Please wait a minute...
Shopping Cart Email List Chinese
Advanced Search Fig/Tab
Frontiers of Medicine

ISSN 2095-0217

ISSN 2095-0225(Online)

CN 11-5983/R

Postal Subscription Code 80-967

2018 Impact Factor: 1.847

Featured Articles  |  Most Downloaded  |  Most Reading
Online Submission Subscribe to Our RSS Content Feed
Front. Med.    2022, Vol. 16 Issue (3) : 496-506    https://doi.org/10.1007/s11684-021-0828-7
RESEARCH ARTICLE
Hybrid deep learning model for risk prediction of fracture in patients with diabetes and osteoporosis
Yaxin Chen1,2, Tianyi Yang2, Xiaofeng Gao2(), Ajing Xu1,3()
1. Department of Pharmacy, Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai 200240, China
2. Shanghai Key Laboratory of Scalable Computing and Systems, Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
3. Clinical Research Unit, Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai 200240, China
 Download: PDF(702 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

The fracture risk of patients with diabetes is higher than those of patients without diabetes due to hyperglycemia, usage of diabetes drugs, changes in insulin levels, and excretion, and this risk begins as early as adolescence. Many factors including demographic data (such as age, height, weight, and gender), medical history (such as smoking, drinking, and menopause), and examination (such as bone mineral density, blood routine, and urine routine) may be related to bone metabolism in patients with diabetes. However, most of the existing methods are qualitative assessments and do not consider the interactions of the physiological factors of humans. In addition, the fracture risk of patients with diabetes and osteoporosis has not been further studied previously. In this paper, a hybrid model combining XGBoost with deep neural network is used to predict the fracture risk of patients with diabetes and osteoporosis, and investigate the effect of patients’ physiological factors on fracture risk. A total of 147 raw input features are considered in our model. The presented model is compared with several benchmarks based on various metrics to prove its effectiveness. Moreover, the top 18 influencing factors of fracture risks of patients with diabetes are determined.
Keywords XGBoost      deep neural network      healthcare      risk prediction     
Corresponding Author(s): Xiaofeng Gao,Ajing Xu   
Just Accepted Date: 12 July 2021   Online First Date: 26 August 2021    Issue Date: 18 July 2022
 Cite this article:   
Yaxin Chen,Tianyi Yang,Xiaofeng Gao, et al. Hybrid deep learning model for risk prediction of fracture in patients with diabetes and osteoporosis[J]. Front. Med., 2022, 16(3): 496-506.
 URL:  
https://academic.hep.com.cn/fmd/EN/10.1007/s11684-021-0828-7
https://academic.hep.com.cn/fmd/EN/Y2022/V16/I3/496
Fig.1  Overview of the model.
  
Variables Fracture
No (N=1480) Yes (N=123)
Gender
Male, n 902 98
Female, n 578 25
Smoking History
Yes, n 278 18
No, n 1202 105
Vaccination
Yes, n 446 74
No, n 1034 49
Chemotherapy
Yes, n 65 6
No, n 1415 117
Age, year, median (IQR) 66.14 (21–99) 72.70 (29–94)
Weight, kg, median (IQR) 65.99 (31–120) 63.76 (30–108)
Height, cm, median (IQR) 163.11 (139–193) 159.08 (143–185)
Creatinine, umol/L, median (IQR) 68.57 (16–1379) 78.84 (26–1177)
Ca (in 24-hour urine), mmol/L, median (IQR) 198.96 (0.92–785.17) 164.94 (22.44–328.26)
Urea nitrogen, mmol/L, median (IQR) 6.09 (1.67–33.17) 6.45 (2.14–26.42)
Alkaline phosphatase, U/L, median (IQR) 83.49 (17–511) 112.28 (27–1779)
C-reactive protein, mg/L, median (IQR) 5.62 (0.01–162.46) 5.34 (0.1–134.47)
Aspartate aminotransferase, U/L, median (IQR) 14.28 (3.5–302.4) 12.61 (5.5–33.8)
Albumin (ALB), g/L, median (IQR) 39.07 (17.2–56.3) (39.91 (22.6–54.4)
Glycated hemoglobin, %, median (IQR) 8.47 (4.5–17.7) 8.28 (4.8–15.8)
Prealbumin (PA), mg/L, median (IQR) 237.28 (2–460) 220.09 (55–471)
Apolipoprotein B, g/L, median (IQR) 0.97 (0.10–2.53) 0.94 (0.29–2.12)
DXA value of L3–L4, median (IQR) -0.12 (−6.08–9.28) -1.06 (-4.51–5.11)
DXA value of L2–L3, median (IQR) -0.38 (-5.54–9.39) -1.23 (-4.43–5.77)
Glucose, mmol/L, median (IQR) 7.84 (1.8–22.81) 7.25 (3.66–20.145)
Insulin, pmol/L, median (IQR) 113.43 (4.19–1294.56) 86.67 (9.45–391.17)
C-peptide, μmol/L, median (IQR) 3.57 (0–29.31) 4.34 (0.01–28.76)
Total cholesterol, mmol/L, median(IQR) 4.65 (1.64–13.14) 4.46 (1.97–8)
Triglyceride, mmol/L, median (IQR) 1.9 (0.41–18.38) 1.74 (0.48–4.39)
Microalbumin, mg/L, median (IQR) 255.48 (6.68–11 573.33) 303.15 (9.907–6760)
Ratio of microalbumin to creatinine, mg/g, median (IQR) 214.83 (0.3–10 444.04) 118.39 (0.89–917.64)
Retinol binding protein, mg/L, median (IQR) 43.30 (0–140.7) 39.66 (11–92.6)
Total protein, g/L, median (IQR) 65.56 (34.6–113.1) 66.26 (47.6–78.9)
Basophilic granulocytes count, 109/L, median (IQR) 0.012 (0–0.15) 0.008 (0–0.10)
Tab.1  Demographics of the major explanatory variables in 1603 patients with diabetes combined with osteoporosis [33,34]
Size of hidden layers L2 regularization strength Train steps
[30,10] 100 800
Tab.2  Parameters of neural network
Method Parameters
LR L2 regularization strength= 1.0
SVM Rbf Gaussian kernel, C= 1.0, g = 0.01
DT Max depth= 20, min sample leaf= 10, min sample split= 10
KNN Number of neighbors= 3
RF Number of trees= 25
ERT Max depth= 8, max number of features= 50
GBDT Max number of trees= 500
AdaBoost Max number of trees= 200
CatBoost Depth= 6, L2 regularization strength= 5.0
XGBoost Max number of trees= 400
MLP Hidden layer size= [300, 300, 30], L2 Regularization strength= 2
Tab.3  Parameters of benchmarks
Method Accuracy Precision Recall F1-score
LR 67.76 70.41 60.81 64.94
SVM 83.54 84.90 83.57 83.39
DT 73.93 78.01 66.47 71.44
KNN 78.23 78.23 78.23 78.23
RF 80.51 84.19 80.55 79.98
ERT 83.04 83.05 83.04 83.04
GBDT 84.81 86.79 84.94 84.61
AdaBoost 82.28 83.14 82.30 82.17
CatBoost 83.80 86.10 83.83 83.54
XGBoost 86.08 87.69 86.10 85.93
MLP 82.78 84.18 82.81 82.62
XGBoost+ MLP 90.38 90.52 90.39 90.37
Tab.4  Performance comparison between benchmarks and our hybrid model (XGBoost+ MLP) on test set (%)
Fig.2  (A) Trend of AUC versus number of trees. (B) Trend of AUC versus sample size.
Fig.3  18 influencing factors of fracture risks of patients with diabetes.
1 G Leidig-Bruckner, S Grobholz, T Bruckner, C Scheidt-Nave, P Nawroth, JG Schneider. Prevalence and determinants of osteoporosis in patients with type 1 and type 2 diabetes mellitus. BMC Endocr Disord 2014; 14(1): 33
https://doi.org/10.1186/1472-6823-14-33 pmid: 24721668
2 S Epstein, G Defeudis, S Manfrini, N Napoli, P Pozzilli; Scientific Committee of the First International Symposium on Diabetes and Bone.Diabetes and disordered bone metabolism (diabetic osteodystrophy): time for recognition. Osteoporos Int 2016; 27(6): 1931–1951
https://doi.org/10.1007/s00198-015-3454-x pmid: 26980458
3 Z Zhao. Correlation analysis of urine proteins and inflammatory cytokines with osteoporosis in patients with diabetic nephropathy. J Musculoskelet Neuronal Interact 2018; 18(3): 348–353
pmid: 30179212
4 NB Khazai, GR Beck Jr, GE Umpierrez. Diabetes and fractures: an overshadowed association. Curr Opin Endocrinol Diabetes Obes 2009; 16(6): 435–445
https://doi.org/10.1097/MED.0b013e328331c7eb pmid: 19779334
5 JL Pecina, L Romanovsky, SP Merry, KA Kennel, TD Thacher. Comparison of clinical risk tools for predicting osteoporosis in women ages 50–64. J Am Board Fam Med 2016; 29(2): 233–239
https://doi.org/10.3122/jabfm.2016.02.150237 pmid: 26957380
6 JM Liu, DL Zhu, YM Mu, WB Xia; Chinese Society of Osteoporosis and Bone Mineral Research, the Chinese Society of Endocrinology, Chinese Diabetes Society, Chinese Medical Association; Chinese Endocrinologist Association, Chinese Medical Doctor Association.Management of fracture risk in patients with diabetes—Chinese Expert Consensus. J Diabetes 2019; 11(11): 906–919
https://doi.org/10.1111/1753-0407.12962 pmid: 31219236
7 S Roux, F Cabana, N Carrier, M Beaulieu, PM April, MC Beaulieu, G Boire. The World Health Organization Fracture Risk Assessment Tool (FRAX) underestimates incident and recurrent fractures in consecutive patients with fragility fractures. J Clin Endocrinol Metab 2014; 99(7): 2400–2408
https://doi.org/10.1210/jc.2013-4507 pmid: 24780062
8 A Esteva, B Kuprel, RA Novoa, J Ko, SM Swetter, HM Blau, S Thrun. Dermatologist-level classification of skin cancer with deep neural networks. Nature 2017; 542(7639): 115–118
https://doi.org/10.1038/nature21056 pmid: 28117445
9 HA Haenssle, C Fink, R Schneiderbauer, F Toberer, T Buhl, A Blum, A Kalloo, ABH Hassen, L Thomas, A Enk, L Uhlmann, Reader study level-I and level-II Groups; C Alt, M Arenbergerova, R Bakos, A Baltzer, I Bertlich, A Blum, T Bokor-Billmann, J Bowling, N Braghiroli, R Braun, K Buder-Bakhaya, T Buhl, H Cabo, L Cabrijan, N Cevic, A Classen, D Deltgen, C Fink, I Georgieva, LE Hakim-Meibodi, S Hanner, F Hartmann, J Hartmann, G Haus, E Hoxha, R Karls, H Koga, J Kreusch, A Lallas, P Majenka, A Marghoob, C Massone, L Mekokishvili, D Mestel, V Meyer, A Neuberger, K Nielsen, M Oliviero, R Pampena, J Paoli, E Pawlik, B Rao, A Rendon, T Russo, A Sadek, K Samhaber, R Schneiderbauer, A Schweizer, F Toberer, L Trennheuser, L Vlahova, A Wald, J Winkler, P Wölbing, I Zalaudek. Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists. Ann Oncol 2018; 29(8): 1836–1842
https://doi.org/10.1093/annonc/mdy166 pmid: 29846502
10 JZ Cheng, D Ni, YH Chou, J Qin, CM Tiu, YC Chang, CS Huang, D Shen, CM Chen. Computer-aided diagnosis with deep learning architecture: applications to breast lesions in US images and pulmonary nodules in CT scans. Sci Rep 2016; 6(1): 24454
https://doi.org/10.1038/srep24454 pmid: 27079888
11 M Cicero, A Bilbily, E Colak, T Dowdell, B Gray, K Perampaladas, J Barfett. Training and validating a deep convolutional neural network for computer-aided detection and classification of abnormalities on frontal chest radiographs. Invest Radiol 2017; 52(5): 281–287
https://doi.org/10.1097/RLI.0000000000000341 pmid: 27922974
12 T Kooi, G Litjens, B van Ginneken, A Gubern-Mérida, CI Sánchez, R Mann, A den Heeten, N Karssemeijer. Large scale deep learning for computer aided detection of mammographic lesions. Med Image Anal 2017; 35: 303–312
https://doi.org/10.1016/j.media.2016.07.007 pmid: 27497072
13 CM Barreira, M Bouslama, DC Haussen, JA Grossberg, B Baxter, T Devlin, M Frankel, RG Nogueira. Abstract WP61: automated large artery occlusion detection IN stroke imaging-ALADIN study. Stroke 2018; 49(Suppl_1): AWP61
14 V Gulshan, L Peng, M Coram, MC Stumpe, D Wu, A Narayanaswamy, S Venugopalan, K Widner, T Madams, J Cuadros, R Kim, R Raman, PC Nelson, JL Mega, DR Webster. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA 2016; 316(22): 2402–2410
https://doi.org/10.1001/jama.2016.17216 pmid: 27898976
15 R Poplin, AV Varadarajan, K Blumer, Y Liu, MV McConnell, GS Corrado, L Peng, DR Webster. Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning. Nat Biomed Eng 2018; 2(3): 158–164
https://doi.org/10.1038/s41551-018-0195-0 pmid: 31015713
16 J De Fauw, JR Ledsam, B Romera-Paredes, S Nikolov, N Tomasev, S Blackwell, H Askham, X Glorot, B O’Donoghue, D Visentin, G van den Driessche, B Lakshminarayanan, C Meyer, F Mackinder, S Bouton, K Ayoub, R Chopra, D King, A Karthikesalingam, CO Hughes, R Raine, J Hughes, DA Sim, C Egan, A Tufail, H Montgomery, D Hassabis, G Rees, T Back, PT Khaw, M Suleyman, J Cornebise, PA Keane, O Ronneberger. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat Med 2018; 24(9): 1342–1350
https://doi.org/10.1038/s41591-018-0107-6 pmid: 30104768
17 DC Cireşan, A Giusti, LM Gambardella, J Schmidhuber. Mitosis detection in breast cancer histology images with deep neural networks. International Conference on Medical Image Computing and Computer-assisted Intervention 2013: 411–418
18 Y Liu, K Gadepalli, M Norouzi, GE Dahl, T Kohlberger, A Boyko, S Venugopalan, A Timofeev, PQ Nelson, GS Corrado. Detecting cancer metastases on gigapixel pathology images. 2017. arXiv:1703.02442
19 P Charoentong, F Finotello, M Angelova, C Mayer, M Efremova, D Rieder, H Hackl, Z Trajanoski. Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade. Cell Rep 2017; 18(1): 248–262
https://doi.org/10.1016/j.celrep.2016.12.019 pmid: 28052254
20 AH Beck, AR Sangoi, S Leung, RJ Marinelli, TO Nielsen, MJ Van De Vijver, RB West, M Van De Rijn, D Koller. Systematic analysis of breast cancer morphology uncovers stromal features associated with survival. Sci Transl Med 2011; 3(108): 108ra113 PMID: 22072638
https://doi.org/DOI: 10.1126/scitranslmed.3002564
21 B Shickel, PJ Tighe, A Bihorac, P Rashidi. Deep EHR: a survey of recent advances in deep learning techniques for electronic health record (EHR) analysis. IEEE J Biomed Health Inform 2018; 22(5): 1589–1604
https://doi.org/10.1109/JBHI.2017.2767063 pmid: 29989977
22 A Rajkomar, E Oren, K Chen, AM Dai, N Hajaj, M Hardt, PJ Liu, X Liu, J Marcus, M Sun, P Sundberg, H Yee, K Zhang, Y Zhang, G Flores, GE Duggan, J Irvine, Q Le, K Litsch, A Mossin, J Tansuwan, D Wang, J Wexler, J Wilson, D Ludwig, SL Volchenboum, K Chou, M Pearson, S Madabushi, NH Shah, AJ Butte, MD Howell, C Cui, GS Corrado, J Dean. Scalable and accurate deep learning with electronic health records. NPJ Digit Med 2018; 1(1): 18
https://doi.org/10.1038/s41746-018-0029-1 pmid: 31304302
23 V Liu, P Kipnis, MK Gould, GJ Escobar. Length of stay predictions: improvements through the use of automated laboratory and comorbidity variables. Med Care 2010; 48(8): 739–744
https://doi.org/10.1097/MLR.0b013e3181e359f3 pmid: 20613662
24 Ed Choi, MT Bahadori, A Schuetz, WF Stewart, J Sun. Doctor AI: predicting clinical events via recurrent neural networks. Machine Learning for Healthcare Conference 2016: 301–318
25 Z Che, S Purushotham, K Cho, D Sontag, Y Liu. Recurrent neural networks for multivariate time series with missing values. Sci Rep 2018; 8(1): 6085
https://doi.org/10.1038/s41598-018-24271-9 pmid: 29666385
26 H Suresh, N Hunt, A Johnson, LA Celi, P Szolovits, M Ghassemi. Clinical intervention prediction and understanding with deep neural networks. Machine Learning for Healthcare Conference 2017: 322–337
27 P Abbeel, AY Ng. Apprenticeship learning via inverse reinforcement learning. Proceedings of the 21st International Conference on Machine Learning 2004: 1
28 ND Ratliff, D Silver, JA Bagnell. Learning to search: Functional gradient techniques for imitation learning. Auton Robots 2009; 27(1): 25–53
https://doi.org/10.1007/s10514-009-9121-3
29 J Schulman, A Gupta, S Venkatesan, M Tayson-Frederick, P Abbeel. A case study of trajectory transfer through non-rigid registration for a simplified suturing scenario. IEEE/RSJ International Conference on Intelligent Robots and Systems 2013: 4111–4117
30 X He, J Pan, O Jin, T Xu, B Liu, T Xu, Y Shi, A Atallah, R Herbrich, S Bowers, JQ Candela. Practical lessons from predicting clicks on Ads at Facebook. Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 2014: 1–9
31 T Chen, C Guestrin. XGBoost: a scalable tree boosting system. KDD '16: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 2016: 785–794
32 DP Kingma, J Ba. Adam: a method for stochastic optimization. 2014. arXiv:1412.6980
33 G Defeudis, R Mazzilli, D Gianfrilli, A Lenzi, AM Isidori. The CATCH checklist to investigate adult–onset hypogonadism. Andrology 2018; 6(5):665–679 PMID: 29888533
https://doi.org/10.1111/andr.12506
34 C Poiana, C Capatina. Fracture risk assessment in patients with diabetes mellitus. J Clin Densitom 2017; 20(3): 432–443
https://doi.org/10.1016/j.jocd.2017.06.011 pmid: 28716499
35 CI Li, CS Liu, WY Lin, NH Meng, CC Chen, SY Yang, HJ Chen, CC Lin, TC Li. Glycated hemoglobin level and risk of hip fracture in older people with type 2 diabetes: a competing risk analysis of Taiwan diabetes cohort study. J Bone Miner Res 2015; 30(7): 1338–1346
https://doi.org/10.1002/jbmr.2462 pmid: 25598134
36 YY Lee, HB Kim, JW Lee, GM Lee, SY Kim, JA Hur, HC Cho. The association between urine albumin to creatinine ratio and osteoporosis in postmenopausal women with type 2 diabetes. J Bone Metab 2016; 23(1): 1–7
https://doi.org/10.11005/jbm.2016.23.1.1 pmid: 26981514
37 SK Jassal, D von Muhlen, E Barrett-Connor. Measures of renal function, BMD, bone loss, and osteoporotic fracture in older adults: the Rancho Bernardo study. J Bone Miner Res 2007; 22(2): 203–210
https://doi.org/10.1359/jbmr.061014 pmid: 17059370
38 C Liu, H Li. Correlation of the severity of chronic kidney disease with serum inflammation, osteoporosis and vitamin D deficiency. Exp Ther Med 2019; 17(1): 368–372
pmid: 30651805
39 F Afshinnia, S Pennathur. Association of hypoalbuminemia with osteoporosis: analysis of the National Health and Nutrition Examination Survey. J Clin Endocrinol Metab 2016; 101(6): 2468–2474
https://doi.org/10.1210/jc.2016-1099 pmid: 27144935
40 S Xiu, JK Chhetri, L Sun, Z Mu, L Wang. Association of serum prealbumin with risk of osteoporosis in older adults with type 2 diabetes mellitus: a cross-sectional study. Ther Adv Chronic Dis 2019; 10: 2040622319857361
https://doi.org/10.1177/2040622319857361 pmid: 31223465
41 HJ Do, JS Shin, J Lee, YJ Lee, MR Kim, D Nam, EJ Kim, Y Park, K Suhr, IH Ha. Association between liver enzymes and bone mineral density in Koreans: a cross-sectional study. BMC Musculoskelet Disord 2018; 19(1): 410
https://doi.org/10.1186/s12891-018-2322-1 pmid: 30474546
42 T Noguchi, K Ebina, M Hirao, S Otsuru, AJ Guess, R Kawase, T Ohama, S Yamashita, Y Etani, G Okamura, H Yoshikawa. Apolipoprotein E plays crucial roles in maintaining bone mass by promoting osteoblast differentiation via ERK1/2 pathway and by suppressing osteoclast differentiation via c-Fos, NFATc1, and NF-kB pathway. Biochem Biophys Res Commun 2018; 503(2): 644–650
https://doi.org/10.1016/j.bbrc.2018.06.055 pmid: 29906458
[1] Zhiruo Zhang, Shelan Liu, Mi Xiang, Shijian Li, Dahai Zhao, Chaolin Huang, Saijuan Chen. Protecting healthcare personnel from 2019-nCoV infection risks: lessons and suggestions[J]. Front. Med., 2020, 14(2): 229-231.
[2] Zhi-Ruo ZHANG PhD, Jian-Qing MI MD, Zhao-Jun WEN MA, Sai-Juan CHEN MD, PhD, Zhu CHEN PhD, Long-Jun GU MD, Jing-Yan TANG MD, PhD, Shu-Hong SHEN MD, PhD, . Using sound Clinical Paths and Diagnosis-related Groups (DRGs)-based payment reform to bring benefits to patient care: A case study of leukemia therapy[J]. Front. Med., 2010, 4(1): 8-15.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed