Variability of the Positive Predictive Value of PI-RADS for Prostate MRI across 26 Centers: Experience of the Society of Abdominal Radiology Prostate Cancer Disease-focused Panel
Bibliographic record
Abstract
Background Prostate MRI is used widely in clinical care for guiding tissue sampling, active surveillance, and staging. The Prostate Imaging Reporting and Data System (PI-RADS) helps provide a standardized probabilistic approach for identifying clinically significant prostate cancer. Despite widespread use, the variability in performance of prostate MRI across practices remains unknown. Purpose To estimate the positive predictive value (PPV) of PI-RADS for the detection of high-grade prostate cancer across imaging centers. Materials and Methods This retrospective cross-sectional study was compliant with the HIPAA. Twenty-six centers with members in the Society of Abdominal Radiology Prostate Cancer Disease-focused Panel submitted data from men with suspected or biopsy-proven untreated prostate cancer. MRI scans were obtained between January 2015 and April 2018. This was followed with targeted biopsy. Only men with at least one MRI lesion assigned a PI-RADS score of 2–5 were included. Outcome was prostate cancer with Gleason score (GS) greater than or equal to 3+4 (International Society of Urological Pathology grade group ≥2). A mixed-model logistic regression with institution and individuals as random effects was used to estimate overall PPVs. The variability of observed PPV of PI-RADS across imaging centers was described by using the median and interquartile range. Results The authors evaluated 3449 men (mean age, 65 years ± 8 [standard deviation]) with 5082 lesions. Biopsy results showed 1698 cancers with GS greater than or equal to 3+4 (International Society of Urological Pathology grade group ≥2) in 2082 men. Across all centers, the estimated PPV was 35% (95% confidence interval [CI]: 27%, 43%) for a PI-RADS score greater than or equal to 3 and 49% (95% CI: 40%, 58%) for a PI-RADS score greater than or equal to 4. The interquartile ranges of PPV at these same PI-RADS score thresholds were 27%–44% and 27%–48%, respectively. Conclusion The positive predictive value of the Prostate Imaging and Reporting Data System was low and varied widely across centers. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Milot in this issue.
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 | 0.016 | 0.036 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| 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".