Inter-observer variation in LV analysis in a dedicated CMR unit: the impact of audit and consensus guideline on reproducibility
Bibliographic record
Abstract
Cardiac MRI (CMRI) is reported to be accurate and reproducible in the assessment of left ventricular function. This, however, is generally by experienced observers using automated or semi-automated software. Many departments still rely on manual contour tracing with trainees increasingly performing analysis but is reproducibility still good? This study assess the impact of the observer experience on LV function assessment and the role of consensus guidelines in raising standards among trainees. 20 LV data sets of varying volumes and ejection fractions (EF) were anonymized. Each data set comprised 2 and 4 chamber long axis cine SSFPs and contiguous short axis cine SSFPs from base to apex. LV volumes (LVEDV, LVESV), end diastolic muscle mass (EDMM), and EF were manually evaluated using Argus software (Siemens Medical Solutions, Erlangen) by all those regularly analyzing data in our department (7 experienced operators (> 2 years CMR experience) and 4 inexperienced operators (< 1 year CMR experience)). Inter-observer variability for all parameters was assessed, using the mean of all expert observers as the reference. Analysis of saved contours for all observers showed a small number of common causes of variability. Based on these, consensus guidelines were agreed and instituted. 4 of the experienced and all the inexperienced observers repeated the analysis of the 10 most problematic data sets after 3 months. Inter-operator variances for analyses before and after introduction of guidelines were compared. The department as a whole showed wide inter-observer variation for all parameters (mean standard deviation for EF, LVEDV, LVESV and EDMM were 3.8%, 10.8 mls, 10.5 mls and 23.6 gms respectively). As expected, there was greater variation between inexperienced observers than experienced observers (mean SD of variation in EF for inexperienced was 4.9% compared with 2.7% for experienced, LVEDV 12.3 mls and 7.6 mls, LVESV 11.9 mls and 7.4 mls, EDMM 29.2 gms and 16.0 gms). Following introduction of consensus guidelines, mean SD for EF fell to 2.7%, LVEDV to 7.2 mls, LVESV to 6.7 mls. There was little change in mean SD for EDMM (18.8 gms). The use of guidelines eliminated differences between experienced and inexperienced observers for all parameters. Reported reproducibility of LV function measurements by CMRI is high for experienced observers but this may not be true in large departments or when observers are inexperienced. Internal audit should be routine for validating practice and consensus guidelines can help in raising standards to meet published values.
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 | 0.251 | 0.433 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| 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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".