Multi-modal AI for comprehensive breast cancer prognostication
Treatment selection in breast cancer is guided by molecular subtypes and clinical characteristics. Recurrence risk assessment plays a crucial role in personalizing treatment. Current methods, including genomic assays, have limited accuracy and clinical utility, leading to suboptimal decisions for ma...
Gespeichert in:
Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
---|---|
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 | |
container_start_page | |
container_title | |
container_volume | |
creator | Witowski, Jan Zeng, Ken Cappadona, Joseph Elayoubi, Jailan Chiru, Elena Diana Chan, Nancy Kang, Young-Joon Howard, Frederick Ostrovnaya, Irina Fernandez-Granda, Carlos Schnabel, Freya Ozerdem, Ugur Liu, Kangning Steinsnyder, Zoe Thakore, Nitya Sadic, Mohammad Yeung, Frank Liu, Elisa Hill, Theodore Swett, Benjamin Rigau, Danielle Clayburn, Andrew Speirs, Valerie Vetter, Marcus Sojak, Lina Soysal, Simone Muenst Baumhoer, Daniel Choucair, Khalil Zong, Yu Daoud, Lina Saad, Anas Abdulsattar, Waleed Beydoun, Rafic Pan, Jia-Wern Makmur, Haslina Teo, Soo-Hwang Pak, Linda Ma Angel, Victor Zilenaite-Petrulaitiene, Dovile Laurinavicius, Arvydas Klar, Natalie Piening, Brian D Bifulco, Carlo Jun, Sun-Young Yi, Jae Pak Lim, Su Hyun Brufsky, Adam Esteva, Francisco J Pusztai, Lajos LeCun, Yann Geras, Krzysztof J |
description | Treatment selection in breast cancer is guided by molecular subtypes and
clinical characteristics. Recurrence risk assessment plays a crucial role in
personalizing treatment. Current methods, including genomic assays, have
limited accuracy and clinical utility, leading to suboptimal decisions for many
patients. We developed a test for breast cancer patient stratification based on
digital pathology and clinical characteristics using novel AI methods.
Specifically, we utilized a vision transformer-based pan-cancer foundation
model trained with self-supervised learning to extract features from digitized
H&E-stained slides. These features were integrated with clinical data to form a
multi-modal AI test predicting cancer recurrence and death. The test was
developed and evaluated using data from a total of 8,161 breast cancer patients
across 15 cohorts originating from seven countries. Of these, 3,502 patients
from five cohorts were used exclusively for evaluation, while the remaining
patients were used for training. Our test accurately predicted our primary
endpoint, disease-free interval, in the five external cohorts (C-index: 0.71
[0.68-0.75], HR: 3.63 [3.02-4.37, p |
doi_str_mv | 10.48550/arxiv.2410.21256 |
format | Article |
fullrecord | <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2410_21256</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2410_21256</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2410_212563</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMgEKGBkamZpxMtj5luaUZOrm5qck5ig4eiqk5RcpJOfnFhSlZqTmFWeWpSokFaUmFpcoJCfmJacWKRQU5afn5ReXZCYnlmTm5_EwsKYl5hSn8kJpbgZ5N9cQZw9dsE3xBUWZuYlFlfEgG-PBNhoTVgEAlLI2VQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Multi-modal AI for comprehensive breast cancer prognostication</title><source>arXiv.org</source><creator>Witowski, Jan ; Zeng, Ken ; Cappadona, Joseph ; Elayoubi, Jailan ; Chiru, Elena Diana ; Chan, Nancy ; Kang, Young-Joon ; Howard, Frederick ; Ostrovnaya, Irina ; Fernandez-Granda, Carlos ; Schnabel, Freya ; Ozerdem, Ugur ; Liu, Kangning ; Steinsnyder, Zoe ; Thakore, Nitya ; Sadic, Mohammad ; Yeung, Frank ; Liu, Elisa ; Hill, Theodore ; Swett, Benjamin ; Rigau, Danielle ; Clayburn, Andrew ; Speirs, Valerie ; Vetter, Marcus ; Sojak, Lina ; Soysal, Simone Muenst ; Baumhoer, Daniel ; Choucair, Khalil ; Zong, Yu ; Daoud, Lina ; Saad, Anas ; Abdulsattar, Waleed ; Beydoun, Rafic ; Pan, Jia-Wern ; Makmur, Haslina ; Teo, Soo-Hwang ; Pak, Linda Ma ; Angel, Victor ; Zilenaite-Petrulaitiene, Dovile ; Laurinavicius, Arvydas ; Klar, Natalie ; Piening, Brian D ; Bifulco, Carlo ; Jun, Sun-Young ; Yi, Jae Pak ; Lim, Su Hyun ; Brufsky, Adam ; Esteva, Francisco J ; Pusztai, Lajos ; LeCun, Yann ; Geras, Krzysztof J</creator><creatorcontrib>Witowski, Jan ; Zeng, Ken ; Cappadona, Joseph ; Elayoubi, Jailan ; Chiru, Elena Diana ; Chan, Nancy ; Kang, Young-Joon ; Howard, Frederick ; Ostrovnaya, Irina ; Fernandez-Granda, Carlos ; Schnabel, Freya ; Ozerdem, Ugur ; Liu, Kangning ; Steinsnyder, Zoe ; Thakore, Nitya ; Sadic, Mohammad ; Yeung, Frank ; Liu, Elisa ; Hill, Theodore ; Swett, Benjamin ; Rigau, Danielle ; Clayburn, Andrew ; Speirs, Valerie ; Vetter, Marcus ; Sojak, Lina ; Soysal, Simone Muenst ; Baumhoer, Daniel ; Choucair, Khalil ; Zong, Yu ; Daoud, Lina ; Saad, Anas ; Abdulsattar, Waleed ; Beydoun, Rafic ; Pan, Jia-Wern ; Makmur, Haslina ; Teo, Soo-Hwang ; Pak, Linda Ma ; Angel, Victor ; Zilenaite-Petrulaitiene, Dovile ; Laurinavicius, Arvydas ; Klar, Natalie ; Piening, Brian D ; Bifulco, Carlo ; Jun, Sun-Young ; Yi, Jae Pak ; Lim, Su Hyun ; Brufsky, Adam ; Esteva, Francisco J ; Pusztai, Lajos ; LeCun, Yann ; Geras, Krzysztof J</creatorcontrib><description>Treatment selection in breast cancer is guided by molecular subtypes and
clinical characteristics. Recurrence risk assessment plays a crucial role in
personalizing treatment. Current methods, including genomic assays, have
limited accuracy and clinical utility, leading to suboptimal decisions for many
patients. We developed a test for breast cancer patient stratification based on
digital pathology and clinical characteristics using novel AI methods.
Specifically, we utilized a vision transformer-based pan-cancer foundation
model trained with self-supervised learning to extract features from digitized
H&E-stained slides. These features were integrated with clinical data to form a
multi-modal AI test predicting cancer recurrence and death. The test was
developed and evaluated using data from a total of 8,161 breast cancer patients
across 15 cohorts originating from seven countries. Of these, 3,502 patients
from five cohorts were used exclusively for evaluation, while the remaining
patients were used for training. Our test accurately predicted our primary
endpoint, disease-free interval, in the five external cohorts (C-index: 0.71
[0.68-0.75], HR: 3.63 [3.02-4.37, p<0.01]). In a direct comparison (N=858), the
AI test was more accurate than Oncotype DX, the standard-of-care 21-gene assay,
with a C-index of 0.67 [0.61-0.74] versus 0.61 [0.49-0.73], respectively.
Additionally, the AI test added independent information to Oncotype DX in a
multivariate analysis (HR: 3.11 [1.91-5.09, p<0.01)]). The test demonstrated
robust accuracy across all major breast cancer subtypes, including TNBC
(C-index: 0.71 [0.62-0.81], HR: 3.81 [2.35-6.17, p=0.02]), where no diagnostic
tools are currently recommended by clinical guidelines. These results suggest
that our AI test can improve accuracy, extend applicability to a wider range of
patients, and enhance access to treatment selection tools.</description><identifier>DOI: 10.48550/arxiv.2410.21256</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2024-10</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2410.21256$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2410.21256$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Witowski, Jan</creatorcontrib><creatorcontrib>Zeng, Ken</creatorcontrib><creatorcontrib>Cappadona, Joseph</creatorcontrib><creatorcontrib>Elayoubi, Jailan</creatorcontrib><creatorcontrib>Chiru, Elena Diana</creatorcontrib><creatorcontrib>Chan, Nancy</creatorcontrib><creatorcontrib>Kang, Young-Joon</creatorcontrib><creatorcontrib>Howard, Frederick</creatorcontrib><creatorcontrib>Ostrovnaya, Irina</creatorcontrib><creatorcontrib>Fernandez-Granda, Carlos</creatorcontrib><creatorcontrib>Schnabel, Freya</creatorcontrib><creatorcontrib>Ozerdem, Ugur</creatorcontrib><creatorcontrib>Liu, Kangning</creatorcontrib><creatorcontrib>Steinsnyder, Zoe</creatorcontrib><creatorcontrib>Thakore, Nitya</creatorcontrib><creatorcontrib>Sadic, Mohammad</creatorcontrib><creatorcontrib>Yeung, Frank</creatorcontrib><creatorcontrib>Liu, Elisa</creatorcontrib><creatorcontrib>Hill, Theodore</creatorcontrib><creatorcontrib>Swett, Benjamin</creatorcontrib><creatorcontrib>Rigau, Danielle</creatorcontrib><creatorcontrib>Clayburn, Andrew</creatorcontrib><creatorcontrib>Speirs, Valerie</creatorcontrib><creatorcontrib>Vetter, Marcus</creatorcontrib><creatorcontrib>Sojak, Lina</creatorcontrib><creatorcontrib>Soysal, Simone Muenst</creatorcontrib><creatorcontrib>Baumhoer, Daniel</creatorcontrib><creatorcontrib>Choucair, Khalil</creatorcontrib><creatorcontrib>Zong, Yu</creatorcontrib><creatorcontrib>Daoud, Lina</creatorcontrib><creatorcontrib>Saad, Anas</creatorcontrib><creatorcontrib>Abdulsattar, Waleed</creatorcontrib><creatorcontrib>Beydoun, Rafic</creatorcontrib><creatorcontrib>Pan, Jia-Wern</creatorcontrib><creatorcontrib>Makmur, Haslina</creatorcontrib><creatorcontrib>Teo, Soo-Hwang</creatorcontrib><creatorcontrib>Pak, Linda Ma</creatorcontrib><creatorcontrib>Angel, Victor</creatorcontrib><creatorcontrib>Zilenaite-Petrulaitiene, Dovile</creatorcontrib><creatorcontrib>Laurinavicius, Arvydas</creatorcontrib><creatorcontrib>Klar, Natalie</creatorcontrib><creatorcontrib>Piening, Brian D</creatorcontrib><creatorcontrib>Bifulco, Carlo</creatorcontrib><creatorcontrib>Jun, Sun-Young</creatorcontrib><creatorcontrib>Yi, Jae Pak</creatorcontrib><creatorcontrib>Lim, Su Hyun</creatorcontrib><creatorcontrib>Brufsky, Adam</creatorcontrib><creatorcontrib>Esteva, Francisco J</creatorcontrib><creatorcontrib>Pusztai, Lajos</creatorcontrib><creatorcontrib>LeCun, Yann</creatorcontrib><creatorcontrib>Geras, Krzysztof J</creatorcontrib><title>Multi-modal AI for comprehensive breast cancer prognostication</title><description>Treatment selection in breast cancer is guided by molecular subtypes and
clinical characteristics. Recurrence risk assessment plays a crucial role in
personalizing treatment. Current methods, including genomic assays, have
limited accuracy and clinical utility, leading to suboptimal decisions for many
patients. We developed a test for breast cancer patient stratification based on
digital pathology and clinical characteristics using novel AI methods.
Specifically, we utilized a vision transformer-based pan-cancer foundation
model trained with self-supervised learning to extract features from digitized
H&E-stained slides. These features were integrated with clinical data to form a
multi-modal AI test predicting cancer recurrence and death. The test was
developed and evaluated using data from a total of 8,161 breast cancer patients
across 15 cohorts originating from seven countries. Of these, 3,502 patients
from five cohorts were used exclusively for evaluation, while the remaining
patients were used for training. Our test accurately predicted our primary
endpoint, disease-free interval, in the five external cohorts (C-index: 0.71
[0.68-0.75], HR: 3.63 [3.02-4.37, p<0.01]). In a direct comparison (N=858), the
AI test was more accurate than Oncotype DX, the standard-of-care 21-gene assay,
with a C-index of 0.67 [0.61-0.74] versus 0.61 [0.49-0.73], respectively.
Additionally, the AI test added independent information to Oncotype DX in a
multivariate analysis (HR: 3.11 [1.91-5.09, p<0.01)]). The test demonstrated
robust accuracy across all major breast cancer subtypes, including TNBC
(C-index: 0.71 [0.62-0.81], HR: 3.81 [2.35-6.17, p=0.02]), where no diagnostic
tools are currently recommended by clinical guidelines. These results suggest
that our AI test can improve accuracy, extend applicability to a wider range of
patients, and enhance access to treatment selection tools.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMgEKGBkamZpxMtj5luaUZOrm5qck5ig4eiqk5RcpJOfnFhSlZqTmFWeWpSokFaUmFpcoJCfmJacWKRQU5afn5ReXZCYnlmTm5_EwsKYl5hSn8kJpbgZ5N9cQZw9dsE3xBUWZuYlFlfEgG-PBNhoTVgEAlLI2VQ</recordid><startdate>20241028</startdate><enddate>20241028</enddate><creator>Witowski, Jan</creator><creator>Zeng, Ken</creator><creator>Cappadona, Joseph</creator><creator>Elayoubi, Jailan</creator><creator>Chiru, Elena Diana</creator><creator>Chan, Nancy</creator><creator>Kang, Young-Joon</creator><creator>Howard, Frederick</creator><creator>Ostrovnaya, Irina</creator><creator>Fernandez-Granda, Carlos</creator><creator>Schnabel, Freya</creator><creator>Ozerdem, Ugur</creator><creator>Liu, Kangning</creator><creator>Steinsnyder, Zoe</creator><creator>Thakore, Nitya</creator><creator>Sadic, Mohammad</creator><creator>Yeung, Frank</creator><creator>Liu, Elisa</creator><creator>Hill, Theodore</creator><creator>Swett, Benjamin</creator><creator>Rigau, Danielle</creator><creator>Clayburn, Andrew</creator><creator>Speirs, Valerie</creator><creator>Vetter, Marcus</creator><creator>Sojak, Lina</creator><creator>Soysal, Simone Muenst</creator><creator>Baumhoer, Daniel</creator><creator>Choucair, Khalil</creator><creator>Zong, Yu</creator><creator>Daoud, Lina</creator><creator>Saad, Anas</creator><creator>Abdulsattar, Waleed</creator><creator>Beydoun, Rafic</creator><creator>Pan, Jia-Wern</creator><creator>Makmur, Haslina</creator><creator>Teo, Soo-Hwang</creator><creator>Pak, Linda Ma</creator><creator>Angel, Victor</creator><creator>Zilenaite-Petrulaitiene, Dovile</creator><creator>Laurinavicius, Arvydas</creator><creator>Klar, Natalie</creator><creator>Piening, Brian D</creator><creator>Bifulco, Carlo</creator><creator>Jun, Sun-Young</creator><creator>Yi, Jae Pak</creator><creator>Lim, Su Hyun</creator><creator>Brufsky, Adam</creator><creator>Esteva, Francisco J</creator><creator>Pusztai, Lajos</creator><creator>LeCun, Yann</creator><creator>Geras, Krzysztof J</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20241028</creationdate><title>Multi-modal AI for comprehensive breast cancer prognostication</title><author>Witowski, Jan ; Zeng, Ken ; Cappadona, Joseph ; Elayoubi, Jailan ; Chiru, Elena Diana ; Chan, Nancy ; Kang, Young-Joon ; Howard, Frederick ; Ostrovnaya, Irina ; Fernandez-Granda, Carlos ; Schnabel, Freya ; Ozerdem, Ugur ; Liu, Kangning ; Steinsnyder, Zoe ; Thakore, Nitya ; Sadic, Mohammad ; Yeung, Frank ; Liu, Elisa ; Hill, Theodore ; Swett, Benjamin ; Rigau, Danielle ; Clayburn, Andrew ; Speirs, Valerie ; Vetter, Marcus ; Sojak, Lina ; Soysal, Simone Muenst ; Baumhoer, Daniel ; Choucair, Khalil ; Zong, Yu ; Daoud, Lina ; Saad, Anas ; Abdulsattar, Waleed ; Beydoun, Rafic ; Pan, Jia-Wern ; Makmur, Haslina ; Teo, Soo-Hwang ; Pak, Linda Ma ; Angel, Victor ; Zilenaite-Petrulaitiene, Dovile ; Laurinavicius, Arvydas ; Klar, Natalie ; Piening, Brian D ; Bifulco, Carlo ; Jun, Sun-Young ; Yi, Jae Pak ; Lim, Su Hyun ; Brufsky, Adam ; Esteva, Francisco J ; Pusztai, Lajos ; LeCun, Yann ; Geras, Krzysztof J</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2410_212563</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Witowski, Jan</creatorcontrib><creatorcontrib>Zeng, Ken</creatorcontrib><creatorcontrib>Cappadona, Joseph</creatorcontrib><creatorcontrib>Elayoubi, Jailan</creatorcontrib><creatorcontrib>Chiru, Elena Diana</creatorcontrib><creatorcontrib>Chan, Nancy</creatorcontrib><creatorcontrib>Kang, Young-Joon</creatorcontrib><creatorcontrib>Howard, Frederick</creatorcontrib><creatorcontrib>Ostrovnaya, Irina</creatorcontrib><creatorcontrib>Fernandez-Granda, Carlos</creatorcontrib><creatorcontrib>Schnabel, Freya</creatorcontrib><creatorcontrib>Ozerdem, Ugur</creatorcontrib><creatorcontrib>Liu, Kangning</creatorcontrib><creatorcontrib>Steinsnyder, Zoe</creatorcontrib><creatorcontrib>Thakore, Nitya</creatorcontrib><creatorcontrib>Sadic, Mohammad</creatorcontrib><creatorcontrib>Yeung, Frank</creatorcontrib><creatorcontrib>Liu, Elisa</creatorcontrib><creatorcontrib>Hill, Theodore</creatorcontrib><creatorcontrib>Swett, Benjamin</creatorcontrib><creatorcontrib>Rigau, Danielle</creatorcontrib><creatorcontrib>Clayburn, Andrew</creatorcontrib><creatorcontrib>Speirs, Valerie</creatorcontrib><creatorcontrib>Vetter, Marcus</creatorcontrib><creatorcontrib>Sojak, Lina</creatorcontrib><creatorcontrib>Soysal, Simone Muenst</creatorcontrib><creatorcontrib>Baumhoer, Daniel</creatorcontrib><creatorcontrib>Choucair, Khalil</creatorcontrib><creatorcontrib>Zong, Yu</creatorcontrib><creatorcontrib>Daoud, Lina</creatorcontrib><creatorcontrib>Saad, Anas</creatorcontrib><creatorcontrib>Abdulsattar, Waleed</creatorcontrib><creatorcontrib>Beydoun, Rafic</creatorcontrib><creatorcontrib>Pan, Jia-Wern</creatorcontrib><creatorcontrib>Makmur, Haslina</creatorcontrib><creatorcontrib>Teo, Soo-Hwang</creatorcontrib><creatorcontrib>Pak, Linda Ma</creatorcontrib><creatorcontrib>Angel, Victor</creatorcontrib><creatorcontrib>Zilenaite-Petrulaitiene, Dovile</creatorcontrib><creatorcontrib>Laurinavicius, Arvydas</creatorcontrib><creatorcontrib>Klar, Natalie</creatorcontrib><creatorcontrib>Piening, Brian D</creatorcontrib><creatorcontrib>Bifulco, Carlo</creatorcontrib><creatorcontrib>Jun, Sun-Young</creatorcontrib><creatorcontrib>Yi, Jae Pak</creatorcontrib><creatorcontrib>Lim, Su Hyun</creatorcontrib><creatorcontrib>Brufsky, Adam</creatorcontrib><creatorcontrib>Esteva, Francisco J</creatorcontrib><creatorcontrib>Pusztai, Lajos</creatorcontrib><creatorcontrib>LeCun, Yann</creatorcontrib><creatorcontrib>Geras, Krzysztof J</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Witowski, Jan</au><au>Zeng, Ken</au><au>Cappadona, Joseph</au><au>Elayoubi, Jailan</au><au>Chiru, Elena Diana</au><au>Chan, Nancy</au><au>Kang, Young-Joon</au><au>Howard, Frederick</au><au>Ostrovnaya, Irina</au><au>Fernandez-Granda, Carlos</au><au>Schnabel, Freya</au><au>Ozerdem, Ugur</au><au>Liu, Kangning</au><au>Steinsnyder, Zoe</au><au>Thakore, Nitya</au><au>Sadic, Mohammad</au><au>Yeung, Frank</au><au>Liu, Elisa</au><au>Hill, Theodore</au><au>Swett, Benjamin</au><au>Rigau, Danielle</au><au>Clayburn, Andrew</au><au>Speirs, Valerie</au><au>Vetter, Marcus</au><au>Sojak, Lina</au><au>Soysal, Simone Muenst</au><au>Baumhoer, Daniel</au><au>Choucair, Khalil</au><au>Zong, Yu</au><au>Daoud, Lina</au><au>Saad, Anas</au><au>Abdulsattar, Waleed</au><au>Beydoun, Rafic</au><au>Pan, Jia-Wern</au><au>Makmur, Haslina</au><au>Teo, Soo-Hwang</au><au>Pak, Linda Ma</au><au>Angel, Victor</au><au>Zilenaite-Petrulaitiene, Dovile</au><au>Laurinavicius, Arvydas</au><au>Klar, Natalie</au><au>Piening, Brian D</au><au>Bifulco, Carlo</au><au>Jun, Sun-Young</au><au>Yi, Jae Pak</au><au>Lim, Su Hyun</au><au>Brufsky, Adam</au><au>Esteva, Francisco J</au><au>Pusztai, Lajos</au><au>LeCun, Yann</au><au>Geras, Krzysztof J</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Multi-modal AI for comprehensive breast cancer prognostication</atitle><date>2024-10-28</date><risdate>2024</risdate><abstract>Treatment selection in breast cancer is guided by molecular subtypes and
clinical characteristics. Recurrence risk assessment plays a crucial role in
personalizing treatment. Current methods, including genomic assays, have
limited accuracy and clinical utility, leading to suboptimal decisions for many
patients. We developed a test for breast cancer patient stratification based on
digital pathology and clinical characteristics using novel AI methods.
Specifically, we utilized a vision transformer-based pan-cancer foundation
model trained with self-supervised learning to extract features from digitized
H&E-stained slides. These features were integrated with clinical data to form a
multi-modal AI test predicting cancer recurrence and death. The test was
developed and evaluated using data from a total of 8,161 breast cancer patients
across 15 cohorts originating from seven countries. Of these, 3,502 patients
from five cohorts were used exclusively for evaluation, while the remaining
patients were used for training. Our test accurately predicted our primary
endpoint, disease-free interval, in the five external cohorts (C-index: 0.71
[0.68-0.75], HR: 3.63 [3.02-4.37, p<0.01]). In a direct comparison (N=858), the
AI test was more accurate than Oncotype DX, the standard-of-care 21-gene assay,
with a C-index of 0.67 [0.61-0.74] versus 0.61 [0.49-0.73], respectively.
Additionally, the AI test added independent information to Oncotype DX in a
multivariate analysis (HR: 3.11 [1.91-5.09, p<0.01)]). The test demonstrated
robust accuracy across all major breast cancer subtypes, including TNBC
(C-index: 0.71 [0.62-0.81], HR: 3.81 [2.35-6.17, p=0.02]), where no diagnostic
tools are currently recommended by clinical guidelines. These results suggest
that our AI test can improve accuracy, extend applicability to a wider range of
patients, and enhance access to treatment selection tools.</abstract><doi>10.48550/arxiv.2410.21256</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | DOI: 10.48550/arxiv.2410.21256 |
ispartof | |
issn | |
language | eng |
recordid | cdi_arxiv_primary_2410_21256 |
source | arXiv.org |
subjects | Computer Science - Artificial Intelligence Computer Science - Computer Vision and Pattern Recognition |
title | Multi-modal AI for comprehensive breast cancer prognostication |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T13%3A27%3A42IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Multi-modal%20AI%20for%20comprehensive%20breast%20cancer%20prognostication&rft.au=Witowski,%20Jan&rft.date=2024-10-28&rft_id=info:doi/10.48550/arxiv.2410.21256&rft_dat=%3Carxiv_GOX%3E2410_21256%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |