A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications

Abstract Drug response prediction is an important problem in personalized cancer therapy. Among various newly developed models, significant improvement in prediction performance has been reported using deep learning methods. However, systematic comparisons of deep learning methods, especially of the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Briefings in bioinformatics 2023-01, Vol.24 (1)
Hauptverfasser: Shen, Bihan, Feng, Fangyoumin, Li, Kunshi, Lin, Ping, Ma, Liangxiao, Li, Hong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue 1
container_start_page
container_title Briefings in bioinformatics
container_volume 24
creator Shen, Bihan
Feng, Fangyoumin
Li, Kunshi
Lin, Ping
Ma, Liangxiao
Li, Hong
description Abstract Drug response prediction is an important problem in personalized cancer therapy. Among various newly developed models, significant improvement in prediction performance has been reported using deep learning methods. However, systematic comparisons of deep learning methods, especially of the transferability from preclinical models to clinical cohorts, are currently lacking. To provide a more rigorous assessment, the performance of six representative deep learning methods for drug response prediction using nine evaluation metrics, including the overall prediction accuracy, predictability of each drug, potential associated factors and transferability to clinical cohorts, in multiple application scenarios was benchmarked. Most methods show promising prediction within cell line datasets, and TGSA, with its lower time cost and better performance, is recommended. Although the performance metrics decrease when applying models trained on cell lines to patients, a certain amount of power to distinguish clinical response on some drugs can be maintained using CRDNN and TGSA. With these assessments, we provide a guidance for researchers to choose appropriate methods, as well as insights into future directions for the development of more effective methods in clinical scenarios.
doi_str_mv 10.1093/bib/bbac605
format Article
fullrecord <record><control><sourceid>proquest_TOX</sourceid><recordid>TN_cdi_proquest_miscellaneous_2759002743</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><oup_id>10.1093/bib/bbac605</oup_id><sourcerecordid>2759002743</sourcerecordid><originalsourceid>FETCH-LOGICAL-c348t-8665a9691433674ac2c80a3331c2c49c1766d1049f1e82d1c2cb1850fe6430733</originalsourceid><addsrcrecordid>eNp90d9L3TAUB_AwJtOpT3uXgDCEUU2aX61vF3FTuOCLPpc0PXWRNulyUsH_fi33ugcf9pRD-PDlcL6EfOPskrNaXLW-vWpb6zRTn8gRl8YUkin5eZ21KZTU4pB8RXxhrGSm4l_IodDKqKrUR2TeUHzDDKPN3lGLCIgjhExjTzuAiQ5gU_DhmY6Qf8cOaR8T7dL8TBPgFAMCnRJ03mUfwzXtUxypD_TV5xRpjtQNPnhnB2qnaViGleEJOejtgHC6f4_J08_bx5u7Yvvw6_5msy2ckFUuKq2VrXXNpRDaSOtKVzErhODLJGvHjdYdZ7LuOVRlt_62vFKsBy0FM0Ick4td7pTinxkwN6NHB8NgA8QZm9KoermKkSs9_0Bf4pzCsl0jOBdSc8PUon7slEsRMUHfTMmPNr01nDVrG83SRrNvY9Fn-8y5HaH7Z9_Pv4DvOxDn6b9JfwGd-ZMU</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3113461705</pqid></control><display><type>article</type><title>A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications</title><source>Oxford Journals Open Access Collection</source><creator>Shen, Bihan ; Feng, Fangyoumin ; Li, Kunshi ; Lin, Ping ; Ma, Liangxiao ; Li, Hong</creator><creatorcontrib>Shen, Bihan ; Feng, Fangyoumin ; Li, Kunshi ; Lin, Ping ; Ma, Liangxiao ; Li, Hong</creatorcontrib><description>Abstract Drug response prediction is an important problem in personalized cancer therapy. Among various newly developed models, significant improvement in prediction performance has been reported using deep learning methods. However, systematic comparisons of deep learning methods, especially of the transferability from preclinical models to clinical cohorts, are currently lacking. To provide a more rigorous assessment, the performance of six representative deep learning methods for drug response prediction using nine evaluation metrics, including the overall prediction accuracy, predictability of each drug, potential associated factors and transferability to clinical cohorts, in multiple application scenarios was benchmarked. Most methods show promising prediction within cell line datasets, and TGSA, with its lower time cost and better performance, is recommended. Although the performance metrics decrease when applying models trained on cell lines to patients, a certain amount of power to distinguish clinical response on some drugs can be maintained using CRDNN and TGSA. With these assessments, we provide a guidance for researchers to choose appropriate methods, as well as insights into future directions for the development of more effective methods in clinical scenarios.</description><identifier>ISSN: 1467-5463</identifier><identifier>EISSN: 1477-4054</identifier><identifier>DOI: 10.1093/bib/bbac605</identifier><identifier>PMID: 36575826</identifier><language>eng</language><publisher>England: Oxford University Press</publisher><subject>Cancer therapies ; Cell culture ; Cell Line ; Deep Learning ; Drug development ; Humans ; In vitro methods and tests ; Performance evaluation ; Performance measurement ; Performance prediction ; Predictions</subject><ispartof>Briefings in bioinformatics, 2023-01, Vol.24 (1)</ispartof><rights>The Author(s) 2022. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com 2022</rights><rights>The Author(s) 2022. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.</rights><rights>The Author(s) 2022. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c348t-8665a9691433674ac2c80a3331c2c49c1766d1049f1e82d1c2cb1850fe6430733</citedby><cites>FETCH-LOGICAL-c348t-8665a9691433674ac2c80a3331c2c49c1766d1049f1e82d1c2cb1850fe6430733</cites><orcidid>0000-0002-9103-4913 ; 0000-0003-4592-2554</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,776,780,1598,27903,27904</link.rule.ids><linktorsrc>$$Uhttps://dx.doi.org/10.1093/bib/bbac605$$EView_record_in_Oxford_University_Press$$FView_record_in_$$GOxford_University_Press</linktorsrc><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/36575826$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Shen, Bihan</creatorcontrib><creatorcontrib>Feng, Fangyoumin</creatorcontrib><creatorcontrib>Li, Kunshi</creatorcontrib><creatorcontrib>Lin, Ping</creatorcontrib><creatorcontrib>Ma, Liangxiao</creatorcontrib><creatorcontrib>Li, Hong</creatorcontrib><title>A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications</title><title>Briefings in bioinformatics</title><addtitle>Brief Bioinform</addtitle><description>Abstract Drug response prediction is an important problem in personalized cancer therapy. Among various newly developed models, significant improvement in prediction performance has been reported using deep learning methods. However, systematic comparisons of deep learning methods, especially of the transferability from preclinical models to clinical cohorts, are currently lacking. To provide a more rigorous assessment, the performance of six representative deep learning methods for drug response prediction using nine evaluation metrics, including the overall prediction accuracy, predictability of each drug, potential associated factors and transferability to clinical cohorts, in multiple application scenarios was benchmarked. Most methods show promising prediction within cell line datasets, and TGSA, with its lower time cost and better performance, is recommended. Although the performance metrics decrease when applying models trained on cell lines to patients, a certain amount of power to distinguish clinical response on some drugs can be maintained using CRDNN and TGSA. With these assessments, we provide a guidance for researchers to choose appropriate methods, as well as insights into future directions for the development of more effective methods in clinical scenarios.</description><subject>Cancer therapies</subject><subject>Cell culture</subject><subject>Cell Line</subject><subject>Deep Learning</subject><subject>Drug development</subject><subject>Humans</subject><subject>In vitro methods and tests</subject><subject>Performance evaluation</subject><subject>Performance measurement</subject><subject>Performance prediction</subject><subject>Predictions</subject><issn>1467-5463</issn><issn>1477-4054</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>EIF</sourceid><recordid>eNp90d9L3TAUB_AwJtOpT3uXgDCEUU2aX61vF3FTuOCLPpc0PXWRNulyUsH_fi33ugcf9pRD-PDlcL6EfOPskrNaXLW-vWpb6zRTn8gRl8YUkin5eZ21KZTU4pB8RXxhrGSm4l_IodDKqKrUR2TeUHzDDKPN3lGLCIgjhExjTzuAiQ5gU_DhmY6Qf8cOaR8T7dL8TBPgFAMCnRJ03mUfwzXtUxypD_TV5xRpjtQNPnhnB2qnaViGleEJOejtgHC6f4_J08_bx5u7Yvvw6_5msy2ckFUuKq2VrXXNpRDaSOtKVzErhODLJGvHjdYdZ7LuOVRlt_62vFKsBy0FM0Ick4td7pTinxkwN6NHB8NgA8QZm9KoermKkSs9_0Bf4pzCsl0jOBdSc8PUon7slEsRMUHfTMmPNr01nDVrG83SRrNvY9Fn-8y5HaH7Z9_Pv4DvOxDn6b9JfwGd-ZMU</recordid><startdate>20230119</startdate><enddate>20230119</enddate><creator>Shen, Bihan</creator><creator>Feng, Fangyoumin</creator><creator>Li, Kunshi</creator><creator>Lin, Ping</creator><creator>Ma, Liangxiao</creator><creator>Li, Hong</creator><general>Oxford University Press</general><general>Oxford Publishing Limited (England)</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>7SC</scope><scope>8FD</scope><scope>FR3</scope><scope>JQ2</scope><scope>K9.</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>P64</scope><scope>RC3</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0002-9103-4913</orcidid><orcidid>https://orcid.org/0000-0003-4592-2554</orcidid></search><sort><creationdate>20230119</creationdate><title>A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications</title><author>Shen, Bihan ; Feng, Fangyoumin ; Li, Kunshi ; Lin, Ping ; Ma, Liangxiao ; Li, Hong</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c348t-8665a9691433674ac2c80a3331c2c49c1766d1049f1e82d1c2cb1850fe6430733</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Cancer therapies</topic><topic>Cell culture</topic><topic>Cell Line</topic><topic>Deep Learning</topic><topic>Drug development</topic><topic>Humans</topic><topic>In vitro methods and tests</topic><topic>Performance evaluation</topic><topic>Performance measurement</topic><topic>Performance prediction</topic><topic>Predictions</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Shen, Bihan</creatorcontrib><creatorcontrib>Feng, Fangyoumin</creatorcontrib><creatorcontrib>Li, Kunshi</creatorcontrib><creatorcontrib>Lin, Ping</creatorcontrib><creatorcontrib>Ma, Liangxiao</creatorcontrib><creatorcontrib>Li, Hong</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>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>ProQuest Health &amp; Medical Complete (Alumni)</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>Biotechnology and BioEngineering Abstracts</collection><collection>Genetics Abstracts</collection><collection>MEDLINE - Academic</collection><jtitle>Briefings in bioinformatics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Shen, Bihan</au><au>Feng, Fangyoumin</au><au>Li, Kunshi</au><au>Lin, Ping</au><au>Ma, Liangxiao</au><au>Li, Hong</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications</atitle><jtitle>Briefings in bioinformatics</jtitle><addtitle>Brief Bioinform</addtitle><date>2023-01-19</date><risdate>2023</risdate><volume>24</volume><issue>1</issue><issn>1467-5463</issn><eissn>1477-4054</eissn><abstract>Abstract Drug response prediction is an important problem in personalized cancer therapy. Among various newly developed models, significant improvement in prediction performance has been reported using deep learning methods. However, systematic comparisons of deep learning methods, especially of the transferability from preclinical models to clinical cohorts, are currently lacking. To provide a more rigorous assessment, the performance of six representative deep learning methods for drug response prediction using nine evaluation metrics, including the overall prediction accuracy, predictability of each drug, potential associated factors and transferability to clinical cohorts, in multiple application scenarios was benchmarked. Most methods show promising prediction within cell line datasets, and TGSA, with its lower time cost and better performance, is recommended. Although the performance metrics decrease when applying models trained on cell lines to patients, a certain amount of power to distinguish clinical response on some drugs can be maintained using CRDNN and TGSA. With these assessments, we provide a guidance for researchers to choose appropriate methods, as well as insights into future directions for the development of more effective methods in clinical scenarios.</abstract><cop>England</cop><pub>Oxford University Press</pub><pmid>36575826</pmid><doi>10.1093/bib/bbac605</doi><orcidid>https://orcid.org/0000-0002-9103-4913</orcidid><orcidid>https://orcid.org/0000-0003-4592-2554</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1467-5463
ispartof Briefings in bioinformatics, 2023-01, Vol.24 (1)
issn 1467-5463
1477-4054
language eng
recordid cdi_proquest_miscellaneous_2759002743
source Oxford Journals Open Access Collection
subjects Cancer therapies
Cell culture
Cell Line
Deep Learning
Drug development
Humans
In vitro methods and tests
Performance evaluation
Performance measurement
Performance prediction
Predictions
title A systematic assessment of deep learning methods for drug response prediction: from in vitro to clinical applications
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-25T18%3A56%3A40IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_TOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=A%20systematic%20assessment%20of%20deep%20learning%20methods%20for%20drug%20response%20prediction:%20from%20in%20vitro%20to%20clinical%20applications&rft.jtitle=Briefings%20in%20bioinformatics&rft.au=Shen,%20Bihan&rft.date=2023-01-19&rft.volume=24&rft.issue=1&rft.issn=1467-5463&rft.eissn=1477-4054&rft_id=info:doi/10.1093/bib/bbac605&rft_dat=%3Cproquest_TOX%3E2759002743%3C/proquest_TOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3113461705&rft_id=info:pmid/36575826&rft_oup_id=10.1093/bib/bbac605&rfr_iscdi=true