Abstract PR001: Risk prediction modeling for colorectal adenomas in persons under age 50: a risk-stratified approach to early onset colorectal cancer prevention
Bibliographic record
Abstract
Abstract Background: As early-onset colorectal cancer (EOCRC) incidence rises, identifying colorectal adenomas (CRAs) under age 50 remains a key prevention strategy. We previously published a model to predict CRA risk in individuals under 50, achieving an area under the curve (AUC) of 0.71 (PMID: 39280910). Here, we validate this in a separate cohort. We also explore whether accelerated aging could be associated with CRA risk, as prior studies have suggested. PhenoAge, a proposed biological aging metric derived from clinical lab values and associated with chronic disease and mortality risk, may capture physiological decline not reflected in chronological age. Methods: Retrospective cohort study of adults under age 50 who had a colonoscopy between 2014-2024 for benign indications (excluding those with high-risk indications, alarm symptoms, or inadequate bowel preparation). The primary outcome was pathologically confirmed CRA. The previous model was applied without refitting. Performance was assessed using area under the curve (AUC), Hosmer–Lemeshow goodness-of-fit test, and diagnostic metrics (sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV)) at a pre-specified probability cutoff of 0.16. For the PhenoAge analysis, we included patients within the cohort who had available labs within 60 days of colonoscopy for the calculation of PhenoAge. We evaluated whether age acceleration (PhenoAge – chronological age) were different between persons with and without CRAs. Results: We identified 2874 persons who met inclusion criteria for inclusion in the validation cohort, median age 44 years, 230 were male (49.0%). Of these, 469 (16%) had at least 1 CRA. Applying the prior model, AUC was 0.67 (95% CI: 0.64–0.69). The overall sensitivity and specificity were 67.0% and 55.2%, respectively, with a PPV of 22.6% and a NPV of 89.5%. Performance differed across age groups: for participants under 45 years, sensitivity was 29.0% and specificity was 80.6%, while for the 45-49 age group, sensitivity was 85.7% and specificity was 21.6%. A total of 202 persons had labs available to calculate PhenoAge (median age 44 years, 32% male). Of these, 25 (12%) had >1 CRA. CRA group did not show statistically significant degrees of age acceleration (-0.3 years vs -2.8 years, p=0.12), but did show higher BMI (30 vs 26, p=0.01). There were no significant differences in race, ethnicity, marital status, or aspirin use between groups. Conclusion: External validation of our previously published risk prediction model demonstrates moderate discrimination, with higher sensitivity and precision for adults under age 45. Model refinement is needed to improve performance, but this shows excellent promise for a risk-stratified approach to EOCRC prevention in persons outside screening age. Given the median age of EOCRC diagnosis is 44, such strategies are essential. Biological age as measured by PhenoAge does not appear to improve model performance, but future studies should continue to explore risk factors and strategies for personalized EOCRC prevention. Citation Format: Ritika Modi, Ryan Hood, Divya Dasani, Catherine Blandon, Shria Kumar. Risk prediction modeling for colorectal adenomas in persons under age 50: a risk-stratified approach to early onset colorectal cancer prevention [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: The Rise in Early-Onset Cancers—Knowledge Gaps and Research Opportunities; 2025 Dec 10-13; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(23_Suppl):Abstract nr PR001.
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 | 0.003 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| 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 | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
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".