Applying machine learning methods to enable automatic customisation of knee replacement implants from CT data

Abstract

The aim of this study was to develop an automated pipeline capable of designing custom total knee replacement implants from CT scans. The developed pipeline firstly utilised a series of machine learning methods including classification, object detection, and image segmentation models, to extract geometrical information from inputted DICOM files. Statistical shape models then used the information to create femur and tibia 3D surface model predictions which were ultimately used by computer aided design scripts to generate customised implant designs. The developed pipeline was trained and tested using CT scan images, along with segmented 3D models, obtained for 98 Korean Asian subjects. The performance of the pipeline was tested computationally by virtually fitting outputted implant designs with ‘ground truth’ 3D models for each test subject’s bones. This demonstrated the pipeline was capable of repeatably producing highly accurate designs, and its performance was not impacted by subject sex, height, age, or knee side. In conclusion, a robust, accurate and automatic, CT-based total knee replacement customisation pipeline was shown to be feasible and could afford significant time and cost advantages over conventional methods. The pipeline framework could also be adapted to enable customisation of other medical implants.

Towards skin-on-a-chip for screening the dermal absorption of cosmetics

“…As highlighted by recent reports, these trends support the notion that the landscape is steadily shifting towards a point where CAD processes can be fully automated or derived from a simple sketch, scan or photograph supplied by the user. 83 This highlights a growing challenge in design engineering where future users cannot rely on graphical interfaces but must instead be proficient in computer programming to drive automation. Current CAD packages would also require users to spend an unrealistic amount of time to accurately model the multi-scale heterogeneity and complexity of artificial cells, especially if nanofeatures such as lipids, proteins, and nucleic acids were to be included.…”

Towards automation in non-invasive measurement of knee implant displacement

“…In recent years, deep learning (DL) approaches have been developed to generate automatic segmentation of various bone structures in CT imaging, which can alleviate user interaction. [6][7][8] The aim of this study is to automate segmentation tasks in the knee prosthesis displacement analysis by Kievit et al using DL. This is another step towards reducing the amount of user interaction and therefore observer variation in knee prosthesis displacement analysis.…”

Customized‐individually‐made origin® implants in total knee arthroplasty allow a reliable solution for accurate reproduction of planned implant positioning

“…These individualized patient's specific implants, created from cross-sectional imaging of the lower limb, are designed with the triple target to respect the patient's anatomy, minimize bone cuts, and restore the most physiological joint kinematics possible [5,7]. The incorporation of machine learning in the automatic customization process has further contributed to the accuracy and success of CIM TKR [18], where these prosthesis are now available for primary as well as for revision hinged arthroplasty [19][20][21].…”

“…Although their potential clinical benefits is not well established in literature [ 11 , 16 ], recent investigation of the new ORIGIN® CIM prosthesis showed significant improvement in various clinical and patient-reported outcomes including the Knee Society Score (KSS), range of motion, and the Forgotten Joint Score (FJS) at 1-year follow-up [ 17 ]. These individualized patient’s specific implants, created from cross-sectional imaging of the lower limb, are designed with the triple target to respect the patient's anatomy, minimize bone cuts, and restore the most physiological joint kinematics possible [ 5 , 7 ].The incorporation of machine learning in the automatic customization process has further contributed to the accuracy and success of CIM TKR [ 18 ], where these prosthesis are now available for primary as well as for revision hinged arthroplasty [ 19 – 21 ].…”