Journal article 535 views 201 downloads
Objective assessment of stored blood quality by deep learning
Proceedings of the National Academy of Sciences, Volume: 117, Issue: 35, Pages: 21381 - 21390
Swansea University Author:
Paul Rees
-
PDF | Version of Record
Copyright © 2020 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).
Download (1.85MB)
DOI (Published version): 10.1073/pnas.2001227117
Abstract
Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we...
Published in: | Proceedings of the National Academy of Sciences |
---|---|
ISSN: | 0027-8424 1091-6490 |
Published: |
Proceedings of the National Academy of Sciences
2020
|
Online Access: |
Check full text
|
URI: | https://cronfa.swan.ac.uk/Record/cronfa55200 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
first_indexed |
2020-09-17T16:05:43Z |
---|---|
last_indexed |
2021-12-03T04:14:06Z |
id |
cronfa55200 |
recordtype |
SURis |
fullrecord |
<?xml version="1.0"?><rfc1807><datestamp>2021-12-02T08:31:02.1817840</datestamp><bib-version>v2</bib-version><id>55200</id><entry>2020-09-17</entry><title>Objective assessment of stored blood quality by deep learning</title><swanseaauthors><author><sid>537a2fe031a796a3bde99679ee8c24f5</sid><ORCID>0000-0002-7715-6914</ORCID><firstname>Paul</firstname><surname>Rees</surname><name>Paul Rees</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2020-09-17</date><deptcode>MEDE</deptcode><abstract>Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we demonstrate the use of label-free imaging flow cytometry and deep learning to characterize RBC lesions. Using brightfield images, a trained neural network achieved 76.7% agreement with experts in classifying seven clinically relevant RBC morphologies associated with storage lesions, comparable to 82.5% agreement between different experts. Given that human observation and classification may not optimally discern RBC quality, we went further and eliminated subjective human annotation in the training step by training a weakly supervised neural network using only storage duration times. The feature space extracted by this network revealed a chronological progression of morphological changes that better predicted blood quality, as measured by physiological hemolytic assay readouts, than the conventional expert-assessed morphology classification system. With further training and clinical testing across multiple sites, protocols, and instruments, deep learning and label-free imaging flow cytometry might be used to routinely and objectively assess RBC storage lesions. This would automate a complex protocol, minimize laboratory sample handling and preparation, and reduce the impact of procedural errors and discrepancies between facilities and blood donors. The chronology-based machine-learning approach may also improve upon humans’ assessment of morphological changes in other biomedically important progressions, such as differentiation and metastasis.</abstract><type>Journal Article</type><journal>Proceedings of the National Academy of Sciences</journal><volume>117</volume><journalNumber>35</journalNumber><paginationStart>21381</paginationStart><paginationEnd>21390</paginationEnd><publisher>Proceedings of the National Academy of Sciences</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>0027-8424</issnPrint><issnElectronic>1091-6490</issnElectronic><keywords/><publishedDay>1</publishedDay><publishedMonth>9</publishedMonth><publishedYear>2020</publishedYear><publishedDate>2020-09-01</publishedDate><doi>10.1073/pnas.2001227117</doi><url/><notes/><college>COLLEGE NANME</college><department>Biomedical Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MEDE</DepartmentCode><institution>Swansea University</institution><apcterm/><funders>BBSRC, CIHR</funders><projectreference>BB/N005163, BB/P026818/1, 315271,</projectreference><lastEdited>2021-12-02T08:31:02.1817840</lastEdited><Created>2020-09-17T17:03:30.7532104</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Engineering and Applied Sciences - Biomedical Engineering</level></path><authors><author><firstname>Minh</firstname><surname>Doan</surname><order>1</order></author><author><firstname>Joseph A.</firstname><surname>Sebastian</surname><order>2</order></author><author><firstname>Juan C.</firstname><surname>Caicedo</surname><order>3</order></author><author><firstname>Stefanie</firstname><surname>Siegert</surname><order>4</order></author><author><firstname>Aline</firstname><surname>Roch</surname><order>5</order></author><author><firstname>Tracey R.</firstname><surname>Turner</surname><order>6</order></author><author><firstname>Olga</firstname><surname>Mykhailova</surname><order>7</order></author><author><firstname>Ruben N.</firstname><surname>Pinto</surname><order>8</order></author><author><firstname>Claire</firstname><surname>McQuin</surname><order>9</order></author><author><firstname>Allen</firstname><surname>Goodman</surname><order>10</order></author><author><firstname>Michael J.</firstname><surname>Parsons</surname><order>11</order></author><author><firstname>Olaf</firstname><surname>Wolkenhauer</surname><order>12</order></author><author><firstname>Holger</firstname><surname>Hennig</surname><order>13</order></author><author><firstname>Shantanu</firstname><surname>Singh</surname><order>14</order></author><author><firstname>Anne</firstname><surname>Wilson</surname><order>15</order></author><author><firstname>Jason P.</firstname><surname>Acker</surname><order>16</order></author><author><firstname>Paul</firstname><surname>Rees</surname><orcid>0000-0002-7715-6914</orcid><order>17</order></author><author><firstname>Michael C.</firstname><surname>Kolios</surname><order>18</order></author><author><firstname>Anne E.</firstname><surname>Carpenter</surname><order>19</order></author></authors><documents><document><filename>55200__18194__f75d240f05054dcea22792a2fc8de151.pdf</filename><originalFilename>55200.pdf</originalFilename><uploaded>2020-09-17T17:05:03.3516525</uploaded><type>Output</type><contentLength>1937701</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>Copyright © 2020 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by-nc-nd/4.0/</licence></document></documents><OutputDurs/></rfc1807> |
spelling |
2021-12-02T08:31:02.1817840 v2 55200 2020-09-17 Objective assessment of stored blood quality by deep learning 537a2fe031a796a3bde99679ee8c24f5 0000-0002-7715-6914 Paul Rees Paul Rees true false 2020-09-17 MEDE Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we demonstrate the use of label-free imaging flow cytometry and deep learning to characterize RBC lesions. Using brightfield images, a trained neural network achieved 76.7% agreement with experts in classifying seven clinically relevant RBC morphologies associated with storage lesions, comparable to 82.5% agreement between different experts. Given that human observation and classification may not optimally discern RBC quality, we went further and eliminated subjective human annotation in the training step by training a weakly supervised neural network using only storage duration times. The feature space extracted by this network revealed a chronological progression of morphological changes that better predicted blood quality, as measured by physiological hemolytic assay readouts, than the conventional expert-assessed morphology classification system. With further training and clinical testing across multiple sites, protocols, and instruments, deep learning and label-free imaging flow cytometry might be used to routinely and objectively assess RBC storage lesions. This would automate a complex protocol, minimize laboratory sample handling and preparation, and reduce the impact of procedural errors and discrepancies between facilities and blood donors. The chronology-based machine-learning approach may also improve upon humans’ assessment of morphological changes in other biomedically important progressions, such as differentiation and metastasis. Journal Article Proceedings of the National Academy of Sciences 117 35 21381 21390 Proceedings of the National Academy of Sciences 0027-8424 1091-6490 1 9 2020 2020-09-01 10.1073/pnas.2001227117 COLLEGE NANME Biomedical Engineering COLLEGE CODE MEDE Swansea University BBSRC, CIHR BB/N005163, BB/P026818/1, 315271, 2021-12-02T08:31:02.1817840 2020-09-17T17:03:30.7532104 Faculty of Science and Engineering School of Engineering and Applied Sciences - Biomedical Engineering Minh Doan 1 Joseph A. Sebastian 2 Juan C. Caicedo 3 Stefanie Siegert 4 Aline Roch 5 Tracey R. Turner 6 Olga Mykhailova 7 Ruben N. Pinto 8 Claire McQuin 9 Allen Goodman 10 Michael J. Parsons 11 Olaf Wolkenhauer 12 Holger Hennig 13 Shantanu Singh 14 Anne Wilson 15 Jason P. Acker 16 Paul Rees 0000-0002-7715-6914 17 Michael C. Kolios 18 Anne E. Carpenter 19 55200__18194__f75d240f05054dcea22792a2fc8de151.pdf 55200.pdf 2020-09-17T17:05:03.3516525 Output 1937701 application/pdf Version of Record true Copyright © 2020 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND). true eng https://creativecommons.org/licenses/by-nc-nd/4.0/ |
title |
Objective assessment of stored blood quality by deep learning |
spellingShingle |
Objective assessment of stored blood quality by deep learning Paul Rees |
title_short |
Objective assessment of stored blood quality by deep learning |
title_full |
Objective assessment of stored blood quality by deep learning |
title_fullStr |
Objective assessment of stored blood quality by deep learning |
title_full_unstemmed |
Objective assessment of stored blood quality by deep learning |
title_sort |
Objective assessment of stored blood quality by deep learning |
author_id_str_mv |
537a2fe031a796a3bde99679ee8c24f5 |
author_id_fullname_str_mv |
537a2fe031a796a3bde99679ee8c24f5_***_Paul Rees |
author |
Paul Rees |
author2 |
Minh Doan Joseph A. Sebastian Juan C. Caicedo Stefanie Siegert Aline Roch Tracey R. Turner Olga Mykhailova Ruben N. Pinto Claire McQuin Allen Goodman Michael J. Parsons Olaf Wolkenhauer Holger Hennig Shantanu Singh Anne Wilson Jason P. Acker Paul Rees Michael C. Kolios Anne E. Carpenter |
format |
Journal article |
container_title |
Proceedings of the National Academy of Sciences |
container_volume |
117 |
container_issue |
35 |
container_start_page |
21381 |
publishDate |
2020 |
institution |
Swansea University |
issn |
0027-8424 1091-6490 |
doi_str_mv |
10.1073/pnas.2001227117 |
publisher |
Proceedings of the National Academy of Sciences |
college_str |
Faculty of Science and Engineering |
hierarchytype |
|
hierarchy_top_id |
facultyofscienceandengineering |
hierarchy_top_title |
Faculty of Science and Engineering |
hierarchy_parent_id |
facultyofscienceandengineering |
hierarchy_parent_title |
Faculty of Science and Engineering |
department_str |
School of Engineering and Applied Sciences - Biomedical Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Engineering and Applied Sciences - Biomedical Engineering |
document_store_str |
1 |
active_str |
0 |
description |
Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we demonstrate the use of label-free imaging flow cytometry and deep learning to characterize RBC lesions. Using brightfield images, a trained neural network achieved 76.7% agreement with experts in classifying seven clinically relevant RBC morphologies associated with storage lesions, comparable to 82.5% agreement between different experts. Given that human observation and classification may not optimally discern RBC quality, we went further and eliminated subjective human annotation in the training step by training a weakly supervised neural network using only storage duration times. The feature space extracted by this network revealed a chronological progression of morphological changes that better predicted blood quality, as measured by physiological hemolytic assay readouts, than the conventional expert-assessed morphology classification system. With further training and clinical testing across multiple sites, protocols, and instruments, deep learning and label-free imaging flow cytometry might be used to routinely and objectively assess RBC storage lesions. This would automate a complex protocol, minimize laboratory sample handling and preparation, and reduce the impact of procedural errors and discrepancies between facilities and blood donors. The chronology-based machine-learning approach may also improve upon humans’ assessment of morphological changes in other biomedically important progressions, such as differentiation and metastasis. |
published_date |
2020-09-01T04:09:14Z |
_version_ |
1763753648467214336 |
score |
11.021648 |