Evaluation of deep learning training strategies for the classification of bone marrow cell images.
Deep learning
Hematopathology
Hematopoiesis
In-domain pre-training
Journal
Computer methods and programs in biomedicine
ISSN: 1872-7565
Titre abrégé: Comput Methods Programs Biomed
Pays: Ireland
ID NLM: 8506513
Informations de publication
Date de publication:
Jan 2024
Jan 2024
Historique:
received:
11
05
2023
revised:
28
09
2023
accepted:
06
11
2023
medline:
4
12
2023
pubmed:
19
11
2023
entrez:
18
11
2023
Statut:
ppublish
Résumé
The classification of bone marrow (BM) cells by light microscopy is an important cornerstone of hematological diagnosis, performed thousands of times a day by highly trained specialists in laboratories worldwide. As the manual evaluation of blood or BM smears is very time-consuming and prone to inter-observer variation, new reliable automated systems are needed. We aim to improve the automatic classification performance of hematological cell types. Therefore, we evaluate four state-of-the-art Convolutional Neural Network (CNN) architectures on a dataset of 171,374 microscopic cytological single-cell images obtained from BM smears from 945 patients diagnosed with a variety of hematological diseases. We further evaluate the effect of an in-domain vs. out-of-domain pre-training, and assess whether class activation maps provide human-interpretable explanations for the models' predictions. The best performing pre-trained model (Regnet_y_32gf) yields a mean precision, recall, and F1 scores of 0.787±0.060, 0.755±0.061, and 0.762±0.050, respectively. This is a 53.5% improvement in precision and 7.3% improvement in recall over previous results with CNNs (ResNeXt-50) that were trained from scratch. The out-of-domain pre-training apparently yields general feature extractors/filters that apply very well to the BM cell classification use case. The class activation maps on cell types with characteristic morphological features were found to be consistent with the explanations of a human domain expert. For example, the Auer rods in the cytoplasm were the predictive cellular feature for correctly classified images of faggot cells. Our study provides data that can help hematology laboratories to choose the optimal training strategy for blood cell classification deep learning models to improve computer-assisted blood and bone marrow cell identification. It also highlights the need for more specific training data, i.e. images of difficult-to-classify classes, including cells labeled with disease information.
Sections du résumé
BACKGROUND AND OBJECTIVE
OBJECTIVE
The classification of bone marrow (BM) cells by light microscopy is an important cornerstone of hematological diagnosis, performed thousands of times a day by highly trained specialists in laboratories worldwide. As the manual evaluation of blood or BM smears is very time-consuming and prone to inter-observer variation, new reliable automated systems are needed.
METHODS
METHODS
We aim to improve the automatic classification performance of hematological cell types. Therefore, we evaluate four state-of-the-art Convolutional Neural Network (CNN) architectures on a dataset of 171,374 microscopic cytological single-cell images obtained from BM smears from 945 patients diagnosed with a variety of hematological diseases. We further evaluate the effect of an in-domain vs. out-of-domain pre-training, and assess whether class activation maps provide human-interpretable explanations for the models' predictions.
RESULTS
RESULTS
The best performing pre-trained model (Regnet_y_32gf) yields a mean precision, recall, and F1 scores of 0.787±0.060, 0.755±0.061, and 0.762±0.050, respectively. This is a 53.5% improvement in precision and 7.3% improvement in recall over previous results with CNNs (ResNeXt-50) that were trained from scratch. The out-of-domain pre-training apparently yields general feature extractors/filters that apply very well to the BM cell classification use case. The class activation maps on cell types with characteristic morphological features were found to be consistent with the explanations of a human domain expert. For example, the Auer rods in the cytoplasm were the predictive cellular feature for correctly classified images of faggot cells.
CONCLUSIONS
CONCLUSIONS
Our study provides data that can help hematology laboratories to choose the optimal training strategy for blood cell classification deep learning models to improve computer-assisted blood and bone marrow cell identification. It also highlights the need for more specific training data, i.e. images of difficult-to-classify classes, including cells labeled with disease information.
Identifiants
pubmed: 37979517
pii: S0169-2607(23)00590-4
doi: 10.1016/j.cmpb.2023.107924
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
107924Informations de copyright
Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.
Déclaration de conflit d'intérêts
Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Stefan Balabanov reports a relationship with Alexion that includes: board membership. Stefan Balabanov reports a relationship with Amgen Europe GmbH that includes: board membership. Stefan Balabanov reports a relationship with Blueprint Medicines (Switzerland) GmbH that includes: speaking and lecture fees. Stefan Balabanov reports a relationship with Incyte Biosciences Germany GmbH that includes: speaking and lecture fees. Stefan Balabanov reports a relationship with Novartis that includes: speaking and lecture fees. Stefan Balabanov reports a relationship with Takeda Oncology that includes: speaking and lecture fees. Viktor Hendrik Koelzer reports a relationship with Indica Labs that includes: invited speaker. Viktor Hendrik Koelzer reports a relationship with Image Analysis Group that includes: funding grants. Viktor Hendrik Koelzer reports a relationship with SPCC that includes: speaker fees. Viktor Hendrik Koelzer reports a relationship with Roche that includes: funding grants; advisory board. Viktor Hendrik Koelzer reports a relationship with Takeda that includes: advisory board.