Quantification ofin vivoimplant wear in total knee replacement from dynamic single plane radiography

Joint Track Machine Learning: An Autonomous Method of Measuring Total Knee Arthroplasty Kinematics From Single-Plane X-Ray Images

BACKGROUND

Dynamic radiographic measurements of 3-dimensional (3-D) total knee arthroplasty (TKA) kinematics have provided important information for implant design and surgical technique for over 30 years. However, current methods of measuring TKA kinematics are too cumbersome, inaccurate, or time-consuming for practical clinical application. Even state-of-the-art techniques require human-supervision to obtain clinically reliable kinematics. Eliminating human supervision could potentially make this technology practical for clinical use.

METHODS

We demonstrate a fully autonomous pipeline for quantifying 3D-TKA kinematics from single-plane radiographic imaging. First, a convolutional neural network (CNN) segmented the femoral and tibial implants from the image. Second, those segmented images were compared to precomputed shape libraries for initial pose estimates. Lastly, a numerical optimization routine aligned 3D implant contours and fluoroscopic images to obtain the final implant poses.

RESULTS

The autonomous technique reliably produces kinematic measurements comparable to human-supervised measures, with root-mean-squared differences of less than 0.7 mm and 4° for our test data, and 0.8 mm and 1.7° for external validation studies.

CONCLUSION

A fully autonomous method to measure 3D-TKA kinematics from single-plane radiographic images produces results equivalent to a human-supervised method, and may soon make it practical to perform these measurements in a clinical setting.

Less polyethylene wear in monobloc compared to modular ultra-high-molecular-weight-polyethylene inlays in hybrid total knee arthroplasty: A 5-year randomized radiostereometry study

BACKGROUND

A modular polyethylene (PE) inlay in total knee arthroplasty (TKA) may wear on both sides. PE particles may induce osteolysis, which can lead to implant loosening. The aim of this study was to determine if PE wear in monobloc TKA differs from that ofmodular TKA at 60-month follow-up.

PATIENTS AND METHODS

In a prospective, patient-blinded trial, 50 patients were randomized to hybrid TKA surgery with either acementless high-porosity trabecular-metal tibial component with a monobloc UHMWPE inlay (MONO-TM) or a cementless low-porosity screw-augmented titanium fiber-mesh tibial component with a modular UHMWPE inlay (MODULAR-FM). Radiostereometry was used to measure PE wear and tibial component migration.

RESULTS

At 60-monthfollow-up, mean PE wear of the medial compartment was 0.24 mm and 0.61 mm and mean PE wear of the lateral compartment was 0.31 mm and 0.82 mm for the MONO-TM and the MODULAR-FM groups, respectively (p < 0.01). The PE wear-rate was 0.05 mm (95% CI 0.03-0.08) in the MONO-TM group and 0.14 mm (95% CI 0.12-0.17) in the MODULAR-FM group (p < 0.01). Total translation at 60 months was mean 0.30 mm (95% CI 0.10-0.51) less (p < 0.01) for MONO-TM compared with MODULAR-FM tibial components. The majority of tibial components were stable (<0.2 mm MTPM) from 12 to 24-month and 24 to 60-month follow-up.

CONCLUSION

At mid-term follow-up, monobloc PE inlay wear was approximately 40% of that of the modular PE inlay wear, which suggest that back-side wear of modular PE inlays is a significant contributor of PE wear in hybrid TKA.

In vivo volumetric and linear wear measurement of reverse shoulder arthroplasty at minimum 5-year follow-up

BACKGROUND

Reverse shoulder arthroplasty is quickly becoming the most frequently performed glenohumeral joint replacement. The purpose of this study was to evaluate the volumetric and linear wear rates of ultrahigh-molecular-weight polyethylene humeral liners in vivo at a minimum 5-year follow-up.

METHODS

Radiostereometric analysis was used to image 15 patients at terminal range of motion in forward flexion, abduction, external rotation, and internal rotation and with the arm at the side. The relative position and orientation of the glenosphere and polyethylene were identified for each arm position. The apparent intersection of the glenosphere into the polyethylene was recorded as wear. Mean volumetric and linear wear rates were recorded, and Pearson correlation coefficients were applied to the 36-mm liners to assess the relationship between the wear rate and term of service.

RESULTS

The mean reverse shoulder arthroplasty term of service at the time of imaging was 8 ± 1 years (range, 6-11 years). The mean volumetric and linear wear rates for the 36-mm liners (n = 13) were 42 ± 22 mm³/yr (r = 0.688, P = .009) and 0.11 ± 0.03 mm/yr (r = 0.767, P = .002), respectively. The mean volumetric and linear wear rates for the 42-mm liners (n = 2) were 114 ± 44 mm³/yr and 0.17 ± 0.01 mm/yr, respectively. No single arm position was able to capture all recorded wear individually.

CONCLUSION

This study showed volumetric and linear wear rates of approximately 40 mm³/yr and 0.1 mm/yr, respectively, for the 36-mm polyethylene liners. The 42-mm liners showed higher wear rates, although a greater number of subjects is required for conclusive results. In vivo wear of reverse total shoulder arthroplasty is multidirectional and perceptible.

Validation of In Vivo Linear and Volumetric Wear Measurement for Reverse Total Shoulder Arthroplasty Using Model-Based Radiostereometric Analysis

Presently, polyethylene wear measurement of reverse total shoulder arthroplasty (rTSA) is restricted to in vitro, in silico, and retrieval analysis, with no method for the quantification of in vivo wear of well-functioning implants. The purpose of this study was to validate the use of model-based radiostereometric analysis (MBRSA) as a measurement tool for in vivo rTSA wear using a phantom setup. Six additively manufactured polyethylene inserts were fabricated, one unworn control and five to represent known wear patterns, and individually fit within the rTSA components. Each insert was imaged using standard radiostereometric techniques and analyzed using MBRSA. From the position and orientation estimation provided by MBRSA, a micro-computed tomography model of the control insert was virtually placed within the metaphyseal tray. The apparent intersection of the glenosphere into the insert was recorded as wear. This method enables wear measurements with a linear precision of 0.21 mm and a bias of 0.36 ± 0.13 mm, and a volumetric precision of 49.3 mm³ , with a bias of 48.9 ± 24.3 mm³ . This technique allows for the in vivo measurement of polyethylene wear without the requirement of marker beads or baseline radiographs, expanding the potential for in vivo wear measurements to larger populations and retrospective analysis. © 2019 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 37:1620-1627, 2019.

Radiostereometric Analysis Permits In Vivo Measurement of Very Small Levels of Wear in TKA

BACKGROUND

Revision of TKA as a result of polyethylene wear is decreasing, but long-term wear performance of polyethylene is still a topic of interest to surgeons and device manufacturers seeking to improve longevity. Measuring wear of modern, wear-resistant implants has been described using radiostereometric analysis (RSA). Performing in vivo measurements would establish whether implant retrieval studies are representative of wear in well-performing knees.

QUESTIONS/PURPOSES

For a single knee implant system, we sought to determine (1) the linear wear rate using RSA; (2) the association between demographic factors and wear rate; and (3) the association between limb alignment and wear rate.

METHODS

A total of 49 patients with a minimum followup of 10 years (median, 12 years; range, 10-20 years) were retrospectively selected. During the examined period, 4082 TKAs were performed of which 2085 were the implant examined in this study. There were 71 of these patients who met the criteria including an available full-leg radiograph postoperatively, and 34 of these patients returned for examination along with 15 additional from a separate RSA study that also met the criteria. All patients received a posterior-stabilized, cobalt-chromium-on-conventional polyethylene total knee implant from a single implant system, which was the most commonly used at our institution at the time. Patients underwent standing RSA examinations from 0° to 120° of flexion at a single time point without the use of marker beads. Linear wear rates (including creep) were measured based on intersections between the femoral component and tibial insert models. Associations between wear and patient age at surgery, sex, height, weight, body mass index, tibial insert size, and limb alignment were examined.

RESULTS

Using the maximum linear wear rate from any flexion angle, the lateral rate was 0.047 mm/year (interquartile range [IQR], 0.034-0.066 mm/year) and the medial rate was 0.052 mm/year (IQR, 0.040-0.069 mm/year). Using the median of the linear wear rates across all flexion angles, the lateral rate was 0.027 mm/year (IQR, 0.017-0.046 mm/year) and the medial rate was 0.038 mm/year (IQR, 0.022-0.054 mm/year). This rate for males was 0.049 mm/year medially (IQR, 0.042-0.077 mm/year) and 0.032 mm/year laterally (IQR, 0.026-0.059 mm/year), and for females was 0.027 mm/year medially (0.016-0.039 mm/year) and 0.020 mm/year laterally (IQR, 0.013-0.032 mm/year). The wear rate for males was greater medially (difference = 0.022 mm/year, p < 0.001) and laterally (difference = 0.012 mm/year, p = 0.008). There were associations between greater wear and increasing height (ρ = 0.48, p < 0.001 medially and ρ = 0.30, p = 0.04 laterally), decreasing body mass index (ρ = -0.31, p = 0.03 medially), and greater implant size (ρ = 0.34, p = 0.02 medially). Increasingly varus leg alignment was associated with greater medial wear (ρ = 0.33, p = 0.02).

CONCLUSIONS

Greater wear rates were associated with demographic factors and leg alignment. Further RSA wear studies of other modern implant systems would provide complementary information to retrieval studies and valuable data on wear resistance.

CLINICAL RELEVANCE

Good wear resistance was demonstrated by well-performing implants in patients at long-term followup with wear magnitudes in agreement with reported values from retrieval studies.

Additively manufactured implant components for imaging validation studies

Radiographic imaging is the current standard for evaluating postoperative joint replacements. Prior to application, such imaging methods need to be validated to determine the lower limits of performance under ideal conditions, using either a phantom or cadaver setup. Conventionally manufactured orthopedic implants for use in such studies are not always accessible and may be cost-prohibitive to purchase. We propose the use of additively manufactured implants as a cheaper, more accessible alternative for use in radiographic imaging validation studies. Bias and repeatability were compared between conventionally manufactured and additively manufactured reverse total shoulder implant sets under a standard model-based radiostereometric analysis phantom study environment. Measurements were compared using the humeral stem or glenosphere model relative to reference bone beads, and the humeral stem relative to the glenosphere model to measure implant relative displacement. Compared to the conventionally manufactured implants, the additively manufactured implants had less bias along the internal-external rotation axis (p < 0.001), but greater bias along the abduction-adduction and flexion-extension rotation axes (p = 0.005, 0.011). Additively manufactured implants had greater repeatability along the internal-external rotation axis (p < 0.001), but worse repeatability along the medial-lateral translation axis (p = 0.001) and the abduction-adduction rotation axis (p < 0.001). Differences were on the orders of 0.01 mm and 0.5°. For the purpose of validating two-dimensional-three-dimensional radiographic imaging techniques of orthopedic implants, additively manufactured implants can be used in place of conventionally manufactured implants, assuming they are fabricated to the manufacturer's specifications. Observed differences were within the errors of the measurement technique and not clinically meaningful.

How Does Wear Rate Compare in Well-functioning Total Hip and Knee Replacements? A Postmortem Polyethylene Liner Study

BACKGROUND

The longevity of total hip (THR) and knee replacements (TKR) that used historical bearing materials of gamma-in-air sterilized UHMWPE was affected more by osteolysis in THRs than in TKRs, although osteolysis remains a concern in TKRs. Therefore, the study of polyethylene wear is still of interest for the knee, particularly because few studies have investigated volumetric material loss in tibial knee inserts. For this study, a unique collection of autopsy-retrieved TKR and THR components that were well-functioning at the time of retrieval was used to compare volumetric wear differences between hip and knee polyethylene components made from identical material.

QUESTIONS/PURPOSES

The following questions were addressed: (1) How much did the hip liners wear and what wear patterns did they exhibit? (2) How much did the knee inserts wear and what wear patterns did they exhibit? (3) What is the ratio between TKR and THR wear after controlling for implantation time and patient age?

METHODS

We compared 23 THR components (Harris-Galante [HG] and HG II) and 20 TKR components (Miller-Galante [MG II]) that were retrieved postmortem. The components were made from the same polyethylene formulation and with similar manufacturing and sterilization (gamma-in-air) processes. Twenty-one patients (12 males, nine females) had THRs and 16 (four males, 12 females) had TKRs. Patients who had TKRs had an older (p = 0.001) average age than patients who had THRs (age, 75 years; SD, 10, versus 66 years; SD, 12, respectively). Only well-functioning components were included in this study. Therefore, implants retrieved postmortem from physically active patients and implanted for at least 2 years were considered. In addition, only normally wearing TKR components were considered, ie, those with fatigue wear (delamination) were excluded. The wear volume of each component was measured using metrology. For the tibial inserts an autonomous mathematic reconstruction method was used for quantification.

RESULTS

The acetabular liners of the THR group had a wear rate of 38 mm(3) per year (95% CI, 29-47 mm(3)/year). Excluding patients with low-activity, the wear rate was 47 mm(3) per year (95% CI, 37-56 mm(3)/year). The wear rate of normally wearing tibial inserts was 17 mm(3) per year (95% CI, -6 to 40 mm(3)/year). After controlling for the relevant confounding variable of age, we found a TKR/THR wear rate ratio of 0.5 (95% CI, 0.29-0.77) at 70 years of age with a slightly increasing difference with increasing age.

CONCLUSIONS

Excluding delamination, TKRs exhibited lower articular wear rates than THRs for historical polyethylene in these two unique cohorts of postmortem retrievals.

CLINICAL RELEVANCE

The lower TKR wear rate is in line with the lower incidence of osteolysis in TKRs compared with THRs.