Senescence and Genotoxicity Signatures Across Clinical and Laboratory Outcomes in Patients at High Cardiovascular Risk: An Automated Machine Learning Study

Background. Cellular senescence, chromosomal instability, endothelial injury, and altered immune composition may jointly reflect biological aging and cardiometabolic dysfunction. We evaluated whether a fixed 12-biomarker panel provided reproducible feature-level clues for 11 pre-selected clinical and laboratory outcomes in a three-group cohort at high cardiovascular risk. Methods. The dataset included 90 patients suffered from acute coronary syndrome (ACS), divided equally into three groups (n = 30 each): pre-pandemic ACS (Group 1), ACS with concomitant laboratory-confirmed acute COVID-19 during the pandemic (Group 2), and ACS without concomitant acute COVID-19 during the COVID-19 pandemic period (Group 3). Four model families were evaluated separately using 5-fold cross-fitting repeated five times. All imputation, scaling, and model fitting were restricted to training folds. Feature evidence integrated held-out permutation importance, cross-fitted SHAP values, linear coefficient stability, Group+SEX-adjusted partial rank associations, within-group sensitivity analyses, and leave-one-group-out transportability. Results. The largest median improvements over Group+SEX baselines occurred for history of acute myocardial infarction with coronary revascularization procedures (Δ=0.161), glucose level (Δ=0.118), heart rate (Δ=0.096), and glycated hemoglobin percentage (Δ=0.091). Recurrent feature themes included senescence-associated heterochromatin foci and CD34 + progenitor cells for glycemic outcomes, micronuclei for heart rate and prior COVID-19, nucleoplasmic bridges for cardiac-history and inflammatory/immune outcomes, and CD31 hi circulating endothelial cells for leukocyte and granulocyte counts. Absolute performance remained weak for several outcomes despite stable feature rankings. Conclusions. Rather than predicting any single outcome with strong absolute accuracy, the same senescence and genotoxicity markers recurred as leading clues across metabolically, cardiac, and immunologically distinct outcomes, and did so consistently across three pathogenetic different cohorts. This cross-outcome stability suggests these markers may index a shared, stressor-independent dimension of cellular burden in ACS patients rather than outcome-specific biology, a hypothesis that warrants mechanistic follow-up.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-20
DOI
https://doi.org/10.21203/rs.3.rs-11524324/v1
Primary Topic
Telomeres, Telomerase, and Senescence
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Senescence and Genotoxicity Signatures Across Clinical and Laboratory Outcomes in Patients at High Cardiovascular Risk: An Automated Machine Learning Study

Alina Ustymenko, Dmytro Krasnienkov, Oleksandr Parkhomenko, Kostiantyn Midlovets et al.
2 links
Research Square
Telomeres, Telomerase, and Senescence
preprint

Senescence and Genotoxicity Signatures Across Clinical and Laboratory Outcomes in Patients at High Cardiovascular Risk: An Automated Machine Learning Study

Alina Ustymenko, Dmytro Krasnienkov, Oleksandr Parkhomenko, Kostiantyn Midlovets, Vitalii Kyryk, Oleg Tsupykov, Dmytro Khomiakov, Oleksandr Skarzhevskyi
preprint en

Abstract

Background. Cellular senescence, chromosomal instability, endothelial injury, and altered immune composition may jointly reflect biological aging and cardiometabolic dysfunction. We evaluated whether a fixed 12-biomarker panel provided reproducible feature-level clues for 11 pre-selected clinical and laboratory outcomes in a three-group cohort at high cardiovascular risk. Methods. The dataset included 90 patients suffered from acute coronary syndrome (ACS), divided equally into three groups (n = 30 each): pre-pandemic ACS (Group 1), ACS with concomitant laboratory-confirmed acute COVID-19 during the pandemic (Group 2), and ACS without concomitant acute COVID-19 during the COVID-19 pandemic period (Group 3). Four model families were evaluated separately using 5-fold cross-fitting repeated five times. All imputation, scaling, and model fitting were restricted to training folds. Feature evidence integrated held-out permutation importance, cross-fitted SHAP values, linear coefficient stability, Group+SEX-adjusted partial rank associations, within-group sensitivity analyses, and leave-one-group-out transportability. Results. The largest median improvements over Group+SEX baselines occurred for history of acute myocardial infarction with coronary revascularization procedures (Δ=0.161), glucose level (Δ=0.118), heart rate (Δ=0.096), and glycated hemoglobin percentage (Δ=0.091). Recurrent feature themes included senescence-associated heterochromatin foci and CD34 + progenitor cells for glycemic outcomes, micronuclei for heart rate and prior COVID-19, nucleoplasmic bridges for cardiac-history and inflammatory/immune outcomes, and CD31 hi circulating endothelial cells for leukocyte and granulocyte counts. Absolute performance remained weak for several outcomes despite stable feature rankings. Conclusions. Rather than predicting any single outcome with strong absolute accuracy, the same senescence and genotoxicity markers recurred as leading clues across metabolically, cardiac, and immunologically distinct outcomes, and did so consistently across three pathogenetic different cohorts. This cross-outcome stability suggests these markers may index a shared, stressor-independent dimension of cellular burden in ACS patients rather than outcome-specific biology, a hypothesis that warrants mechanistic follow-up.

Research Square
Bogomoletz Institute of Physiology (UA), D. F. Chebotarev State Institute of Gerontology of the National Academy of Medical Sciences of Ukraine (UA), National Academy of Medical Sciences of Ukraine (UA), National Scientific Center "M.D. Strazhesko Institute of Cardiology" (UA)
Good health and well-being
Telomeres, Telomerase, and Senescence
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon