Influence of patients’ sex on radiomics-based predictions of lung artery thrombus
- 1Institute of Biomedical Engineering and Informatics, Technische Universität Ilmenau, Ilmenau, Germany
- 2Department of Medical Informatics and Statistics, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany
- 3Department of Radiology, Neuroradiology and Nuclear Medicine, Johannes Wesling University Hospital, Ruhr University Bochum, Minden, Germany
Abstract
Purpose: This study investigates possible differences of female and male patients’ data in regard to the prediction of mortality and troponin level of acute pulmonary embolism (APE) patients. Radiomics features are utilized to gain insights of sex-specific image properties, which should support APE diagnosis and treatment.
Methods: A database of 200 manually segmented APE patients’ computed tomographic data (100 female and 100 male) was used for the prediction of 30-day mortality and biomarker troponin level. The radiomics features were extracted from the segmented thrombi and reduced via correlation analysis and 12 feature selection followed each by 12 classification methods. For comparison, the weighted F1-Score, Matthew’s Correlation Coefficient, and balanced accuracy were used. For assessment of generalizability, the study was extended to 427 automatically selected datasets.
Results: Sex-dependent differences have been found in the resulting metrics and highest achieving feature selection and classification methods. Not stratifying the dataset by sex resulted in lower performance compared to the sex-stratified male and female only calculations. The subgroup of female patients achieved up to 31.04 % higher metric values compared to the other cohorts. Different thrombus properties contributed to the prediction based on the patients’ sex as well.
Conclusion: In this study, we analyzed possible correlations between sex of APE patients and mortality and troponin levels as well as radiomics features. Our analysis identified an influence of sex on the prediction task and image feature selection. Female patient data especially had higher predictive potential, which could help for further sex-stratified analysis of APE.
Related
- Influence of Perioperative Variables on Surgical Site Infection in Patients With Frailty and Prefrailty Undergoing Coronary Artery Bypass Grafting: Retrospective Multicenter Study
- Giant Aneurysm Involving the Sinus of Valsalva and the Right Coronary Artery with a Massive Intramural Thrombus
- Large, Life-Threatening Right Ventricular and Pulmonary Artery Thrombus Revealing Antiphospholipid Syndrome: The Role of Risk Stratification and Thrombolytic Therapy.
- Development and internal validation of CTA-based radiomics models for differentiating symptomatic from asymptomatic unruptured spontaneous intracranial vertebrobasilar artery dissection
- Radiomics performance to identify high-risk carotid artery plaques: a systematic review
BibTeX
@article{ehrhardt2026influence,
title = {{Influence of patients’ sex on radiomics-based predictions of lung artery thrombus}},
author = {Lea Ehrhardt and Georg Hille and Ehsan Amini and Alexey Surov and Jan Borggrefe and Patrique Fiedler and Sylvia Saalfeld},
journal = {International Journal of Computer Assisted Radiology and Surgery},
year = {2026},
month = jun,
doi = {10.1007/s11548-026-03704-1},
url = {https://doi.org/10.1007/s11548-026-03704-1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Influence of patients’ sex on radiomics-based predictions of lung artery thrombus
AU - Lea Ehrhardt
AU - Georg Hille
AU - Ehsan Amini
AU - Alexey Surov
AU - Jan Borggrefe
AU - Patrique Fiedler
AU - Sylvia Saalfeld
PY - 2026
DA - 2026/06/04
JO - International Journal of Computer Assisted Radiology and Surgery
DO - 10.1007/s11548-026-03704-1
UR - https://doi.org/10.1007/s11548-026-03704-1
AB - Purpose: This study investigates possible differences of female and male patients’ data in regard to the prediction of mortality and troponin level of acute pulmonary embolism (APE) patients. Radiomics features are utilized to gain insights of sex-specific image properties, which should support APE diagnosis and treatment. Methods: A database of 200 manually segmented APE patients’ computed tomographic data (100 female and 100 male) was used for the prediction of 30-day mortality and biomarker troponin level. The radiomics features were extracted from the segmented thrombi and reduced via correlation analysis and 12 feature selection followed each by 12 classification methods. For comparison, the weighted F1-Score, Matthew’s Correlation Coefficient, and balanced accuracy were used. For assessment of generalizability, the study was extended to 427 automatically selected datasets. Results: Sex-dependent differences have been found in the resulting metrics and highest achieving feature selection and classification methods. Not stratifying the dataset by sex resulted in lower performance compared to the sex-stratified male and female only calculations. The subgroup of female patients achieved up to 31.04 % higher metric values compared to the other cohorts. Different thrombus properties contributed to the prediction based on the patients’ sex as well. Conclusion: In this study, we analyzed possible correlations between sex of APE patients and mortality and troponin levels as well as radiomics features. Our analysis identified an influence of sex on the prediction task and image feature selection. Female patient data especially had higher predictive potential, which could help for further sex-stratified analysis of APE.
SN - 1861-6410
SN - 1861-6429
ER -