Deep Active Learning Model for Adaptive PET Attenuation and Scatter Correction in Multi-Centric Studies
Bibliographic record
Abstract
Quantitative PET imaging requires multiple steps including attenuation and scatter correction (ASC), commonly carried out using anatomical images (CT, MRI). Clinical centers are equipped with different scanners and use different acquisition/reconstruction protocols, and these may vary over time in a given center. As new scanners and protocols emerge, deep learning model performance may deteriorate significantly, even for models built using large datasets, hence the need to update the models. The aim of the current study was to apply a deep active learning model for adaptive PET attenuation and scatter correction in multi-centric studies. We enrolled 111018F-FDG and 95068Ga-PSMA PET/CT images from 3 and 4 different centers, respectively. We implemented a deep residual network architecture with 20 blocks for different-level feature extractions. First, the deep neural network was trained on18F-FDG PET images. Since the radiotracer distribution is different from68Ga-PSMA PET, we used body fine tuning transfer learning to transfer ASC knowledge between radiotracers. We build deep learning-based ASC models on68Ga-PSMA and then applied active learning approaches to build center-specific ASC model. We trained a 2D deep neural network for direct generation of CT-based attenuation corrected PET images from non-attenuation-scatter corrected (NAC) images of68Ga-PSMA patients. For model evaluation, voxel-wise mean error (ME), mean absolute error (MAE), relative error (RE%), absolute relative error (ARE%) and structural similarity index (SSIM) were calculated between ground truth CT-based attenuation/scatter corrected and predicted PET images using the deep learning algorithm.We achieved a ME of 0.22±0.05, MAE of 0.80±0.05, RE of 2.72±7.5%, ARE of 10.0±4.5% and SSIM of 0. 98±0.02 in the test set. Overall, we applied transfer learning to transfer knowledge between different PET radiotracers and built specific models for each center separately using active learning approaches.
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".