Identification and interpretation of gait analysis features and foot conditions by explainable AI.

Clinical gait analysis is a crucial step for identifying foot disorders and planning surgery. Automating this process is essential for efficiently assessing the substantial amount of gait data. In this study, we explored the potential of state-of-the-art machine learning (ML) and explainable artificial intelligence (XAI) algorithms to automate all various steps involved in gait analysis for six specific foot conditions. To address the complexity of gait data, we manually created new features, followed by recursive feature elimination using Support Vector Machines (SVM) and Random Forests (RF) to eliminate low-variance features. SVM, RF, K-nearest Neighbor (KNN), and Logistic Regression (LREGR) were compared for classification, with a Majority Voting (MV) model combining trained models. KNN and MV achieved mean balanced accuracy, recall, precision, and F1 score of 0.87. All models were interpreted using Local Interpretable Model-agnostic Explanation (LIME) method and the five most relevant features were identified for each foot condition. High success scores indicate a strong relationship between selected features and foot conditions, potentially indicating clinical relevance. The proposed ML pipeline, adaptable for other foot conditions, showcases its potential in aiding experts in foot condition identification and planning surgeries.

Machine Learning Models for Classification of Cushing's Syndrome Using Retrospective Data.

Accurate classification of Cushing's Syndrome (CS) plays a critical role in providing the early and correct diagnosis of CS that may facilitate treatment and improve patient outcomes. Diagnosis of CS is a complex process, which requires careful and concurrent interpretation of signs and symptoms, multiple biochemical test results, and findings of medical imaging by physicians with a high degree of specialty and knowledge to make correct judgments. In this article, we explore the state of the art machine learning algorithms to demonstrate their potential as a clinical decision support system to analyze and classify CS to facilitate the diagnosis, prognosis, and treatment of CS. Prominent algorithms are compared using nested cross-validation and various class comparison strategies including multiclass, one vs. all, and one vs. one binary classification. Our findings show that Random Forest (RF) algorithm is most suitable for the classification of CS. We demonstrate that the proposed approach can classify CS with an average accuracy of 92% and an average F1 score of 91.5%, depending on the class comparison strategy and selected features. RF-based one vs. all binary classification model achieves sensitivity of 97.6%, precision of 91.1%, and specificity of 87.1% to discriminate CS from non-CS on the test dataset. RF-based multiclass classification model achieves average per class sensitivity of 91.8%, average per class specificity of 97.1%, and average per class precision of 92.1% to classify different subtypes of CS on the test dataset. Clinical performance evaluation suggests that the developed models can help improve physicians' judgment in diagnosing CS.

iBEX: Modular Open-Source Software for Digital Radiography.

A device-independent software package, named iBEX, is developed to accelerate the research and development efforts for X-ray imaging setups such as chest radiography, linear and multidirectional tomography, and dental and skeletal radiography. Its extension mechanism makes the software adaptable for a wide range of digital X-ray imaging hardware combinations and provides capabilities for researchers to develop image processing plug-ins. Independent of the X-ray sensor technology, iBEX could integrate with heterogeneous communication channels of digital detectors. iBEX is a freeware option for preclinical and early clinical testing of radiography devices. It provides tools to calibrate the device, integrate to health information infrastructure, acquire image, store studies on local storage, and send them to Picture Archiving and Communication System (PACS). iBEX is a unique open-source project bringing X-ray imaging devices' software into the scope of the open-source community to reduce the X-ray scanners' research effort, potentially increase the image quality, and cut down the production and testing costs of radiography devices.

Changes in serum asymmetric dimethylarginine and endothelial markers levels with varying periods of hemodialysis.

Asymmetric dimethylarginine (ADMA) as a uremia toxin is accumulated in end-stage renal disease (ESRD) patients. Elevated ADMA level has been shown to be predictive of cardiovascular diseases (CVDs) and all-cause mortality in ESRD. Therefore, we investigated the effect of prolonged hemodialysis (HD) treatment on the levels of serum ADMA, arginine, nitric oxide (NO), soluble intercellular adhesion molecule-1 (sICAM-1) and soluble vascular cell adhesion molecule-1 (sVCAM-1). Seventy-five patients (M/F = 40/35) with chronic renal failure (CRF) and who were on HD were divided into five groups with differing treatment periods of HD; from 6 to 24 months to 97-120 months. Fifteen apparently healthy subjects acted as controls. The serum levels of ADMA, sICAM-1 and sVCAM-1 were increased in all patient groups compared to the control group. No significant difference was observed when the patient groups were compared in terms of HD treatment periods. Nitric oxide levels were lower in the three groups who were treated for periods of 49-72, 73-96, 97-120 months compared to the control group. The L-arginine to ADMA ratio was decreased in all patient groups compared to controls. Consequently, our investigations have shown that in HD continued for more than 4 years NO levels began to decrease significantly and the levels of serum ADMA, sICAM-1 and sVCAM-1 levels increased although this increase was not affected by the period in which hemodialysis treatment was applied.

Magnetic resonance imaging assessment of mechanical interactions between human lower leg muscles in vivo.

Evidence on epimuscular myofascial force transmission (EMFT) was shown for undissected muscle in situ. We hypothesize that global length changes of gastrocnemius muscle-tendon complex in vivo will cause sizable and heterogeneous local strains within all muscles of the human lower leg. Our goal is to test this hypothesis. A method was developed and validated using high-resolution 3D magnetic resonance image sets and Demons nonrigid registration algorithm for performing large deformation analyses. Calculation of strain tensors per voxel in human muscles in vivo allowed quantifying local heterogeneous tissue deformations and volume changes. After hip and knee movement (Δ knee angle ≈ 25 deg) but without any ankle movement, local lengthening within m. gastrocnemius was shown to occur simultaneously with local shortening (maximally by +34.2% and -32.6%, respectively) at different locations. Moreover, similar local strains occur also within other muscles, despite being kept at constant muscle-tendon complex length. This is shown for synergistic m. soleus and deep flexors, as well as for antagonistic anterior crural and peroneal muscle groups: minimum peak lengthening and shortening equaled 23.3% and 25.54%, respectively despite global isometric conditions. These findings confirm our hypothesis and show that in vivo, muscles are in principle not independent mechanically.

Effects of knee joint angle on global and local strains within human triceps surae muscle: MRI analysis indicating in vivo myofascial force transmission between synergistic muscles.

PURPOSE: Mechanical interactions between muscles have been shown for in situ conditions. In vivo data for humans is unavailable. Global and local length changes of calf muscles were studied to test the hypothesis that local strains may occur also within muscle for which global strain equals zero. METHODS: For determination of globally induced strain in m. gastrocnemius in dissected human cadavers several knee joint angles were imposed, while keeping ankle joint angle constant and measuring its muscle-tendon complex length changes. In vivo local strains in both gastrocnemius and soleus muscles were calculated using MRI techniques in healthy human volunteers comparing images taken at static knee angles of 173° and 150°. RESULTS: Imposed global strains on gastrocnemius were much smaller than local strains. High distributions of strains were encountered, e.g. overall lengthened muscle contains locally lengthened, as well as shortened areas within it. Substantial strains were not limited to gastrocnemius, but were found also in synergistic soleus muscle, despite the latter muscle-tendon complex length remaining isometric (constant ankle angle: i.e. global strain = 0), as it does not cross the knee. Based on results of animal experiments this effect is ascribed to myofascial connections between these synergistic muscles. The most likely pathway is the neurovascular tract within the anterior crural compartment (i.e. the collagen reinforcements of blood vessels, lymphatics and nerves). However, direct intermuscular transmission of force may also occur via the perimysium shared between the two muscles. CONCLUSIONS: Global strains imposed on muscle (joint movement) are not good estimators of in vivo local strains within it: differing in magnitude, as well as direction of length change. Substantial mechanical interaction occurs between calf muscles, which is mediated by myofascial force transmission between these synergistic muscles. This confirms conclusions of previous in situ studies in experimental animals and human patients, for in vivo conditions in healthy human subjects.