Evaluation of Quality of Biomarker Data Capture by Peripheral Sites in an International, Cooperative Study: Analysis of 104 Cases From the T-Cell Project.
Bibliographic record
Abstract
Abstract Abstract 2650 Background and Rationale: The T-Cell Project (TCP) aims at verifying if a prospective collection of data in patients with Peripheral T-cell lymphomas (PTCLs) provides more accurate information to better define their prognosis. So far 885 patients have been accrued and diagnosis confirmed in 84% of the 350 reviewed cases. A dedicated online pathology form collecting a detailed biomarkers profile is filled out by research staff at peripheral sites with data from the local pathologist report. However, it is very difficult for people who are not accustomed to routinely reading pathology reports to interpret the findings correctly. To guarantee the data capture of biomarkers reproduces the pathology reports a control of quality of information entered in the pathology forms of the TCP was carried out. A similar evaluation was performed by the COMPLETE Registry, presenting the results in a parallel abstract. Methods: Biomarker data quality assessment concerns the review by an expert pathologist of peripheral capture of a suggested quite wide panel of biomarkers (54) used to diagnose patients registered in the TCP. The first 104 patients enrolled having complete registration data, availability of the original pathology report at the Trial Office and its data entered at the website constitute the sample of this analysis. A single mismatch between the site-entered data and the reviewer's findings, recorded on a separate forms, counts as an error. Results: On the whole, 5740 entries were reviewed, with a mean number of 9 immunophenotypic markers (range 3–22) and a mean number of 0.3 (range 0–4) gene rearrangement tests. In 35 (34%) cases out of the 104 reviewed no conflicting entry between what the site entered and what the reviewer determined was found. Patient disagreement of different extent was determined for the remaining cases: 1–2 errors, 35 (34%); 3–5 errors, 20 (19%); 6–10 errors, 8 (8%); 11–19 errors, 6 (5%). Where the site noted a finding was positive the reviewer was in agreement 73% of the time, noted they were negative in 15%, indeterminate in 1% and not assessed in 11%. For cases where the site noted a finding was negative the reviewer agreed with 97% of cases, noted they were positive in 1%, indeterminate in 0% and not assessed in 2%. The reviewer was in agreement with the site in 96% of the cases when the marker was indicated peripherally as not assessed, and for the remaining tests found the marker was positive, negative or indeterminate in 1%, 2% and 0% of cases respectively. The markers most difficult to interpret (at least 5% of total errors) are listed in the Table, reporting also the types of errors noted. With respect to the T-cell markers, the misreporting is mainly due to the difficulty in recognizing and thus entering the findings on the pathology report. For the suggested B-cell markers, a high rate of errors concerns the coding of the findings of the pathology report noted in the non-neoplastic populations surrounding the neoplastic cells. Of relevance, the very frequent mistaken interpretation of the EBV in situ hybridization (ISH), recorded as the result of the EBV immunoistochemistry test. Gene rearrangement studies are often missing and if present almost totally misinterpreted by the site, and reported as an immunoistochemistry test. The review by the expert pathologist accomplished a 11% of errors due to an ambiguous noting of the findings on the pathology report. Conclusion: The results of the biomarker quality assessment for the TCP confirm the difficulty for a correct interpretation of the pathology report in a relatively high rate of cases. For international projects the need for a periodic review emerges to guide site training and improve the accuracy of biomarker data capture in order to ensure database quality. Disclosures: No relevant conflicts of interest to declare.
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 | 0.039 | 0.052 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.006 | 0.006 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 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".