Abstract 6095: Lung cancer risk assessment with the INTEGRAL protein panel: Preliminary results from development and validation in the Lung Cancer Cohort Consortium
Bibliographic record
Abstract
Abstract Background: In the NCI-funded Integrative Analysis of Lung Cancer Etiology and Risk (INTEGRAL) program, the Lung Cancer Cohort Consortium (LC3) recently identified 36 proteins associated with lung cancer risk in pre-diagnostic blood samples after evaluating 1,161 proteins in 731 case-control pairs (Nature Communications, 2023). Based on these data, we designed the INTEGRAL panel, which measures absolute concentrations of 21 proteins, to optimize lung cancer risk assessment for screening beyond prediction models (e.g. PLCOm2012) or categorical screening criteria (e.g. USPSTF-2021). Methods: To evaluate the performance of the INTEGRAL panel, we assayed blood from participants selected as a representative case-cohort sample from 14 LC3 cohorts, divided into development and validation sets (Annals of Epidemiology, 2023). Here, we present preliminary results from the 7 cohorts in the development set, including 807 lung cancer cases diagnosed within 3 years of blood draw and 1,144 sub-cohort representatives, who were weighted to represent all current and former smokers in their corresponding cohorts. We first fit a flexible parametric survival model in 4 cohorts (n=478 cases) including a subset of proteins and age, smoking duration, and smoking intensity. The model was subsequently evaluated in the remaining 3 cohorts (n=329 cases) and benchmarked against the PLCOm2012 risk score and USPSTF2021 screening criteria. Results: The preliminary model includes 4 proteins (CEACAM5, MMP12, SCF, LPL). Compared with the PLCOm2012 score, the model improved discrimination of future lung cancer for cases occurring over 3 years (AUC=0.81 vs. 0.76, p<0.0001) and over 1 year (AUC=0.86 vs. 0.78, p<0.0001). At the USPSTF-2021 specificity of 76%, the protein model increased sensitivity by an absolute difference of 14% over 3 years and 25% over 1 year compared to USPSTF-2021, and by 9% over 3 years and 17% over 1 year compared to the PLCOm2012 model. When screening the same number of participants as PLCOm2012, the protein model identified more future cases (223 vs 193), and the cases identified only by the protein model were predicted to gain more life-years from 3 LDCT screens compared with the PLCOm2012-identified cases (4.2 vs. 3.6 life-years). Conclusions: Preliminary results suggest that the INTEGRAL protein panel can improve risk discrimination beyond questionnaire-based risk prediction models, as well as identify future cases who have more life-years to gain from screening. Final validation is planned for completion in early 2024, by analysis of fully independent data from 7 additional LC3 cohorts (n=560 cases and 1,126 sub-cohort representatives). Citation Format: Hana Zahed, Karine Alcala, David C. Muller, Rayjean Hung, Mattias Johansson, Hilary A. Robbins, The Lung Cancer Cohort Consortium. Lung cancer risk assessment with the INTEGRAL protein panel: Preliminary results from development and validation in the Lung Cancer Cohort Consortium [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6095.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 teacher head, 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".