Analysis of prognostic (prog) Web-based models for stage II and III colon cancer (CC): A population-based validation of <i>Numeracy</i> (NUM) and <i>ADJUVANT! Online</i> (ADJ!)
Bibliographic record
Abstract
4044 Background: To aid in decisions regarding adjuvant therapy (AT) for resected high-risk CC, two prog models are in common use: the Mayo Clinic NUM calculator developed from a pooled data analysis of 7 randomized 5FU-based AT trials, and ADJ! developed using SEER data. This study examines the accuracy of NUM and ADJ! utilizing a cohort of patients (pts) referred to the BC Cancer Agency (BCCA). Methods: Demographic, disease and treatment data for pts with stage II/III CC referred to the BCCA from 1995–1996 + 1999–2003 were collected. Observed (obs) 5-year relapse free survival (RFS) and overall survival (OS) were compared to predicted estimates (pred) using NUM and ADJ!, both overall and for all prog subgroups with ≥ 10pts, as stratified by T stage, N stage, tumor grade and age. Data are presented in a descriptive manner and using confidence intervals. Results: Median follow-up was 5.6 yrs for 2,033 pts - 53% male, median age 68y, 40% N0. The mean percentages of 5 year pred outcomes for each of the two models and the actual Kaplan Meier mean survivals are presented in the table . The percentage correct predictions of 5 y status is also presented, with correctness deemed accurate if the pt was alive and predicted to be alive by ≥ 50% as determined by each model or dead while the respective tool predicted < 50% possibility of being alive. For surgery alone pts, ADJ!pred were more often closer to what was observed, as compared to NUMpred, in the prog subgroups (for RFS 56%, OS 88%). For surgery + 5-FU pts, within these subgroups, NUMpred were more often closer to what was observed, as compared to ADJ!pred, for RFS (62%) and for OS (55%). Conclusions: In this independent population-based validation, NUM and ADJ! have acceptable and similar reliability with modest over-estimations of 5y RFS and OS. Both models thus appear to be useful adjuvant decision-aids. [Table: see text] No significant financial relationships to disclose.
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 | 0.037 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".