Tags

Type your tag names separated by a space and hit enter

Comparison of deep learning, radiomics and subjective assessment of chest CT findings in SARS-CoV-2 pneumonia.
Clin Imaging. 2021 Dec; 80:58-66.CI

Abstract

PURPOSE

Comparison of deep learning algorithm, radiomics and subjective assessment of chest CT for predicting outcome (death or recovery) and intensive care unit (ICU) admission in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.

METHODS

The multicenter, ethical committee-approved, retrospective study included non-contrast-enhanced chest CT of 221 SARS-CoV-2 positive patients from Italy (n = 196 patients; mean age 64 ± 16 years) and Denmark (n = 25; mean age 69 ± 13 years). A thoracic radiologist graded presence, type and extent of pulmonary opacities and severity of motion artifacts in each lung lobe on all chest CTs. Thin-section CT images were processed with CT Pneumonia Analysis Prototype (Siemens Healthineers) which yielded segmentation masks from a deep learning (DL) algorithm to derive features of lung abnormalities such as opacity scores, mean HU, as well as volume and percentage of all-attenuation and high-attenuation (opacities >-200 HU) opacities. Separately, whole lung radiomics were obtained for all CT exams. Analysis of variance and multiple logistic regression were performed for data analysis.

RESULTS

Moderate to severe respiratory motion artifacts affected nearly one-quarter of chest CTs in patients. Subjective severity assessment, DL-based features and radiomics predicted patient outcome (AUC 0.76 vs AUC 0.88 vs AUC 0.83) and need for ICU admission (AUC 0.77 vs AUC 0.0.80 vs 0.82). Excluding chest CT with motion artifacts, the performance of DL-based and radiomics features improve for predicting ICU admission.

CONCLUSION

DL-based and radiomics features of pulmonary opacities from chest CT were superior to subjective assessment for differentiating patients with favorable and adverse outcomes.

Authors+Show Affiliations

Massachusetts General Hospital, 55 Fruit Street, Boston, MA, USA.Massachusetts General Hospital, 55 Fruit Street, Boston, MA, USA. Electronic address: sebrahimian@mgh.harvard.edu.Ospedale Maggiore della Carita', Novara, Italy.Department of Cardiology, Department of Clinical Medicine, Aarhus University Hospital, Palle Juul Jensens Boulevard 99, 8200 Aarhus N, Denmark. Electronic address: jvh@clin.au.dk.Ospedale Maggiore della Carita', Novara, Italy.Department of Cardiology, Department of Clinical Medicine, Aarhus University Hospital, Palle Juul Jensens Boulevard 99, 8200 Aarhus N, Denmark. Electronic address: mads.dam@clin.au.dk.Siemens Healthcare GmbH, Diagnostic Imaging, Erlangen, Germany. Electronic address: mathis.zimmermann@siemens-healthineers.com.Siemens Healthcare GmbH, Diagnostic Imaging, Erlangen, Germany. Electronic address: felix.durlak.ext@siemens-healthineers.com.Siemens Healthcare GmbH, Diagnostic Imaging, Erlangen, Germany. Electronic address: matthias.mitschke@siemens-healthineers.com.Ospedale Maggiore della Carita', Novara, Italy.Department of Cardiology, Department of Clinical Medicine, Aarhus University Hospital, Palle Juul Jensens Boulevard 99, 8200 Aarhus N, Denmark.Massachusetts General Hospital, 55 Fruit Street, Boston, MA, USA. Electronic address: mkalra@mgh.harvard.edu.Azienda Ospedaliera Universitaria di Cagliari, Cagliari, Italy.

Pub Type(s)

Journal Article
Multicenter Study

Language

eng

PubMed ID

34246044

Citation

Arru, Chiara, et al. "Comparison of Deep Learning, Radiomics and Subjective Assessment of Chest CT Findings in SARS-CoV-2 Pneumonia." Clinical Imaging, vol. 80, 2021, pp. 58-66.
Arru C, Ebrahimian S, Falaschi Z, et al. Comparison of deep learning, radiomics and subjective assessment of chest CT findings in SARS-CoV-2 pneumonia. Clin Imaging. 2021;80:58-66.
Arru, C., Ebrahimian, S., Falaschi, Z., Hansen, J. V., Pasche, A., Lyhne, M. D., Zimmermann, M., Durlak, F., Mitschke, M., Carriero, A., Nielsen-Kudsk, J. E., Kalra, M. K., & Saba, L. (2021). Comparison of deep learning, radiomics and subjective assessment of chest CT findings in SARS-CoV-2 pneumonia. Clinical Imaging, 80, 58-66. https://doi.org/10.1016/j.clinimag.2021.06.036
Arru C, et al. Comparison of Deep Learning, Radiomics and Subjective Assessment of Chest CT Findings in SARS-CoV-2 Pneumonia. Clin Imaging. 2021;80:58-66. PubMed PMID: 34246044.
* Article titles in AMA citation format should be in sentence-case
TY - JOUR T1 - Comparison of deep learning, radiomics and subjective assessment of chest CT findings in SARS-CoV-2 pneumonia. AU - Arru,Chiara, AU - Ebrahimian,Shadi, AU - Falaschi,Zeno, AU - Hansen,Jacob Valentin, AU - Pasche,Alessio, AU - Lyhne,Mads Dam, AU - Zimmermann,Mathis, AU - Durlak,Felix, AU - Mitschke,Matthias, AU - Carriero,Alessandro, AU - Nielsen-Kudsk,Jens Erik, AU - Kalra,Mannudeep K, AU - Saba,Luca, Y1 - 2021/07/01/ PY - 2021/02/22/received PY - 2021/06/23/revised PY - 2021/06/28/accepted PY - 2021/7/11/pubmed PY - 2021/11/16/medline PY - 2021/7/10/entrez KW - Chest CT KW - Deep learning KW - Pneumonia KW - Radiomics KW - SARS-CoV-2 SP - 58 EP - 66 JF - Clinical imaging JO - Clin Imaging VL - 80 N2 - PURPOSE: Comparison of deep learning algorithm, radiomics and subjective assessment of chest CT for predicting outcome (death or recovery) and intensive care unit (ICU) admission in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. METHODS: The multicenter, ethical committee-approved, retrospective study included non-contrast-enhanced chest CT of 221 SARS-CoV-2 positive patients from Italy (n = 196 patients; mean age 64 ± 16 years) and Denmark (n = 25; mean age 69 ± 13 years). A thoracic radiologist graded presence, type and extent of pulmonary opacities and severity of motion artifacts in each lung lobe on all chest CTs. Thin-section CT images were processed with CT Pneumonia Analysis Prototype (Siemens Healthineers) which yielded segmentation masks from a deep learning (DL) algorithm to derive features of lung abnormalities such as opacity scores, mean HU, as well as volume and percentage of all-attenuation and high-attenuation (opacities >-200 HU) opacities. Separately, whole lung radiomics were obtained for all CT exams. Analysis of variance and multiple logistic regression were performed for data analysis. RESULTS: Moderate to severe respiratory motion artifacts affected nearly one-quarter of chest CTs in patients. Subjective severity assessment, DL-based features and radiomics predicted patient outcome (AUC 0.76 vs AUC 0.88 vs AUC 0.83) and need for ICU admission (AUC 0.77 vs AUC 0.0.80 vs 0.82). Excluding chest CT with motion artifacts, the performance of DL-based and radiomics features improve for predicting ICU admission. CONCLUSION: DL-based and radiomics features of pulmonary opacities from chest CT were superior to subjective assessment for differentiating patients with favorable and adverse outcomes. SN - 1873-4499 UR - https://www.unboundmedicine.com/medline/citation/34246044/Comparison_of_deep_learning_radiomics_and_subjective_assessment_of_chest_CT_findings_in_SARS_CoV_2_pneumonia_ L2 - https://linkinghub.elsevier.com/retrieve/pii/S0899-7071(21)00282-5 DB - PRIME DP - Unbound Medicine ER -