The impact of the combat method on radiomics feature compensation and analysis of scanners from different manufacturers.

BACKGROUND: This study investigated whether the Combat compensation method can remove the variability of radiomic features extracted from different scanners, while also examining its impact on the subsequent predictive performance of machine learning models. MATERIALS AND METHODS: 135 CT images of Credence Cartridge Radiomic phantoms were collected and screened from three scanners manufactured by Siemens, Philips, and GE. 100 radiomic features were extracted and 20 radiomic features were screened according to the Lasso regression method. The radiomic features extracted from the rubber and resin-filled regions in the cartridges were labeled into different categories for evaluating the performance of the machine learning model. Radiomics features were divided into three groups based on the different scanner manufacturers. The radiomic features were randomly divided into training and test sets with a ratio of 8:2. Five machine learning models (lasso, logistic regression, random forest, support vector machine, neural network) were employed to evaluate the impact of Combat on radiomic features. The variability among radiomic features were assessed using analysis of variance (ANOVA) and principal component analysis (PCA). Accuracy, precision, recall, and area under the receiver curve (AUC) were used as evaluation metrics for model classification. RESULTS: The principal component and ANOVA analysis results show that the variability of different scanner manufacturers in radiomic features was removed (P˃0.05). After harmonization with the Combat algorithm, the distributions of radiomic features were aligned in terms of location and scale. The performance of machine learning models for classification improved, with the Random Forest model showing the most significant enhancement. The AUC value increased from 0.88 to 0.92. CONCLUSIONS: The Combat algorithm has reduced variability in radiomic features from different scanners. In the phantom CT dataset, it appears that the machine learning model's classification performance may have improved after Combat harmonization. However, further investigation and validation are required to fully comprehend Combat's impact on radiomic features in medical imaging.

Deep learning PET/CT-based radiomics integrates clinical data: A feasibility study to distinguish between tuberculosis nodules and lung cancer.

BACKGROUND: Radiomic diagnosis models generally consider only a single dimension of information, leading to limitations in their diagnostic accuracy and reliability. The integration of multiple dimensions of information into the deep learning model have the potential to improve its diagnostic capabilities. The purpose of study was to evaluate the performance of deep learning model in distinguishing tuberculosis (TB) nodules and lung cancer (LC) based on deep learning features, radiomic features, and clinical information. METHODS: Positron emission tomography (PET) and computed tomography (CT) image data from 97 patients with LC and 77 patients with TB nodules were collected. One hundred radiomic features were extracted from both PET and CT imaging using the pyradiomics platform, and 2048 deep learning features were obtained through a residual neural network approach. Four models included traditional machine learning model with radiomic features as input (traditional radiomics), a deep learning model with separate input of image features (deep convolutional neural networks [DCNN]), a deep learning model with two inputs of radiomic features and deep learning features (radiomics-DCNN) and a deep learning model with inputs of radiomic features and deep learning features and clinical information (integrated model). The models were evaluated using area under the curve (AUC), sensitivity, accuracy, specificity, and F1-score metrics. RESULTS: The results of the classification of TB nodules and LC showed that the integrated model achieved an AUC of 0.84 (0.82-0.88), sensitivity of 0.85 (0.80-0.88), and specificity of 0.84 (0.83-0.87), performing better than the other models. CONCLUSION: The integrated model was found to be the best classification model in the diagnosis of TB nodules and solid LC.

3D Printed Robot Hand Structure Using Four-Bar Linkage Mechanism for Prosthetic Application.

Trans-radial prosthesis is a wearable device that intends to help amputees under the elbow to replace the function of the missing anatomical segment that resembles an actual human hand. However, there are some challenging aspects faced mainly on the robot hand structural design itself. Improvements are needed as this is closely related to structure efficiency. This paper proposes a robot hand structure with improved features (four-bar linkage mechanism) to overcome the deficiency of using the cable-driven actuated mechanism that leads to less structure durability and inaccurate motion range. Our proposed robot hand structure also took into account the existing design problems such as bulky structure, unindividual actuated finger, incomplete fingers and a lack of finger joints compared to the actual finger in its design. This paper presents the improvements achieved by applying the proposed design such as the use of a four-bar linkage mechanism instead of using the cable-driven mechanism, the size of an average human hand, five-fingers with completed joints where each finger is moved by motor individually, joint protection using a mechanical stopper, detachable finger structure from the palm frame, a structure that has sufficient durability for everyday use and an easy to fabricate structure using 3D printing technology. The four-bar linkage mechanism is the use of the solid linkage that connects the actuator with the structure to allow the structure to move. The durability was investigated using static analysis simulation. The structural details and simulation results were validated through motion capture analysis and load test. The motion analyses towards the 3D printed robot structure show 70-98% similar motion range capability to the designed structure in the CAD software, and it can withstand up to 1.6 kg load in the simulation and the real test. The improved robot hand structure with optimum durability for prosthetic uses was successfully developed.

Fused multivariate empirical mode decomposition (MEMD) and inverse solution method for EEG source localization.

EEG source localization is determining possible cortical sources of brain activities with scalp EEG. Generally, every step of the data processing sequence affects the accuracy of EEG source localization. In this paper, we introduce a fused multivariate empirical mode decomposing (MEMD) and inverse solution algorithm with an embedded unsupervised eye blink remover in order to localize the epileptogenic zone accurately. For this purpose, we constructed realistic forward models using MRI and boundary element method (BEM) for each patient to obtain results that are more realistic. We also developed an unsupervised algorithm utilizing a wavelet method to remove eye blink artifacts. Additionally, we applied MEMD, which is one of the recent and suitable feature extraction methods for non-linear, non-stationary, and multivariate signals such as EEG, to extract the signal of interest. We examined the localization results using the two most reliable linear distributed inverse methods in the literature: weighted minimum norm estimation (wMN) and standardized low resolution tomography (sLORETA). Results affirm the success of the proposed algorithm with the highest agreement compared to MRI reference by a specialist. Fusion of MEMD and sLORETA results in approximately zero localization error in terms of spatial difference with the validated MRI reference. High accuracy results of proposed algorithm using non-invasive and low-resolution EEG provide the potential of using this work for pre-surgical evaluation towards epileptogenic zone localization in clinics.