Description and assessment of a registration-based approach to include bones for attenuation correction of whole-body PET/MRI
Bibliographic record
Abstract
PURPOSE: Attenuation correction for whole-body PET/MRI is challenging. Most commercial systems compute the attenuation map from MRI using a four-tissue segmentation approach. Bones, the most electron-dense tissue, are neglected because they are difficult to segment. In this work, the authors build on this segmentation approach by adding bones using a registration technique and assessing its performance on human PET images. METHODS: Twelve oncology patients were imaged with FDG PET/CT and MRI using a Turbo-FLASH pulse sequence. A database of 121 attenuation correction quality CT scans was also collected. Each patient MRI was compared to the CT database via weighted heuristic measures to find the "most similar" CT in terms of body geometry. The similar CT was aligned to the MRI with a deformable registration method. Two MRI-based attenuation maps were computed. One was a standard four-tissue segmentation (air, lung, fat, and lean tissue) using basic image processing techniques. The other was identical, except the bones from the aligned CT were added. The PET data were reconstructed with the patient's CT-based attenuation map (the silver standard) and both MRI-based attenuation maps. The relative errors of the MRI-based attenuation corrections were computed in 14 standardized volumes of interest, in lesions, and over whole tissues. The squared Pearson correlation coefficient was also calculated over whole tissues. Statistical testing was done with ANOVAs and paired t-tests. RESULTS: The MRI-based attenuation correction ignoring bone had relative errors ranging from -37% to -8% in volumes of interest containing bone. By including bone, the magnitude of the relative error was reduced in all cases (p<0.001), ranging from -3% to 4%. Further, the relative error in volumes of interest adjacent to bone was improved from a mean of -7.5% to 2% (p<0.001). In the other seven volumes of interest, including bone reduced the magnitude of relative error in three cases (p<0.001), had no effect in three cases, and increased relative error in one case. There was no statistically significant difference in the relative error in lesions. Over whole tissues, including bone slightly increased relative error in lung from 7.7% to 8.0% (p=0.002), in fat from 8.5% to 9.2% (p<0.001), and in lean tissue from -2.1% to 2.6% (p<0.001), but reduced the magnitude of relative error in bone from -14.6% to 1.3% (p<0.001). The correlation coefficient was essentially unchanged in all tissues regardless of whether bone was included or not. CONCLUSIONS: The approach to include bones in MRI-based attenuation maps described in this work improves quantification of whole-body PET images in and around bony anatomy. The reduction in error is often large (tens of percents), and could alter image interpretation and subsequent patient care. Changes in other parts of the PET image are minimal and likely not of clinical significance.
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 | 0.000 | 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".