Diagnostic accuracy of an AI-based model for quantifying COVID-19 lung involvement on chest CT: A cross-sectional study
Bibliographic record
Abstract
Rationale: Rapid and accurate assessment of pulmonary involvement in COVID-19 is essential for clinical management. This study aimed to evaluate the accuracy of AI-based software in quantifying lung involvement in COVID-19. Methods: This observational, cross-sectional study analyzed chest CT scans from patients suspected of SARS-CoV-2 pneumonia between 2020 and 2023 in Salvador, Brazil. Scans were assessed using AI software Chest-CT Siemens® and compared with radiologist reports. Pulmonary involvement was categorized as mild (<25%), moderate (25-50%), and severe (>50%). The agreement between methods was evaluated using the Kappa coefficient. Results: A total of 1,143 CT scans were assessed for eligibility, and 14 patients were excluded for the absence of a clinical history of COVID-19. Participants' mean age was 49 (±15) years, with an equal distribution between sexes. Hypertension (41%), diabetes (15%), and asthma (14%) were the most prevalent comorbidities. Absent COVID-19 lung involvement was reported by the IA model and radiologists in 25 (7.8%) and 158 (23.6%) of the patients, respectively. Pneumonia was identified in 58% of cases, while incidental findings were observed in 74% of CT scans. The AI software demonstrated low overall concordance with radiologists (Kappa=0.159), particularly underestimating lung involvement with ground-glass opacities. The model exhibited moderate diagnostic performance for mild lung involvement (PPV=79%) but showed poor predictive accuracy for moderate (PPV=8%) and severe (PPV=23%) involvement. Conclusion: AI-based CT analysis shows potential as an adjunctive tool but lacks sufficient accuracy for standalone diagnosis.
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 | 0.012 | 0.031 |
| 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 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| 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; 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".