Back to the list
Congress: ECR25
Poster Number: C-13515
Type: Poster: EPOS Radiologist (scientific)
Authorblock: M. Correia De Verdier1, R. Saluja2, L. Gagnon3, U. Baid4, A. Abayazeed5, R. Huang6, S. Bakas4, E. Calabrese7, J. D. Rudie8; 1Uppsala/SE, 2New York, NY/US, 3Quebec City, QC/CA, 4Indianapolis, IN/US, 5Stanford, CA/US, 6Boston, MA/US, 7Durham, NC/US, 8La Jolla, CA/US
Disclosures:
Maria Correia De Verdier: Nothing to disclose
Rachit Saluja: Nothing to disclose
Louis Gagnon: Nothing to disclose
Ujjwal Baid: Nothing to disclose
Aly Abayazeed: Nothing to disclose
Raymond Huang: Nothing to disclose
Spyridon Bakas: Nothing to disclose
Evan Calabrese: Nothing to disclose
Jeffrey D. Rudie: Nothing to disclose
Keywords: Artificial Intelligence, Neuroradiology brain, MR, Neural networks, Segmentation, Cancer
Purpose Diffuse gliomas are the most common malignant primary brain tumors in adults and one of the deadliest types of cancer [1]. There are many challenges in treatment and monitoring due to genetic diversity and high intrinsic heterogeneity in appearance, shape, histology, and treatment response. Treatments include surgery, radiation, and systemic therapies, with magnetic resonance imaging (MRI) playing a key role in treatment planning and post-treatment longitudinal assessment. Post-treatment imaging of diffuse gliomas is therefore a fundamental part of patient management...
Read more Methods and materials DataThe 2024 BraTS-PTG dataset consists of retrospective, multi-institutional cohorts of patients diagnosed with diffuse gliomas having already undergone treatment, which may include surgery, radiation, and/or systemic therapy. The patients have been clinically scanned with multiparametric MRI (mpMRI) acquisition protocols, including the following sequences: pre-contrast T1-weighted (T1) contrast-enhanced T1-weighted (T1-Gd) T2-weighted (T2) T2-weighted fluid-attenuated inversion recovery (FLAIR) Data Preprocessing The preprocessing pipeline (Figure 1) was similar to the one evaluated and followed by the BraTS 2017-2023 challenges [6]. [fig 1]  The T1, T1-Gd, T2, and FLAIR sequences were extracted...
Read more Results The dataset comprises 1936 cases, and data was contributed from six different academic medical centers (Figure 5, Table 1). [fig 5] [fig 1] This dataset has been used by BraTS-PTG challenge participants to develop, containerize, and evaluate their automated segmentation models, predicting the six sub-regions on validation data from April 2024 through July 2024. Thirty-seven teams participated in the validation phase and evaluated their models on the validation data. Six teams participated in the test phase in August 2024, and their models were...
Read more Conclusion The BraTS-PTG establishes a benchmark and defines a community standard for automated segmentation on post-treatment MRI, utilizing the largest, publicly available, expert-annotated post-treatment glioma MRI dataset. Compared with previous BraTS challenges, it comprises an entirely new dataset of exclusively post-treatment diffuse gliomas and includes a novel tissue class, the RC. Results show promising performance in segmenting sub-regions within the test dataset, indicating the potential of these models in capturing complex tumor structures. Given the significant effort required to annotate multilabel...
Read more References [1] M. Price, C. Ballard, J. Benedetti, C. Neff, G. Cioffi, K.A. Waite, C. Kruchko, J.S. Barnholtz-Sloan, Q.T. Ostrom, CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2017-2021., Neuro-Oncol. 26 (2024) vi1–vi85. https://doi.org/10.1093/neuonc/noae145.[2] P. Kickingereder, F. Isensee, I. Tursunova, J. Petersen, U. Neuberger, D. Bonekamp, G. Brugnara, M. Schell, T. Kessler, M. Foltyn, Automated quantitative tumour response assessment of MRI in neuro-oncology with artificial neural networks: a multicentre, retrospective study, Lancet...
Read more
GALLERY