Pixel-Wise Prediction of Glioblastoma Recurrence Using Mid-Treatment MRI and Deep Learning
4 affiliations
- 1Physical Sciences, Sunnybrook Research Institute, Toronto, ON M4N 3M5, Canada
- 2Department of Radiation Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, ON M4N 3M5, Canada
- 3Department of Medical Biophysics, University of Toronto, Toronto, ON M5G 2C4, Canada
- 4Department of Radiation Oncology, University of Toronto, Toronto, ON M4N 3M5, Canada
Abstract
Introduction: Glioblastoma (GBM) commonly recurs close to the original gross tumor volume (GTV), suggesting treatment-resistant subregions may be present before radiologically visible progression. This study investigated whether deep learning on apparent diffusion coefficient (ADC) maps could generate pixel-wise probability maps of treatment-resistant regions within the GTV. Methods: The training/validation set included images of 64 in-house patients after radiation fraction 10 (Fx10) and 40 external patients before treatment used as an auxiliary cross-domain training cohort; the held-out test set included 16 in-house Fx10 patients. Resistant regions were defined as the overlap between the GTV from scans acquired at radiation fraction 10 and subsequent recurrence after finishing treatment. Models were trained on 2D GTV-cropped slices with patient-level separation to prevent data leakage. Single- and multimodal MRI inputs were compared, and ADC-based models (including baseline UNet, pretrained ResNet34, BiomedCLIP-assisted UNet, and a BLIP image-text UNet) were later evaluated. Performance was evaluated using conventional metrics, including ROC AUC, PR AUC, F1-score, recall, specificity, and precision. Results: ADC was the most informative single modality, achieving 69% and 80% for Receiver Operating Characteristic (ROC)- and Precision-Recall (PR)-Area Under the Curve (AUC), respectively. In the ADC-based architecture comparison, no model was consistently superior across all metrics. The baseline UNet achieved the highest accuracy, F1-score, and recall, whereas pretrained ResNet34 and BLIP-assisted models showed slightly higher ROC AUC than the baseline UNet (both 70%). Error analysis showed that performance decreased substantially for small or sparse resistant regions, and false-negative errors were enriched near resistant-region and GTV boundaries. Conclusions: ADC-based deep learning showed feasibility for pixel-wise spatial mapping of treatment-resistant glioblastoma regions within the GTV. A UNet trained from scratch provided the strongest sensitivity/F1 trade-off, whereas pretrained and VLM-assisted models provided only metric-specific benefits. Model comparisons showed that BiomedCLIP-assisted and BLIP image-text UNet models did not clearly improve fine spatial segmentation performance over the baseline UNet.
Related
- Decoding glioblastoma recurrence: harnessing radiomics and machine learning for early prediction
- Development and internal validation of a Cox-based recurrence risk prediction model after endovenous laser treatment for lower-extremity varicose veins.
- Explainable DSC Perfusion Radiomics for Post-Treatment Glioblastoma Assessment Using a Soft-Voting Machine-Learning Ensemble
- Neoplastic and non-neoplastic astrocyte states converge on mesenchymal and injury-response programs during glioblastoma recurrence
- Classification of recurrence status after surgical treatment of chronic subdural hemorrhage – A machine learning approach
BibTeX
@article{rezvanjou2026pixelwise,
title = {{Pixel-Wise Prediction of Glioblastoma Recurrence Using Mid-Treatment MRI and Deep Learning}},
author = {Sara Rezvanjou and Laurentius Oscar Osapoetra and Daniel Moore-Palhares and Liam S. P. Lawrence and Angus Z. Lau and Arjun Sahgal and Gregory J. Czarnota},
journal = {Cancers},
year = {2026},
month = oct,
doi = {10.3390/cancers18213465},
url = {https://www.mdpi.com/2072-6694/18/21/3465},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Pixel-Wise Prediction of Glioblastoma Recurrence Using Mid-Treatment MRI and Deep Learning
AU - Sara Rezvanjou
AU - Laurentius Oscar Osapoetra
AU - Daniel Moore-Palhares
AU - Liam S. P. Lawrence
AU - Angus Z. Lau
AU - Arjun Sahgal
AU - Gregory J. Czarnota
PY - 2026
DA - 2026/10/27
JO - Cancers
DO - 10.3390/cancers18213465
UR - https://www.mdpi.com/2072-6694/18/21/3465
AB - Introduction: Glioblastoma (GBM) commonly recurs close to the original gross tumor volume (GTV), suggesting treatment-resistant subregions may be present before radiologically visible progression. This study investigated whether deep learning on apparent diffusion coefficient (ADC) maps could generate pixel-wise probability maps of treatment-resistant regions within the GTV. Methods: The training/validation set included images of 64 in-house patients after radiation fraction 10 (Fx10) and 40 external patients before treatment used as an auxiliary cross-domain training cohort; the held-out test set included 16 in-house Fx10 patients. Resistant regions were defined as the overlap between the GTV from scans acquired at radiation fraction 10 and subsequent recurrence after finishing treatment. Models were trained on 2D GTV-cropped slices with patient-level separation to prevent data leakage. Single- and multimodal MRI inputs were compared, and ADC-based models (including baseline UNet, pretrained ResNet34, BiomedCLIP-assisted UNet, and a BLIP image-text UNet) were later evaluated. Performance was evaluated using conventional metrics, including ROC AUC, PR AUC, F1-score, recall, specificity, and precision. Results: ADC was the most informative single modality, achieving 69% and 80% for Receiver Operating Characteristic (ROC)- and Precision-Recall (PR)-Area Under the Curve (AUC), respectively. In the ADC-based architecture comparison, no model was consistently superior across all metrics. The baseline UNet achieved the highest accuracy, F1-score, and recall, whereas pretrained ResNet34 and BLIP-assisted models showed slightly higher ROC AUC than the baseline UNet (both 70%). Error analysis showed that performance decreased substantially for small or sparse resistant regions, and false-negative errors were enriched near resistant-region and GTV boundaries. Conclusions: ADC-based deep learning showed feasibility for pixel-wise spatial mapping of treatment-resistant glioblastoma regions within the GTV. A UNet trained from scratch provided the strongest sensitivity/F1 trade-off, whereas pretrained and VLM-assisted models provided only metric-specific benefits. Model comparisons showed that BiomedCLIP-assisted and BLIP image-text UNet models did not clearly improve fine spatial segmentation performance over the baseline UNet.
SN - 2072-6694
ER -