Abstract PR17: Comprehensive colorectal cancer risk prediction to inform personalized screening and intervention
Bibliographic record
Abstract
Abstract Background and Aims: Colorectal cancer (CRC) is one of the most preventable and treatable cancers when detected early via screening. The current screening guidelines for CRC recommend exams only based on age, family history, and previous screening results. Multiple environmental and lifestyle risk factors, however, have been established or suspected for CRC, as have many common genetic susceptibility loci. It is critical to utilize this information to better stratify individuals into low- and high-risk groups for optimized and personalized screening and intervention recommendations. Methods: Using data from two large consortia (8421 CRC cases and 9767 controls): the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO) and the Colorectal Transdisciplinary study (CORECT), we developed risk prediction models for men and women based on family history, environmental and lifestyle risk factors, and known CRC susceptibility loci identified through genome-wide association studies. We constructed an environmental risk score (E-score) as a weighted sum of 19 established or potential environmental and lifestyle risk factors for CRC with weights obtained from a multivariate logistic regression analysis. Similarly, we also constructed a genetic risk score (G-score) using 64 common variants associated with CRC risk. We evaluated the discriminatory accuracy of risk prediction models by calculating the area under the Receiver Operating Characteristic curve (AUC), correcting for potential overestimating by using the training data set. Our models also estimate absolute risk of developing CRC given various risk profiles, and provide recommended ages for the first endoscopic screening exam. Results: Both the E-score and the G-score are independent predictors of CRC risk, and models that incorporate both scores improve the discriminatory accuracy significantly compared to family history-only models. Compared to the model that includes only family history, the E-score significantly improves the discriminatory accuracy for both men (AUC = 0.62 vs. 0.53, p-value < 1e-5 ) and women (AUC = 0.60 vs. 0.52, p-value < 1e-5 ). The G-score also significantly improves the discriminatory accuracy for both men (AUC = 0.60 vs. 0.53, p-value < 1e-5 ) and women (AUC = 0.60 vs. 0.52, p-value < 1e-5 ) over the family history-only model. Compared to the model with family history and E-score, the inclusion of the G-score in the model further improves the discriminatory accuracy for both men (AUC = 0.65 vs. 0.62, p-value = 0.0152) and women (AUC = 0.63 vs. 0.60, p-value = 0.0005). Based on the 10-year risk estimates of developing CRC, the difference in recommended age to start screening for the top 90% and the bottom 10% of risk score ranges from 12 to 14 years depending on sex and status of CRC family history. Conclusions: Our risk prediction models incorporating both comprehensive environmental and lifestyle risk factors, and known CRC common genetic variants provide more accurate estimation of CRC risk. These models will be useful for recommending individually tailored screening and intervention strategies to prevent this common cancer. This abstract is also being presented as Poster B17. Citation Format: Jihyoun Jeon, Sonja I. Berndt, Hermann Brenner, Peter T. Campbell, Andrew T. Chan, Jenny Chang-Claude, Mengmeng Du, Graham Giles, Jian Gong, Stephen B. Gruber, Tabitha A. Harrison, Michael Hoffmeister, Loic LeMarchand, Li Li, John D. Potter, Gad Rennert, Robert E. Schoen, Martha L. Slattery, Emily White, Michael O. Woods, Ulrike Peters, Li Hsu. Comprehensive colorectal cancer risk prediction to inform personalized screening and intervention. [abstract]. In: Proceedings of the AACR Special Conference: Improving Cancer Risk Prediction for Prevention and Early Detection; Nov 16-19, 2016; Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2017;26(5 Suppl):Abstract nr PR17.
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 | 0.004 | 0.014 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 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".