Toral R. Patel
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Toral R. Patel has a total of 226 co-authors. Here are the top co-authors by number of shared publications:
- Joseph A. Maldjian
Magnetoencephalography Center of Excellence, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390, USA; Department of Radiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390, USA
12 shared publications - Baowei Fei
University of Texas at Dallas, Center for Imaging and Surgical Innovation, Richardson, Texas, United States
10 shared publications - Ananth J. Madhuranthakam
Department of Radiology, UT Southwestern Medical Center, Dallas, USA
8 shared publications - Benjamin C. Wagner
Magnetoencephalography Center of Excellence, University of Texas Southwestern Medical Center (UTSW), Dallas, USA
7 shared publications - Marco C. Pinho
University of Texas Southwestern Medical Center, Dallas, USA
7 shared publications - Chandan Ganesh Bangalore Yogananda
University of Texas Southwestern Medical Center
6 shared publications
Latest Publications
- 8 July 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Translational Medicine
- Vol. 18 (857) , eadx8398
EGFR inhibition down-regulates MGMT and enhances responsiveness to temozolomide in glioblastoma
Glioblastoma (GBM) is a devastating cancer with a dismal prognosis. Current treatment includes temozolomide (TMZ), which is more effective in about 50% of GBMs that have O 6 - methylguanine DNA methyltransferase ( MGMT ) promoter methylation. MGMT is a DNA repair protein that reverses TMZ-induced DNA damage. EGFR is a prime oncogene in GBM. Here, we report that EGFR inhibition induced the down-regulation of MGMT in GBM cells, revealing a previously unidentified link between EGFR signaling and response to TMZ. EGFR inhibition led to activation of two transcription factors, activator protein-1 (AP-1), which repressed MGMT transcription, and nuclear factor κB (NF-κB), which up-regulated MGMT transcription. EGFR inhibition also induced AP-1–mediated transcription of miR-616. miR-616 inhibited both MGMT translation and NF-κB activation. Thus, the overall effect of EGFR inhibition was down-regulation of MGMT expression. In addition, we provided an explanation for the prior failure of clinical trials that used concomitant EGFR tyrosine kinase inhibitors (TKIs) and TMZ. TMZ up-regulated MGMT, and concomitant treatment with EGFR TKIs and TMZ failed to down-regulate MGMT. However, pretreatment with EGFR TKIs followed by TMZ efficiently down-regulated MGMT and enhanced TMZ sensitivity in experimental models. Posttreatment tumor tissues from two clinical trials were used to validate these findings. We demonstrated that EGFR inhibitors induced down-regulation of MGMT in posttreatment resected tumor tissues from patients with GBM and the failure of EGFR inhibition to down-regulate MGMT if TMZ was used concomitantly. These data support using EGFR TKIs before TMZ treatment as a therapeutic approach in MGMT unmethylated GBM.
- 8 July 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Translational Medicine
- Vol. 18 (857) , eadx8398
EGFR inhibition down-regulates MGMT and enhances responsiveness to temozolomide in glioblastoma
Glioblastoma (GBM) is a devastating cancer with a dismal prognosis. Current treatment includes temozolomide (TMZ), which is more effective in about 50% of GBMs that have O 6 - methylguanine DNA methyltransferase ( MGMT ) promoter methylation. MGMT is a DNA repair protein that reverses TMZ-induced DNA damage. EGFR is a prime oncogene in GBM. Here, we report that EGFR inhibition induced the down-regulation of MGMT in GBM cells, revealing a previously unidentified link between EGFR signaling and response to TMZ. EGFR inhibition led to activation of two transcription factors, activator protein-1 (AP-1), which repressed MGMT transcription, and nuclear factor κB (NF-κB), which up-regulated MGMT transcription. EGFR inhibition also induced AP-1–mediated transcription of miR-616. miR-616 inhibited both MGMT translation and NF-κB activation. Thus, the overall effect of EGFR inhibition was down-regulation of MGMT expression. In addition, we provided an explanation for the prior failure of clinical trials that used concomitant EGFR tyrosine kinase inhibitors (TKIs) and TMZ. TMZ up-regulated MGMT, and concomitant treatment with EGFR TKIs and TMZ failed to down-regulate MGMT. However, pretreatment with EGFR TKIs followed by TMZ efficiently down-regulated MGMT and enhanced TMZ sensitivity in experimental models. Posttreatment tumor tissues from two clinical trials were used to validate these findings. We demonstrated that EGFR inhibitors induced down-regulation of MGMT in posttreatment resected tumor tissues from patients with GBM and the failure of EGFR inhibition to down-regulate MGMT if TMZ was used concomitantly. These data support using EGFR TKIs before TMZ treatment as a therapeutic approach in MGMT unmethylated GBM.
- 7 May 2026
- Published by American Society of Neuroradiology (ASNR) in American Journal of Neuroradiology
- Vol. 47 (11) , 2759-2766
Enhancing 1p/19q Classification in Brain Gliomas Using Isocitrate Dehydrogenase Status: A Deep Learning Study
IDH mutation & 1p/19q codeletion are critical biomarkers for glioma diagnosis & therapy. 1p/19q codeletion occurs exclusively in IDH-mutated gliomas. In this study, we developed a 2-stage, non-invasive, MRI-based deep learning method that leverages IDH status to enhance 1p/19q predictions. Multi-contrast brain tumor MRI & genomic information were obtained from five publicly available (TCIA, UCSF, EGD, UPenn & LGG), and three in-house/collaborator institutions (UTSW, NYU, UWM). Subjects were screened for the availability of IDH & 1p/19q status as well as T1, T1CE, T2, FLAIR MR images. For training purposes, missing T1 and FLAIR contrasts for the LGG database were generated using an in-house multi-contrast simulator. Two separate U-Nets were developed for 1p/19q-classification: a multi-contrast network (MC-Net) and a T2w-only network (T2-Net). A separate U-Net was developed for IDH classification (IDH-net). A total of 2044 subjects were used in training and testing IDH-Net, and 1426 subjects were used in training and testing the 1p/19q models. The IDH-Net was trained using subjects from TCIA, UTSW, and UPenn. The 1p/19q networks were trained using subjects from TCIA, UTSW, and LGG. The trained networks were tested on true held-out cases from NYU, UWM, EGD, and UCSF. In the 2-stage approach, subjects were initially classified for IDH status using IDH-Net. Predicted IDH-wildtype cases default to 1p/19q non-codeleted. Then the IDH-mutated cases were further classified for 1p/19q status using the 1p/19q-networks. IDH-Net achieved a classification accuracy of 93.7%. 1p/19q MC-Net & T2-Net achieved classification accuracies of 86.5% & 86.0%, respectively. In the 2-stage approach, 1p/19q MC-Net and T2-Net achieved accuracies of 91.5% & 91.2% respectively, improving the classification accuracy by ∼5%. This study demonstrates the effectiveness of leveraging IDH status to enhance 1p/19q classification. A ∼5% increase in classification accuracy was achieved when using the 2-stage approach, using IDH-Net to gate 1p/19q predictions. The developed method offers a reliable, non-invasive approach to determine important biomarkers for glioma diagnosis.
- 7 May 2026
- Published by American Society of Neuroradiology (ASNR) in American Journal of Neuroradiology
- Vol. 47 (11) , 2759-2766
Enhancing 1p/19q Classification in Brain Gliomas Using Isocitrate Dehydrogenase Status: A Deep Learning Study
IDH mutation & 1p/19q codeletion are critical biomarkers for glioma diagnosis & therapy. 1p/19q codeletion occurs exclusively in IDH-mutated gliomas. In this study, we developed a 2-stage, non-invasive, MRI-based deep learning method that leverages IDH status to enhance 1p/19q predictions. Multi-contrast brain tumor MRI & genomic information were obtained from five publicly available (TCIA, UCSF, EGD, UPenn & LGG), and three in-house/collaborator institutions (UTSW, NYU, UWM). Subjects were screened for the availability of IDH & 1p/19q status as well as T1, T1CE, T2, FLAIR MR images. For training purposes, missing T1 and FLAIR contrasts for the LGG database were generated using an in-house multi-contrast simulator. Two separate U-Nets were developed for 1p/19q-classification: a multi-contrast network (MC-Net) and a T2w-only network (T2-Net). A separate U-Net was developed for IDH classification (IDH-net). A total of 2044 subjects were used in training and testing IDH-Net, and 1426 subjects were used in training and testing the 1p/19q models. The IDH-Net was trained using subjects from TCIA, UTSW, and UPenn. The 1p/19q networks were trained using subjects from TCIA, UTSW, and LGG. The trained networks were tested on true held-out cases from NYU, UWM, EGD, and UCSF. In the 2-stage approach, subjects were initially classified for IDH status using IDH-Net. Predicted IDH-wildtype cases default to 1p/19q non-codeleted. Then the IDH-mutated cases were further classified for 1p/19q status using the 1p/19q-networks. IDH-Net achieved a classification accuracy of 93.7%. 1p/19q MC-Net & T2-Net achieved classification accuracies of 86.5% & 86.0%, respectively. In the 2-stage approach, 1p/19q MC-Net and T2-Net achieved accuracies of 91.5% & 91.2% respectively, improving the classification accuracy by ∼5%. This study demonstrates the effectiveness of leveraging IDH status to enhance 1p/19q classification. A ∼5% increase in classification accuracy was achieved when using the 2-stage approach, using IDH-Net to gate 1p/19q predictions. The developed method offers a reliable, non-invasive approach to determine important biomarkers for glioma diagnosis.
- 5 April 2026
- Published by Elsevier in Biomaterials
- Vol. 333, 124198
Enhanced delivery of low-density lipoprotein-based nanoparticles to the mouse glioblastoma using focused ultrasound as a novel therapy
- 5 April 2026
- Published by Elsevier in Biomaterials
- Vol. 333, 124198
Enhanced delivery of low-density lipoprotein-based nanoparticles to the mouse glioblastoma using focused ultrasound as a novel therapy
- Open Access
- 23 October 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-11
Categorical and phenotypic image synthetic learning as an alternative to federated learning
Multi-center collaborations are crucial in developing robust and generalizable machine learning models in medical imaging. Traditional methods, such as centralized data sharing or federated learning (FL), face challenges, including privacy issues, communication burdens, and synchronization complexities. We present CATegorical and PHenotypic Image SyntHetic learnING (CATphishing), an alternative to FL using Latent Diffusion Models (LDM) to generate synthetic multi-contrast three-dimensional magnetic resonance imaging data for downstream tasks, eliminating the need for raw data sharing or iterative inter-site communication. Each institution trains an LDM to capture site-specific data distributions, producing synthetic samples aggregated at a central server. We evaluate CATphishing using data from 2491 patients across seven institutions for isocitrate dehydrogenase mutation classification and three-class tumor-type classification. CATphishing achieves accuracy comparable to centralized training and FL, with synthetic data exhibiting high fidelity. This method addresses privacy, scalability, and communication challenges, offering a promising alternative for collaborative artificial intelligence development in medical imaging.
- 23 October 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-11
Open AccessCategorical and phenotypic image synthetic learning as an alternative to federated learning
Multi-center collaborations are crucial in developing robust and generalizable machine learning models in medical imaging. Traditional methods, such as centralized data sharing or federated learning (FL), face challenges, including privacy issues, communication burdens, and synchronization complexities. We present CATegorical and PHenotypic Image SyntHetic learnING (CATphishing), an alternative to FL using Latent Diffusion Models (LDM) to generate synthetic multi-contrast three-dimensional magnetic resonance imaging data for downstream tasks, eliminating the need for raw data sharing or iterative inter-site communication. Each institution trains an LDM to capture site-specific data distributions, producing synthetic samples aggregated at a central server. We evaluate CATphishing using data from 2491 patients across seven institutions for isocitrate dehydrogenase mutation classification and three-class tumor-type classification. CATphishing achieves accuracy comparable to centralized training and FL, with synthetic data exhibiting high fidelity. This method addresses privacy, scalability, and communication challenges, offering a promising alternative for collaborative artificial intelligence development in medical imaging.
- Open Access
Gliomas phenocopy an inborn error of metabolism to drive neuronal activity and tumor growth
SUMMARY: The metabolic hallmarks of high-grade glioma (HGG) are not fully understood. Human brain tissue metabolomics revealed that the creatine synthesis pathway intermediate guanidinoacetate (GAA) accumulated ∼100-fold in HGGs relative to controls, which was caused by imbalanced activities of enzymes in this pathway. Glioma cells secreted GAA rather than using it to produce creatine, implicating an unexpected function. GAA accumulates in GAMT deficiency, an inborn error of metabolism, and elevates neuronal excitability. Neuronal excitability is also increased in glioma and drives tumor growth through neuron-glioma interactions. We hypothesized that glioma-generated GAA excites surrounding neurons. Indeed, GAA induced neuronal hyperactivity by activating GABA A receptors and causing depolarizing GABA currents in glioma-associated neurons with dysregulated chloride homeostasis. Depleting tumoral GAA decreased electrochemical activity, neuron-glioma interactions, and tumor aggressiveness. Our findings unveil a new mechanism linking cancer metabolism with cancer neuroscience and leverage human genetics to nominate GAA synthesis as a target in gliomas.
Open AccessGliomas phenocopy an inborn error of metabolism to drive neuronal activity and tumor growth
SUMMARY: The metabolic hallmarks of high-grade glioma (HGG) are not fully understood. Human brain tissue metabolomics revealed that the creatine synthesis pathway intermediate guanidinoacetate (GAA) accumulated ∼100-fold in HGGs relative to controls, which was caused by imbalanced activities of enzymes in this pathway. Glioma cells secreted GAA rather than using it to produce creatine, implicating an unexpected function. GAA accumulates in GAMT deficiency, an inborn error of metabolism, and elevates neuronal excitability. Neuronal excitability is also increased in glioma and drives tumor growth through neuron-glioma interactions. We hypothesized that glioma-generated GAA excites surrounding neurons. Indeed, GAA induced neuronal hyperactivity by activating GABA A receptors and causing depolarizing GABA currents in glioma-associated neurons with dysregulated chloride homeostasis. Depleting tumoral GAA decreased electrochemical activity, neuron-glioma interactions, and tumor aggressiveness. Our findings unveil a new mechanism linking cancer metabolism with cancer neuroscience and leverage human genetics to nominate GAA synthesis as a target in gliomas.
- 23 June 2025
- Published by SPIE-Intl Soc Optical Eng
- Vol. 13411, 200-343
Mitigating data scarcity in the classification of glioma molecular subtypes: the power of generative imaging (Erratum)
Publisher's Note: This paper, originally published on 10 April 2025, was replaced with a corrected/revised version on 23 June 2025. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance.
- 23 June 2025
- Published by SPIE-Intl Soc Optical Eng
- Vol. 13411, 200-343
Mitigating data scarcity in the classification of glioma molecular subtypes: the power of generative imaging (Erratum)
Publisher's Note: This paper, originally published on 10 April 2025, was replaced with a corrected/revised version on 23 June 2025. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance.
- 10 April 2025
- Published by SPIE-Intl Soc Optical Eng in Proceedings of SPIE
- Vol. 13411, 20-121
Mitigating data scarcity in the classification of glioma molecular subtypes: the power of generative imaging
Isocitrate dehydrogenase (IDH) mutation status is a critical prognostic indicator in glioma patients. Numerous studies have focused on developing non-invasive methodologies to classify IDH status using pre-operative MRI scans. However, the challenge lies in data scarcity and class imbalance in IDH mutations. This study explores generative AI methods to augment training data and enhance IDH classification accuracy. We developed a 3D conditional latent diffusion model (LDM) for generating 3D multi-contrast brain tumor MRI data (128×128×64 with a voxel spacing of 1.5×1.5×2.0mm) with whole tumor mask and IDH mutation status as conditions. The LDM comprises a 3D autoencoder for perceptual compression and a conditional 3D diffusion model (DM) for generating multi-contrast synthetic samples guided by tumor masks and the IDH mutation status. We incorporated two types of attention modules within the denoising UNet of the LDM to capture the semantic class-dependent data distribution driven by the provided whole tumor mask and IDH status. The LDM was trained using two brain tumor datasets: The Cancer Genome Atlas dataset and an internal dataset from the University of Texas Southwestern Medical Center. The synthetic images generated by the LDM were then used to train IDH classification models, which were subsequently tested on real brain tumor data comprising 327 mutated and 1,394 wild-type cases from the University of California San Francisco Preoperative Diffuse Glioma MRI dataset, the Erasmus Glioma Database, the University of Pennsylvania glioblastoma, and two held-out internal datasets. The IDH classification models, trained on synthetic images and tested on real data, achieved an excellent overall classification accuracy of 94.02%. This approach has the potential to be extended to other molecular markers where data scarcity presents a challenge.
- 10 April 2025
- Published by SPIE-Intl Soc Optical Eng in Proceedings of SPIE
- Vol. 13411, 20-121
Mitigating data scarcity in the classification of glioma molecular subtypes: the power of generative imaging
Isocitrate dehydrogenase (IDH) mutation status is a critical prognostic indicator in glioma patients. Numerous studies have focused on developing non-invasive methodologies to classify IDH status using pre-operative MRI scans. However, the challenge lies in data scarcity and class imbalance in IDH mutations. This study explores generative AI methods to augment training data and enhance IDH classification accuracy. We developed a 3D conditional latent diffusion model (LDM) for generating 3D multi-contrast brain tumor MRI data (128×128×64 with a voxel spacing of 1.5×1.5×2.0mm) with whole tumor mask and IDH mutation status as conditions. The LDM comprises a 3D autoencoder for perceptual compression and a conditional 3D diffusion model (DM) for generating multi-contrast synthetic samples guided by tumor masks and the IDH mutation status. We incorporated two types of attention modules within the denoising UNet of the LDM to capture the semantic class-dependent data distribution driven by the provided whole tumor mask and IDH status. The LDM was trained using two brain tumor datasets: The Cancer Genome Atlas dataset and an internal dataset from the University of Texas Southwestern Medical Center. The synthetic images generated by the LDM were then used to train IDH classification models, which were subsequently tested on real brain tumor data comprising 327 mutated and 1,394 wild-type cases from the University of California San Francisco Preoperative Diffuse Glioma MRI dataset, the Erasmus Glioma Database, the University of Pennsylvania glioblastoma, and two held-out internal datasets. The IDH classification models, trained on synthetic images and tested on real data, achieved an excellent overall classification accuracy of 94.02%. This approach has the potential to be extended to other molecular markers where data scarcity presents a challenge.
- Top Cited PapersOpen Access
- 7 March 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
Convection enhanced delivery of Rhenium (186Re) Obisbemeda (186RNL) in recurrent glioma: a multicenter, single arm, phase 1 clinical trial
Rhenium (186Re) Obisbemeda (186RNL), chelated-186Re encapsulated in nanoliposomes and delivered to brain tumors via convection enhanced delivery (CED), was evaluated in a Phase 1 dose escalation trial (NCT01906385). The primary objective was to determine the maximum tolerated dose (MTD). Secondary objectives included safety and tolerability, dose distribution, the overall response rate (ORR), disease-specific progression-free survival (PFS), and overall survival (OS). 21 patients received up to 22.3 mCi 186RNL over 6 dosing cohorts. Most adverse events (AEs) were unrelated to 186RNL and the MTD was not reached. Although not predefined outcomes, the mOS and mPFS were 11 and 4 months, respectively, and found to correlate with radiation absorbed dose to the tumor and percent tumor treated. When dichotomized by absorbed dose of 100 Gy, the mOS and mPFS were 17 months and 6 months, respectively, for >100 Gy, compared to 6 (mOS) and 2 (mPFS) months, respectively, for <100 Gy. For ORR, 57.1% exhibited stable disease (SD), 4.8% partial response, and 38.1% progressive disease. Overall, patients received radiation absorbed doses without significant toxicity higher than possible with external beam radiation therapy (EBRT) and demonstrated mOS beyond standard of care for recurrent glioblastoma (~8 months).
- 7 March 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
Top Cited PapersOpen AccessConvection enhanced delivery of Rhenium (186Re) Obisbemeda (186RNL) in recurrent glioma: a multicenter, single arm, phase 1 clinical trial
Rhenium (186Re) Obisbemeda (186RNL), chelated-186Re encapsulated in nanoliposomes and delivered to brain tumors via convection enhanced delivery (CED), was evaluated in a Phase 1 dose escalation trial (NCT01906385). The primary objective was to determine the maximum tolerated dose (MTD). Secondary objectives included safety and tolerability, dose distribution, the overall response rate (ORR), disease-specific progression-free survival (PFS), and overall survival (OS). 21 patients received up to 22.3 mCi 186RNL over 6 dosing cohorts. Most adverse events (AEs) were unrelated to 186RNL and the MTD was not reached. Although not predefined outcomes, the mOS and mPFS were 11 and 4 months, respectively, and found to correlate with radiation absorbed dose to the tumor and percent tumor treated. When dichotomized by absorbed dose of 100 Gy, the mOS and mPFS were 17 months and 6 months, respectively, for >100 Gy, compared to 6 (mOS) and 2 (mPFS) months, respectively, for <100 Gy. For ORR, 57.1% exhibited stable disease (SD), 4.8% partial response, and 38.1% progressive disease. Overall, patients received radiation absorbed doses without significant toxicity higher than possible with external beam radiation therapy (EBRT) and demonstrated mOS beyond standard of care for recurrent glioblastoma (~8 months).
- Open Access
ProFit-1D for quantifying J-difference edited data at 3T
Reproducible and accurate fitting of the magnetic resonance spectrum is critical for estimating metabolite concentrations. We have previously developed a fitting software called ProFit-1D which was shown to fit 9.4T semi-LASER data from the human brain with high accuracy and precision. In this study, we adapted ProFit-1D to fit J-difference edited spectra acquired at a clinical field strength of 3T and to assessed its performance in simulated and in vivo data. ProFit-1D was adapted to fit J-difference edited data with alterations to the fitting range to exclude the 1.3 ppm lipid resonance, starting T 2 relaxation constants, initial fit parameters, and adaptive spectral baseline determination. The accuracy of ProFit-1D was systematically evaluated on simulated GABA-edited and 2-hydroxyglutarate-edited (2HG-edited) data with different types of in vivo parameter variations and compared to that of LCModel and Gannet, two software commonly used to fit J-difference edited data. The precision of ProFit-1D was also evaluated in GABA-edited spectra acquired in vivo in the occipital cortex (OCC) and medial prefrontal cortex (mPFC) of healthy participants at 3T using subsets of averages and compared to that of LCModel and Gannet. The 2HG fit error was also evaluated for ProFit-1D in 2HG-edited spectra acquired in glioma patients and compared to that of LCModel. Overall, it was found that ProFit-1D generally produced fits with low parameter fit errors across a variety of parameter variations. GABA, glutamate plus glutamine (Glx), and 2HG levels were also more accurately estimated with ProFit-1D than with LCModel and Gannet across different spectral disturbances and simulated concentrations. ProFit-1D was found to be as precise as LCModel and more precise than Gannet in estimating GABA and Glx. 2HG fit errors were 45% lower with ProFit-1D than with LCModel. Thus, ProFit-1D was found to produce high-quality fits to J- difference edited data with high accuracy and precision.
Open AccessProFit-1D for quantifying J-difference edited data at 3T
Reproducible and accurate fitting of the magnetic resonance spectrum is critical for estimating metabolite concentrations. We have previously developed a fitting software called ProFit-1D which was shown to fit 9.4T semi-LASER data from the human brain with high accuracy and precision. In this study, we adapted ProFit-1D to fit J-difference edited spectra acquired at a clinical field strength of 3T and to assessed its performance in simulated and in vivo data. ProFit-1D was adapted to fit J-difference edited data with alterations to the fitting range to exclude the 1.3 ppm lipid resonance, starting T 2 relaxation constants, initial fit parameters, and adaptive spectral baseline determination. The accuracy of ProFit-1D was systematically evaluated on simulated GABA-edited and 2-hydroxyglutarate-edited (2HG-edited) data with different types of in vivo parameter variations and compared to that of LCModel and Gannet, two software commonly used to fit J-difference edited data. The precision of ProFit-1D was also evaluated in GABA-edited spectra acquired in vivo in the occipital cortex (OCC) and medial prefrontal cortex (mPFC) of healthy participants at 3T using subsets of averages and compared to that of LCModel and Gannet. The 2HG fit error was also evaluated for ProFit-1D in 2HG-edited spectra acquired in glioma patients and compared to that of LCModel. Overall, it was found that ProFit-1D generally produced fits with low parameter fit errors across a variety of parameter variations. GABA, glutamate plus glutamine (Glx), and 2HG levels were also more accurately estimated with ProFit-1D than with LCModel and Gannet across different spectral disturbances and simulated concentrations. ProFit-1D was found to be as precise as LCModel and more precise than Gannet in estimating GABA and Glx. 2HG fit errors were 45% lower with ProFit-1D than with LCModel. Thus, ProFit-1D was found to produce high-quality fits to J- difference edited data with high accuracy and precision.
- Open Access
- 1 July 2024
- Published by Radiological Society of North America (RSNA) in Radiology: Artificial Intelligence
- Vol. 6 (4) , e230218
Two-Stage Training Framework Using Multicontrast MRI Radiomics for IDH Mutation Status Prediction in Glioma
"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To develop a radiomics framework for preoperative MRI-based prediction of IDH mutation status, a crucial glioma prognostic indicator. Materials and Methods Radiomics features (shape, first-order statistics, and texture) were extracted from the whole tumor or the combination of nonenhancing, necrosis, and edema regions. Segmentation masks were obtained via the federated tumor segmentation tool or the original data source. Boruta, a wrapper-based feature selection algorithm, identified relevant features. Addressing the imbalance between mutated and wild-type cases, multiple prediction models were trained on balanced data subsets using Random Forest or XGBoost and assembled to build the final classifier. The framework was evaluated using retrospective MRI scans from three public datasets (The Cancer Imaging Archive (TCIA, 227 patients), the University of California San Francisco Preoperative Diffuse Glioma MRI dataset (UCSF, 495 patients), and the Erasmus Glioma Database (EGD, 456 patients)) and internal datasets collected from UT Southwestern Medical Center (UTSW, 356 patients), New York University (NYU, 136 patients), and University of Wisconsin-Madison (UWM, 174 patients). TCIA and UTSW served as separate training sets, while the remaining data constituted the test set (1617 or 1488 testing cases, respectively). Results The best-performing models trained on the TCIA dataset achieved area under the receiver operating characteristic curve (AUC) values of 0.89 for UTSW, 0.86 for NYU, 0.93 for UWM, 0.94 for UCSF, and 0.88 for EGD test sets. The best-performing models trained on the UTSW dataset achieved slightly higher AUCs: 0.92 for TCIA, 0.88 for NYU, 0.96 for UWM, 0.93 for UCSF, and 0.90 for EGD. Conclusion This MRI radiomics-based framework shows promise for accurate preoperative prediction of IDH mutation status in patients with glioma. Published under a CC BY 4.0 license.
- 1 July 2024
- Published by Radiological Society of North America (RSNA) in Radiology: Artificial Intelligence
- Vol. 6 (4) , e230218
Open AccessTwo-Stage Training Framework Using Multicontrast MRI Radiomics for IDH Mutation Status Prediction in Glioma
"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To develop a radiomics framework for preoperative MRI-based prediction of IDH mutation status, a crucial glioma prognostic indicator. Materials and Methods Radiomics features (shape, first-order statistics, and texture) were extracted from the whole tumor or the combination of nonenhancing, necrosis, and edema regions. Segmentation masks were obtained via the federated tumor segmentation tool or the original data source. Boruta, a wrapper-based feature selection algorithm, identified relevant features. Addressing the imbalance between mutated and wild-type cases, multiple prediction models were trained on balanced data subsets using Random Forest or XGBoost and assembled to build the final classifier. The framework was evaluated using retrospective MRI scans from three public datasets (The Cancer Imaging Archive (TCIA, 227 patients), the University of California San Francisco Preoperative Diffuse Glioma MRI dataset (UCSF, 495 patients), and the Erasmus Glioma Database (EGD, 456 patients)) and internal datasets collected from UT Southwestern Medical Center (UTSW, 356 patients), New York University (NYU, 136 patients), and University of Wisconsin-Madison (UWM, 174 patients). TCIA and UTSW served as separate training sets, while the remaining data constituted the test set (1617 or 1488 testing cases, respectively). Results The best-performing models trained on the TCIA dataset achieved area under the receiver operating characteristic curve (AUC) values of 0.89 for UTSW, 0.86 for NYU, 0.93 for UWM, 0.94 for UCSF, and 0.88 for EGD test sets. The best-performing models trained on the UTSW dataset achieved slightly higher AUCs: 0.92 for TCIA, 0.88 for NYU, 0.96 for UWM, 0.93 for UCSF, and 0.90 for EGD. Conclusion This MRI radiomics-based framework shows promise for accurate preoperative prediction of IDH mutation status in patients with glioma. Published under a CC BY 4.0 license.