Abstract B058: An informed machine learning-based blood test for risk assessment of lethal and common cancers
Bibliographic record
Abstract
Abstract Multi-cancer detection tests represent a promising advancement, with the potential to detect multiple cancer types through a single test, at earlier more treatable stages. Early detection is associated with improved treatment outcomes, and reduced healthcare costs. These tests may focus on a few common cancers or encompass a broader spectrum, including those for which no current screening guidelines exist. We have developed a machine learning–based blood test, the Multi-Cancer Risk Assessment Test (MCaST), designed to estimate an individual’s risk of harboring one or more of the following cancers: breast, ovarian, colorectal, esophageal, gastric, liver, pancreatic, prostate, and lung cancer. Our method utilizes measurements of protein biomarkers as inputs into predictive algorithms. Elevated levels in one or more biomarkers yield a composite score that is subsequently translated into an absolute 6-year cancer risk. Importantly, the algorithm also integrates longitudinal data from two or more biomarker measurements taken at different time points, allowing dynamic changes in biomarker levels to further refine risk classification. An individual is classified as high risk if their estimated 6-year absolute risk exceeds a predefined threshold of 1%. An important advantage of this approach is to tailor screening according to risk for cancers for which screening is available and to bring into personalized screening subjects at increased risk for cancers for which screening is not recommended for the general population. Citation Format: Ehsan Irajizad, Sam Hanash. An informed machine learning-based blood test for risk assessment of lethal and common cancers [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Artificial Intelligence and Machine Learning; 2025 Jul 10-12; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(13_Suppl):Abstract nr B058.
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 | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| 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) | 0.004 | 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".