Deep learning identifies Acute Promyelocytic Leukemia in bone marrow smears.
Acute myeloid leukemia
Acute promyelocytic leukemia
Artificial intelligence
Bone marrow smear
Deep learning
Journal
BMC cancer
ISSN: 1471-2407
Titre abrégé: BMC Cancer
Pays: England
ID NLM: 100967800
Informations de publication
Date de publication:
22 Feb 2022
22 Feb 2022
Historique:
received:
15
09
2021
accepted:
16
02
2022
entrez:
23
2
2022
pubmed:
24
2
2022
medline:
11
3
2022
Statut:
epublish
Résumé
Acute promyelocytic leukemia (APL) is considered a hematologic emergency due to high risk of bleeding and fatal hemorrhages being a major cause of death. Despite lower death rates reported from clinical trials, patient registry data suggest an early death rate of 20%, especially for elderly and frail patients. Therefore, reliable diagnosis is required as treatment with differentiation-inducing agents leads to cure in the majority of patients. However, diagnosis commonly relies on cytomorphology and genetic confirmation of the pathognomonic t(15;17). Yet, the latter is more time consuming and in some regions unavailable. In recent years, deep learning (DL) has been evaluated for medical image recognition showing outstanding capabilities in analyzing large amounts of image data and provides reliable classification results. We developed a multi-stage DL platform that automatically reads images of bone marrow smears, accurately segments cells, and subsequently predicts APL using image data only. We retrospectively identified 51 APL patients from previous multicenter trials and compared them to 1048 non-APL acute myeloid leukemia (AML) patients and 236 healthy bone marrow donor samples, respectively. Our DL platform segments bone marrow cells with a mean average precision and a mean average recall of both 0.97. Further, it achieves high accuracy in detecting APL by distinguishing between APL and non-APL AML as well as APL and healthy donors with an area under the receiver operating characteristic of 0.8575 and 0.9585, respectively, using visual image data only. Our study underlines not only the feasibility of DL to detect distinct morphologies that accompany a cytogenetic aberration like t(15;17) in APL, but also shows the capability of DL to abstract information from a small medical data set, i. e. 51 APL patients, and infer correct predictions. This demonstrates the suitability of DL to assist in the diagnosis of rare cancer entities. As our DL platform predicts APL from bone marrow smear images alone, this may be used to diagnose APL in regions were molecular or cytogenetic subtyping is not routinely available and raise attention to suspected cases of APL for expert evaluation.
Sections du résumé
BACKGROUND
BACKGROUND
Acute promyelocytic leukemia (APL) is considered a hematologic emergency due to high risk of bleeding and fatal hemorrhages being a major cause of death. Despite lower death rates reported from clinical trials, patient registry data suggest an early death rate of 20%, especially for elderly and frail patients. Therefore, reliable diagnosis is required as treatment with differentiation-inducing agents leads to cure in the majority of patients. However, diagnosis commonly relies on cytomorphology and genetic confirmation of the pathognomonic t(15;17). Yet, the latter is more time consuming and in some regions unavailable.
METHODS
METHODS
In recent years, deep learning (DL) has been evaluated for medical image recognition showing outstanding capabilities in analyzing large amounts of image data and provides reliable classification results. We developed a multi-stage DL platform that automatically reads images of bone marrow smears, accurately segments cells, and subsequently predicts APL using image data only. We retrospectively identified 51 APL patients from previous multicenter trials and compared them to 1048 non-APL acute myeloid leukemia (AML) patients and 236 healthy bone marrow donor samples, respectively.
RESULTS
RESULTS
Our DL platform segments bone marrow cells with a mean average precision and a mean average recall of both 0.97. Further, it achieves high accuracy in detecting APL by distinguishing between APL and non-APL AML as well as APL and healthy donors with an area under the receiver operating characteristic of 0.8575 and 0.9585, respectively, using visual image data only.
CONCLUSIONS
CONCLUSIONS
Our study underlines not only the feasibility of DL to detect distinct morphologies that accompany a cytogenetic aberration like t(15;17) in APL, but also shows the capability of DL to abstract information from a small medical data set, i. e. 51 APL patients, and infer correct predictions. This demonstrates the suitability of DL to assist in the diagnosis of rare cancer entities. As our DL platform predicts APL from bone marrow smear images alone, this may be used to diagnose APL in regions were molecular or cytogenetic subtyping is not routinely available and raise attention to suspected cases of APL for expert evaluation.
Identifiants
pubmed: 35193533
doi: 10.1186/s12885-022-09307-8
pii: 10.1186/s12885-022-09307-8
pmc: PMC8864866
doi:
Types de publication
Evaluation Study
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
201Informations de copyright
© 2022. The Author(s).
Références
Cell. 1993 Aug 13;74(3):423-31
pubmed: 8394219
Blood. 2019 Apr 11;133(15):1630-1643
pubmed: 30803991
Acta Med Scand. 1957 Nov 29;159(3):189-94
pubmed: 13508085
Clin Chim Acta. 2021 Jan;512:1-6
pubmed: 33159948
Eur J Cancer. 2014 Apr;50(6):1159-68
pubmed: 24440088
Curr Hematol Malig Rep. 2020 Jun;15(3):203-210
pubmed: 32239350
Blood. 2017 Jan 26;129(4):424-447
pubmed: 27895058
Cancer Cell. 2010 Feb 17;17(2):173-85
pubmed: 20159609
Leuk Lymphoma. 2019 Dec;60(13):3107-3115
pubmed: 31842650
Br J Haematol. 1976 Aug;33(4):451-8
pubmed: 188440
Blood Rev. 2018 Mar;32(2):89-95
pubmed: 29033137
IEEE Trans Pattern Anal Mach Intell. 2017 Jun;39(6):1137-1149
pubmed: 27295650
Leukemia. 2022 Jan;36(1):111-118
pubmed: 34497326
Blood. 2011 Aug 4;118(5):1248-54
pubmed: 21653939
Ann Hematol. 2015 Apr;94(4):557-63
pubmed: 25366167
Best Pract Res Clin Haematol. 2003 Sep;16(3):357-67
pubmed: 12935956
N Engl J Med. 2005 Aug 4;353(5):498-507
pubmed: 16079373
J Clin Oncol. 2017 Feb 20;35(6):605-612
pubmed: 27400939
Blood. 1988 Aug;72(2):567-72
pubmed: 3165295
Cancers (Basel). 2019 Oct 18;11(10):
pubmed: 31635329
Nature. 1998 Feb 19;391(6669):815-8
pubmed: 9486655
Eur J Cancer Prev. 2006 Aug;15(4):367-70
pubmed: 16835508
Blood. 2016 May 19;127(20):2391-405
pubmed: 27069254
Transl Oncol. 2020 Jan;13(1):11-16
pubmed: 31733590
Carcinogenesis. 2019 Jul 4;40(5):651-660
pubmed: 30715157
Oncol Ther. 2019 Jun;7(1):33-65
pubmed: 32700196
Blood Adv. 2020 Dec 8;4(23):6077-6085
pubmed: 33290546
J Microbiol Methods. 2000 Dec 1;43(1):3-31
pubmed: 11084225
Biomed Opt Express. 2019 Dec 23;11(1):480-491
pubmed: 32010529
Blood. 1996 Aug 1;88(3):1052-61
pubmed: 8704214
Lab Invest. 2020 Jan;100(1):98-109
pubmed: 31570774
Oncotarget. 2020 Mar 17;11(11):992-1003
pubmed: 32215187
Nature. 1990 Oct 11;347(6293):558-61
pubmed: 2170850