% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@ARTICLE{Salome:274490,
author = {P. Salome$^*$ and F. Sforazzini$^*$ and G. Grugnara and A.
Kudak$^*$ and M. Dostal$^*$ and C. Herold-Mende and S.
Heiland and J. Debus$^*$ and A. Abdollahi$^*$ and M.
Knoll$^*$},
title = {{MR}-{C}lass: {A} {P}ython {T}ool for {B}rain {MR} {I}mage
{C}lassification {U}tilizing {O}ne-vs-{A}ll {DCNN}s to
{D}eal with the {O}pen-{S}et {R}ecognition {P}roblem.},
journal = {Cancers},
volume = {15},
number = {6},
issn = {2072-6694},
address = {Basel},
publisher = {MDPI},
reportid = {DKFZ-2023-00633},
pages = {1820},
year = {2023},
note = {#EA:E210#LA:E210#},
abstract = {MR image classification in datasets collected from multiple
sources is complicated by inconsistent and missing DICOM
metadata. Therefore, we aimed to establish a method for the
efficient automatic classification of MR brain
sequences.Deep convolutional neural networks (DCNN) were
trained as one-vs-all classifiers to differentiate between
six classes: T1 weighted (w), contrast-enhanced T1w, T2w,
T2w-FLAIR, ADC, and SWI. Each classifier yields a
probability, allowing threshold-based and relative
probability assignment while excluding images with low
probability (label: unknown, open-set recognition problem).
Data from three high-grade glioma (HGG) cohorts was
assessed; C1 (320 patients, 20,101 MRI images) was used for
training, while C2 (197, 11,333) and C3 (256, 3522) were for
testing. Two raters manually checked images through an
interactive labeling tool. Finally, MR-Class' added value
was evaluated via radiomics model performance for
progression-free survival (PFS) prediction in C2, utilizing
the concordance index (C-I).Approximately $10\%$ of
annotation errors were observed in each cohort between the
DICOM series descriptions and the derived labels. MR-Class
accuracy was $96.7\%$ $[95\%$ Cl: 95.8, 97.3] for C2 and
$94.4\%$ [93.6, 96.1] for C3. A total of 620 images were
misclassified; manual assessment of those frequently showed
motion artifacts or alterations of anatomy by large tumors.
Implementation of MR-Class increased the PFS model C-I by
$14.6\%$ on average, compared to a model trained without
MR-Class.We provide a DCNN-based method for the sequence
classification of brain MR images and demonstrate its
usability in two independent HGG datasets.},
keywords = {artificial intelligence (AI) (Other) / content-based image
classification (Other) / convolutional neural networks (CNN)
(Other) / data curation and preparation (Other) / deep
learning (Other)},
cin = {E210 / HD01 / E050},
ddc = {610},
cid = {I:(DE-He78)E210-20160331 / I:(DE-He78)HD01-20160331 /
I:(DE-He78)E050-20160331},
pnm = {315 - Bildgebung und Radioonkologie (POF4-315)},
pid = {G:(DE-HGF)POF4-315},
typ = {PUB:(DE-HGF)16},
pubmed = {pmid:36980707},
doi = {10.3390/cancers15061820},
url = {https://inrepo02.dkfz.de/record/274490},
}