Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI
Background Quantification of aortic morphology plays an important role in the evaluation and follow‐up assessment of patients with aortic diseases, but often requires labor‐intensive and operator‐dependent measurements. Automatic solutions would help enhance their quality and reproducibility. Purpos...
Gespeichert in:
| Veröffentlicht in: | Journal of magnetic resonance imaging 2024-10, Vol.60 (4), p.1565-1576 |
|---|---|
| Hauptverfasser: | , , , , , , , , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 1576 |
|---|---|
| container_issue | 4 |
| container_start_page | 1565 |
| container_title | Journal of magnetic resonance imaging |
| container_volume | 60 |
| creator | Guo, Jia Bouaou, Kevin Houriez‐‐Gombaud‐Saintonge, Sophia Gueda, Moussa Gencer, Umit Nguyen, Vincent Charpentier, Etienne Soulat, Gilles Redheuil, Alban Mousseaux, Elie Kachenoura, Nadjia Dietenbeck, Thomas |
| description | Background
Quantification of aortic morphology plays an important role in the evaluation and follow‐up assessment of patients with aortic diseases, but often requires labor‐intensive and operator‐dependent measurements. Automatic solutions would help enhance their quality and reproducibility.
Purpose
To design a deep learning (DL)‐based automated approach for aortic landmarks and lumen detection derived from three‐dimensional (3D) MRI.
Study Type
Retrospective.
Population
Three hundred ninety‐one individuals (female: 47%, age = 51.9 ± 18.4) from three sites, including healthy subjects and patients (hypertension, aortic dilation, Turner syndrome), randomly divided into training/validation/test datasets (N = 236/77/78). Twenty‐five subjects were randomly selected and analyzed by three operators with different levels of expertise.
Field Strength/Sequence
1.5‐T and 3‐T, 3D spoiled gradient‐recalled or steady‐state free precession sequences.
Assessment
Reinforcement learning and a two‐stage network trained using reference landmarks and segmentation from an existing semi‐automatic software were used for aortic landmark detection and segmentation from sinotubular junction to coeliac trunk. Aortic segments were defined using the detected landmarks while the aortic centerline was extracted from the segmentation and morphological indices (length, aortic diameter, and volume) were computed for both the reference and the proposed segmentations.
Statistical Tests
Segmentation: Dice similarity coefficient (DSC), Hausdorff distance (HD), average symmetrical surface distance (ASSD); landmark detection: Euclidian distance (ED); model robustness: Spearman correlation, Bland–Altman analysis, Kruskal–Wallis test for comparisons between reference and DL‐derived aortic indices; inter‐observer study: Williams index (WI). A WI 95% confidence interval (CI) lower bound >1 indicates that the method is within the inter‐observer variability. A P‐value 0.95, mean bias |
| doi_str_mv | 10.1002/jmri.29236 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_hal_p</sourceid><recordid>TN_cdi_hal_primary_oai_HAL_hal_04393779v1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3128154851</sourcerecordid><originalsourceid>FETCH-LOGICAL-c3866-c9897bd0f91cdfac70f57198f565c688c0968e4d16a8f0c552b705faf9f1b6d73</originalsourceid><addsrcrecordid>eNp9kctOAyEUhonReKlufAAziRs1GeUyMLCsd02NxuiaUAYszcxQodV05yP4jD6J1FEXLlxByHe-cP4fgG0EDxGE-GjcBHeIBSZsCawjinGOKWfL6Q4pyRGH5RrYiHEMIRSioKtgjXCMGC3YOrg7NWaSDYwKrWufPt7ej1U0VdZvVT2PLmbeZn0fpk5nNz5MRr72T_PsPPgmexgFY9LAqWtMG51PE9nN_dUmWLGqjmbr--yBx_Ozh5PLfHB7cXXSH-SacMZyLbgohxW0AunKKl1CS0skuKWMasa5hoJxU1SIKW6hphQPS0itssKiIatK0gP7nXekajkJrlFhLr1y8rI_kIs3WBBBylK8oMTudewk-OeZiVPZuKhNXavW-FmUKTsBCSEprh7Y_YOO_Syk3aIkCHNEC04XwoOO0sHHGIz9_QGCctGJXHQivzpJ8M63cjZsTPWL_pSQANQBr642839U8joF3Ek_AZEIllE</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3128154851</pqid></control><display><type>article</type><title>Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI</title><source>MEDLINE</source><source>Wiley Journals</source><creator>Guo, Jia ; Bouaou, Kevin ; Houriez‐‐Gombaud‐Saintonge, Sophia ; Gueda, Moussa ; Gencer, Umit ; Nguyen, Vincent ; Charpentier, Etienne ; Soulat, Gilles ; Redheuil, Alban ; Mousseaux, Elie ; Kachenoura, Nadjia ; Dietenbeck, Thomas</creator><creatorcontrib>Guo, Jia ; Bouaou, Kevin ; Houriez‐‐Gombaud‐Saintonge, Sophia ; Gueda, Moussa ; Gencer, Umit ; Nguyen, Vincent ; Charpentier, Etienne ; Soulat, Gilles ; Redheuil, Alban ; Mousseaux, Elie ; Kachenoura, Nadjia ; Dietenbeck, Thomas</creatorcontrib><description>Background
Quantification of aortic morphology plays an important role in the evaluation and follow‐up assessment of patients with aortic diseases, but often requires labor‐intensive and operator‐dependent measurements. Automatic solutions would help enhance their quality and reproducibility.
Purpose
To design a deep learning (DL)‐based automated approach for aortic landmarks and lumen detection derived from three‐dimensional (3D) MRI.
Study Type
Retrospective.
Population
Three hundred ninety‐one individuals (female: 47%, age = 51.9 ± 18.4) from three sites, including healthy subjects and patients (hypertension, aortic dilation, Turner syndrome), randomly divided into training/validation/test datasets (N = 236/77/78). Twenty‐five subjects were randomly selected and analyzed by three operators with different levels of expertise.
Field Strength/Sequence
1.5‐T and 3‐T, 3D spoiled gradient‐recalled or steady‐state free precession sequences.
Assessment
Reinforcement learning and a two‐stage network trained using reference landmarks and segmentation from an existing semi‐automatic software were used for aortic landmark detection and segmentation from sinotubular junction to coeliac trunk. Aortic segments were defined using the detected landmarks while the aortic centerline was extracted from the segmentation and morphological indices (length, aortic diameter, and volume) were computed for both the reference and the proposed segmentations.
Statistical Tests
Segmentation: Dice similarity coefficient (DSC), Hausdorff distance (HD), average symmetrical surface distance (ASSD); landmark detection: Euclidian distance (ED); model robustness: Spearman correlation, Bland–Altman analysis, Kruskal–Wallis test for comparisons between reference and DL‐derived aortic indices; inter‐observer study: Williams index (WI). A WI 95% confidence interval (CI) lower bound >1 indicates that the method is within the inter‐observer variability. A P‐value <0.05 was considered statistically significant.
Results
DSC was 0.90 ± 0.05, HD was 12.11 ± 7.79 mm, and ASSD was 1.07 ± 0.63 mm. ED was 5.0 ± 6.1 mm. A good agreement was found between all DL‐derived and reference aortic indices (r >0.95, mean bias <7%). Our segmentation and landmark detection performances were within the inter‐observer variability except the sinotubular junction landmark (CI = 0.96;1.04).
Data Conclusion
A DL‐based aortic segmentation and anatomical landmark detection approach was developed and applied to 3D MRI data for achieve aortic morphology evaluation.
Evidence Level
3
Technical Efficacy
Stage 2</description><identifier>ISSN: 1053-1807</identifier><identifier>ISSN: 1522-2586</identifier><identifier>EISSN: 1522-2586</identifier><identifier>DOI: 10.1002/jmri.29236</identifier><identifier>PMID: 38216546</identifier><language>eng</language><publisher>Hoboken, USA: John Wiley & Sons, Inc</publisher><subject>3D cardiac MRI ; Adult ; Aged ; Aorta ; Aorta - diagnostic imaging ; Aortic Diseases - diagnostic imaging ; Artificial Intelligence ; Computer Science ; Confidence intervals ; Deep Learning ; Dimensional analysis ; Engineering Sciences ; Evaluation ; Female ; Field strength ; Humans ; Hypertension ; Hypertension - diagnostic imaging ; Image Interpretation, Computer-Assisted - methods ; Image processing ; Image Processing, Computer-Assisted - methods ; Image segmentation ; Imaging, Three-Dimensional - methods ; Lower bounds ; Magnetic resonance imaging ; Magnetic Resonance Imaging - methods ; Male ; Medical Imaging ; Metric space ; Middle Aged ; Morphology ; Operators ; Population studies ; Reproducibility of Results ; Retrospective Studies ; Segmentation ; Signal and Image processing ; Statistical analysis ; Statistical tests ; Turner Syndrome - diagnostic imaging ; Turner's syndrome</subject><ispartof>Journal of magnetic resonance imaging, 2024-10, Vol.60 (4), p.1565-1576</ispartof><rights>2024 International Society for Magnetic Resonance in Medicine.</rights><rights>2024 International Society for Magnetic Resonance in Medicine</rights><rights>Distributed under a Creative Commons Attribution 4.0 International License</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c3866-c9897bd0f91cdfac70f57198f565c688c0968e4d16a8f0c552b705faf9f1b6d73</cites><orcidid>0009-0005-0870-9844 ; 0000-0003-4165-6113 ; 0000-0001-5643-0497 ; 0000-0002-8076-1445 ; 0000-0002-9893-9694 ; 0000-0002-1002-9902</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://onlinelibrary.wiley.com/doi/pdf/10.1002%2Fjmri.29236$$EPDF$$P50$$Gwiley$$H</linktopdf><linktohtml>$$Uhttps://onlinelibrary.wiley.com/doi/full/10.1002%2Fjmri.29236$$EHTML$$P50$$Gwiley$$H</linktohtml><link.rule.ids>230,314,780,784,885,1417,27924,27925,45574,45575</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/38216546$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink><backlink>$$Uhttps://hal.sorbonne-universite.fr/hal-04393779$$DView record in HAL$$Hfree_for_read</backlink></links><search><creatorcontrib>Guo, Jia</creatorcontrib><creatorcontrib>Bouaou, Kevin</creatorcontrib><creatorcontrib>Houriez‐‐Gombaud‐Saintonge, Sophia</creatorcontrib><creatorcontrib>Gueda, Moussa</creatorcontrib><creatorcontrib>Gencer, Umit</creatorcontrib><creatorcontrib>Nguyen, Vincent</creatorcontrib><creatorcontrib>Charpentier, Etienne</creatorcontrib><creatorcontrib>Soulat, Gilles</creatorcontrib><creatorcontrib>Redheuil, Alban</creatorcontrib><creatorcontrib>Mousseaux, Elie</creatorcontrib><creatorcontrib>Kachenoura, Nadjia</creatorcontrib><creatorcontrib>Dietenbeck, Thomas</creatorcontrib><title>Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI</title><title>Journal of magnetic resonance imaging</title><addtitle>J Magn Reson Imaging</addtitle><description>Background
Quantification of aortic morphology plays an important role in the evaluation and follow‐up assessment of patients with aortic diseases, but often requires labor‐intensive and operator‐dependent measurements. Automatic solutions would help enhance their quality and reproducibility.
Purpose
To design a deep learning (DL)‐based automated approach for aortic landmarks and lumen detection derived from three‐dimensional (3D) MRI.
Study Type
Retrospective.
Population
Three hundred ninety‐one individuals (female: 47%, age = 51.9 ± 18.4) from three sites, including healthy subjects and patients (hypertension, aortic dilation, Turner syndrome), randomly divided into training/validation/test datasets (N = 236/77/78). Twenty‐five subjects were randomly selected and analyzed by three operators with different levels of expertise.
Field Strength/Sequence
1.5‐T and 3‐T, 3D spoiled gradient‐recalled or steady‐state free precession sequences.
Assessment
Reinforcement learning and a two‐stage network trained using reference landmarks and segmentation from an existing semi‐automatic software were used for aortic landmark detection and segmentation from sinotubular junction to coeliac trunk. Aortic segments were defined using the detected landmarks while the aortic centerline was extracted from the segmentation and morphological indices (length, aortic diameter, and volume) were computed for both the reference and the proposed segmentations.
Statistical Tests
Segmentation: Dice similarity coefficient (DSC), Hausdorff distance (HD), average symmetrical surface distance (ASSD); landmark detection: Euclidian distance (ED); model robustness: Spearman correlation, Bland–Altman analysis, Kruskal–Wallis test for comparisons between reference and DL‐derived aortic indices; inter‐observer study: Williams index (WI). A WI 95% confidence interval (CI) lower bound >1 indicates that the method is within the inter‐observer variability. A P‐value <0.05 was considered statistically significant.
Results
DSC was 0.90 ± 0.05, HD was 12.11 ± 7.79 mm, and ASSD was 1.07 ± 0.63 mm. ED was 5.0 ± 6.1 mm. A good agreement was found between all DL‐derived and reference aortic indices (r >0.95, mean bias <7%). Our segmentation and landmark detection performances were within the inter‐observer variability except the sinotubular junction landmark (CI = 0.96;1.04).
Data Conclusion
A DL‐based aortic segmentation and anatomical landmark detection approach was developed and applied to 3D MRI data for achieve aortic morphology evaluation.
Evidence Level
3
Technical Efficacy
Stage 2</description><subject>3D cardiac MRI</subject><subject>Adult</subject><subject>Aged</subject><subject>Aorta</subject><subject>Aorta - diagnostic imaging</subject><subject>Aortic Diseases - diagnostic imaging</subject><subject>Artificial Intelligence</subject><subject>Computer Science</subject><subject>Confidence intervals</subject><subject>Deep Learning</subject><subject>Dimensional analysis</subject><subject>Engineering Sciences</subject><subject>Evaluation</subject><subject>Female</subject><subject>Field strength</subject><subject>Humans</subject><subject>Hypertension</subject><subject>Hypertension - diagnostic imaging</subject><subject>Image Interpretation, Computer-Assisted - methods</subject><subject>Image processing</subject><subject>Image Processing, Computer-Assisted - methods</subject><subject>Image segmentation</subject><subject>Imaging, Three-Dimensional - methods</subject><subject>Lower bounds</subject><subject>Magnetic resonance imaging</subject><subject>Magnetic Resonance Imaging - methods</subject><subject>Male</subject><subject>Medical Imaging</subject><subject>Metric space</subject><subject>Middle Aged</subject><subject>Morphology</subject><subject>Operators</subject><subject>Population studies</subject><subject>Reproducibility of Results</subject><subject>Retrospective Studies</subject><subject>Segmentation</subject><subject>Signal and Image processing</subject><subject>Statistical analysis</subject><subject>Statistical tests</subject><subject>Turner Syndrome - diagnostic imaging</subject><subject>Turner's syndrome</subject><issn>1053-1807</issn><issn>1522-2586</issn><issn>1522-2586</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>EIF</sourceid><recordid>eNp9kctOAyEUhonReKlufAAziRs1GeUyMLCsd02NxuiaUAYszcxQodV05yP4jD6J1FEXLlxByHe-cP4fgG0EDxGE-GjcBHeIBSZsCawjinGOKWfL6Q4pyRGH5RrYiHEMIRSioKtgjXCMGC3YOrg7NWaSDYwKrWufPt7ej1U0VdZvVT2PLmbeZn0fpk5nNz5MRr72T_PsPPgmexgFY9LAqWtMG51PE9nN_dUmWLGqjmbr--yBx_Ozh5PLfHB7cXXSH-SacMZyLbgohxW0AunKKl1CS0skuKWMasa5hoJxU1SIKW6hphQPS0itssKiIatK0gP7nXekajkJrlFhLr1y8rI_kIs3WBBBylK8oMTudewk-OeZiVPZuKhNXavW-FmUKTsBCSEprh7Y_YOO_Syk3aIkCHNEC04XwoOO0sHHGIz9_QGCctGJXHQivzpJ8M63cjZsTPWL_pSQANQBr642839U8joF3Ek_AZEIllE</recordid><startdate>202410</startdate><enddate>202410</enddate><creator>Guo, Jia</creator><creator>Bouaou, Kevin</creator><creator>Houriez‐‐Gombaud‐Saintonge, Sophia</creator><creator>Gueda, Moussa</creator><creator>Gencer, Umit</creator><creator>Nguyen, Vincent</creator><creator>Charpentier, Etienne</creator><creator>Soulat, Gilles</creator><creator>Redheuil, Alban</creator><creator>Mousseaux, Elie</creator><creator>Kachenoura, Nadjia</creator><creator>Dietenbeck, Thomas</creator><general>John Wiley & Sons, Inc</general><general>Wiley Subscription Services, Inc</general><general>Wiley-Blackwell</general><scope>CGR</scope><scope>CUY</scope><scope>CVF</scope><scope>ECM</scope><scope>EIF</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7QO</scope><scope>7TK</scope><scope>8FD</scope><scope>FR3</scope><scope>K9.</scope><scope>P64</scope><scope>7X8</scope><scope>1XC</scope><scope>VOOES</scope><orcidid>https://orcid.org/0009-0005-0870-9844</orcidid><orcidid>https://orcid.org/0000-0003-4165-6113</orcidid><orcidid>https://orcid.org/0000-0001-5643-0497</orcidid><orcidid>https://orcid.org/0000-0002-8076-1445</orcidid><orcidid>https://orcid.org/0000-0002-9893-9694</orcidid><orcidid>https://orcid.org/0000-0002-1002-9902</orcidid></search><sort><creationdate>202410</creationdate><title>Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI</title><author>Guo, Jia ; Bouaou, Kevin ; Houriez‐‐Gombaud‐Saintonge, Sophia ; Gueda, Moussa ; Gencer, Umit ; Nguyen, Vincent ; Charpentier, Etienne ; Soulat, Gilles ; Redheuil, Alban ; Mousseaux, Elie ; Kachenoura, Nadjia ; Dietenbeck, Thomas</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c3866-c9897bd0f91cdfac70f57198f565c688c0968e4d16a8f0c552b705faf9f1b6d73</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>3D cardiac MRI</topic><topic>Adult</topic><topic>Aged</topic><topic>Aorta</topic><topic>Aorta - diagnostic imaging</topic><topic>Aortic Diseases - diagnostic imaging</topic><topic>Artificial Intelligence</topic><topic>Computer Science</topic><topic>Confidence intervals</topic><topic>Deep Learning</topic><topic>Dimensional analysis</topic><topic>Engineering Sciences</topic><topic>Evaluation</topic><topic>Female</topic><topic>Field strength</topic><topic>Humans</topic><topic>Hypertension</topic><topic>Hypertension - diagnostic imaging</topic><topic>Image Interpretation, Computer-Assisted - methods</topic><topic>Image processing</topic><topic>Image Processing, Computer-Assisted - methods</topic><topic>Image segmentation</topic><topic>Imaging, Three-Dimensional - methods</topic><topic>Lower bounds</topic><topic>Magnetic resonance imaging</topic><topic>Magnetic Resonance Imaging - methods</topic><topic>Male</topic><topic>Medical Imaging</topic><topic>Metric space</topic><topic>Middle Aged</topic><topic>Morphology</topic><topic>Operators</topic><topic>Population studies</topic><topic>Reproducibility of Results</topic><topic>Retrospective Studies</topic><topic>Segmentation</topic><topic>Signal and Image processing</topic><topic>Statistical analysis</topic><topic>Statistical tests</topic><topic>Turner Syndrome - diagnostic imaging</topic><topic>Turner's syndrome</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Guo, Jia</creatorcontrib><creatorcontrib>Bouaou, Kevin</creatorcontrib><creatorcontrib>Houriez‐‐Gombaud‐Saintonge, Sophia</creatorcontrib><creatorcontrib>Gueda, Moussa</creatorcontrib><creatorcontrib>Gencer, Umit</creatorcontrib><creatorcontrib>Nguyen, Vincent</creatorcontrib><creatorcontrib>Charpentier, Etienne</creatorcontrib><creatorcontrib>Soulat, Gilles</creatorcontrib><creatorcontrib>Redheuil, Alban</creatorcontrib><creatorcontrib>Mousseaux, Elie</creatorcontrib><creatorcontrib>Kachenoura, Nadjia</creatorcontrib><creatorcontrib>Dietenbeck, Thomas</creatorcontrib><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Biotechnology Research Abstracts</collection><collection>Neurosciences Abstracts</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>ProQuest Health & Medical Complete (Alumni)</collection><collection>Biotechnology and BioEngineering Abstracts</collection><collection>MEDLINE - Academic</collection><collection>Hyper Article en Ligne (HAL)</collection><collection>Hyper Article en Ligne (HAL) (Open Access)</collection><jtitle>Journal of magnetic resonance imaging</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Guo, Jia</au><au>Bouaou, Kevin</au><au>Houriez‐‐Gombaud‐Saintonge, Sophia</au><au>Gueda, Moussa</au><au>Gencer, Umit</au><au>Nguyen, Vincent</au><au>Charpentier, Etienne</au><au>Soulat, Gilles</au><au>Redheuil, Alban</au><au>Mousseaux, Elie</au><au>Kachenoura, Nadjia</au><au>Dietenbeck, Thomas</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI</atitle><jtitle>Journal of magnetic resonance imaging</jtitle><addtitle>J Magn Reson Imaging</addtitle><date>2024-10</date><risdate>2024</risdate><volume>60</volume><issue>4</issue><spage>1565</spage><epage>1576</epage><pages>1565-1576</pages><issn>1053-1807</issn><issn>1522-2586</issn><eissn>1522-2586</eissn><abstract>Background
Quantification of aortic morphology plays an important role in the evaluation and follow‐up assessment of patients with aortic diseases, but often requires labor‐intensive and operator‐dependent measurements. Automatic solutions would help enhance their quality and reproducibility.
Purpose
To design a deep learning (DL)‐based automated approach for aortic landmarks and lumen detection derived from three‐dimensional (3D) MRI.
Study Type
Retrospective.
Population
Three hundred ninety‐one individuals (female: 47%, age = 51.9 ± 18.4) from three sites, including healthy subjects and patients (hypertension, aortic dilation, Turner syndrome), randomly divided into training/validation/test datasets (N = 236/77/78). Twenty‐five subjects were randomly selected and analyzed by three operators with different levels of expertise.
Field Strength/Sequence
1.5‐T and 3‐T, 3D spoiled gradient‐recalled or steady‐state free precession sequences.
Assessment
Reinforcement learning and a two‐stage network trained using reference landmarks and segmentation from an existing semi‐automatic software were used for aortic landmark detection and segmentation from sinotubular junction to coeliac trunk. Aortic segments were defined using the detected landmarks while the aortic centerline was extracted from the segmentation and morphological indices (length, aortic diameter, and volume) were computed for both the reference and the proposed segmentations.
Statistical Tests
Segmentation: Dice similarity coefficient (DSC), Hausdorff distance (HD), average symmetrical surface distance (ASSD); landmark detection: Euclidian distance (ED); model robustness: Spearman correlation, Bland–Altman analysis, Kruskal–Wallis test for comparisons between reference and DL‐derived aortic indices; inter‐observer study: Williams index (WI). A WI 95% confidence interval (CI) lower bound >1 indicates that the method is within the inter‐observer variability. A P‐value <0.05 was considered statistically significant.
Results
DSC was 0.90 ± 0.05, HD was 12.11 ± 7.79 mm, and ASSD was 1.07 ± 0.63 mm. ED was 5.0 ± 6.1 mm. A good agreement was found between all DL‐derived and reference aortic indices (r >0.95, mean bias <7%). Our segmentation and landmark detection performances were within the inter‐observer variability except the sinotubular junction landmark (CI = 0.96;1.04).
Data Conclusion
A DL‐based aortic segmentation and anatomical landmark detection approach was developed and applied to 3D MRI data for achieve aortic morphology evaluation.
Evidence Level
3
Technical Efficacy
Stage 2</abstract><cop>Hoboken, USA</cop><pub>John Wiley & Sons, Inc</pub><pmid>38216546</pmid><doi>10.1002/jmri.29236</doi><tpages>12</tpages><orcidid>https://orcid.org/0009-0005-0870-9844</orcidid><orcidid>https://orcid.org/0000-0003-4165-6113</orcidid><orcidid>https://orcid.org/0000-0001-5643-0497</orcidid><orcidid>https://orcid.org/0000-0002-8076-1445</orcidid><orcidid>https://orcid.org/0000-0002-9893-9694</orcidid><orcidid>https://orcid.org/0000-0002-1002-9902</orcidid><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1053-1807 |
| ispartof | Journal of magnetic resonance imaging, 2024-10, Vol.60 (4), p.1565-1576 |
| issn | 1053-1807 1522-2586 1522-2586 |
| language | eng |
| recordid | cdi_hal_primary_oai_HAL_hal_04393779v1 |
| source | MEDLINE; Wiley Journals |
| subjects | 3D cardiac MRI Adult Aged Aorta Aorta - diagnostic imaging Aortic Diseases - diagnostic imaging Artificial Intelligence Computer Science Confidence intervals Deep Learning Dimensional analysis Engineering Sciences Evaluation Female Field strength Humans Hypertension Hypertension - diagnostic imaging Image Interpretation, Computer-Assisted - methods Image processing Image Processing, Computer-Assisted - methods Image segmentation Imaging, Three-Dimensional - methods Lower bounds Magnetic resonance imaging Magnetic Resonance Imaging - methods Male Medical Imaging Metric space Middle Aged Morphology Operators Population studies Reproducibility of Results Retrospective Studies Segmentation Signal and Image processing Statistical analysis Statistical tests Turner Syndrome - diagnostic imaging Turner's syndrome |
| title | Deep Learning‐Based Analysis of Aortic Morphology From Three‐Dimensional MRI |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T07%3A55%3A48IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_hal_p&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Deep%20Learning%E2%80%90Based%20Analysis%20of%20Aortic%20Morphology%20From%20Three%E2%80%90Dimensional%20MRI&rft.jtitle=Journal%20of%20magnetic%20resonance%20imaging&rft.au=Guo,%20Jia&rft.date=2024-10&rft.volume=60&rft.issue=4&rft.spage=1565&rft.epage=1576&rft.pages=1565-1576&rft.issn=1053-1807&rft.eissn=1522-2586&rft_id=info:doi/10.1002/jmri.29236&rft_dat=%3Cproquest_hal_p%3E3128154851%3C/proquest_hal_p%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3128154851&rft_id=info:pmid/38216546&rfr_iscdi=true |