optimization, brain tumor AI algorithms trained on adults and used on paediatric patients may yield acceptable results depending on the clinical scenario.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2022 |
---|---|
Erschienen: |
2022 |
Enthalten in: |
Zur Gesamtaufnahme - volume:151 |
---|---|
Enthalten in: |
European journal of radiology - 151(2022) vom: 15. Juni, Seite 110291 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Drai, Maxime [VerfasserIn] |
---|
Links: |
---|
Themen: |
AI |
---|
Anmerkungen: |
Date Completed 17.05.2022 Date Revised 17.05.2022 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1016/j.ejrad.2022.110291 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM339351861 |
---|
LEADER | 01000naa a22002652 4500 | ||
---|---|---|---|
001 | NLM339351861 | ||
003 | DE-627 | ||
005 | 20231226002716.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2022 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1016/j.ejrad.2022.110291 |2 doi | |
028 | 5 | 2 | |a pubmed24n1131.xml |
035 | |a (DE-627)NLM339351861 | ||
035 | |a (NLM)35405580 | ||
035 | |a (PII)S0720-048X(22)00141-3 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Drai, Maxime |e verfasserin |4 aut | |
245 | 1 | 0 | |a Borrowing strength from adults |b Transferability of AI algorithms for paediatric brain and tumour segmentation |
264 | 1 | |c 2022 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ƒaComputermedien |b c |2 rdamedia | ||
338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
500 | |a Date Completed 17.05.2022 | ||
500 | |a Date Revised 17.05.2022 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a Copyright © 2022 Elsevier B.V. All rights reserved. | ||
520 | |a PURPOSE: AI brain tumour segmentation and brain extraction algorithms promise better diagnostic and follow-up of brain tumours in adults. The development of such tools for paediatric populations is restricted by limited training data but careful adaption of adult algorithms to paediatric population might be a solution. Here, we aim exploring the transferability of algorithms for brain (HD-BET) and tumour segmentation (HD-GLIOMA) in adults to paediatric imaging studies | ||
520 | |a METHOD: In a retrospective cohort, we compared automated segmentation with expert masks. We used the dice coefficient for evaluating the similarity and multivariate regressions for the influence of covariates. We explored the feasibility of automatic tumor classification based on diffusion data | ||
520 | |a RESULTS: In 42 patients (mean age 7 years, 9 below 2 years, 26 males), segmentation was excellent for brain extraction (mean dice 0.99, range 0.85-1), moderate for segmentation of contrast-enhancing tumours (mean dice 0.67, range 0-1), and weak for non-enhancing T2-signal abnormalities (mean dice 0.41). Precision was better for enhancing tumour parts (p < 0.001) and for malignant histology (p = 0.006 and p = 0.012) but independent from myelinisation as indicated by the age (p = 0.472). Automated tumour grading based on mean diffusivity (MD) values from automated masks was good (AUC = 0.86) but tended to be less accurate than MD values from expert masks (AUC = 1, p = 0.208) | ||
520 | |a CONCLUSION: HD-BET provides a reliable extraction of the paediatric brain. HD-GLIOMA works moderately for contrast-enhancing tumours parts. Without optimization, brain tumor AI algorithms trained on adults and used on paediatric patients may yield acceptable results depending on the clinical scenario | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a AI | |
650 | 4 | |a Grading | |
650 | 4 | |a MRI | |
650 | 4 | |a Paediatric brain tumour | |
650 | 4 | |a Segmentation | |
700 | 1 | |a Testud, Benoit |e verfasserin |4 aut | |
700 | 1 | |a Brun, Gilles |e verfasserin |4 aut | |
700 | 1 | |a Hak, Jean-François |e verfasserin |4 aut | |
700 | 1 | |a Scavarda, Didier |e verfasserin |4 aut | |
700 | 1 | |a Girard, Nadine |e verfasserin |4 aut | |
700 | 1 | |a Stellmann, Jan-Patrick |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t European journal of radiology |d 1993 |g 151(2022) vom: 15. Juni, Seite 110291 |w (DE-627)NLM012952397 |x 1872-7727 |7 nnns |
773 | 1 | 8 | |g volume:151 |g year:2022 |g day:15 |g month:06 |g pages:110291 |
856 | 4 | 0 | |u http://dx.doi.org/10.1016/j.ejrad.2022.110291 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 151 |j 2022 |b 15 |c 06 |h 110291 |