Auto-segmentation of the tibia and femur from knee MR images via deep learning and its application to cartilage strain and recovery

A comparison of three methods for establishing an ACL reference length in vivo

As anterior cruciate ligament (ACL) injuries are highly prevalent among active individuals, it is vital to better understand the loading conditions which lead to injury. One method for doing so is through measurement of dynamic, in vivo ACL strain. To measure strain, it is necessary to normalize elongation of the ACL to a 'reference length' which corresponds to the point at which the ligament transitions from being unloaded to carrying tension. The purpose of this study was to compare the length of the ACL in three different positions to evaluate their utility for establishing a reference (or zero-strain) length of the ACL. ACL reference length was determined using three different methods for each of ten healthy participants. Using magnetic resonance and biplanar radiographic imaging techniques, we measured the length of the ACL during supine resting, quiet standing, and anterior/posterior (AP) drawer testing. During the AP drawer testing, the slack-taut transition point was defined as the inflection point of the AP translation vs ACL elongation curve. There was good consistency between the three ACL length measurements (ICC=0.80). Differences in mean ACL length between the three methods were within 1 mm. While determining the precise zero-strain length of the ACL in vivo remains a challenge, the reference positions utilized in this study produce consistent measurements of ACL length. These findings are important because reliable measurements of in vivo ACL strain have the potential to serve as indicators of propensity for injury.

A scoping review of human skeletal kinematics research using biplane radiography

Biplane radiography has emerged as the gold standard for accurately measuring in vivo skeletal kinematics during physiological loading. The purpose of this scoping review was to map the extent, range, and nature of biplane radiography research on humans from 2004 through 2022. A literature search was performed using the terms biplane radiography, dual fluoroscopy, dynamic stereo X-ray, and biplane videoradiography. All articles referenced in included publications were also assessed for inclusion. A secondary search was then performed using the names of the most frequently appearing principal investigators among included papers. A total of 379 manuscripts were identified and included. The first studies published in 2004 focused on the native knee, followed by studies of the ankle joint complex in 2006, the shoulder in 2007, and the spine in 2008. Nearly half (180, 47.5%) of all manuscripts investigated knee kinematics. The average number of publications increased from 21.6 per year from 2012 to 2017 to 34.6 per year from 2017 to 2022. The average number of participants per study was 16, with a range from 1 to 101. A total of 90.2% of studies featured cohorts of 30 or less. The most prolific research groups for each joint were: Mass General Hospital (lumbar spine and knee), Henry Ford Hospital (shoulder), the University of Utah (ankle and hip), The University of Pittsburgh (cervical spine), and Brown University (hand/wrist/elbow). Future advancements in biplane radiography research are dependent upon increased availability of these imaging systems, standardization of data collection protocols, and the development of automated approaches to expedite data processing.

Development and validation of a semi-automated and unsupervised method for femur segmentation from CT

Quantitative computed tomography (QCT)-based in silico models have demonstrated improved accuracy in predicting hip fractures with respect to the current gold standard, the areal bone mineral density. These models require that the femur bone is segmented as a first step. This task can be challenging, and in fact, it is often almost fully manual, which is time-consuming, operator-dependent, and hard to reproduce. This work proposes a semi-automated procedure for femur bone segmentation from CT images. The proposed procedure is based on the bone and joint enhancement filter and graph-cut algorithms. The semi-automated procedure performances were assessed on 10 subjects through comparison with the standard manual segmentation. Metrics based on the femur geometries and the risk of fracture assessed in silico resulting from the two segmentation procedures were considered. The average Hausdorff distance (0.03 ± 0.01 mm) and the difference union ratio (0.06 ± 0.02) metrics computed between the manual and semi-automated segmentations were significantly higher than those computed within the manual segmentations (0.01 ± 0.01 mm and 0.03 ± 0.02). Besides, a blind qualitative evaluation revealed that the semi-automated procedure was significantly superior (p < 0.001) to the manual one in terms of fidelity to the CT. As for the hip fracture risk assessed in silico starting from both segmentations, no significant difference emerged between the two (R2 = 0.99). The proposed semi-automated segmentation procedure overcomes the manual one, shortening the segmentation time and providing a better segmentation. The method could be employed within CT-based in silico methodologies and to segment large volumes of images to train and test fully automated and supervised segmentation methods.

Integration of Swin UNETR and statistical shape modeling for a semi-automated segmentation of the knee and biomechanical modeling of articular cartilage

Simulation studies, such as finite element (FE) modeling, provide insight into knee joint mechanics without patient involvement. Generic FE models mimic the biomechanical behavior of the tissue, but overlook variations in geometry, loading, and material properties of a population. Conversely, subject-specific models include these factors, resulting in enhanced predictive precision, but are laborious and time intensive. The present study aimed to enhance subject-specific knee joint FE modeling by incorporating a semi-automated segmentation algorithm using a 3D Swin UNETR for an initial segmentation of the femur and tibia, followed by a statistical shape model (SSM) adjustment to improve surface roughness and continuity. For comparison, a manual FE model was developed through manual segmentation (i.e., the de-facto standard approach). Both FE models were subjected to gait loading and the predicted mechanical response was compared. The semi-automated segmentation achieved a Dice similarity coefficient (DSC) of over 98% for both the femur and tibia. Hausdorff distance (mm) between the semi-automated and manual segmentation was 1.4 mm. The mechanical results (max principal stress and strain, fluid pressure, fibril strain, and contact area) showed no significant differences between the manual and semi-automated FE models, indicating the effectiveness of the proposed semi-automated segmentation in creating accurate knee joint FE models. We have made our semi-automated models publicly accessible to support and facilitate biomechanical modeling and medical image segmentation efforts ( https://data.mendeley.com/datasets/k5hdc9cz7w/1 ).

Use of a novel magnetically actuated compression system to study the temporal dynamics of axial and lateral strain in human osteochondral plugs

The high water content of articular cartilage allows this biphasic tissue to withstand large compressive loads through fluid pressurization. The system presented here, termed the "MagnaSquish", provides new capabilities for quantifying the effect of rehydration on cartilage behavior during cyclic loading. An imbalanced rate of fluid exudation during load and fluid re-entry during recovery can lead to the accumulation of strain during successive loading cycles - a phenomenon known as ratcheting. Typical experimental systems for cartilage biomechanics use continuous contact between the platen and sample, which may affect tissue rehydration by compressing the top layer of cartilage and slowing fluid re-entry. To address this limitation, we developed a magnetically actuated device that provides full lift-off of the platen in between loading cycles. We investigated strain accumulation in cadaveric human osteochondral plugs during 750 loading cycles, with two dimensional profiles of the cartilage captured at 30 frames per second throughout loading and 10 min of additional free swelling recovery. Axial and lateral strain measurements were extracted from the tissue profiles using a UNet-based deep learning algorithm to circumvent manual tracing. We observed increased axial strain accumulation with shorter inter-cycle recovery, with static loading serving as the extreme case of zero recovery. The loading waveform during the 750 cycles dictated the pace of the recovery during the extended free swelling period, as shorter inter-cycle recovery led to more persistent axial strain accumulation for up to five minutes. This work showcases the importance of fluid re-entry in resisting strain accumulation during cyclical compression.