Continual learning strategies for cancer-independent detection of lymph node metastases
Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved..
Recently, large, high-quality public datasets have led to the development of convolutional neural networks that can detect lymph node metastases of breast cancer at the level of expert pathologists. Many cancers, regardless of the site of origin, can metastasize to lymph nodes. However, collecting and annotating high-volume, high-quality datasets for every cancer type is challenging. In this paper we investigate how to leverage existing high-quality datasets most efficiently in multi-task settings for closely related tasks. Specifically, we will explore different training and domain adaptation strategies, including prevention of catastrophic forgetting, for breast, colon and head-and-neck cancer metastasis detection in lymph nodes. Our results show state-of-the-art performance on colon and head-and-neck cancer metastasis detection tasks. We show the effectiveness of adaptation of networks from one cancer type to another to obtain multi-task metastasis detection networks. Furthermore, we show that leveraging existing high-quality datasets can significantly boost performance on new target tasks and that catastrophic forgetting can be effectively mitigated.Last, we compare different mitigation strategies.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2023 |
---|---|
Erschienen: |
2023 |
Enthalten in: |
Zur Gesamtaufnahme - volume:85 |
---|---|
Enthalten in: |
Medical image analysis - 85(2023) vom: 16. Apr., Seite 102755 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Bándi, Péter [VerfasserIn] |
---|
Links: |
---|
Themen: |
Cancer |
---|
Anmerkungen: |
Date Completed 28.02.2023 Date Revised 17.03.2023 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1016/j.media.2023.102755 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM352349255 |
---|
LEADER | 01000naa a22002652 4500 | ||
---|---|---|---|
001 | NLM352349255 | ||
003 | DE-627 | ||
005 | 20231226053523.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2023 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1016/j.media.2023.102755 |2 doi | |
028 | 5 | 2 | |a pubmed24n1174.xml |
035 | |a (DE-627)NLM352349255 | ||
035 | |a (NLM)36724605 | ||
035 | |a (PII)S1361-8415(23)00016-6 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Bándi, Péter |e verfasserin |4 aut | |
245 | 1 | 0 | |a Continual learning strategies for cancer-independent detection of lymph node metastases |
264 | 1 | |c 2023 | |
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 28.02.2023 | ||
500 | |a Date Revised 17.03.2023 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved. | ||
520 | |a Recently, large, high-quality public datasets have led to the development of convolutional neural networks that can detect lymph node metastases of breast cancer at the level of expert pathologists. Many cancers, regardless of the site of origin, can metastasize to lymph nodes. However, collecting and annotating high-volume, high-quality datasets for every cancer type is challenging. In this paper we investigate how to leverage existing high-quality datasets most efficiently in multi-task settings for closely related tasks. Specifically, we will explore different training and domain adaptation strategies, including prevention of catastrophic forgetting, for breast, colon and head-and-neck cancer metastasis detection in lymph nodes. Our results show state-of-the-art performance on colon and head-and-neck cancer metastasis detection tasks. We show the effectiveness of adaptation of networks from one cancer type to another to obtain multi-task metastasis detection networks. Furthermore, we show that leveraging existing high-quality datasets can significantly boost performance on new target tasks and that catastrophic forgetting can be effectively mitigated.Last, we compare different mitigation strategies | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a Research Support, Non-U.S. Gov't | |
650 | 4 | |a Cancer | |
650 | 4 | |a Continual learning | |
650 | 4 | |a Convolutional neural network | |
650 | 4 | |a Deep learning | |
650 | 4 | |a Lymph node | |
700 | 1 | |a Balkenhol, Maschenka |e verfasserin |4 aut | |
700 | 1 | |a van Dijk, Marcory |e verfasserin |4 aut | |
700 | 1 | |a Kok, Michel |e verfasserin |4 aut | |
700 | 1 | |a van Ginneken, Bram |e verfasserin |4 aut | |
700 | 1 | |a van der Laak, Jeroen |e verfasserin |4 aut | |
700 | 1 | |a Litjens, Geert |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t Medical image analysis |d 1996 |g 85(2023) vom: 16. Apr., Seite 102755 |w (DE-627)NLM096527188 |x 1361-8423 |7 nnns |
773 | 1 | 8 | |g volume:85 |g year:2023 |g day:16 |g month:04 |g pages:102755 |
856 | 4 | 0 | |u http://dx.doi.org/10.1016/j.media.2023.102755 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 85 |j 2023 |b 16 |c 04 |h 102755 |