Abstract B047: Development and validation of an improved risk stratification model for ovarian cancer
Bibliographic record
Abstract
Abstract Based on previous models, among individuals who are not known to carry a pathogenic variant, lifetime risk of ovarian cancer ranges between ~0.1% and ~11%. Risk stratification to identify the people at the higher end of this lifetime risk range is of paramount importance for prevention efforts. Previous risk stratification models for ovarian cancer were based on a limited number of risk/protective factors. Further, we have also shown that risk estimates differ by menopausal status which most models have not considered. We aimed to develop and internally validate a risk stratification model for ovarian cancer that considers 15 unequivocal risk/protective factors and properly accounts for effect modification by menopausal status. We used data from nine studies (7,984 cases, 12,260 controls) participating in the Ovarian Cancer Association Consortium (OCAC). The data were split into a training set and a test set that comprised 80% and 20% of the OCAC dataset, respectively. Seven risk factors (body mass index, height, later age at menopause, menopausal hormonal therapy use, first-degree family history of ovarian cancer, endometriosis, and a polygenic score of 36 common genetic variants) and eight protective factors (later age at menarche, parity, breastfeeding, incomplete pregnancy, later age at last pregnancy, tubal ligation, combined oral contraceptive use and depot-medroxyprogesterone acetate use) were included. Other risk/protective factors such as talcum powder or aspirin use were not included due to high proportions of missing values. We fit multiplicative logistic regression models separately by menopausal status group in the training set to determine the associations between the factors and ovarian cancer. All models were adjusted for race/ethnicity, education level, age and OCAC study. In the test set, we calculated a summary relative risk for every combination of the 15 risk/protective factors (hereafter called a risk profile) based on the estimates from the training set. The summary relative risk was then translated into an absolute risk: the frequency-weighted average of all the profile-specific relative risks was scaled to the average absolute risk, and then this scaling factor was applied to each profile-specific relative risk and its confidence interval (CI). The range of absolute lifetime risks observed in the test set was 0.1%-7.2% using 15 factors, accounting for menopausal status. The area under the receiving operating curve (AUC) was 0.67 (95% CI 0.65-0.68). This is slightly higher than the previous risk stratification models (AUC=0.55-0.66). External validation of our risk stratification model in a longitudinal cohort is warranted, as our findings suggest that there is a subset of individuals at the higher end of the risk range who would be potential candidates for primary prevention strategies including salpingectomy. Citation Format: Minh Tung Phung, Alice W. Lee, Karen McLean, Lilah Khoja, Hoda Anton-Culver, Elisa V. Bandera, Jennifer Anne Doherty, Renee T. Fortner, Marc T. Goodman, Francesmary Modugno, Paul D. P. Pharoah, Kathryn L. Terry, Penelope M. Webb, Anna H. Wu, Andrew Berchuck, Gillian E. Hanley, Bhramar Mukherjee, Malcolm C. Pike, Celeste Leigh Pearce, Britton Trabert. Development and validation of an improved risk stratification model for ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Ovarian Cancer; 2023 Oct 5-7; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(5 Suppl_2):Abstract nr B047.
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 | 0.009 | 0.015 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
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".