Flore Belmans, Laura Seldeslachts, Eliane Vanhoffelen, Birger Tielemans, Wim Vos, Frederik Maes, Greetje Vande Velde
{"title":"利用深度学习在不同疾病模型的微ct上自动量化肺病理。","authors":"Flore Belmans, Laura Seldeslachts, Eliane Vanhoffelen, Birger Tielemans, Wim Vos, Frederik Maes, Greetje Vande Velde","doi":"10.1016/j.ebiom.2025.105904","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Micro-CT significantly enhances the efficiency, predictive power and translatability of animal studies to human clinical trials for respiratory diseases. However, the analysis of large micro-CT datasets remains a bottleneck.</p><p><strong>Methods: </strong>We developed a generic deep learning (DL)-based lung segmentation model using longitudinal micro-CT images from studies of Down syndrome, viral and fungal infections, and exacerbation with variable lung pathology and degree of disease burden. 2D models were trained with cross-validation on axial, coronal and sagittal slices. Predictions from these single-orientation models were combined to create a 2.5D model using majority voting or probability averaging. The generalisability of these models to other studies (COVID-19, lung inflammation and fibrosis), scanner configurations and rodent species (rats, hamsters, degus) was tested, including a publicly available database.</p><p><strong>Findings: </strong>On the internal validation data, the highest mean Dice Similarity Coefficient (DSC) was found for the 2.5D probability averaging model (0.953 ± 0.023), further improving the output of the 2D models by removing erroneous voxels outside the lung region. The models demonstrated good generalisability with average DSC values ranging from 0.89 to 0.94 across different lung pathologies and scanner configurations. The biomarkers extracted from manual and automated segmentations are well in agreement and proved that our proposed solution effectively monitors longitudinal lung pathology development and response to treatment in real-world preclinical studies.</p><p><strong>Interpretation: </strong>Our DL-based pipeline for lung pathology quantification offers efficient analysis of large micro-CT datasets, is widely applicable across rodent disease models and acquisition protocols and enables real-time insights into therapy efficacy.</p><p><strong>Funding: </strong>This research was supported by the Service Public de Wallonie (AEROVID grant to FB, WV) and The Flemish Research Foundation (FWO, doctoral mandate 1SF2224N to EV and 1186121N/1186123N to LS, infrastructure grant I006524N to GVV).</p>","PeriodicalId":11494,"journal":{"name":"EBioMedicine","volume":"119 ","pages":"105904"},"PeriodicalIF":10.8000,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12419106/pdf/","citationCount":"0","resultStr":"{\"title\":\"Automated quantification of lung pathology on micro-CT in diverse disease models using deep learning.\",\"authors\":\"Flore Belmans, Laura Seldeslachts, Eliane Vanhoffelen, Birger Tielemans, Wim Vos, Frederik Maes, Greetje Vande Velde\",\"doi\":\"10.1016/j.ebiom.2025.105904\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>Micro-CT significantly enhances the efficiency, predictive power and translatability of animal studies to human clinical trials for respiratory diseases. However, the analysis of large micro-CT datasets remains a bottleneck.</p><p><strong>Methods: </strong>We developed a generic deep learning (DL)-based lung segmentation model using longitudinal micro-CT images from studies of Down syndrome, viral and fungal infections, and exacerbation with variable lung pathology and degree of disease burden. 2D models were trained with cross-validation on axial, coronal and sagittal slices. Predictions from these single-orientation models were combined to create a 2.5D model using majority voting or probability averaging. The generalisability of these models to other studies (COVID-19, lung inflammation and fibrosis), scanner configurations and rodent species (rats, hamsters, degus) was tested, including a publicly available database.</p><p><strong>Findings: </strong>On the internal validation data, the highest mean Dice Similarity Coefficient (DSC) was found for the 2.5D probability averaging model (0.953 ± 0.023), further improving the output of the 2D models by removing erroneous voxels outside the lung region. The models demonstrated good generalisability with average DSC values ranging from 0.89 to 0.94 across different lung pathologies and scanner configurations. The biomarkers extracted from manual and automated segmentations are well in agreement and proved that our proposed solution effectively monitors longitudinal lung pathology development and response to treatment in real-world preclinical studies.</p><p><strong>Interpretation: </strong>Our DL-based pipeline for lung pathology quantification offers efficient analysis of large micro-CT datasets, is widely applicable across rodent disease models and acquisition protocols and enables real-time insights into therapy efficacy.</p><p><strong>Funding: </strong>This research was supported by the Service Public de Wallonie (AEROVID grant to FB, WV) and The Flemish Research Foundation (FWO, doctoral mandate 1SF2224N to EV and 1186121N/1186123N to LS, infrastructure grant I006524N to GVV).</p>\",\"PeriodicalId\":11494,\"journal\":{\"name\":\"EBioMedicine\",\"volume\":\"119 \",\"pages\":\"105904\"},\"PeriodicalIF\":10.8000,\"publicationDate\":\"2025-09-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12419106/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"EBioMedicine\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1016/j.ebiom.2025.105904\",\"RegionNum\":1,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2025/8/30 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q1\",\"JCRName\":\"MEDICINE, RESEARCH & EXPERIMENTAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"EBioMedicine","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.ebiom.2025.105904","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/8/30 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"MEDICINE, RESEARCH & EXPERIMENTAL","Score":null,"Total":0}
Automated quantification of lung pathology on micro-CT in diverse disease models using deep learning.
Background: Micro-CT significantly enhances the efficiency, predictive power and translatability of animal studies to human clinical trials for respiratory diseases. However, the analysis of large micro-CT datasets remains a bottleneck.
Methods: We developed a generic deep learning (DL)-based lung segmentation model using longitudinal micro-CT images from studies of Down syndrome, viral and fungal infections, and exacerbation with variable lung pathology and degree of disease burden. 2D models were trained with cross-validation on axial, coronal and sagittal slices. Predictions from these single-orientation models were combined to create a 2.5D model using majority voting or probability averaging. The generalisability of these models to other studies (COVID-19, lung inflammation and fibrosis), scanner configurations and rodent species (rats, hamsters, degus) was tested, including a publicly available database.
Findings: On the internal validation data, the highest mean Dice Similarity Coefficient (DSC) was found for the 2.5D probability averaging model (0.953 ± 0.023), further improving the output of the 2D models by removing erroneous voxels outside the lung region. The models demonstrated good generalisability with average DSC values ranging from 0.89 to 0.94 across different lung pathologies and scanner configurations. The biomarkers extracted from manual and automated segmentations are well in agreement and proved that our proposed solution effectively monitors longitudinal lung pathology development and response to treatment in real-world preclinical studies.
Interpretation: Our DL-based pipeline for lung pathology quantification offers efficient analysis of large micro-CT datasets, is widely applicable across rodent disease models and acquisition protocols and enables real-time insights into therapy efficacy.
Funding: This research was supported by the Service Public de Wallonie (AEROVID grant to FB, WV) and The Flemish Research Foundation (FWO, doctoral mandate 1SF2224N to EV and 1186121N/1186123N to LS, infrastructure grant I006524N to GVV).
EBioMedicineBiochemistry, Genetics and Molecular Biology-General Biochemistry,Genetics and Molecular Biology
CiteScore
17.70
自引率
0.90%
发文量
579
审稿时长
5 weeks
期刊介绍:
eBioMedicine is a comprehensive biomedical research journal that covers a wide range of studies that are relevant to human health. Our focus is on original research that explores the fundamental factors influencing human health and disease, including the discovery of new therapeutic targets and treatments, the identification of biomarkers and diagnostic tools, and the investigation and modification of disease pathways and mechanisms. We welcome studies from any biomedical discipline that contribute to our understanding of disease and aim to improve human health.