SurvMaximin : Robust federated approach to transporting survival risk prediction models
Copyright © 2022. Published by Elsevier Inc..
OBJECTIVE: For multi-center heterogeneous Real-World Data (RWD) with time-to-event outcomes and high-dimensional features, we propose the SurvMaximin algorithm to estimate Cox model feature coefficients for a target population by borrowing summary information from a set of health care centers without sharing patient-level information.
MATERIALS AND METHODS: For each of the centers from which we want to borrow information to improve the prediction performance for the target population, a penalized Cox model is fitted to estimate feature coefficients for the center. Using estimated feature coefficients and the covariance matrix of the target population, we then obtain a SurvMaximin estimated set of feature coefficients for the target population. The target population can be an entire cohort comprised of all centers, corresponding to federated learning, or a single center, corresponding to transfer learning.
RESULTS: Simulation studies and a real-world international electronic health records application study, with 15 participating health care centers across three countries (France, Germany, and the U.S.), show that the proposed SurvMaximin algorithm achieves comparable or higher accuracy compared with the estimator using only the information of the target site and other existing methods. The SurvMaximin estimator is robust to variations in sample sizes and estimated feature coefficients between centers, which amounts to significantly improved estimates for target sites with fewer observations.
CONCLUSIONS: The SurvMaximin method is well suited for both federated and transfer learning in the high-dimensional survival analysis setting. SurvMaximin only requires a one-time summary information exchange from participating centers. Estimated regression vectors can be very heterogeneous. SurvMaximin provides robust Cox feature coefficient estimates without outcome information in the target population and is privacy-preserving.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2022 |
---|---|
Erschienen: |
2022 |
Enthalten in: |
Zur Gesamtaufnahme - volume:134 |
---|---|
Enthalten in: |
Journal of biomedical informatics - 134(2022) vom: 01. Okt., Seite 104176 |
Sprache: |
Englisch |
---|
Links: |
---|
Themen: |
Journal Article |
---|
Anmerkungen: |
Date Completed 13.10.2022 Date Revised 16.04.2024 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1016/j.jbi.2022.104176 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM345261925 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | NLM345261925 | ||
003 | DE-627 | ||
005 | 20240416232156.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2022 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1016/j.jbi.2022.104176 |2 doi | |
028 | 5 | 2 | |a pubmed24n1377.xml |
035 | |a (DE-627)NLM345261925 | ||
035 | |a (NLM)36007785 | ||
035 | |a (PII)S1532-0464(22)00187-3 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Wang, Xuan |e verfasserin |4 aut | |
245 | 1 | 0 | |a SurvMaximin |b Robust federated approach to transporting survival risk prediction models |
264 | 1 | |c 2022 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ƒaComputermedien |b c |2 rdamedia | ||
338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
500 | |a Date Completed 13.10.2022 | ||
500 | |a Date Revised 16.04.2024 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a Copyright © 2022. Published by Elsevier Inc. | ||
520 | |a OBJECTIVE: For multi-center heterogeneous Real-World Data (RWD) with time-to-event outcomes and high-dimensional features, we propose the SurvMaximin algorithm to estimate Cox model feature coefficients for a target population by borrowing summary information from a set of health care centers without sharing patient-level information | ||
520 | |a MATERIALS AND METHODS: For each of the centers from which we want to borrow information to improve the prediction performance for the target population, a penalized Cox model is fitted to estimate feature coefficients for the center. Using estimated feature coefficients and the covariance matrix of the target population, we then obtain a SurvMaximin estimated set of feature coefficients for the target population. The target population can be an entire cohort comprised of all centers, corresponding to federated learning, or a single center, corresponding to transfer learning | ||
520 | |a RESULTS: Simulation studies and a real-world international electronic health records application study, with 15 participating health care centers across three countries (France, Germany, and the U.S.), show that the proposed SurvMaximin algorithm achieves comparable or higher accuracy compared with the estimator using only the information of the target site and other existing methods. The SurvMaximin estimator is robust to variations in sample sizes and estimated feature coefficients between centers, which amounts to significantly improved estimates for target sites with fewer observations | ||
520 | |a CONCLUSIONS: The SurvMaximin method is well suited for both federated and transfer learning in the high-dimensional survival analysis setting. SurvMaximin only requires a one-time summary information exchange from participating centers. Estimated regression vectors can be very heterogeneous. SurvMaximin provides robust Cox feature coefficient estimates without outcome information in the target population and is privacy-preserving | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a Research Support, Non-U.S. Gov't | |
650 | 4 | |a Research Support, N.I.H., Extramural | |
650 | 4 | |a Research Support, U.S. Gov't, Non-P.H.S. | |
700 | 1 | |a Zhang, Harrison G |e verfasserin |4 aut | |
700 | 1 | |a Xiong, Xin |e verfasserin |4 aut | |
700 | 1 | |a Hong, Chuan |e verfasserin |4 aut | |
700 | 1 | |a Weber, Griffin M |e verfasserin |4 aut | |
700 | 1 | |a Brat, Gabriel A |e verfasserin |4 aut | |
700 | 1 | |a Bonzel, Clara-Lea |e verfasserin |4 aut | |
700 | 1 | |a Luo, Yuan |e verfasserin |4 aut | |
700 | 1 | |a Duan, Rui |e verfasserin |4 aut | |
700 | 1 | |a Palmer, Nathan P |e verfasserin |4 aut | |
700 | 1 | |a Hutch, Meghan R |e verfasserin |4 aut | |
700 | 1 | |a Gutiérrez-Sacristán, Alba |e verfasserin |4 aut | |
700 | 1 | |a Bellazzi, Riccardo |e verfasserin |4 aut | |
700 | 1 | |a Chiovato, Luca |e verfasserin |4 aut | |
700 | 1 | |a Cho, Kelly |e verfasserin |4 aut | |
700 | 1 | |a Dagliati, Arianna |e verfasserin |4 aut | |
700 | 1 | |a Estiri, Hossein |e verfasserin |4 aut | |
700 | 1 | |a García-Barrio, Noelia |e verfasserin |4 aut | |
700 | 1 | |a Griffier, Romain |e verfasserin |4 aut | |
700 | 1 | |a Hanauer, David A |e verfasserin |4 aut | |
700 | 1 | |a Ho, Yuk-Lam |e verfasserin |4 aut | |
700 | 1 | |a Holmes, John H |e verfasserin |4 aut | |
700 | 1 | |a Keller, Mark S |e verfasserin |4 aut | |
700 | 1 | |a Klann MEng, Jeffrey G |e verfasserin |4 aut | |
700 | 1 | |a L'Yi, Sehi |e verfasserin |4 aut | |
700 | 1 | |a Lozano-Zahonero, Sara |e verfasserin |4 aut | |
700 | 1 | |a Maidlow, Sarah E |e verfasserin |4 aut | |
700 | 1 | |a Makoudjou, Adeline |e verfasserin |4 aut | |
700 | 1 | |a Malovini, Alberto |e verfasserin |4 aut | |
700 | 1 | |a Moal, Bertrand |e verfasserin |4 aut | |
700 | 1 | |a Moore, Jason H |e verfasserin |4 aut | |
700 | 1 | |a Morris, Michele |e verfasserin |4 aut | |
700 | 1 | |a Mowery, Danielle L |e verfasserin |4 aut | |
700 | 1 | |a Murphy, Shawn N |e verfasserin |4 aut | |
700 | 1 | |a Neuraz, Antoine |e verfasserin |4 aut | |
700 | 1 | |a Yuan Ngiam, Kee |e verfasserin |4 aut | |
700 | 1 | |a Omenn, Gilbert S |e verfasserin |4 aut | |
700 | 1 | |a Patel, Lav P |e verfasserin |4 aut | |
700 | 1 | |a Pedrera-Jiménez, Miguel |e verfasserin |4 aut | |
700 | 1 | |a Prunotto, Andrea |e verfasserin |4 aut | |
700 | 1 | |a Jebathilagam Samayamuthu, Malarkodi |e verfasserin |4 aut | |
700 | 1 | |a Sanz Vidorreta, Fernando J |e verfasserin |4 aut | |
700 | 1 | |a Schriver, Emily R |e verfasserin |4 aut | |
700 | 1 | |a Schubert, Petra |e verfasserin |4 aut | |
700 | 1 | |a Serrano-Balazote, Pablo |e verfasserin |4 aut | |
700 | 1 | |a South, Andrew M |e verfasserin |4 aut | |
700 | 1 | |a Tan, Amelia L M |e verfasserin |4 aut | |
700 | 1 | |a Tan, Byorn W L |e verfasserin |4 aut | |
700 | 1 | |a Tibollo, Valentina |e verfasserin |4 aut | |
700 | 1 | |a Tippmann, Patric |e verfasserin |4 aut | |
700 | 1 | |a Visweswaran, Shyam |e verfasserin |4 aut | |
700 | 1 | |a Xia, Zongqi |e verfasserin |4 aut | |
700 | 1 | |a Yuan, William |e verfasserin |4 aut | |
700 | 1 | |a Zöller, Daniela |e verfasserin |4 aut | |
700 | 1 | |a Kohane, Isaac S |e verfasserin |4 aut | |
700 | 1 | |a Avillach, Paul |e verfasserin |4 aut | |
700 | 1 | |a Guo, Zijian |e verfasserin |4 aut | |
700 | 1 | |a Cai, Tianxi |e verfasserin |4 aut | |
700 | 0 | |a Consortium for Clinical Characterization of COVID-19 by EHR 4CE |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t Journal of biomedical informatics |d 2001 |g 134(2022) vom: 01. Okt., Seite 104176 |w (DE-627)NLM112821766 |x 1532-0480 |7 nnns |
773 | 1 | 8 | |g volume:134 |g year:2022 |g day:01 |g month:10 |g pages:104176 |
856 | 4 | 0 | |u http://dx.doi.org/10.1016/j.jbi.2022.104176 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 134 |j 2022 |b 01 |c 10 |h 104176 |