A Clinical Risk Model for Personalized Screening and Prevention of Breast Cancer.
artificial intelligence
breast cancer
image-derived risk model
individualized screening
long-term risk
primary prevention
risk model
Journal
Cancers
ISSN: 2072-6694
Titre abrégé: Cancers (Basel)
Pays: Switzerland
ID NLM: 101526829
Informations de publication
Date de publication:
19 Jun 2023
19 Jun 2023
Historique:
received:
07
03
2023
revised:
10
06
2023
accepted:
18
06
2023
medline:
28
6
2023
pubmed:
28
6
2023
entrez:
28
6
2023
Statut:
epublish
Résumé
Image-derived artificial intelligence (AI) risk models have shown promise in identifying high-risk women in the short term. The long-term performance of image-derived risk models expanded with clinical factors has not been investigated. We performed a case-cohort study of 8110 women aged 40-74 randomly selected from a Swedish mammography screening cohort initiated in 2010 together with 1661 incident BCs diagnosed before January 2022. The imaging-only AI risk model extracted mammographic features and age at screening. Additional lifestyle/familial risk factors were incorporated into the lifestyle/familial-expanded AI model. Absolute risks were calculated using the two models and the clinical Tyrer-Cuzick v8 model. Age-adjusted model performances were compared across the 10-year follow-up. The AUCs of the lifestyle/familial-expanded AI risk model ranged from 0.75 (95%CI: 0.70-0.80) to 0.68 (95%CI: 0.66-0.69) 1-10 years after study entry. Corresponding AUCs were 0.72 (95%CI: 0.66-0.78) to 0.65 (95%CI: 0.63-0.66) for the imaging-only model and 0.62 (95%CI: 0.55-0.68) to 0.60 (95%CI: 0.58-0.61) for Tyrer-Cuzick v8. The increased performances were observed in multiple risk subgroups and cancer subtypes. Among the 5% of women at highest risk, the PPV was 5.8% using the lifestyle/familial-expanded model compared with 5.3% using the imaging-only model, The lifestyle/familial-expanded AI risk model showed higher performance for both long-term and short-term risk assessment compared with imaging-only and Tyrer-Cuzick models.
Sections du résumé
BACKGROUND
BACKGROUND
Image-derived artificial intelligence (AI) risk models have shown promise in identifying high-risk women in the short term. The long-term performance of image-derived risk models expanded with clinical factors has not been investigated.
METHODS
METHODS
We performed a case-cohort study of 8110 women aged 40-74 randomly selected from a Swedish mammography screening cohort initiated in 2010 together with 1661 incident BCs diagnosed before January 2022. The imaging-only AI risk model extracted mammographic features and age at screening. Additional lifestyle/familial risk factors were incorporated into the lifestyle/familial-expanded AI model. Absolute risks were calculated using the two models and the clinical Tyrer-Cuzick v8 model. Age-adjusted model performances were compared across the 10-year follow-up.
RESULTS
RESULTS
The AUCs of the lifestyle/familial-expanded AI risk model ranged from 0.75 (95%CI: 0.70-0.80) to 0.68 (95%CI: 0.66-0.69) 1-10 years after study entry. Corresponding AUCs were 0.72 (95%CI: 0.66-0.78) to 0.65 (95%CI: 0.63-0.66) for the imaging-only model and 0.62 (95%CI: 0.55-0.68) to 0.60 (95%CI: 0.58-0.61) for Tyrer-Cuzick v8. The increased performances were observed in multiple risk subgroups and cancer subtypes. Among the 5% of women at highest risk, the PPV was 5.8% using the lifestyle/familial-expanded model compared with 5.3% using the imaging-only model,
CONCLUSIONS
CONCLUSIONS
The lifestyle/familial-expanded AI risk model showed higher performance for both long-term and short-term risk assessment compared with imaging-only and Tyrer-Cuzick models.
Identifiants
pubmed: 37370856
pii: cancers15123246
doi: 10.3390/cancers15123246
pmc: PMC10296673
pii:
doi:
Types de publication
Journal Article
Langues
eng
Subventions
Organisme : CSRD VA
ID : 1
Pays : United States
Organisme : CSRD VA
ID : 1
Pays : United States
Références
J Natl Cancer Inst. 2018 Sep 1;110(9):994-1002
pubmed: 29490057
Breast Cancer Res Treat. 2020 Apr;180(2):481-490
pubmed: 32056055
Cochrane Database Syst Rev. 2019 Apr 29;4:CD012191
pubmed: 31032883
J Natl Cancer Inst. 2005 Nov 16;97(22):1652-62
pubmed: 16288118
Breast Cancer Res Treat. 2018 Jun;169(2):371-379
pubmed: 29392583
Stat Med. 2000 May 15;19(9):1141-64
pubmed: 10797513
R Soc Open Sci. 2017 Feb 1;4(2):160254
pubmed: 28386409
Ann Intern Med. 2016 Feb 16;164(4):279-96
pubmed: 26757170
J Med Genet. 2022 Dec;59(12):1196-1205
pubmed: 36162852
Sci Transl Med. 2021 Jan 27;13(578):
pubmed: 33504648
CA Cancer J Clin. 2019 May;69(3):184-210
pubmed: 30875085
Breast Cancer Res Treat. 2022 Apr;192(2):375-383
pubmed: 34994879
J Am Coll Radiol. 2017 Sep;14(9):1137-1143
pubmed: 28648873
AJR Am J Roentgenol. 2021 Jul;217(1):48-55
pubmed: 33978450
J Clin Epidemiol. 1999 Dec;52(12):1165-72
pubmed: 10580779
Lancet Oncol. 2015 Jan;16(1):67-75
pubmed: 25497694
Int J Equity Health. 2015 Dec 30;14:157
pubmed: 26715453
Stat Med. 2004 Apr 15;23(7):1111-30
pubmed: 15057881
Radiology. 2020 Nov;297(2):327-333
pubmed: 32897160
Sci Transl Med. 2022 May 11;14(644):eabn3971
pubmed: 35544593
J Natl Compr Canc Netw. 2015 Jul;13(7):880-915
pubmed: 26150582
BMJ. 1994 Jul 9;309(6947):102
pubmed: 8038641
Biometrika. 2009 Jun;96(2):371-382
pubmed: 22822245
Br J Cancer. 2019 Jul;121(1):76-85
pubmed: 31114019
Breast Cancer Res. 2017 Mar 14;19(1):29
pubmed: 28288659
Nature. 2020 Jan;577(7788):89-94
pubmed: 31894144
IEEE Trans Vis Comput Graph. 2007 Nov-Dec;13(6):1121-8
pubmed: 17968055
Nat Genet. 2010 Jul;42(7):558-60
pubmed: 20581876
Int J Epidemiol. 2017 Dec 1;46(6):1740-1741g
pubmed: 28180256
CA Cancer J Clin. 2017 Mar;67(2):93-99
pubmed: 28094848
Nat Rev Cancer. 2020 Aug;20(8):417-436
pubmed: 32528185
J Natl Cancer Inst. 2007 Feb 21;99(4):283-90
pubmed: 17312305
PLoS One. 2014 Jun 27;9(6):e101176
pubmed: 24972092