Accurate diagnosis of colorectal cancer based on histopathology images using artificial intelligence.
Cancer diagnosis
Colorectal cancer
Deep learning
Histopathology image
Journal
BMC medicine
ISSN: 1741-7015
Titre abrégé: BMC Med
Pays: England
ID NLM: 101190723
Informations de publication
Date de publication:
23 03 2021
23 03 2021
Historique:
received:
22
10
2020
accepted:
16
02
2021
entrez:
23
3
2021
pubmed:
24
3
2021
medline:
16
10
2021
Statut:
epublish
Résumé
Accurate and robust pathological image analysis for colorectal cancer (CRC) diagnosis is time-consuming and knowledge-intensive, but is essential for CRC patients' treatment. The current heavy workload of pathologists in clinics/hospitals may easily lead to unconscious misdiagnosis of CRC based on daily image analyses. Based on a state-of-the-art transfer-learned deep convolutional neural network in artificial intelligence (AI), we proposed a novel patch aggregation strategy for clinic CRC diagnosis using weakly labeled pathological whole-slide image (WSI) patches. This approach was trained and validated using an unprecedented and enormously large number of 170,099 patches, > 14,680 WSIs, from > 9631 subjects that covered diverse and representative clinical cases from multi-independent-sources across China, the USA, and Germany. Our innovative AI tool consistently and nearly perfectly agreed with (average Kappa statistic 0.896) and even often better than most of the experienced expert pathologists when tested in diagnosing CRC WSIs from multicenters. The average area under the receiver operating characteristics curve (AUC) of AI was greater than that of the pathologists (0.988 vs 0.970) and achieved the best performance among the application of other AI methods to CRC diagnosis. Our AI-generated heatmap highlights the image regions of cancer tissue/cells. This first-ever generalizable AI system can handle large amounts of WSIs consistently and robustly without potential bias due to fatigue commonly experienced by clinical pathologists. It will drastically alleviate the heavy clinical burden of daily pathology diagnosis and improve the treatment for CRC patients. This tool is generalizable to other cancer diagnosis based on image recognition.
Sections du résumé
BACKGROUND
Accurate and robust pathological image analysis for colorectal cancer (CRC) diagnosis is time-consuming and knowledge-intensive, but is essential for CRC patients' treatment. The current heavy workload of pathologists in clinics/hospitals may easily lead to unconscious misdiagnosis of CRC based on daily image analyses.
METHODS
Based on a state-of-the-art transfer-learned deep convolutional neural network in artificial intelligence (AI), we proposed a novel patch aggregation strategy for clinic CRC diagnosis using weakly labeled pathological whole-slide image (WSI) patches. This approach was trained and validated using an unprecedented and enormously large number of 170,099 patches, > 14,680 WSIs, from > 9631 subjects that covered diverse and representative clinical cases from multi-independent-sources across China, the USA, and Germany.
RESULTS
Our innovative AI tool consistently and nearly perfectly agreed with (average Kappa statistic 0.896) and even often better than most of the experienced expert pathologists when tested in diagnosing CRC WSIs from multicenters. The average area under the receiver operating characteristics curve (AUC) of AI was greater than that of the pathologists (0.988 vs 0.970) and achieved the best performance among the application of other AI methods to CRC diagnosis. Our AI-generated heatmap highlights the image regions of cancer tissue/cells.
CONCLUSIONS
This first-ever generalizable AI system can handle large amounts of WSIs consistently and robustly without potential bias due to fatigue commonly experienced by clinical pathologists. It will drastically alleviate the heavy clinical burden of daily pathology diagnosis and improve the treatment for CRC patients. This tool is generalizable to other cancer diagnosis based on image recognition.
Identifiants
pubmed: 33752648
doi: 10.1186/s12916-021-01942-5
pii: 10.1186/s12916-021-01942-5
pmc: PMC7986569
doi:
Types de publication
Journal Article
Multicenter Study
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
76Subventions
Organisme : NIDDK NIH HHS
ID : R01 DK115679
Pays : United States
Organisme : NIAMS NIH HHS
ID : R01 AR069055
Pays : United States
Organisme : NIMH NIH HHS
ID : R01 MH104680
Pays : United States
Organisme : NIAMS NIH HHS
ID : R01 AR059781
Pays : United States
Organisme : NIA NIH HHS
ID : U19 AG055373
Pays : United States
Organisme : NIMHD NIH HHS
ID : U54 MD007595
Pays : United States
Organisme : NIMH NIH HHS
ID : R01 MH107354
Pays : United States
Organisme : NIGMS NIH HHS
ID : R01 GM109068
Pays : United States
Références
N Engl J Med. 2016 Sep 22;375(12):1109-12
pubmed: 27653561
Nature. 2017 Feb 2;542(7639):115-118
pubmed: 28117445
JAMA. 2017 Dec 12;318(22):2199-2210
pubmed: 29234806
Proc Natl Acad Sci U S A. 2018 Mar 27;115(13):E2970-E2979
pubmed: 29531073
PeerJ. 2017 Oct 3;5:e3874
pubmed: 29018612
Gut. 2017 Apr;66(4):683-691
pubmed: 26818619
IEEE Trans Med Imaging. 2019 May;38(5):1139-1149
pubmed: 30403624
Med Image Anal. 2017 Dec;42:60-88
pubmed: 28778026
J Gastrointest Oncol. 2012 Sep;3(3):153-73
pubmed: 22943008
CA Cancer J Clin. 2018 Nov;68(6):394-424
pubmed: 30207593
IEEE Trans Med Imaging. 2017 Apr;36(4):994-1004
pubmed: 28026754
Med Image Anal. 2015 Feb;20(1):237-48
pubmed: 25547073
Nat Med. 2018 Oct;24(10):1559-1567
pubmed: 30224757
Lancet Oncol. 2020 Feb;21(2):222-232
pubmed: 31926806
JAMA. 2016 Dec 13;316(22):2402-2410
pubmed: 27898976
Anal Cell Pathol (Amst). 2017;2017:8428102
pubmed: 28331793
IEEE/ACM Trans Comput Biol Bioinform. 2020 Jan-Feb;17(1):91-101
pubmed: 30040652
J Pathol Inform. 2017 Feb 28;8:1
pubmed: 28400990
Nat Med. 2019 Aug;25(8):1301-1309
pubmed: 31308507
Lancet. 2020 Feb 1;395(10221):350-360
pubmed: 32007170
Acad Pathol. 2016 Sep 28;3:2374289516665393
pubmed: 28725776
EBioMedicine. 2018 Jan;27:317-328
pubmed: 29292031
Onco Targets Ther. 2015 Aug 04;8:2015-22
pubmed: 26346558
IEEE J Biomed Health Inform. 2020 May;24(5):1413-1426
pubmed: 31567104
Neuroimage. 2006 Nov 1;33(2):599-608
pubmed: 16952467
JAMA Netw Open. 2019 May 3;2(5):e194337
pubmed: 31150073
J Glob Oncol. 2015 Sep 23;1(1):3-6
pubmed: 28804765
Sci Rep. 2018 Feb 21;8(1):3395
pubmed: 29467373
Cancer Discov. 2020 Mar;10(3):OF1
pubmed: 31953243
Lancet Oncol. 2020 Feb;21(2):233-241
pubmed: 31926805
J Med Ethics. 2005 Jun;31(6):333-7
pubmed: 15923480
IEEE Trans Med Imaging. 2016 May;35(5):1196-1206
pubmed: 26863654
BMC Bioinformatics. 2017 May 26;18(1):281
pubmed: 28549410
Comput Struct Biotechnol J. 2018 Feb 09;16:34-42
pubmed: 30275936
PLoS Med. 2019 Jan 24;16(1):e1002730
pubmed: 30677016
PLoS One. 2017 Jun 1;12(6):e0177544
pubmed: 28570557
Cell. 2018 Feb 22;172(5):1122-1131.e9
pubmed: 29474911