TY - JOUR
T1 - Can deep learning on retinal images augment known risk factors for cardiovascular disease prediction in diabetes?
T2 - A prospective cohort study from the national screening programme in Scotland
AU - Mellor, Joseph
AU - Jiang, Wenhua
AU - Fleming, Alan
AU - McGurnaghan, Stuart J
AU - Blackbourn, Luke
AU - Styles, Caroline
AU - Storkey, Amos J
AU - McKeigue, Paul M
AU - Colhoun, Helen M
AU - Scottish Diabetes Research Network Epidemiology Group
N1 - Funding Information:
We thank the Scottish Diabetes Research Network for the role in data generation. This work was supported by JDRF [grant 2-SRA-2019-857-S-B ].
Publisher Copyright:
© 2023 The Authors
PY - 2023/4/18
Y1 - 2023/4/18
N2 - AIMS: This study's objective was to evaluate whether deep learning (DL) on retinal photographs from a diabetic retinopathy screening programme improve prediction of incident cardiovascular disease (CVD).METHODS: DL models were trained to jointly predict future CVD risk and CVD risk factors and used to output a DL score. Poisson regression models including clinical risk factors with and without a DL score were fitted to study cohorts with 2,072 and 38,730 incident CVD events in type 1 (T1DM) and type 2 diabetes (T2DM) respectively.RESULTS: DL scores were independently associated with incident CVD with adjusted standardised incidence rate ratios of 1.14 (P = 3 × 10-04 95 % CI (1.06, 1.23)) and 1.16 (P = 4 × 10-33 95 % CI (1.13, 1.18)) in T1DM and T2DM cohorts respectively. The differences in predictive performance between models with and without a DL score were statistically significant (differences in test log-likelihood 6.7 and 51.1 natural log units) but the increments in C-statistics from 0.820 to 0.822 and from 0.709 to 0.711 for T1DM and T2DM respectively, were small.CONCLUSIONS: These results show that in people with diabetes, retinal photographs contain information on future CVD risk. However for this to contribute appreciably to clinical prediction of CVD further approaches, including exploitation of serial images, need to be evaluated.
AB - AIMS: This study's objective was to evaluate whether deep learning (DL) on retinal photographs from a diabetic retinopathy screening programme improve prediction of incident cardiovascular disease (CVD).METHODS: DL models were trained to jointly predict future CVD risk and CVD risk factors and used to output a DL score. Poisson regression models including clinical risk factors with and without a DL score were fitted to study cohorts with 2,072 and 38,730 incident CVD events in type 1 (T1DM) and type 2 diabetes (T2DM) respectively.RESULTS: DL scores were independently associated with incident CVD with adjusted standardised incidence rate ratios of 1.14 (P = 3 × 10-04 95 % CI (1.06, 1.23)) and 1.16 (P = 4 × 10-33 95 % CI (1.13, 1.18)) in T1DM and T2DM cohorts respectively. The differences in predictive performance between models with and without a DL score were statistically significant (differences in test log-likelihood 6.7 and 51.1 natural log units) but the increments in C-statistics from 0.820 to 0.822 and from 0.709 to 0.711 for T1DM and T2DM respectively, were small.CONCLUSIONS: These results show that in people with diabetes, retinal photographs contain information on future CVD risk. However for this to contribute appreciably to clinical prediction of CVD further approaches, including exploitation of serial images, need to be evaluated.
KW - Humans
KW - Diabetes Mellitus, Type 2/diagnosis
KW - Diabetes Mellitus, Type 1/complications
KW - Prospective Studies
KW - Deep Learning
KW - Cardiovascular Diseases/diagnosis
KW - Risk Factors
KW - Scotland/epidemiology
KW - Heart Disease Risk Factors
U2 - 10.1016/j.ijmedinf.2023.105072
DO - 10.1016/j.ijmedinf.2023.105072
M3 - Article
C2 - 37167840
SN - 1386-5056
VL - 175
SP - 105072
JO - International journal of medical informatics
JF - International journal of medical informatics
M1 - 105072
ER -