Evaluation of Processing Software for SPECT Myocardial Blood Flow
Bibliographic record
Abstract
168 Objectives: Stationary cardiac SPECT cameras have greatly simplified the acquisition of dynamic datasets and led to a renewed interest in measuring myocardial blood flow (MBF) with SPECT. Several single-center studies have shown that it is feasible with these cameras and standard myocardial perfusion tracers to accurately measure MBF. MBF data have the potential to identify multi-vessel disease and thus improve risk stratification of patients with coronary artery disease. Clinical use of SPECT MBF will require commercial software for flow processing on the camera workstation. In this study we evaluate the accuracy, intra- and inter-user variability of a recently released software package for SPECT MBF analysis. Methods: The SPECT MBF component of 4DM (v2017, Invia) was compared to PET using a retrospective analysis of 31 patients for whom both SPECT MBF and standard clinical PET MBF studies were acquired within one month. SPECT data were acquired on a multi-pinhole stationary cardiac camera (GE Healthcare) with Tc-99m-tetrofosmin using a rest/stress one-day protocol. Listmode data were obtained starting at injection and continuing for 11min. Vendor online software was used to re-frame data into 9 x 10 s, 6 x 15 s, and 4 x 120 s time frames which were then reconstructed using an iterative algorithm with noise regularization (MAP-EM). The CT from the PET/CT scan was imported and manually co-registered the emission image for attenuation correction (AC). The dynamic image series was imported into 4DM (v2017, Invia) for analysis. Manual motion correction (MC) was applied frame-by-frame to visually align with a late-frame contour of the myocardial uptake. The arterial input function was obtained with a region of interest centered axially on the base of the septum. A 1-tissue compartment model was applied to determine the uptake constant K1. Image processing was performed independently by two users with repeat evaluation by one user to estimate inter- and intra-user variability. Global left-ventricular K1 values were averaged between users and fit to PET MBF values using a repeated two-fold cross-validation approach to estimate parameters of the Renkin-Crone extraction fraction correction for converting K1 to MBF. Images reconstructed with no corrections (NC), AC, MC, and ACMC were considered. Results: There was good correlation between K1 and PET MBF (R2 = 0.57 to 0.71). The extraction-fraction function parameters were found to be different than those determined previously with off-line software but the resulting MBF values were similar. The standard deviation of the percent difference of the 4DM MBF from PET MBF was between 43% (NC) and 33% (ACMC). The mean inter-user difference was 8% ± 18% (NC) and 4% ± 16% (ACMC). The mean intra-user difference was 6% ± 19% (NC) and 2% ± 16% (ACMC). Conclusions: Correlation and standard deviation in the flow difference from PET were similar to those obtained previously with in-house software. The extraction fraction parameters differed from those found previously suggesting that some cross-calibration may be needed between software packages. Inter- and intra-user differences were small compared to overall uncertainty.
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 | 0.009 | 0.022 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.011 | 0.003 |
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".