Deep Learning Framework for the Synthesis of Rest Phase from Stress Phase in Positron Emission Tomography Myocardial Perfusion Imaging
Bibliographic record
Abstract
Positron Emission Tomography Myocardial Perfusion Imaging (PET-MPI) is non-invasive and gold standard for diagnosis of coronary artery disease. MPI using [13N]NH3 is a Rest-Stress process and involves acquiring myocardial images at two different phases of the cardiac cycle. However, the procedure involves two phases that can present challenges in terms of patient comfort, radiation dose, and workload for healthcare facilities. Recent advances in deep learning (DL) have shown promising results in various medical imaging tasks, such as prediction, translation, and synthesis. In this study, we explore the potential of DL in synthesizing the rest phase of PET-MPI scans using stress phase images. To achieve this goal, we developed a DL model using U-Net architecture equipped with self-attention layers that capture long-range dependencies between feature maps and focus on more relevant regions of the image. We further augmented this model with ResNet and trained two separate models on the dataset of 200 patients. We qualitatively and quantitatively evaluated our method. There was a good visual agreement between the synthesized and ground truth images. The quantitative metrics were calculated including peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and normalized mean square error (NMSE) values to quantify the distortion level of the synthesized images and similarity between the ground truth and synthesized images. Our models demonstrated high-quality image reconstruction with an average PSNR of 32.44, a high degree of structural similarity with an average SSIM of 0.96, and high accuracy with an average NMSE of 0.05. In conclusion, our study highlights the potential of DL in improving the accuracy and efficiency of PET-MPI scans, which can have significant clinical implications. Our proposed approach could potentially reduce the need for patients to undergo additional imaging scans, leading to improved patient outcomes and reduced healthcare costs.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".