Echocardiographic Detection of Regional Wall Motion Abnormalities Using Artificial Intelligence Compared to Human Readers

BACKGROUND

Although regional wall motion abnormality (RWMA) detection is foundational to transthoracic echocardiography, current methods are prone to interobserver variability. We aimed to develop a deep learning (DL) model for RWMA assessment and compare it to expert and novice readers.

METHODS

We used 15,746 transthoracic echocardiography studies-including 25,529 apical videos-which were split into training, validation, and test datasets. A convolutional neural network was trained and validated using apical 2-, 3-, and 4-chamber videos to predict the presence of RWMA in 7 regions defined by coronary perfusion territories, using the ground truth derived from clinical transthoracic echocardiography reports. Within the test cohort, DL model accuracy was compared to 6 expert and 3 novice readers using F1 score evaluation, with the ground truth of RWMA defined by expert readers. Significance between the DL model and novices was assessed using the permutation test.

RESULTS

Within the test cohort, the DL model accurately identified any RWMA with an area under the curve of 0.96 (0.92-0.98). The mean F1 scores of the experts and the DL model were numerically similar for 6 of 7 regions: anterior (86 vs 84), anterolateral (80 vs 74), inferolateral (83 vs 87), inferoseptal (86 vs 86), apical (88 vs 87), inferior (79 vs 81), and any RWMA (90 vs 94), respectively, while in the anteroseptal region, the F1 score of the DL model was lower than the experts (75 vs 89). Using F1 scores, the DL model outperformed both novices 1 (P = .002) and 2 (P = .02) for the detection of any RWMA.

CONCLUSIONS

Deep learning provides accurate detection of RWMA, which was comparable to experts and outperformed a majority of novices. Deep learning may improve the efficiency of RWMA assessment and serve as a teaching tool for novices.

Differential knee joint loading patterns during gait for individuals with tibiofemoral and patellofemoral articular cartilage defects in the knee

OBJECTIVE

To determine compartment-specific loading patterns during gait, quantified as joint reaction forces (JRF), of individuals with knee articular cartilage defects (ACD) compared to healthy controls (HC).

METHODS

Individuals with ACDs and HC participated. Individuals with ACDs were divided into groups according to ACD location: PF (only a patellofemoral ACD), TF (only a tibiofemoral ACD), and MIX (both PF and TF ACDs). Participants underwent three-dimensional gait analysis at self-selected speed. TF joint reaction force (TF-JRF) was calculated using inverse dynamics. PF joint reaction force (PF-JRF) was derived from estimated quadriceps force (F_QUAD) and knee flexion angle. Primary variables of interest were the PF- and TF-JRF peaks (body weight [×BW]). Related secondary variables (gait speed, quadriceps strength, knee function, activity level) were evaluated as covariates.

RESULTS

First peak PF-JRF and TF-JRF were similar in the TF and MIX groups (0.75-1.0 ×BW, P = 0.6-0.9). Both peaks were also similar in the PF and HC groups (1.1-1.3 ×BW, P = 0.7-0.8), and higher than the TF and MIX groups (P = 0.004-0.02). For the second peak PF-JRF, only the HC group was higher than the TF group (P = 0.02). The PF group walked at a similar speed as the HC group; both groups walked faster than the TF and MIX groups (P < 0.001). With gait speed and quadriceps strength as covariates, no differences were observed in JRF peaks.

CONCLUSIONS

The results suggest the presence of a TF ACD (TF and MIX groups), but not a PF ACD (PF group), may affect joint loading patterns during walking. Walking slower may be a protective gait modification to reduce load.

Lower extremity work is associated with club head velocity during the golf swing in experienced golfers

While the golf swing is a complex whole body movement requiring coordination of all joints to achieve maximum ball velocity, the kinetic contribution of the lower extremities to club head velocity has not been quantified, despite the perception that the legs are a primary source of power during the swing. Mechanical power at the hips, knees, and ankles was estimated during the downswing phase of a full swing with a driver using a passive optical motion capture system and 2 force plates for adult males across a range of age and self-reported skill levels. Total work by the lower extremities was calculated by integrating the powers of all 6 joints over the downswing. Regression analyses showed that total lower extremity work was a strong predictor of club head velocity (R=0.63). Secondary analyses showed different relationships to club head velocity in lead and trail leg lower extremity joints, but none of these were as predictive of club head velocity as the total work performed by the lower extremities. These results provide quantitative evidence that the lower body's kinetic contribution may be an important factor in achieving greater club head velocity, contributing to greater driving distance and overall golf performance.

Figure skater level moderates balance training

It was suggested that baseline levels of postural control in figure skaters might influence the effectiveness of neuromuscular training. The aims of the present study were to investigate the baseline association of skater skill level with standard center of pressure metrics and time to stabilization, and to determine if skill level influenced the effectiveness of a 6-week neuromuscular training program. There was no main effect of skill level for any baseline center of pressure metric for either test. There was no main effect of skill level on the percent change in any metric for the single leg stance following training. However, skill level did influence landing test outcome measures. The difference in percent change of root mean squared was evident for the mediolateral (low: 24.5±16.50% vs. high: 2.42±14.99%) and anterior-posterior (low: 6.66±9.21% vs. high: - 4.03±5.91%) axes. Percent change in anterior-posterior time to stabilization also differed by skill level (low: - 0.73%±4.74 vs. high: - 5.61%±2.76). Note that this study was underpowered with 26 subjects and 14 subjects contributing to baseline and post-training assessments, respectively. Though no baseline differences in postural control were observed, compared with low skill levels, high skill levels benefitted more from training.

A case-control study of anterior cruciate ligament volume, tibial plateau slopes and intercondylar notch dimensions in ACL-injured knees

The role played by anatomical factors in ACL injury remains elusive. In this study, objective methods were used to characterize ACL volume, tibial slopes and notch geometry from ACL-injured and matched-control subjects. The study tested four hypotheses: (1) the medial tibial plateau slope is steeper posteriorly in the injured group compared to the non-injured group, (2) the lateral tibial plateau slope is steeper posteriorly in the injured group compared to the non-injured group, (3) the femoral intercondylar notch dimensions are smaller in the injured group compared to the non-injured group and (4) the ACL volume, tibial plateau slopes and intercondylar notch dimensions are all independent of each other. Fifty-four subjects were divided into two groups, those who had suffered a non-contact ACL injury and those who still had two healthy ACLs, matched to the injured subjects by gender, age, height and weight. The lateral tibial plateaus in the uninjured contralateral knees of the injured subjects had a significantly steeper posterior slope (1.8 degrees vs. -0.3 degrees ), a factor that potentially contributed to the ACL injury in the opposite knee. The intercondylar notch dimensions were found to be smaller in the injured subjects, potentially putting the ACL at risk of impingement, and intercondylar notch volume was correlated to ACL volume (r=0.58). Discriminant analysis showed that the notch width at the inlet was the best single predictor of ACL injury.

A markerless motion capture system to study musculoskeletal biomechanics: visual hull and simulated annealing approach

Human motion capture is frequently used to study musculoskeletal biomechanics and clinical problems, as well as to provide realistic animation for the entertainment industry. The most popular technique for human motion capture uses markers placed on the skin, despite some important drawbacks including the impediment to the motion by the presence of skin markers and relative movement between the skin where the markers are placed and the underlying bone. The latter makes it difficult to estimate the motion of the underlying bone, which is the variable of interest for biomechanical and clinical applications. A model-based markerless motion capture system is presented in this study, which does not require the placement of any markers on the subject's body. The described method is based on visual hull reconstruction and an a priori model of the subject. A custom version of adapted fast simulated annealing has been developed to match the model to the visual hull. The tracking capability and a quantitative validation of the method were evaluated in a virtual environment for a complete gait cycle. The obtained mean errors, for an entire gait cycle, for knee and hip flexion are respectively 1.5 degrees (+/-3.9 degrees ) and 2.0 degrees (+/-3.0 degrees ), while for knee and hip adduction they are respectively 2.0 degrees (+/-2.3 degrees ) and 1.1 degrees (+/-1.7 degrees ). Results for the ankle and shoulder joints are also presented. Experimental results captured in a gait laboratory with a real subject are also shown to demonstrate the effectiveness and potential of the presented method in a clinical environment.

Hip abductor weakness in distance runners with iliotibial band syndrome

OBJECTIVE

To examine hip abductor strength in long-distance runners with iliotibial band syndrome (ITBS), comparing their injured-limb strength to their nonaffected limb and to the limbs of a control group of healthy long-distance runners; and to determine whether correction of strength deficits in the hip abductors of the affected runners through a rehabilitation program correlates with a successful return to running.

DESIGN

Case series.

SETTING

Stanford University Sports Medicine Clinics.

PARTICIPANTS

24 distance runners with ITBS (14 female, 10 male) were randomly selected from patients presenting to our Runners' Injury Clinic with history and physical examination findings typical for ITBS. The control group of 30 distance runners (14 females, 16 males) were randomly selected from the Stanford University Cross-Country and Track teams.

MAIN OUTCOME MEASURES

Group differences in hip abductor strength, as measured by torque generated, were analyzed using separate two-tailed t-tests between the injured limb, non-injured limb, and the noninjured limbs of the control group. Prerehabilitation hip abductor torque for the injured runners was then compared with postrehabilitation torque after a 6-week rehabilitation program.

RESULTS

Hip abductor torque was measured with the Nicholas Manual Muscle Tester (kg), and normalized for differences in height and weight among subjects to units of percent body weight times height (%BWh). Average prerehabilitation hip abductor torque of the injured females was 7.82%BWh versus 9.82%BWh for their noninjured limb and 10.19%BWh for the control group of female runners. Average prerehabilitation hip abductor torque of the injured males was 6.86%BWh versus 8.62%BWh for their noninjured limb and 9.73%BWh for the control group of male runners. All prerehabilitation group differences were statistically significant at the p < 0.05 level. The injured runners were then enrolled in a 6-week standardized rehabilitation protocol with special attention directed to strengthening the gluteus medius. After rehabilitation, the females demonstrated an average increase in hip abductor torque of 34.9% in the injured limb, and the males an average increase of 51.4%. After 6 weeks of rehabilitation, 22 of 24 athletes were pain free with all exercises and able to return to running, and at 6-months follow-up there were no reports of recurrence.

CONCLUSIONS

Long distance runners with ITBS have weaker hip abduction strength in the affected leg compared with their unaffected leg and unaffected long-distance runners. Additionally, symptom improvement with a successful return to the preinjury training program parallels improvement in hip abductor strength.