Attention-Guided Deep Learning of Chemical Exchange Saturation Transfer Magnetic Resonance Imaging to Differentiate Between Tumor Progression and Radiation Necrosis in Brain Metastasis
Bibliographic record
Abstract
Purpose Stereotactic radiosurgery (SRS) is a standard treatment for brain metastases; however, it may lead to radiation necrosis (RN). RN can be virtually indistinguishable from tumor progression (TP), which can have significant clinical implications on appropriate, time-sensitive treatment. This study investigated the effectiveness of multimodal chemical exchange saturation transfer magnetic resonance imaging (MRI), combined with T1/T2 mapping and/or conventional structural MRI, in addressing this diagnostic challenge, when analyzed through attention-guided deep learning. Methods and Materials MRI data (3-dimensional amide proton transfer magnetization transfer ratio [Amide MTR ], relayed nuclear Overhauser effect magnetization transfer ratio [rNOE MTR ], T1 and T2 parametric maps, and postcontrast T1-weighted [T1c] and T2-weighted fluid-attenuated inversion recovery [T2-FLAIR] images) were acquired from 93 patients (230 brain metastases lesions) treated with SRS a few months prior. Lesion outcomes (TP/RN) were confirmed via histopathology and/or serial clinical imaging, including the use of perfusion imaging, over a follow-up period of at least 6 months. Data were split into training (47 patients; 184 lesions) and independent testing (46 patients; 46 lesions) sets. A 3-dimensional transformer model with 2 new attention mechanisms was developed to classify lesions using various combinations of multimodal MRI inputs. Results Among dual-channel models, T1c and T2-FLAIR yielded an area under the receiver operating characteristic curve (AUC) of 0.78 ± 0.01, whereas Amide MTR and rNOE MTR maps achieved 0.76 ± 0.01. Integrating Amide MTR and rNOE MTR with either T1/T2 maps or T1c/T2-FLAIR substantially improved performance (AUC = 0.84 ± 0.02 and 0.85 ± 0.02, respectively). The highest performance (AUC = 0.87 ± 0.01) was achieved using all 6 modalities. Conclusions Attention-guided deep-learning analysis of chemical exchange saturation transfer MRI shows strong potential for accurately distinguishing RN from TP, underscoring the significance of multimodal MRI inputs for post-SRS lesion evaluation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".