TY - GEN
T1 - Opportunistic Hip Fracture Risk Prediction in Men from X-ray
T2 - 5th International Workshop on Predictive Intelligence in Medicine, PRIME 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
AU - Schmarje, Lars
AU - Reinhold, Stefan
AU - Damm, Timo
AU - Orwoll, Eric
AU - Glüer, Claus C.
AU - Koch, Reinhard
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Osteoporosis is a common disease that increases fracture risk. Hip fractures, especially in elderly people, lead to increased morbidity, decreased quality of life and increased mortality. Being a silent disease before fracture, osteoporosis often remains undiagnosed and untreated. Areal bone mineral density (aBMD) assessed by dual-energy X-ray absorptiometry (DXA) is the gold-standard method for osteoporosis diagnosis and hence also for future fracture prediction (prognostic). However, the required special equipment is not broadly available everywhere, in particular not to patients in developing countries. We propose a deep learning classification model (FORM) that can directly predict hip fracture risk from either plain radiographs (X-ray) or 2D projection images of computed tomography (CT) data. Our method is fully automated and therefore well suited for opportunistic screening settings, identifying high risk patients in a broader population without additional screening. FORM was trained and evaluated on X-rays and CT projections from the Osteoporosis in Men (MrOS) study. 3108 X-rays (89 incident hip fractures) or 2150 CTs (80 incident hip fractures) with a 80/20 split (training/validation) were used. We show that FORM can correctly predict the 10-year hip fracture risk with a validation AUC of 81.44% ± 3.11%/81.04% ± 5.54% (mean ± STD) including additional information like age, BMI, fall history and health background across a 5-fold cross validation on the X-ray and CT cohort, respectively. Our approach significantly (p < 0.01) outperforms previous methods like Cox Proportional-Hazards Model and with 70.19 ± 6.58 and 74.72 ± 7.21 respectively on the X-ray cohort. Our model outperform on both cohorts hip aBMD based predictions (validation AUC 82.67% ± 0.21% vs. 71.82% ± 0.50% and 78.41% ± 0.33 vs. 76.55% ± 0.89%). We are confident that FORM can contribute on improving osteoporosis diagnosis at an early stage.
AB - Osteoporosis is a common disease that increases fracture risk. Hip fractures, especially in elderly people, lead to increased morbidity, decreased quality of life and increased mortality. Being a silent disease before fracture, osteoporosis often remains undiagnosed and untreated. Areal bone mineral density (aBMD) assessed by dual-energy X-ray absorptiometry (DXA) is the gold-standard method for osteoporosis diagnosis and hence also for future fracture prediction (prognostic). However, the required special equipment is not broadly available everywhere, in particular not to patients in developing countries. We propose a deep learning classification model (FORM) that can directly predict hip fracture risk from either plain radiographs (X-ray) or 2D projection images of computed tomography (CT) data. Our method is fully automated and therefore well suited for opportunistic screening settings, identifying high risk patients in a broader population without additional screening. FORM was trained and evaluated on X-rays and CT projections from the Osteoporosis in Men (MrOS) study. 3108 X-rays (89 incident hip fractures) or 2150 CTs (80 incident hip fractures) with a 80/20 split (training/validation) were used. We show that FORM can correctly predict the 10-year hip fracture risk with a validation AUC of 81.44% ± 3.11%/81.04% ± 5.54% (mean ± STD) including additional information like age, BMI, fall history and health background across a 5-fold cross validation on the X-ray and CT cohort, respectively. Our approach significantly (p < 0.01) outperforms previous methods like Cox Proportional-Hazards Model and with 70.19 ± 6.58 and 74.72 ± 7.21 respectively on the X-ray cohort. Our model outperform on both cohorts hip aBMD based predictions (validation AUC 82.67% ± 0.21% vs. 71.82% ± 0.50% and 78.41% ± 0.33 vs. 76.55% ± 0.89%). We are confident that FORM can contribute on improving osteoporosis diagnosis at an early stage.
KW - fracture risk prediction
KW - opportunistic screening
KW - osteoporosis
UR - http://www.scopus.com/inward/record.url?scp=85140447674&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85140447674&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-16919-9_10
DO - 10.1007/978-3-031-16919-9_10
M3 - Conference contribution
AN - SCOPUS:85140447674
SN - 9783031169182
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 103
EP - 114
BT - Predictive Intelligence in Medicine - 5th International Workshop, PRIME 2022, Held in Conjunction with MICCAI 2022, Proceedings
A2 - Rekik, Islem
A2 - Adeli, Ehsan
A2 - Park, Sang Hyun
A2 - Cintas, Celia
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 22 September 2022 through 22 September 2022
ER -