Abstract A007: Bayesian Counterfactual Machine Learning Individualizes Radiation Modality Selection to Mitigate Immunosuppression
Bibliographic record
Abstract
Abstract Purpose: Lymphocytes play critical roles in cancer immunity and tumor surveillance. Radiation-induced lymphopenia (RIL) is a common side effect observed in cancer patients undergoing chemoradiation therapy (CRT), leading to impaired immunity and worse clinical outcomes. While proton beam therapy (PBT) has been suggested to reduce RIL risk compared to intensity-modulated radiation therapy (IMRT), this study utilized Bayesian counterfactual machine learning to identify distinct patient profiles and inform personalized radiation modality choice. Method: A novel Bayesian causal inferential technique is introduced and applied to a matched retrospective cohort of 510 esophageal cancer patients undergoing chemoradiation therapy to identify patient profiles for which immunosuppression could have been mitigated from radiation modality selection. Results: Body mass index (BMI), age, baseline absolute lymphocyte count (ALC), and planning target volume (PTV) determined the extent to which reductions in ALCs varied by radiation modality. Five patient profiles were identified. Significant variation in ALC nadir between PBT and IMRT was observed in three of the patient subtypes. Notably, older patients (age > 69) with normal weight experienced a two-fold reduction in mean ALC nadir when treated with IMRT versus PBT. Mean ALC nadir was reduced significantly for IMRT patients with lower ALC at baseline (< 1.6 ) who were overweight or obese when compared to PBT, whereas overweight patients with higher baseline ALC showed clinical equipoise between modalities. Conclusion: Individualized radiation therapy selection can be an important tool to minimize immunosuppression for high risk patients. The Bayesian counterfactual modeling techniques presented in this article are flexible enough to capture complex, nonlinear patterns while estimating interpretable patient profiles for translation into clinical protocols. Citation Format: Duo Yu, Michael J. Kane, Yiqing Chen, Steven H. Lin, Radhe Mohan, Brian P. Hobbs. Bayesian Counterfactual Machine Learning Individualizes Radiation Modality Selection to Mitigate Immunosuppression [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 A007.
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 | 0.025 | 0.061 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 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".