Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study
The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutiona...
Gespeichert in:
| Veröffentlicht in: | The lancet oncology 2024-03, Vol.25 (3), p.400-410 |
|---|---|
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 410 |
|---|---|
| container_issue | 3 |
| container_start_page | 400 |
| container_title | The lancet oncology |
| container_volume | 25 |
| creator | Rastogi, Aditya Brugnara, Gianluca Foltyn-Dumitru, Martha Mahmutoglu, Mustafa Ahmed Preetha, Chandrakanth J Kobler, Erich Pflüger, Irada Schell, Marianne Deike-Hofmann, Katerina Kessler, Tobias van den Bent, Martin J Idbaih, Ahmed Platten, Michael Brandes, Alba A Nabors, Burt Stupp, Roger Bernhardt, Denise Debus, Jürgen Abdollahi, Amir Gorlia, Thierry Tonn, Jörg-Christian Weller, Michael Maier-Hein, Klaus H Radbruch, Alexander Wick, Wolfgang Bendszus, Martin Meredig, Hagen Kurz, Felix T Vollmuth, Philipp |
| description | The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutional neural network (dCNN) optimisation method for MRI reconstruction and to reduce scan times and evaluate its effect on image quality and accuracy of oncological imaging biomarkers.
In this multicentre, retrospective, cohort study, MRI data from patients with glioblastoma treated at Heidelberg University Hospital (775 patients and 775 examinations) and from the phase 2 CORE trial (260 patients, 1083 examinations, and 58 institutions) and the phase 3 CENTRIC trial (505 patients, 3147 examinations, and 139 institutions) were used to develop, train, and test dCNN for reconstructing MRI from highly undersampled single-coil k-space data with various acceleration rates (R=2, 4, 6, 8, 10, and 15). Independent testing was performed with MRIs from the phase 2/3 EORTC-26101 trial (528 patients with glioblastoma, 1974 examinations, and 32 institutions). The similarity between undersampled dCNN-reconstructed and original MRIs was quantified with various image quality metrics, including structural similarity index measure (SSIM) and the accuracy of undersampled dCNN-reconstructed MRI on downstream radiological assessment of imaging biomarkers in oncology (automated artificial intelligence-based quantification of tumour burden and treatment response) was performed in the EORTC-26101 test dataset. The public NYU Langone Health fastMRI brain test dataset (558 patients and 558 examinations) was used to validate the generalisability and robustness of the dCNN for reconstructing MRIs from available multi-coil (parallel imaging) k-space data.
In the EORTC-26101 test dataset, the median SSIM of undersampled dCNN-reconstructed MRI ranged from 0·88 to 0·99 across different acceleration rates, with 0·92 (95% CI 0·92–0·93) for 10-times acceleration (R=10). The 10-times undersampled dCNN-reconstructed MRI yielded excellent agreement with original MRI when assessing volumes of contrast-enhancing tumour (median DICE for spatial agreement of 0·89 [95% CI 0·88 to 0·89]; median volume difference of 0·01 cm3 [95% CI 0·00 to 0·03] equalling 0·21%; p=0·0036 for equivalence) or non-enhancing tumour or oedema (median DICE of 0·94 [95% CI 0·94 to 0·95]; median volume difference of –0·79 cm3 [95% CI |
| doi_str_mv | 10.1016/S1470-2045(23)00641-1 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2934274720</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S1470204523006411</els_id><sourcerecordid>3039714510</sourcerecordid><originalsourceid>FETCH-LOGICAL-c440t-b80cd781821539dc24915b5f5d19043f35d24c32d663613ec483b004acd0b6823</originalsourceid><addsrcrecordid>eNqFkUlv1TAURiMEoqXwE0CW2BSJwPWUgQ1CZapUhMSwthz7BlwlduqhqP8ev_cKCzasPJ372b6naR5TeEGBdi-_UtFDy0DIU8afAXSCtvROc1y3RSvFMNzdzw_IUfMgpUsA2lOQ95sjPgjGQbLj5tdbxK1dUEfv_I920gktiWiCTzkWk13wJMykeIsx6XVb6vGnL-ckh0rZYpAkoz3JbsX0imiyliU7gz5HfF6JHEPasMZc16UJP0PMJOVibx4292a9JHx0O54039-_-3b2sb34_OH87M1Fa4SA3E4DGNsPdGBU8tEaJkYqJzlLS0cQfObSMmE4s13HO8rRiIFPAEIbC1M3MH7SnB5ytxiuCqasVpcMLov2GEpSbOSC9aJnUNGn_6CXoURfX6c48LGnQtIdJQ-UqV9LEWe1RbfqeKMoqJ0ZtTejdm1XjKu9GUVr3ZPb9DKtaP9W_VFRgdcHAGs7rh1GlYxDb9C66iMrG9x_rvgNdE6c8g</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3039714510</pqid></control><display><type>article</type><title>Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study</title><source>MEDLINE</source><source>Elsevier ScienceDirect Journals Complete</source><source>ProQuest Central UK/Ireland</source><creator>Rastogi, Aditya ; Brugnara, Gianluca ; Foltyn-Dumitru, Martha ; Mahmutoglu, Mustafa Ahmed ; Preetha, Chandrakanth J ; Kobler, Erich ; Pflüger, Irada ; Schell, Marianne ; Deike-Hofmann, Katerina ; Kessler, Tobias ; van den Bent, Martin J ; Idbaih, Ahmed ; Platten, Michael ; Brandes, Alba A ; Nabors, Burt ; Stupp, Roger ; Bernhardt, Denise ; Debus, Jürgen ; Abdollahi, Amir ; Gorlia, Thierry ; Tonn, Jörg-Christian ; Weller, Michael ; Maier-Hein, Klaus H ; Radbruch, Alexander ; Wick, Wolfgang ; Bendszus, Martin ; Meredig, Hagen ; Kurz, Felix T ; Vollmuth, Philipp</creator><creatorcontrib>Rastogi, Aditya ; Brugnara, Gianluca ; Foltyn-Dumitru, Martha ; Mahmutoglu, Mustafa Ahmed ; Preetha, Chandrakanth J ; Kobler, Erich ; Pflüger, Irada ; Schell, Marianne ; Deike-Hofmann, Katerina ; Kessler, Tobias ; van den Bent, Martin J ; Idbaih, Ahmed ; Platten, Michael ; Brandes, Alba A ; Nabors, Burt ; Stupp, Roger ; Bernhardt, Denise ; Debus, Jürgen ; Abdollahi, Amir ; Gorlia, Thierry ; Tonn, Jörg-Christian ; Weller, Michael ; Maier-Hein, Klaus H ; Radbruch, Alexander ; Wick, Wolfgang ; Bendszus, Martin ; Meredig, Hagen ; Kurz, Felix T ; Vollmuth, Philipp</creatorcontrib><description>The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutional neural network (dCNN) optimisation method for MRI reconstruction and to reduce scan times and evaluate its effect on image quality and accuracy of oncological imaging biomarkers.
In this multicentre, retrospective, cohort study, MRI data from patients with glioblastoma treated at Heidelberg University Hospital (775 patients and 775 examinations) and from the phase 2 CORE trial (260 patients, 1083 examinations, and 58 institutions) and the phase 3 CENTRIC trial (505 patients, 3147 examinations, and 139 institutions) were used to develop, train, and test dCNN for reconstructing MRI from highly undersampled single-coil k-space data with various acceleration rates (R=2, 4, 6, 8, 10, and 15). Independent testing was performed with MRIs from the phase 2/3 EORTC-26101 trial (528 patients with glioblastoma, 1974 examinations, and 32 institutions). The similarity between undersampled dCNN-reconstructed and original MRIs was quantified with various image quality metrics, including structural similarity index measure (SSIM) and the accuracy of undersampled dCNN-reconstructed MRI on downstream radiological assessment of imaging biomarkers in oncology (automated artificial intelligence-based quantification of tumour burden and treatment response) was performed in the EORTC-26101 test dataset. The public NYU Langone Health fastMRI brain test dataset (558 patients and 558 examinations) was used to validate the generalisability and robustness of the dCNN for reconstructing MRIs from available multi-coil (parallel imaging) k-space data.
In the EORTC-26101 test dataset, the median SSIM of undersampled dCNN-reconstructed MRI ranged from 0·88 to 0·99 across different acceleration rates, with 0·92 (95% CI 0·92–0·93) for 10-times acceleration (R=10). The 10-times undersampled dCNN-reconstructed MRI yielded excellent agreement with original MRI when assessing volumes of contrast-enhancing tumour (median DICE for spatial agreement of 0·89 [95% CI 0·88 to 0·89]; median volume difference of 0·01 cm3 [95% CI 0·00 to 0·03] equalling 0·21%; p=0·0036 for equivalence) or non-enhancing tumour or oedema (median DICE of 0·94 [95% CI 0·94 to 0·95]; median volume difference of –0·79 cm3 [95% CI –0·87 to –0·72] equalling –1·77%; p=0·023 for equivalence) in the EORTC-26101 test dataset. Automated volumetric tumour response assessment in the EORTC-26101 test dataset yielded an identical median time to progression of 4·27 months (95% CI 4·14 to 4·57) when using 10-times-undersampled dCNN-reconstructed or original MRI (log-rank p=0·80) and agreement in the time to progression in 374 (95·2%) of 393 patients with data. The dCNN generalised well to the fastMRI brain dataset, with significant improvements in the median SSIM when using multi-coil compared with single-coil k-space data (p<0·0001).
Deep-learning-based reconstruction of undersampled MRI allows for a substantial reduction of scan times, with a 10-times acceleration demonstrating excellent image quality while preserving the accuracy of derived imaging biomarkers for the assessment of oncological treatment response. Our developments are available as open source software and hold considerable promise for increasing the accessibility to MRI, pending further prospective validation.
Deutsche Forschungsgemeinschaft (German Research Foundation) and an Else Kröner Clinician Scientist Endowed Professorship by the Else Kröner Fresenius Foundation.</description><identifier>ISSN: 1470-2045</identifier><identifier>EISSN: 1474-5488</identifier><identifier>DOI: 10.1016/S1470-2045(23)00641-1</identifier><identifier>PMID: 38423052</identifier><language>eng</language><publisher>England: Elsevier Ltd</publisher><subject>Accuracy ; Artificial Intelligence ; Biomarkers ; Brain research ; Cancer ; Cohort analysis ; Cohort Studies ; Datasets ; Deep Learning ; Edema ; Fourier transforms ; Glioblastoma ; Glioblastoma - diagnostic imaging ; Glioma ; Humans ; Inverse problems ; Magnetic Resonance Imaging ; Medical imaging ; Neural networks ; Neuroimaging ; Oncology ; Patients ; Retrospective Studies ; Tumors</subject><ispartof>The lancet oncology, 2024-03, Vol.25 (3), p.400-410</ispartof><rights>2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license</rights><rights>Copyright © 2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.</rights><rights>2024. The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. This work is published under https://creativecommons.org/licenses/by/3.0/ (theLicense”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c440t-b80cd781821539dc24915b5f5d19043f35d24c32d663613ec483b004acd0b6823</citedby><cites>FETCH-LOGICAL-c440t-b80cd781821539dc24915b5f5d19043f35d24c32d663613ec483b004acd0b6823</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/3039714510?pq-origsite=primo$$EHTML$$P50$$Gproquest$$H</linktohtml><link.rule.ids>314,780,784,3550,27924,27925,45995,64385,64387,64389,72469</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/38423052$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Rastogi, Aditya</creatorcontrib><creatorcontrib>Brugnara, Gianluca</creatorcontrib><creatorcontrib>Foltyn-Dumitru, Martha</creatorcontrib><creatorcontrib>Mahmutoglu, Mustafa Ahmed</creatorcontrib><creatorcontrib>Preetha, Chandrakanth J</creatorcontrib><creatorcontrib>Kobler, Erich</creatorcontrib><creatorcontrib>Pflüger, Irada</creatorcontrib><creatorcontrib>Schell, Marianne</creatorcontrib><creatorcontrib>Deike-Hofmann, Katerina</creatorcontrib><creatorcontrib>Kessler, Tobias</creatorcontrib><creatorcontrib>van den Bent, Martin J</creatorcontrib><creatorcontrib>Idbaih, Ahmed</creatorcontrib><creatorcontrib>Platten, Michael</creatorcontrib><creatorcontrib>Brandes, Alba A</creatorcontrib><creatorcontrib>Nabors, Burt</creatorcontrib><creatorcontrib>Stupp, Roger</creatorcontrib><creatorcontrib>Bernhardt, Denise</creatorcontrib><creatorcontrib>Debus, Jürgen</creatorcontrib><creatorcontrib>Abdollahi, Amir</creatorcontrib><creatorcontrib>Gorlia, Thierry</creatorcontrib><creatorcontrib>Tonn, Jörg-Christian</creatorcontrib><creatorcontrib>Weller, Michael</creatorcontrib><creatorcontrib>Maier-Hein, Klaus H</creatorcontrib><creatorcontrib>Radbruch, Alexander</creatorcontrib><creatorcontrib>Wick, Wolfgang</creatorcontrib><creatorcontrib>Bendszus, Martin</creatorcontrib><creatorcontrib>Meredig, Hagen</creatorcontrib><creatorcontrib>Kurz, Felix T</creatorcontrib><creatorcontrib>Vollmuth, Philipp</creatorcontrib><title>Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study</title><title>The lancet oncology</title><addtitle>Lancet Oncol</addtitle><description>The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutional neural network (dCNN) optimisation method for MRI reconstruction and to reduce scan times and evaluate its effect on image quality and accuracy of oncological imaging biomarkers.
In this multicentre, retrospective, cohort study, MRI data from patients with glioblastoma treated at Heidelberg University Hospital (775 patients and 775 examinations) and from the phase 2 CORE trial (260 patients, 1083 examinations, and 58 institutions) and the phase 3 CENTRIC trial (505 patients, 3147 examinations, and 139 institutions) were used to develop, train, and test dCNN for reconstructing MRI from highly undersampled single-coil k-space data with various acceleration rates (R=2, 4, 6, 8, 10, and 15). Independent testing was performed with MRIs from the phase 2/3 EORTC-26101 trial (528 patients with glioblastoma, 1974 examinations, and 32 institutions). The similarity between undersampled dCNN-reconstructed and original MRIs was quantified with various image quality metrics, including structural similarity index measure (SSIM) and the accuracy of undersampled dCNN-reconstructed MRI on downstream radiological assessment of imaging biomarkers in oncology (automated artificial intelligence-based quantification of tumour burden and treatment response) was performed in the EORTC-26101 test dataset. The public NYU Langone Health fastMRI brain test dataset (558 patients and 558 examinations) was used to validate the generalisability and robustness of the dCNN for reconstructing MRIs from available multi-coil (parallel imaging) k-space data.
In the EORTC-26101 test dataset, the median SSIM of undersampled dCNN-reconstructed MRI ranged from 0·88 to 0·99 across different acceleration rates, with 0·92 (95% CI 0·92–0·93) for 10-times acceleration (R=10). The 10-times undersampled dCNN-reconstructed MRI yielded excellent agreement with original MRI when assessing volumes of contrast-enhancing tumour (median DICE for spatial agreement of 0·89 [95% CI 0·88 to 0·89]; median volume difference of 0·01 cm3 [95% CI 0·00 to 0·03] equalling 0·21%; p=0·0036 for equivalence) or non-enhancing tumour or oedema (median DICE of 0·94 [95% CI 0·94 to 0·95]; median volume difference of –0·79 cm3 [95% CI –0·87 to –0·72] equalling –1·77%; p=0·023 for equivalence) in the EORTC-26101 test dataset. Automated volumetric tumour response assessment in the EORTC-26101 test dataset yielded an identical median time to progression of 4·27 months (95% CI 4·14 to 4·57) when using 10-times-undersampled dCNN-reconstructed or original MRI (log-rank p=0·80) and agreement in the time to progression in 374 (95·2%) of 393 patients with data. The dCNN generalised well to the fastMRI brain dataset, with significant improvements in the median SSIM when using multi-coil compared with single-coil k-space data (p<0·0001).
Deep-learning-based reconstruction of undersampled MRI allows for a substantial reduction of scan times, with a 10-times acceleration demonstrating excellent image quality while preserving the accuracy of derived imaging biomarkers for the assessment of oncological treatment response. Our developments are available as open source software and hold considerable promise for increasing the accessibility to MRI, pending further prospective validation.
Deutsche Forschungsgemeinschaft (German Research Foundation) and an Else Kröner Clinician Scientist Endowed Professorship by the Else Kröner Fresenius Foundation.</description><subject>Accuracy</subject><subject>Artificial Intelligence</subject><subject>Biomarkers</subject><subject>Brain research</subject><subject>Cancer</subject><subject>Cohort analysis</subject><subject>Cohort Studies</subject><subject>Datasets</subject><subject>Deep Learning</subject><subject>Edema</subject><subject>Fourier transforms</subject><subject>Glioblastoma</subject><subject>Glioblastoma - diagnostic imaging</subject><subject>Glioma</subject><subject>Humans</subject><subject>Inverse problems</subject><subject>Magnetic Resonance Imaging</subject><subject>Medical imaging</subject><subject>Neural networks</subject><subject>Neuroimaging</subject><subject>Oncology</subject><subject>Patients</subject><subject>Retrospective Studies</subject><subject>Tumors</subject><issn>1470-2045</issn><issn>1474-5488</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>EIF</sourceid><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><recordid>eNqFkUlv1TAURiMEoqXwE0CW2BSJwPWUgQ1CZapUhMSwthz7BlwlduqhqP8ev_cKCzasPJ372b6naR5TeEGBdi-_UtFDy0DIU8afAXSCtvROc1y3RSvFMNzdzw_IUfMgpUsA2lOQ95sjPgjGQbLj5tdbxK1dUEfv_I920gktiWiCTzkWk13wJMykeIsx6XVb6vGnL-ckh0rZYpAkoz3JbsX0imiyliU7gz5HfF6JHEPasMZc16UJP0PMJOVibx4292a9JHx0O54039-_-3b2sb34_OH87M1Fa4SA3E4DGNsPdGBU8tEaJkYqJzlLS0cQfObSMmE4s13HO8rRiIFPAEIbC1M3MH7SnB5ytxiuCqasVpcMLov2GEpSbOSC9aJnUNGn_6CXoURfX6c48LGnQtIdJQ-UqV9LEWe1RbfqeKMoqJ0ZtTejdm1XjKu9GUVr3ZPb9DKtaP9W_VFRgdcHAGs7rh1GlYxDb9C66iMrG9x_rvgNdE6c8g</recordid><startdate>20240301</startdate><enddate>20240301</enddate><creator>Rastogi, Aditya</creator><creator>Brugnara, Gianluca</creator><creator>Foltyn-Dumitru, Martha</creator><creator>Mahmutoglu, Mustafa Ahmed</creator><creator>Preetha, Chandrakanth J</creator><creator>Kobler, Erich</creator><creator>Pflüger, Irada</creator><creator>Schell, Marianne</creator><creator>Deike-Hofmann, Katerina</creator><creator>Kessler, Tobias</creator><creator>van den Bent, Martin J</creator><creator>Idbaih, Ahmed</creator><creator>Platten, Michael</creator><creator>Brandes, Alba A</creator><creator>Nabors, Burt</creator><creator>Stupp, Roger</creator><creator>Bernhardt, Denise</creator><creator>Debus, Jürgen</creator><creator>Abdollahi, Amir</creator><creator>Gorlia, Thierry</creator><creator>Tonn, Jörg-Christian</creator><creator>Weller, Michael</creator><creator>Maier-Hein, Klaus H</creator><creator>Radbruch, Alexander</creator><creator>Wick, Wolfgang</creator><creator>Bendszus, Martin</creator><creator>Meredig, Hagen</creator><creator>Kurz, Felix T</creator><creator>Vollmuth, Philipp</creator><general>Elsevier Ltd</general><general>Elsevier Limited</general><scope>6I.</scope><scope>AAFTH</scope><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>0TZ</scope><scope>3V.</scope><scope>7RV</scope><scope>7TO</scope><scope>7X7</scope><scope>7XB</scope><scope>88E</scope><scope>8AO</scope><scope>8C1</scope><scope>8C2</scope><scope>8FI</scope><scope>8FJ</scope><scope>8FK</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>BENPR</scope><scope>CCPQU</scope><scope>FYUFA</scope><scope>GHDGH</scope><scope>H94</scope><scope>K9.</scope><scope>KB0</scope><scope>M0S</scope><scope>M1P</scope><scope>NAPCQ</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>7X8</scope></search><sort><creationdate>20240301</creationdate><title>Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study</title><author>Rastogi, Aditya ; Brugnara, Gianluca ; Foltyn-Dumitru, Martha ; Mahmutoglu, Mustafa Ahmed ; Preetha, Chandrakanth J ; Kobler, Erich ; Pflüger, Irada ; Schell, Marianne ; Deike-Hofmann, Katerina ; Kessler, Tobias ; van den Bent, Martin J ; Idbaih, Ahmed ; Platten, Michael ; Brandes, Alba A ; Nabors, Burt ; Stupp, Roger ; Bernhardt, Denise ; Debus, Jürgen ; Abdollahi, Amir ; Gorlia, Thierry ; Tonn, Jörg-Christian ; Weller, Michael ; Maier-Hein, Klaus H ; Radbruch, Alexander ; Wick, Wolfgang ; Bendszus, Martin ; Meredig, Hagen ; Kurz, Felix T ; Vollmuth, Philipp</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c440t-b80cd781821539dc24915b5f5d19043f35d24c32d663613ec483b004acd0b6823</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Accuracy</topic><topic>Artificial Intelligence</topic><topic>Biomarkers</topic><topic>Brain research</topic><topic>Cancer</topic><topic>Cohort analysis</topic><topic>Cohort Studies</topic><topic>Datasets</topic><topic>Deep Learning</topic><topic>Edema</topic><topic>Fourier transforms</topic><topic>Glioblastoma</topic><topic>Glioblastoma - diagnostic imaging</topic><topic>Glioma</topic><topic>Humans</topic><topic>Inverse problems</topic><topic>Magnetic Resonance Imaging</topic><topic>Medical imaging</topic><topic>Neural networks</topic><topic>Neuroimaging</topic><topic>Oncology</topic><topic>Patients</topic><topic>Retrospective Studies</topic><topic>Tumors</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Rastogi, Aditya</creatorcontrib><creatorcontrib>Brugnara, Gianluca</creatorcontrib><creatorcontrib>Foltyn-Dumitru, Martha</creatorcontrib><creatorcontrib>Mahmutoglu, Mustafa Ahmed</creatorcontrib><creatorcontrib>Preetha, Chandrakanth J</creatorcontrib><creatorcontrib>Kobler, Erich</creatorcontrib><creatorcontrib>Pflüger, Irada</creatorcontrib><creatorcontrib>Schell, Marianne</creatorcontrib><creatorcontrib>Deike-Hofmann, Katerina</creatorcontrib><creatorcontrib>Kessler, Tobias</creatorcontrib><creatorcontrib>van den Bent, Martin J</creatorcontrib><creatorcontrib>Idbaih, Ahmed</creatorcontrib><creatorcontrib>Platten, Michael</creatorcontrib><creatorcontrib>Brandes, Alba A</creatorcontrib><creatorcontrib>Nabors, Burt</creatorcontrib><creatorcontrib>Stupp, Roger</creatorcontrib><creatorcontrib>Bernhardt, Denise</creatorcontrib><creatorcontrib>Debus, Jürgen</creatorcontrib><creatorcontrib>Abdollahi, Amir</creatorcontrib><creatorcontrib>Gorlia, Thierry</creatorcontrib><creatorcontrib>Tonn, Jörg-Christian</creatorcontrib><creatorcontrib>Weller, Michael</creatorcontrib><creatorcontrib>Maier-Hein, Klaus H</creatorcontrib><creatorcontrib>Radbruch, Alexander</creatorcontrib><creatorcontrib>Wick, Wolfgang</creatorcontrib><creatorcontrib>Bendszus, Martin</creatorcontrib><creatorcontrib>Meredig, Hagen</creatorcontrib><creatorcontrib>Kurz, Felix T</creatorcontrib><creatorcontrib>Vollmuth, Philipp</creatorcontrib><collection>ScienceDirect Open Access Titles</collection><collection>Elsevier:ScienceDirect:Open Access</collection><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Pharma and Biotech Premium PRO</collection><collection>ProQuest Central (Corporate)</collection><collection>Nursing & Allied Health Database</collection><collection>Oncogenes and Growth Factors Abstracts</collection><collection>Health & Medical Collection</collection><collection>ProQuest Central (purchase pre-March 2016)</collection><collection>Medical Database (Alumni Edition)</collection><collection>ProQuest Pharma Collection</collection><collection>Public Health Database</collection><collection>Lancet Titles</collection><collection>Hospital Premium Collection</collection><collection>Hospital Premium Collection (Alumni Edition)</collection><collection>ProQuest Central (Alumni) (purchase pre-March 2016)</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central</collection><collection>ProQuest One Community College</collection><collection>Health Research Premium Collection</collection><collection>Health Research Premium Collection (Alumni)</collection><collection>AIDS and Cancer Research Abstracts</collection><collection>ProQuest Health & Medical Complete (Alumni)</collection><collection>Nursing & Allied Health Database (Alumni Edition)</collection><collection>Health & Medical Collection (Alumni Edition)</collection><collection>Medical Database</collection><collection>Nursing & Allied Health Premium</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>MEDLINE - Academic</collection><jtitle>The lancet oncology</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Rastogi, Aditya</au><au>Brugnara, Gianluca</au><au>Foltyn-Dumitru, Martha</au><au>Mahmutoglu, Mustafa Ahmed</au><au>Preetha, Chandrakanth J</au><au>Kobler, Erich</au><au>Pflüger, Irada</au><au>Schell, Marianne</au><au>Deike-Hofmann, Katerina</au><au>Kessler, Tobias</au><au>van den Bent, Martin J</au><au>Idbaih, Ahmed</au><au>Platten, Michael</au><au>Brandes, Alba A</au><au>Nabors, Burt</au><au>Stupp, Roger</au><au>Bernhardt, Denise</au><au>Debus, Jürgen</au><au>Abdollahi, Amir</au><au>Gorlia, Thierry</au><au>Tonn, Jörg-Christian</au><au>Weller, Michael</au><au>Maier-Hein, Klaus H</au><au>Radbruch, Alexander</au><au>Wick, Wolfgang</au><au>Bendszus, Martin</au><au>Meredig, Hagen</au><au>Kurz, Felix T</au><au>Vollmuth, Philipp</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study</atitle><jtitle>The lancet oncology</jtitle><addtitle>Lancet Oncol</addtitle><date>2024-03-01</date><risdate>2024</risdate><volume>25</volume><issue>3</issue><spage>400</spage><epage>410</epage><pages>400-410</pages><issn>1470-2045</issn><eissn>1474-5488</eissn><abstract>The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutional neural network (dCNN) optimisation method for MRI reconstruction and to reduce scan times and evaluate its effect on image quality and accuracy of oncological imaging biomarkers.
In this multicentre, retrospective, cohort study, MRI data from patients with glioblastoma treated at Heidelberg University Hospital (775 patients and 775 examinations) and from the phase 2 CORE trial (260 patients, 1083 examinations, and 58 institutions) and the phase 3 CENTRIC trial (505 patients, 3147 examinations, and 139 institutions) were used to develop, train, and test dCNN for reconstructing MRI from highly undersampled single-coil k-space data with various acceleration rates (R=2, 4, 6, 8, 10, and 15). Independent testing was performed with MRIs from the phase 2/3 EORTC-26101 trial (528 patients with glioblastoma, 1974 examinations, and 32 institutions). The similarity between undersampled dCNN-reconstructed and original MRIs was quantified with various image quality metrics, including structural similarity index measure (SSIM) and the accuracy of undersampled dCNN-reconstructed MRI on downstream radiological assessment of imaging biomarkers in oncology (automated artificial intelligence-based quantification of tumour burden and treatment response) was performed in the EORTC-26101 test dataset. The public NYU Langone Health fastMRI brain test dataset (558 patients and 558 examinations) was used to validate the generalisability and robustness of the dCNN for reconstructing MRIs from available multi-coil (parallel imaging) k-space data.
In the EORTC-26101 test dataset, the median SSIM of undersampled dCNN-reconstructed MRI ranged from 0·88 to 0·99 across different acceleration rates, with 0·92 (95% CI 0·92–0·93) for 10-times acceleration (R=10). The 10-times undersampled dCNN-reconstructed MRI yielded excellent agreement with original MRI when assessing volumes of contrast-enhancing tumour (median DICE for spatial agreement of 0·89 [95% CI 0·88 to 0·89]; median volume difference of 0·01 cm3 [95% CI 0·00 to 0·03] equalling 0·21%; p=0·0036 for equivalence) or non-enhancing tumour or oedema (median DICE of 0·94 [95% CI 0·94 to 0·95]; median volume difference of –0·79 cm3 [95% CI –0·87 to –0·72] equalling –1·77%; p=0·023 for equivalence) in the EORTC-26101 test dataset. Automated volumetric tumour response assessment in the EORTC-26101 test dataset yielded an identical median time to progression of 4·27 months (95% CI 4·14 to 4·57) when using 10-times-undersampled dCNN-reconstructed or original MRI (log-rank p=0·80) and agreement in the time to progression in 374 (95·2%) of 393 patients with data. The dCNN generalised well to the fastMRI brain dataset, with significant improvements in the median SSIM when using multi-coil compared with single-coil k-space data (p<0·0001).
Deep-learning-based reconstruction of undersampled MRI allows for a substantial reduction of scan times, with a 10-times acceleration demonstrating excellent image quality while preserving the accuracy of derived imaging biomarkers for the assessment of oncological treatment response. Our developments are available as open source software and hold considerable promise for increasing the accessibility to MRI, pending further prospective validation.
Deutsche Forschungsgemeinschaft (German Research Foundation) and an Else Kröner Clinician Scientist Endowed Professorship by the Else Kröner Fresenius Foundation.</abstract><cop>England</cop><pub>Elsevier Ltd</pub><pmid>38423052</pmid><doi>10.1016/S1470-2045(23)00641-1</doi><tpages>11</tpages><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1470-2045 |
| ispartof | The lancet oncology, 2024-03, Vol.25 (3), p.400-410 |
| issn | 1470-2045 1474-5488 |
| language | eng |
| recordid | cdi_proquest_miscellaneous_2934274720 |
| source | MEDLINE; Elsevier ScienceDirect Journals Complete; ProQuest Central UK/Ireland |
| subjects | Accuracy Artificial Intelligence Biomarkers Brain research Cancer Cohort analysis Cohort Studies Datasets Deep Learning Edema Fourier transforms Glioblastoma Glioblastoma - diagnostic imaging Glioma Humans Inverse problems Magnetic Resonance Imaging Medical imaging Neural networks Neuroimaging Oncology Patients Retrospective Studies Tumors |
| title | Deep-learning-based reconstruction of undersampled MRI to reduce scan times: a multicentre, retrospective, cohort study |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T01%3A54%3A22IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Deep-learning-based%20reconstruction%20of%20undersampled%20MRI%20to%20reduce%20scan%20times:%20a%20multicentre,%20retrospective,%20cohort%20study&rft.jtitle=The%20lancet%20oncology&rft.au=Rastogi,%20Aditya&rft.date=2024-03-01&rft.volume=25&rft.issue=3&rft.spage=400&rft.epage=410&rft.pages=400-410&rft.issn=1470-2045&rft.eissn=1474-5488&rft_id=info:doi/10.1016/S1470-2045(23)00641-1&rft_dat=%3Cproquest_cross%3E3039714510%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3039714510&rft_id=info:pmid/38423052&rft_els_id=S1470204523006411&rfr_iscdi=true |