Skip to content
DraconiticResearch
Search Sources API
201,683 studies31 sourcesUpdated 2 min ago
Peer-reviewedCC BY 4.0Cancers

Explainable DSC Perfusion Radiomics for Post-Treatment Glioblastoma Assessment Using a Soft-Voting Machine-Learning Ensemble

  • Rafail C. Christodoulou1
  • Giorgos Christofi2
  • Georgios Vamvouras3
  • Mateus A. Esmeraldo4
  • Platon S. Papageorgiou5
  • Evros Vassiliou6
  • Elena E. Solomou7
  • Sokratis G. Papageorgiou8
  • Michalis F. Georgiou9
9 affiliations
  1. 1Division of Neuroimaging and Neurointervention, Department of Radiology, Stanford University, Stanford, CA 94305, USA
  2. 2Faculty of Electrical Engineering, Mathematics and Computer Science, Delft University of Technology, 2628 CD Delft, The Netherlands
  3. 3Department of Electrical and Computer Engineering, National Technical University of Athens (NTUA), 15780 Athens, Greece
  4. 4Department of Radiology, Stanford University, Stanford, CA 94305, USA
  5. 5Department of Medicine, National and Kapodistrian University of Athens, 11527 Athens, Greece
  6. 6Department of Biological Sciences, Kean University, Union, NJ 07083, USA
  7. 7Internal Medicine-Hematology, University of Patras Medical School, 26500 Rion, Greece
  8. 81st Department of Neurology, Eginition Hospital, Medical School, National and Kapodistrian University of Athens, 11528 Athens, Greece
  9. 9Department of Radiology, University of Miami, Miami, FL 33136, USA
Open at sourceDOICite

Abstract

Background/Objectives: After chemoradiation for glioblastoma, residual or recurrent tumor (tumor-positive disease, TP) and treatment-related change (TRC), which includes pseudoprogression and radiation necrosis, look alike on conventional MRI. We tested whether radiomic features from dynamic susceptibility contrast (DSC) perfusion MRI alone can separate them. Methods: We analyzed 224 post-treatment MRI examinations from 168 patients in the public UCSD-PTGBM collection. A machine-learning ensemble was trained on 1223 radiomic features from the leakage-corrected relative cerebral blood volume (rCBV) map, with training (109 patients, 143 examinations) and test (59 patients, 81 examinations) data separated at the patient level and the model fixed before testing. The analysis was repeated on the cerebral blood flow (CBF) map from the same scans, and DSC was compared with arterial spin labeling (ASL) in patients who had both. Results: On the independent test set (patient-level analysis), the model distinguished TP from TRC with an area under the receiver operating characteristic curve (AUC) of 0.79 (95% CI 0.67–0.89). It detected 31 of 49 patients with tumor (sensitivity 63%) and correctly identified all 10 patients with TRC (specificity 100%); at a predefined higher-sensitivity cut-off, sensitivity was 69% and specificity 80%. With the CBF map, the AUC did not differ significantly (0.78; p = 0.89), but there was a different balance of sensitivity and specificity. In an exploratory subset of 36 patients, DSC and ASL did not differ significantly (AUC 0.82 vs. 0.71; p = 0.52). Conclusions: DSC perfusion radiomics alone identified tumor with high specificity but missed about one-third of tumors: a positive result may support a diagnosis of tumor, whereas a negative result cannot exclude it. External, multicenter validation is needed before clinical use.

Related

  1. Ultrasound-based deep learning radiomics with explainable machine learning for predicting benign and malignant thyroid nodules in lymphocytic thyroiditis: a multicenter study
    Ji Zhou, Wenwu Lu, Wei Wei +6
    Quantitative imaging in medicine and surgeryPeer-reviewedCC BY-NC-NDFull textMedical imaging & radiologyEurope PMC
  2. Vascular habitat analysis using dynamic susceptibility contrast-enhanced perfusion MRI for differentiating glioblastoma from solitary brain metastases.
    Xuan Li, Hanwei Wang, Linlan Zeng +6
    Frontiers in oncologyPeer-reviewedCC BYFull textHuman geneticsEurope PMCCrossref
  3. Integrative Single-Cell and Radiomics Analyses Identify N-Acetylneuraminate Synthase (NANS)-High Signature Associated with Adverse Prognosis and Immune Remodeling in Glioblastoma
    Jianan Chen, Qiong Wu, Hannah M. Cardenas +6
    International Journal of Molecular SciencesPeer-reviewedCC BY 4.0Human geneticsCrossref
  4. Decoding glioblastoma recurrence: harnessing radiomics and machine learning for early prediction
    Lijuan Gao, Tao Yuan, Yawu Liu +5
    Frontiers in NeurologyPeer-reviewedCC BY 4.0NeurologyCrossref
  5. Volumetric analysis and MRI radiomics for response assessment in canine glioma immunotherapy: an exploratory feasibility study
    Ana González Aranda, Pablo Delgado-Bonet, Isidro Mateo +2
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyCrossref

Cite

BibTeXRIS
BibTeX
@article{christodoulou2026explainable,
  title = {{Explainable DSC Perfusion Radiomics for Post-Treatment Glioblastoma Assessment Using a Soft-Voting Machine-Learning Ensemble}},
  author = {Rafail C. Christodoulou and Giorgos Christofi and Georgios Vamvouras and Mateus A. Esmeraldo and Platon S. Papageorgiou and Evros Vassiliou and Elena E. Solomou and Sokratis G. Papageorgiou and Michalis F. Georgiou},
  journal = {Cancers},
  year = {2026},
  month = nov,
  doi = {10.3390/cancers18213567},
  url = {https://www.mdpi.com/2072-6694/18/21/3567},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Explainable DSC Perfusion Radiomics for Post-Treatment Glioblastoma Assessment Using a Soft-Voting Machine-Learning Ensemble
AU  - Rafail C. Christodoulou
AU  - Giorgos Christofi
AU  - Georgios Vamvouras
AU  - Mateus A. Esmeraldo
AU  - Platon S. Papageorgiou
AU  - Evros Vassiliou
AU  - Elena E. Solomou
AU  - Sokratis G. Papageorgiou
AU  - Michalis F. Georgiou
PY  - 2026
DA  - 2026/11/06
JO  - Cancers
DO  - 10.3390/cancers18213567
UR  - https://www.mdpi.com/2072-6694/18/21/3567
AB  - Background/Objectives: After chemoradiation for glioblastoma, residual or recurrent tumor (tumor-positive disease, TP) and treatment-related change (TRC), which includes pseudoprogression and radiation necrosis, look alike on conventional MRI. We tested whether radiomic features from dynamic susceptibility contrast (DSC) perfusion MRI alone can separate them. Methods: We analyzed 224 post-treatment MRI examinations from 168 patients in the public UCSD-PTGBM collection. A machine-learning ensemble was trained on 1223 radiomic features from the leakage-corrected relative cerebral blood volume (rCBV) map, with training (109 patients, 143 examinations) and test (59 patients, 81 examinations) data separated at the patient level and the model fixed before testing. The analysis was repeated on the cerebral blood flow (CBF) map from the same scans, and DSC was compared with arterial spin labeling (ASL) in patients who had both. Results: On the independent test set (patient-level analysis), the model distinguished TP from TRC with an area under the receiver operating characteristic curve (AUC) of 0.79 (95% CI 0.67–0.89). It detected 31 of 49 patients with tumor (sensitivity 63%) and correctly identified all 10 patients with TRC (specificity 100%); at a predefined higher-sensitivity cut-off, sensitivity was 69% and specificity 80%. With the CBF map, the AUC did not differ significantly (0.78; p = 0.89), but there was a different balance of sensitivity and specificity. In an exploratory subset of 36 patients, DSC and ASL did not differ significantly (AUC 0.82 vs. 0.71; p = 0.52). Conclusions: DSC perfusion radiomics alone identified tumor with high specificity but missed about one-third of tumors: a positive result may support a diagnosis of tumor, whereas a negative result cannot exclude it. External, multicenter validation is needed before clinical use.
SN  - 2072-6694
ER  -