2023
DOI: 10.1038/s41591-022-02155-w
|View full text |Cite
|
Sign up to set email alerts
|

Federated learning for predicting histological response to neoadjuvant chemotherapy in triple-negative breast cancer

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
31
0
1

Year Published

2023
2023
2024
2024

Publication Types

Select...
6
1
1

Relationship

0
8

Authors

Journals

citations
Cited by 88 publications
(32 citation statements)
references
References 88 publications
0
31
0
1
Order By: Relevance
“…57 In contrast to the conventional data pooling approach, the federated approach trains models locally within each institution. 58 For instance, in forming a federated model, only the models trained in the institutions are shared and aggregated by a central organizing company. Over time, the federated model is then optimized by integrating the insights from all different sources without directly accessing individual patient data from the institutions.…”
Section: Data Sharing Distributed Learning and Synthetic Data Generationmentioning
confidence: 99%
See 1 more Smart Citation
“…57 In contrast to the conventional data pooling approach, the federated approach trains models locally within each institution. 58 For instance, in forming a federated model, only the models trained in the institutions are shared and aggregated by a central organizing company. Over time, the federated model is then optimized by integrating the insights from all different sources without directly accessing individual patient data from the institutions.…”
Section: Data Sharing Distributed Learning and Synthetic Data Generationmentioning
confidence: 99%
“…Federated data access offers a novel way to break these data silos and enable collaborative ML efforts among different data sources (i.e., pharmaceutical companies, clinical researchers, manufacturers, and hospitals) 57 . In contrast to the conventional data pooling approach, the federated approach trains models locally within each institution 58 . For instance, in forming a federated model, only the models trained in the institutions are shared and aggregated by a central organizing company.…”
Section: Challenges and Future Directionmentioning
confidence: 99%
“…Based on the proportion of data in different centers, the FedAvg algorithm weighs and averages the model parameters on the central server and feeds them back to each center. FedAvg achieved preliminary results in the clinical tasks of COVID-19 lung abnormalities detection, 142 rare cancer boundary detection, 143 response prediction to neoadjuvant treatment, 144 and prognosis prediction. 145 However, the FedAvg strategy may exacerbate the negative impact of non-IID data on model training and affect model convergence.…”
Section: Patient Privacy Protectionmentioning
confidence: 99%
“…Itt szövettani teljes digitális képek mély tanulásával (deep learning), azaz a képek több rétegben kivonatolható tulajdonságain alapuló, mesterséges neurális hálózatok rétegeiben leképezett, általánosító következtetésekkel dolgoztak, megkerülve az időigényes szakértői annotációt. Ezzel együtt az értelmezhetőség sem veszett el teljesen, ismert és potenciális biomarkerek -mint az apokrin daganatsejtek, az infiltráló lymphocyták, illetve a fibrosis és a daganatsejtek elrendeződésejelentőségét sikerült számszerűsíteni [28].…”
Section: Adatvagyonleltár éS Együ�működési Célok áLtal Vezérelt Adatg...unclassified