Harmonized Assessment and Quality Control of Quantitative Measures of Lung Structure on Pre-pandemic Cardiac and Lung Computed Tomography (CT) Scans for Large-scale Investigation of Risk of Long COVID-19: The Collaborative Cohort of Cohorts for COVID-19 Research (C4R)
Bibliographic record
Abstract
Abstract RATIONALE: Susceptibility to severe COVID-19 and long COVID may be influenced by pre-existing lung structure. ECG-gated, non-contrast cardiac CT scans, originally acquired for assessing coronary artery calcium, are available in multiple longitudinal cohort studies and have been previously validated against lung CT scans for densitometry; however, protocol variation across cohorts may hinder direct comparison. To address this, we developed a quality control (QC) rubric to grade cardiac CT scans for standardized quantitative lung measures across cohorts. METHODS: C4R assessed SARS-CoV-2 infection and outcomes in 14 NHLBI cohorts, of which 10 (six community-based and four longitudinal case-control studies) included CT scans (12,459 participants with lung and 13,752 with cardiac CTs). Logic rules were applied to categorize scans (Table): Grade “A-B” scans had 2-3.5mm slice thickness, at least 20 axial slices, no lung cropping in axial view and lung and airway volumes greater than the cohort-specific 1st percentile. “C-D” scans had one lung obscured by cropping around the heart, so quantitative measures were restricted to the lung in complete view. “F-H” scans were deemed unusable, suffering from low slice count (<20), severe image cropping, low lung or airway volumes (< 1st%ile). Visual inspection was performed on a sample from each category to ensure accuracy. Percent of lung voxels in low attenuation areas < -950 HU (LAA950), high attenuation areas (-600 to -250 HU, HAA) and 15th percentile HU value (perc15) were measured. Within-participant variability was quantified using the intraclass correlation coefficient (ICC). RESULTS: We performed QC on cardiac CT scans from an initial sample of 6,275 participants from three C4R cohorts (ARIC, CARDIA, MASALA), imaged with GE or Siemens scanners. Of these, 5,125 (81.7%) were graded “A-B”, 496 (7.9%) were “C-D” and 654 (10.4%) were “F-H”. The percentage of “F-H” scans varied widely across cohorts [<1% to 26%] due to image cropping. Among the 5,621 cardiac CT scans graded A-D, the mean±SD for LAA950, HAA and perc15 respectively were 1.1±2.1%, 7.2±4.7% and -887±33 HU. Repeated imaging was available for 107 participants (time interval 5.1±0.7y). For paired scans in grades A-B (n=101/107), the within-participant correlation for LAA950, HAA and perc15 were 0.91, 0.38 and 0.74 respectively. CONCLUSION: Initial application of QC to cardiac CT scans identified scans with high reproducibility, with the aim to provide valid quantitative lung metrics on over 20,000 highly characterized participants in C4R.
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 | 0.153 | 0.157 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.004 | 0.001 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".