% 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{Scheffler:298913,
author = {P. Scheffler and J. Straehle and A. El Rahal and D. Erny
and B. Mizaikoff and I. Vasilikos and M. Prinz and V. A.
Coenen and J. Kühn and F. Scherer$^*$ and D. H. Heiland$^*$
and O. Schnell and R. Roelz and J. Beck and P. C. Reinacher
and N. Neidert$^*$},
title = {{I}ntraoperative classification of {CNS} lymphoma and
glioblastoma by {AI}-based analysis of {S}timulated {R}aman
{H}istology ({SRH})},
journal = {Brain and spine},
volume = {5},
issn = {2772-5294},
address = {[Amsterdam]},
publisher = {Elsevier B.V.},
reportid = {DKFZ-2025-00349},
pages = {104187},
year = {2025},
abstract = {Introduction: Early diagnosis is important to differentiate
central nervous system lymphomas (CNSL) from themain
differential diagnosis, glioblastoma (GBM), because of
different primary treatment modalities for theseentities.
Due to neurological deficits, diagnostic stereotactic
biopsies often need to be performed urgently. In thissetting
the availability of an intraoperative neuropathological
assessment is limited.Research question: This study uses
AI-based analysis of Stimulated Raman Histology (SRH) to
establish a classifierdistinguishing CNSL from glioblastoma
in an intraoperative setting.Material and methods: We
collected 126 intraoperative SRH images from 40 patients
diagnosed with CNSL. TheseSRH images were divided into
patches, measuring 224 x 224 pixels each. Additionally, we
used a comparativedataset of 87 SRH images from 31 patients
with GBM as a control group to train and validate a neural
networkbased on the CTransPath architecture. Two distinct
diagnostic categories were established: “Lymphoma”
and“Glioblastoma".Results: Our model demonstrated an
accuracy rate of $92.5\%$ in distinguishing between lymphoma
and glioblastoma. Analysis of our test dataset showed a
sensitivity of $84.2\%$ and a specificity of $100\%$ in the
detection ofCNSL, demonstrating performance comparable to
standard intraoperative histopathological
analysis.Discussion and conclusion: The use of AI-driven
analysis of SRH images holds promise for intraoperative
tissueexamination of stereotactic biopsies with suspected
CNSL en par with the current gold standard. This study
couldimprove the management of these cases especially in the
emergency setting when conventional
intraoperativeneuropathological evaluation is unavailable.},
cin = {FR01},
ddc = {610},
cid = {I:(DE-He78)FR01-20160331},
pnm = {899 - ohne Topic (POF4-899)},
pid = {G:(DE-HGF)POF4-899},
typ = {PUB:(DE-HGF)16},
doi = {10.1016/j.bas.2025.104187},
url = {https://inrepo02.dkfz.de/record/298913},
}