• Something wrong with this record ?

Preoperative prediction of diffuse glioma type and grade in adults: a gadolinium-free MRI-based decision tree

A. Azizova, Y. Prysiazhniuk, IJHG. Wamelink, M. Cakmak, E. Kaya, P. Wesseling, PC. de Witt Hamer, N. Verburg, J. Petr, F. Barkhof, VC. Keil

. 2025 ; 35 (3) : 1242-1254. [pub] 20241019

Language English Country Germany

Document type Journal Article

Persistent link   https://www.medvik.cz/link/bmc25009761

OBJECTIVES: To develop a gadolinium-free MRI-based diagnosis prediction decision tree (DPDT) for adult-type diffuse gliomas and to assess the added value of gadolinium-based contrast agent (GBCA) enhanced images. MATERIALS AND METHODS: This study included preoperative grade 2-4 adult-type diffuse gliomas (World Health Organization 2021) scanned between 2010 and 2021. The DPDT, incorporating eleven GBCA-free MRI features, was developed using 18% of the dataset based on consensus readings. Diagnosis predictions involved grade (grade 2 vs. grade 3/4) and molecular status (isocitrate dehydrogenase (IDH) and 1p/19q). GBCA-free diagnosis was predicted using DPDT, while GBCA-enhanced diagnosis included post-contrast images. The accuracy of these predictions was assessed by three raters with varying experience levels in neuroradiology using the test dataset. Agreement analyses were applied to evaluate the prediction performance/reproducibility. RESULTS: The test dataset included 303 patients (age (SD): 56.7 (14.2) years, female/male: 114/189, low-grade/high-grade: 54/249, IDH-mutant/wildtype: 82/221, 1p/19q-codeleted/intact: 34/269). Per-rater GBCA-free predictions achieved ≥ 0.85 (95%-CI: 0.80-0.88) accuracy for grade and ≥ 0.75 (95%-CI: 0.70-0.80) for molecular status, while GBCA-enhanced predictions reached ≥ 0.87 (95%-CI: 0.82-0.90) and ≥ 0.77 (95%-CI: 0.71-0.81), respectively. No accuracy difference was observed between GBCA-free and GBCA-enhanced predictions. Group inter-rater agreement was moderate for GBCA-free (0.56 (95%-CI: 0.46-0.66)) and substantial for GBCA-enhanced grade prediction (0.68 (95%-CI: 0.58-0.78), p = 0.008), while substantial for both GBCA-free (0.75 (95%-CI: 0.69-0.80) and GBCA-enhanced (0.77 (95%-CI: 0.71-0.82), p = 0.51) molecular status predictions. CONCLUSION: The proposed GBCA-free diagnosis prediction decision tree performed well, with GBCA-enhanced images adding little to the preoperative diagnostic accuracy of adult-type diffuse gliomas. KEY POINTS: Question Given health and environmental concerns, is there a gadolinium-free imaging protocol to preoperatively evaluate gliomas comparable to the gadolinium-enhanced standard practice? Findings The proposed gadolinium-free diagnosis prediction decision tree for adult-type diffuse gliomas performed well, and gadolinium-enhanced MRI demonstrated only limited improvement in diagnostic accuracy. Clinical relevance Even inexperienced raters effectively classified adult-type diffuse gliomas using the gadolinium-free diagnosis prediction decision tree, which, until further validation, can be used alongside gadolinium-enhanced images to respect standard practice, despite this study showing that gadolinium-enhanced images hardly improved diagnostic accuracy.

References provided by Crossref.org

000      
00000naa a2200000 a 4500
001      
bmc25009761
003      
CZ-PrNML
005      
20250429135040.0
007      
ta
008      
250415s2025 gw f 000 0|eng||
009      
AR
024    7_
$a 10.1007/s00330-024-11140-5 $2 doi
035    __
$a (PubMed)39425768
040    __
$a ABA008 $b cze $d ABA008 $e AACR2
041    0_
$a eng
044    __
$a gw
100    1_
$a Azizova, Aynur $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands $u Cancer Center Amsterdam, Imaging and Biomarkers, Amsterdam, The Netherlands
245    10
$a Preoperative prediction of diffuse glioma type and grade in adults: a gadolinium-free MRI-based decision tree / $c A. Azizova, Y. Prysiazhniuk, IJHG. Wamelink, M. Cakmak, E. Kaya, P. Wesseling, PC. de Witt Hamer, N. Verburg, J. Petr, F. Barkhof, VC. Keil
520    9_
$a OBJECTIVES: To develop a gadolinium-free MRI-based diagnosis prediction decision tree (DPDT) for adult-type diffuse gliomas and to assess the added value of gadolinium-based contrast agent (GBCA) enhanced images. MATERIALS AND METHODS: This study included preoperative grade 2-4 adult-type diffuse gliomas (World Health Organization 2021) scanned between 2010 and 2021. The DPDT, incorporating eleven GBCA-free MRI features, was developed using 18% of the dataset based on consensus readings. Diagnosis predictions involved grade (grade 2 vs. grade 3/4) and molecular status (isocitrate dehydrogenase (IDH) and 1p/19q). GBCA-free diagnosis was predicted using DPDT, while GBCA-enhanced diagnosis included post-contrast images. The accuracy of these predictions was assessed by three raters with varying experience levels in neuroradiology using the test dataset. Agreement analyses were applied to evaluate the prediction performance/reproducibility. RESULTS: The test dataset included 303 patients (age (SD): 56.7 (14.2) years, female/male: 114/189, low-grade/high-grade: 54/249, IDH-mutant/wildtype: 82/221, 1p/19q-codeleted/intact: 34/269). Per-rater GBCA-free predictions achieved ≥ 0.85 (95%-CI: 0.80-0.88) accuracy for grade and ≥ 0.75 (95%-CI: 0.70-0.80) for molecular status, while GBCA-enhanced predictions reached ≥ 0.87 (95%-CI: 0.82-0.90) and ≥ 0.77 (95%-CI: 0.71-0.81), respectively. No accuracy difference was observed between GBCA-free and GBCA-enhanced predictions. Group inter-rater agreement was moderate for GBCA-free (0.56 (95%-CI: 0.46-0.66)) and substantial for GBCA-enhanced grade prediction (0.68 (95%-CI: 0.58-0.78), p = 0.008), while substantial for both GBCA-free (0.75 (95%-CI: 0.69-0.80) and GBCA-enhanced (0.77 (95%-CI: 0.71-0.82), p = 0.51) molecular status predictions. CONCLUSION: The proposed GBCA-free diagnosis prediction decision tree performed well, with GBCA-enhanced images adding little to the preoperative diagnostic accuracy of adult-type diffuse gliomas. KEY POINTS: Question Given health and environmental concerns, is there a gadolinium-free imaging protocol to preoperatively evaluate gliomas comparable to the gadolinium-enhanced standard practice? Findings The proposed gadolinium-free diagnosis prediction decision tree for adult-type diffuse gliomas performed well, and gadolinium-enhanced MRI demonstrated only limited improvement in diagnostic accuracy. Clinical relevance Even inexperienced raters effectively classified adult-type diffuse gliomas using the gadolinium-free diagnosis prediction decision tree, which, until further validation, can be used alongside gadolinium-enhanced images to respect standard practice, despite this study showing that gadolinium-enhanced images hardly improved diagnostic accuracy.
650    _2
$a lidé $7 D006801
650    _2
$a mužské pohlaví $7 D008297
650    _2
$a ženské pohlaví $7 D005260
650    12
$a gliom $x diagnostické zobrazování $7 D005910
650    12
$a rozhodovací stromy $7 D003663
650    _2
$a lidé středního věku $7 D008875
650    12
$a nádory mozku $x diagnostické zobrazování $7 D001932
650    12
$a magnetická rezonanční tomografie $x metody $7 D008279
650    12
$a stupeň nádoru $7 D060787
650    12
$a kontrastní látky $7 D003287
650    _2
$a dospělí $7 D000328
650    _2
$a reprodukovatelnost výsledků $7 D015203
650    _2
$a gadolinium $7 D005682
650    _2
$a senioři $7 D000368
650    _2
$a prediktivní hodnota testů $7 D011237
655    _2
$a časopisecké články $7 D016428
700    1_
$a Prysiazhniuk, Yeva $u Charles University, The Second Faculty of Medicine, Department of Pathophysiology, Prague, Czech Republic $u Motol University Hospital, Prague, Czech Republic
700    1_
$a Wamelink, Ivar J H G $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands $u Cancer Center Amsterdam, Imaging and Biomarkers, Amsterdam, The Netherlands
700    1_
$a Cakmak, Marcus $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands $u Vrije Universiteit Amsterdam, University Medical Center, Amsterdam, The Netherlands
700    1_
$a Kaya, Elif $u Ankara Yıldırım Beyazıt University, Faculty of Medicine, Ankara, Turkey
700    1_
$a Wesseling, Pieter $u Amsterdam UMC location Vrije Universiteit Amsterdam, Department of Pathology, Amsterdam, The Netherlands $u Princess Máxima Center for Pediatric Oncology, Laboratory for Childhood Cancer Pathology, Utrecht, The Netherlands
700    1_
$a de Witt Hamer, Philip C $u Amsterdam UMC location Vrije Universiteit Amsterdam, Department of Neurosurgery, Brain Tumor Center Amsterdam, Amsterdam, The Netherlands
700    1_
$a Verburg, Niels $u Amsterdam UMC location Vrije Universiteit Amsterdam, Department of Neurosurgery, Brain Tumor Center Amsterdam, Amsterdam, The Netherlands
700    1_
$a Petr, Jan $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands $u Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiopharmaceutical Cancer Research, Dresden, Germany
700    1_
$a Barkhof, Frederik $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands $u Amsterdam Neuroscience, Brain Imaging, Amsterdam, The Netherlands $u Queen Square Institute of Neurology and Center for Medical Image Computing, University College London, London, UK
700    1_
$a Keil, Vera C $u Amsterdam UMC Location Vrije Universiteit Amsterdam, Radiology & Nuclear Medicine Department, Amsterdam, The Netherlands. v.c.w.keil@amsterdamumc.nl $u Cancer Center Amsterdam, Imaging and Biomarkers, Amsterdam, The Netherlands. v.c.w.keil@amsterdamumc.nl $u Amsterdam Neuroscience, Brain Imaging, Amsterdam, The Netherlands. v.c.w.keil@amsterdamumc.nl $1 https://orcid.org/0000000186993506
773    0_
$w MED00009646 $t European radiology $x 1432-1084 $g Roč. 35, č. 3 (2025), s. 1242-1254
856    41
$u https://pubmed.ncbi.nlm.nih.gov/39425768 $y Pubmed
910    __
$a ABA008 $b sig $c sign $y - $z 0
990    __
$a 20250415 $b ABA008
991    __
$a 20250429135035 $b ABA008
999    __
$a ok $b bmc $g 2311250 $s 1246842
BAS    __
$a 3
BAS    __
$a PreBMC-MEDLINE
BMC    __
$a 2025 $b 35 $c 3 $d 1242-1254 $e 20241019 $i 1432-1084 $m European radiology $n Eur Radiol $x MED00009646
LZP    __
$a Pubmed-20250415

Find record

Citation metrics

Archiving options