Preoperative Prediction of Optimal Femoral Implant Size by Regularized Regression on 3D Femoral Bone Shape

被引:1
作者
Lambrechts, Adriaan [1 ]
Van Dijck, Christophe [1 ]
Wirix-Speetjens, Roel [1 ]
Vander Sloten, Jos [2 ]
Maes, Frederik [3 ,4 ]
Van Huffel, Sabine [5 ]
机构
[1] Materialise NV, B-3001 Leuven, Belgium
[2] Katholieke Univ Leuven, Biomech Sect, Mech Engn, B-3001 Leuven, Belgium
[3] Katholieke Univ Leuven, Dept Elect Engn ESAT, Proc Speech & Images PSI, B-3001 Leuven, Belgium
[4] UZ Leuven, Med Imaging Res Ctr, B-3000 Leuven, Belgium
[5] Katholieke Univ Leuven, STADIUS Ctr Dynam Syst, Dept Elect Engn ESAT, Signal Proc & Data Analyt, B-3001 Leuven, Belgium
来源
APPLIED SCIENCES-BASEL | 2023年 / 13卷 / 07期
关键词
total knee arthroplasty; templating; machine learning; group lasso; TOTAL KNEE ARTHROPLASTY; COMPONENT; ACCURACY; RISK;
D O I
10.3390/app13074344
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Preoperative determination of implant size for total knee arthroplasty surgery has numerous clinical and logistical benefits. Currently, surgeons use X-ray-based templating to estimate implant size, but this method has low accuracy. Our study aims to improve accuracy by developing a machine learning approach that predicts the required implant size based on a 3D femoral bone mesh, the key factor in determining the correct implant size. A linear regression framework imposing group sparsity on the 3D bone mesh vertex coordinates was proposed based on a dataset of 446 MRI scans. The group sparse regression method was further regularized based on the connectivity of the bone mesh to enforce neighbouring vertices to have similar importance to the model. Our hypergraph regularized group lasso had an accuracy of 70.1% in predicting femoral implant size while the initial implant size prediction provided by the instrumentation manufacturer to the surgeon has an accuracy of 23.1%. Furthermore, our method was capable of predicting the implant size up to one size smaller or larger with an accuracy of 99.1%, thereby surpassing other state-of-the-art methods. The hypergraph regularized group lasso was able to obtain a significantly higher accuracy compared to the implant size prediction provided by the instrumentation manufacturer.
引用
收藏
页数:13
相关论文
共 37 条
[1]   Automated segmentation of knee bone and cartilage combining statistical shape knowledge and convolutional neural networks: Data from the Osteoarthritis Initiative [J].
Ambellan, Felix ;
Tack, Alexander ;
Ehlke, Moritz ;
Zachow, Stefan .
MEDICAL IMAGE ANALYSIS, 2019, 52 :109-118
[2]  
Amberg B, 2007, IEEE I CONF COMP VIS, P1326
[3]  
[Anonymous], 2008, P ICML
[4]  
Bhowmik-Stoker M, 2018, SURG TECHNOL INT, V33
[5]   Predicting implant size in total knee arthroplasty using demographic variables LINEAR REGRESSION AND BAYESIAN MODELLING [J].
Blevins, J. L. ;
Rao, V ;
Chiu, V ;
Man, S. ;
Westrich, G. H. .
BONE & JOINT JOURNAL, 2020, 102B (06) :85-90
[6]   Patient Satisfaction after Total Knee Arthroplasty Who is Satisfied and Who is Not? [J].
Bourne, Robert B. ;
Chesworth, Bert M. ;
Davis, Aileen M. ;
Mahomed, Nizar N. ;
Charron, Kory D. J. .
CLINICAL ORTHOPAEDICS AND RELATED RESEARCH, 2010, 468 (01) :57-63
[7]   Sparse Discriminant Analysis [J].
Clemmensen, Line ;
Hastie, Trevor ;
Witten, Daniela ;
Ersboll, Bjarne .
TECHNOMETRICS, 2011, 53 (04) :406-413
[8]   SPARSE-MATRIX TEST PROBLEMS [J].
DUFF, IS ;
GRIMES, RG ;
LEWIS, JG .
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 1989, 15 (01) :1-14
[9]   2D versus 3D templating in total knee arthroplasty [J].
Ettinger, Max ;
Claassen, Leif ;
Paes, Peter ;
Calliess, Tilman .
KNEE, 2016, 23 (01) :149-151
[10]  
Hastie T., 2015, Statistical Learning with Sparsity: The Lasso and Generalizations, DOI 10.1201/b18401