Multi-Source Domain Adaptation for Medical Image Segmentation.

Unsupervised domain adaptation(UDA) aims to mitigate the performance drop of models tested on the target domain, due to the domain shift from the target to sources. Most UDA segmentation methods focus on the scenario of solely single source domain. However, in practical situations data with gold standard could be available from multiple sources (domains), and the multi-source training data could provide more information for knowledge transfer. How to utilize them to achieve better domain adaptation yet remains to be further explored. This work investigates multi-source UDA and proposes a new framework for medical image segmentation. Firstly, we employ a multi-level adversarial learning scheme to adapt features at different levels between each of the source domains and the target, to improve the segmentation performance. Then, we propose a multi-model consistency loss to transfer the learned multi-source knowledge to the target domain simultaneously. Finally, we validated the proposed framework on two applications, i.e., multi-modality cardiac segmentation and cross-modality liver segmentation. The results showed our method delivered promising performance and compared favorably to state-of-the-art approaches.

MyoPS: A benchmark of myocardial pathology segmentation combining three-sequence cardiac magnetic resonance images.

Assessment of myocardial viability is essential in diagnosis and treatment management of patients suffering from myocardial infarction, and classification of pathology on the myocardium is the key to this assessment. This work defines a new task of medical image analysis, i.e., to perform myocardial pathology segmentation (MyoPS) combining three-sequence cardiac magnetic resonance (CMR) images, which was first proposed in the MyoPS challenge, in conjunction with MICCAI 2020. Note that MyoPS refers to both myocardial pathology segmentation and the challenge in this paper. The challenge provided 45 paired and pre-aligned CMR images, allowing algorithms to combine the complementary information from the three CMR sequences for pathology segmentation. In this article, we provide details of the challenge, survey the works from fifteen participants and interpret their methods according to five aspects, i.e., preprocessing, data augmentation, learning strategy, model architecture and post-processing. In addition, we analyze the results with respect to different factors, in order to examine the key obstacles and explore the potential of solutions, as well as to provide a benchmark for future research. The average Dice scores of submitted algorithms were 0.614±0.231 and 0.644±0.153 for myocardial scars and edema, respectively. We conclude that while promising results have been reported, the research is still in the early stage, and more in-depth exploration is needed before a successful application to the clinics. MyoPS data and evaluation tool continue to be publicly available upon registration via its homepage (www.sdspeople.fudan.edu.cn/zhuangxiahai/0/myops20/).

Mechanical Behaviour Evaluation of Full Iron Tailings Concrete Columns under Large Eccentric Short-Term Loading.

In this study, full iron tailings concrete (FITC) was created using iron tailings from a tailings pond in Qian'an, China. Iron tailings account for 86.8% of the total mass of solid raw materials in the FITC. To enable large-scale use of FITC, a comprehensive investigation of the structural behaviour of full-iron tailing-reinforced concrete (FITRC) specimens is warranted. Therefore, eight rectangular reinforced concrete (RC) columns with conventional reinforced concrete (CRC) as a control were tested to investigate the effects of section dimensions, initial eccentricities, and concrete strengths, on the structural behaviour of FITRC columns under large eccentric short-term loading. The experimental and analytical results indicated that the sectional strain of the FITRC columns satisfied the plane-section assumption under short-term loading, and the lateral deflection curve agreed well with the half-sinusoidal curve. In addition, the FITRC columns exhibited a slightly lower cracking load and lower ultimate load capacity than the CRC columns, and the crack widths were larger than those of the CRC columns. The reduction in the load capacity observed in the FITRC was within the permissible range stated in the design code, thereby satisfying the code requirements. The deformation coefficients of the FITRC and CRC columns were identical, and the cracking and ultimate loads calculated according to the current code and theories were in good agreement with the measured results.

Learning multi-view and centerline topology connectivity information for pulmonary artery-vein separation.

BACKGROUND: Automatic pulmonary artery-vein separation has considerable importance in the diagnosis and treatment of lung diseases. However, insufficient connectivity and spatial inconsistency have always been the problems of artery-vein separation. METHODS: A novel automatic method for artery-vein separation in CT images is presented in this work. Specifically, a multi-scale information aggregated network (MSIA-Net) including multi-scale fusion blocks and deep supervision, is proposed to learn the features of artery-vein and aggregate additional semantic information, respectively. The proposed method integrates nine MSIA-Net models for artery-vein separation, vessel segmentation, and centerline separation tasks along with axial, coronal, and sagittal multi-view slices. First, the preliminary artery-vein separation results are obtained by the proposed multi-view fusion strategy (MVFS). Then, centerline correction algorithm (CCA) is used to correct the preliminary results of artery-vein separation by the centerline separation results. Finally, the vessel segmentation results are utilized to reconstruct the artery-vein morphology. In addition, weighted cross-entropy and dice loss are employed to solve the class imbalance problem. RESULTS: We constructed 50 manually labeled contrast-enhanced computed CT scans for five-fold cross-validation, and experimental results demonstrated that our method achieves superior segmentation performance of 97.7%, 85.1%, and 84.9% on ACC, Pre, and DSC, respectively. Additionally, a series of ablation studies demonstrate the effectiveness of the proposed components. CONCLUSION: The proposed method can effectively solve the problem of insufficient vascular connectivity and correct the spatial inconsistency of artery-vein.

MNC-Net: Multi-task graph structure learning based on node clustering for early Parkinson's disease diagnosis.

PURPOSE: The identification of early-stage Parkinson's disease (PD) is important for the effective management of patients, affecting their treatment and prognosis. Recently, structural brain networks (SBNs) have been used to diagnose PD. However, how to mine abnormal patterns from high-dimensional SBNs has been a challenge due to the complex topology of the brain. Meanwhile, the existing prediction mechanisms of deep learning models are often complicated, and it is difficult to extract effective interpretations. In addition, most works only focus on the classification of imaging and ignore clinical scores in practical applications, which limits the ability of the model. Inspired by the regional modularity of SBNs, we adopted graph learning from the perspective of node clustering to construct an interpretable framework for PD classification. METHODS: In this study, a multi-task graph structure learning framework based on node clustering (MNC-Net) is proposed for the early diagnosis of PD. Specifically, we modeled complex SBNs into modular graphs that facilitated the representation learning of abnormal patterns. Traditional graph neural networks are optimized through graph structure learning based on node clustering, which identifies potentially abnormal brain regions and reduces the impact of irrelevant noise. Furthermore, we employed a regression task to link clinical scores to disease classification, and incorporated latent domain information into model training through multi-task learning. RESULTS: We validated the proposed approach on the Parkinsons Progression Markers Initiative dataset. Experimental results showed that our MNC-Net effectively separated the early-stage PD from healthy controls(HC) with an accuracy of 95.5%. The t-SNE figures have showed that our graph structure learning method can capture more efficient and discriminatory features. Furthermore, node clustering parameters were used as important weights to extract salient task-related brain regions(ROIs). These ROIs are involved in the development of mood disorders, tremors, imbalances and other symptoms, highlighting the importance of memory, language and mild motor function in early PD. In addition, statistical results from clinical scores confirmed that our model could capture abnormal connectivity that was significantly different between PD and HC. These results are consistent with previous studies, demonstrating the interpretability of our methods.

A Lower False Positive Pulmonary Nodule Detection Approach for Early Lung Cancer Screening.

Pulmonary nodule detection with low-dose computed tomography (LDCT) is indispensable in early lung cancer screening. Although existing methods have achieved excellent detection sensitivity, nodule detection still faces challenges such as nodule size variation and uneven distribution, as well as excessive nodule-like false positive candidates in the detection results. We propose a novel two-stage nodule detection (TSND) method. In the first stage, a multi-scale feature detection network (MSFD-Net) is designed to generate nodule candidates. This includes a proposed feature extraction network to learn the multi-scale feature representation of candidates. In the second stage, a candidate scoring network (CS-Net) is built to estimate the score of candidate patches to realize false positive reduction (FPR). Finally, we develop an end-to-end nodule computer-aided detection (CAD) system based on the proposed TSND for LDCT scans. Experimental results on the LUNA16 dataset show that our proposed TSND obtained an excellent average sensitivity of 90.59% at seven predefined false positives (FPs) points: 0.125, 0.25, 0.5, 1, 2, 4, and 8 FPs per scan on the FROC curve introduced in LUNA16. Moreover, comparative experiments indicate that our CS-Net can effectively suppress false positives and improve the detection performance of TSND.

A Time-Frequency Measurement and Evaluation Approach for Body Channel Characteristics in Galvanic Coupling Intrabody Communication.

Intrabody communication (IBC) can achieve better power efficiency and higher levels of security than other traditional wireless communication technologies. Currently, the majority of research on the body channel characteristics of galvanic coupling IBC are motionless and have only been evaluated in the frequency domain. Given the long measuring times of traditional methods, the access to dynamic variations and the simultaneous evaluation of the time-frequency domain remains a challenge for dynamic body channels such as the cardiac channel. To address this challenge, we proposed a parallel measurement methodology with a multi-tone strategy and a time-parameter processing approach to obtain a time-frequency evaluation for dynamic body channels. A group search algorithm has been performed to optimize the crest factor of multitone excitation in the time domain. To validate the proposed methods, in vivo experiments, with both dynamic and motionless conditions were measured using the traditional method and the proposed method. The results indicate that the proposed method is more time efficient (Tmeas = 1 ms) with a consistent performance (ρc > 98%). Most importantly, it is capable of capturing dynamic variations in the body channel and provides a more comprehensive evaluation and richer information for the study of IBC.

Application of a UPLC-MS/MS method to the protein binding study of TM-2 in rat, human and beagle dog plasma.

TM-2 known as a potential antitumor drug is a novel semi-synthetic taxane derivative. As drug-protein interactions contribute to insights into pharmacokinetic and pharmacodynamic properties, we elucidated the binding of TM-2 to plasma protein. In this study, a simple, rapid and reliable method was developed and validated employing equilibrium dialysis for the separation of bound and unbound drugs and ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) for the quantitation. Protein binding reached equilibrium within 24 h of incubation at 37 °C. After liquid-liquid extraction with methyl tert-butyl ether, the samples were separated on Thermo Syncronis UPLC® C18 (2.1 mm×50 mm, 1.7 µm), and acquisition of mass spectrometric data was performed in multiple reaction monitoring (MRM) mode via positive electrospray ionization. The assay was linear over the concentration rang of 5-2000 ng/mL. The intra- and inter-day precisions were 0.1%-14.8%, and the accuracy was from -6.4% to 7.0%. This assay has been successfully applied to a protein binding study of TM-2 in rat, human and beagle dog plasma. TM-2 showed high protein binding of 81.4%±6.5% (rat), 87.9%±3.6% (human) and 79.4%±4.0% (beagle dog). The results revealed that there was an insignificant difference among the three species.

Simultaneous determination and pharmacokinetic study of Atractylenolide I, II and III in rat plasma after intragastric administration of Baizhufuling extract and Atractylodis extract by UPLC-MS/MS.

A simple and rapid ultra high performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) method was developed for the simultaneous determination of Atractylenolide I, II and III in rat plasma. Plasma samples were processed by liquid-liquid extraction with ethyl acetate, using schisandrin as internal standard (IS). Chromatographic separation was accomplished on a Thermo Hypersil GOLD C18 column (2.1mm×50mm, 1.9µm) with mobile phase consisting of acetonitrile and 0.1% formic acid-water (50:50, v/v). The detection was carried out by ESI-MS (positive ionization mode) and low-energy collision dissociation tandem mass spectrometric analyses using the multiple-reaction monitoring (MRM) scan mode. The quantification was performed using the transitions of the protonated molecule→product ion at m/z 231.0→185.1 for Atractylenolide I, at m/z 233.1→187.1 for Atractylenolide II and at m/z 249.1→231.1 for Atractylenolide III, respectively. Method validation revealed excellent linearity over investigated range together with satisfactory intra- and inter-day precision, accuracy, matrix effects and extraction recoveries. This method was successfully applied to the comparative pharmacokinetic study of Atractylenolide I, II and III in rat plasma after intragastric administration of Baizhufuling extract and Atractylodis extract.

Application of an LC-MS/MS method to the pharmacokinetics of TM-2, a potential antitumour agent, in rats.

TM-2 is a novel semi-synthetic taxane derivative, selected for preclinical development based on its greater anticancer activity and lower toxicity compared with docetaxel. In this study, a rapid and sensitive analytical method based on ultra performance liquid chromatography-tandem mass spectrometry (LC-MS/MS) has been developed for the determination of TM-2 in rat plasma. The biological samples were extracted with methyl tert-butyl ether and separated on a C18 column (50 mm × 2.1 mm, 1.7 µm) using a mobile phase consisting of acetonitrile and 2 mM ammonium acetate. The standard curves were linear over the range 5-1000 ng/mL in rat plasma. The precision (relative standard deviation, RSD, %) were within 14.5%, and the accuracy (relative error, RE, %) ranged from -1.56 to 2.36%. Recovery and matrix effect were satisfactory in rat plasma. The validated method was successfully applied to pharmacokinetic studies after intravenous administration of TM-2 to rats. The pharmacokinetics of TM-2 in rats were characterized by a large volume of distribution and a long half-life of elimination after single dose (4, 8, and 16 mg/kg), and a good correlation was observed between AUC and dose. The preclinical data will be useful for the design of subsequent trials of TM-2.

Simultaneous determination of triamcinolone acetonide palmitate and triamcinolone acetonide in beagle dog plasma by UPLC-MS/MS and its application to a long-term pharmacokinetic study of triamcinolone acetonide palmitate lipid emulsion injection.

In order to investigate the pharmacokinetics of triamcinolone acetonide palmitate (TAP) which is a lipid-soluble prodrug of triamcinolone acetonide (TA), a rapid, simple, sensitive and reproducible UPLC-MS/MS method has been developed and validated for the simultaneous determination of TAP and TA in beagle dog plasma. After simple liquid-liquid extraction, the analytes and internal standard (dexamethasone, DEX) were separated on Phenomenex Luna C18 column (50 mm × 2.1mm, 1.7 µm) using a mobile phase consisting of solvent A (acetonitrile) and solvent B (0.1% ammonia solution) at a flow rate of 0.2 ml/min with gradient elution. Acquisition of mass spectrometric data was performed in multiple reaction monitoring (MRM) mode via positive electrospray ionization using the ion transitions of m/z 673.5→397.3, 435.3→415.3 and 393.3→355.3 for TAP, TA and IS, respectively. The method was of satisfactory specificity, sensitivity, precision and accuracy over the concentration range of 1-1,000 ng/ml for TAP and 0.5-500 ng/ml for TA. The intra- and inter-day precisions for both TAP and TA were 3.2% to 18.7% and the accuracy was in the range of -8.4% to 6.8%. The mean recoveries of TAP, TA and IS were 86.7-104.7%. The method was successfully applied to a long-term pharmacokinetic study of TAP and TA after 28-day repeated intravenous administration of TAP lipid emulsion injection to beagle dogs.

Development of a rapid and sensitive UPLC-MS/MS assay for the determination of TM-2 in beagle dog plasma and its application to a pharmacokinetic study.

A simple and sensitive method based on ultra-high-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) has been developed for the determination of TM-2, which was a novel semi-synthetic taxane derivative in beagle dog plasma. Cabazitaxel was chosen as internal standard. Following extraction by methyl tert-butyl ether, the chromatographic separation was achieved on a Thermo Syncronis C18 column (50 × 2.1 mm, 1.7 µm) by gradient elution within a runtime of 3.5 min. The mobile phase consisted of (A) acetonitrile and (B) 2 mmol/L ammonium acetate in water. The detection was accomplished using positive ion electrospray ionization in multiple reaction monitoring mode. The MS/MS ion transitions were monitored at m/z 812.39 → 551.35 for TM-2 and 836.36 → 555.26 for IS, respectively. The method was linear for TM-2 (r = 0.9924) ranging from 2.5 to 1000 ng/mL. The intra-day and inter-day precisions (relative standard deviation) were within 8.0 and 17.6%, respectively, and the accuracy (relative error) was less than 2.3%. The extraction recovery ranged from 83.1 to 97.1%. The reliable method was successfully applied to a pharmacokinetic study of TM-2 in beagle dogs after intravenous drip with different doses of 0.6, 1.2, and 2.4 mg/kg, respectively.

Characterization of in vitro metabolites of TM-2, a potential antitumor drug, in rat, dog and human liver microsomes using liquid chromatography/tandem mass spectrometry.

RATIONALE: TM-2 (13-(N-Boc-3-i-butylisoserinoyl-4,10-ß-diacetoxy-2-α-benzoyloxy-5-ß,20-epoxy-1,13-α-dihydroxy-9-oxo-19-norcyclopropa[g]tax-11-ene) is a novel semi-synthetic taxane derivative. Our previous study demonstrated that it is a promising taxane derivative. The in vitro comparative metabolic profile of a drug between animals and humans is a key issue that should be investigated at early stages of drug development to better select drug candidates. In this study, the in vitro metabolic pathways of TM-2 in rat, dog and human liver microsomes were established and compared. METHODS: TM-2 was incubated with liver microsomes in the presence of NADPH. Two different types of mass spectrometers - a hybrid linear trap quadrupole orbitrap (LC/LTQ-Orbitrap) mass spectrometer and a triple-quadrupole tandem mass spectrometer (LC/QqQ) were employed to acquire structural information of TM-2 metabolites. Accurate mass measurement using LC/LTQ-Orbitrap was used to determine the accurate mass data and elemental compositions of metabolites thereby confirming the proposed structures of the metabolites. For the chemical inhibition study, selective P450 inhibitors were added to incubations to initially characterize the cytochrome P450 (CYP) enzymes involved in the metabolism of TM-2. RESULTS: A total of 12 components (M1-M12) were detected and identified as the metabolites of TM-2 in vitro. M1-M5 were formed by hydroxylation of the taxane ring or the lateral chain. Hydroxylated products can be further oxidized to the dihydroxylated metabolites M6-M10. M11 was a trihydroxylated metabolite. M12 was tentatively identified as a carboxylic acid derivative. The metabolism of TM-2 is much the same in all three species with some differences. The chemical inhibition study initially demonstrated that the formation of M2, the major metabolite of TM-2, is mainly mediated by CYP3A4. CONCLUSIONS: Hydroxylation is the major biotransformation of the TM-2 pathway in vitro. CYP3A4 may play a dominant role in the formation of M2 in liver microsomes. The knowledge of the metabolic pathways of TM-2 is important to support further research of TM-2.

Metabolism of TM-2, a potential antitumor drug, in rats by using LC-MS.

TM-2 (13-(N-Boc-3-i-butylisoserinoyl-4,10-ß-diacetoxy-2-α-benzoyloxy-5-ß-20-epoxy-1,13-α-dihydroxy-9-oxo-19-norcyclopropa[g]tax-11-ene) is a novel semisynthetic taxane derivative. Our previous study suggested that TM-2 is a promising antitumor analogue. In this paper, the metabolism of TM-2 was investigated in rats following intravenous administration. Two different types of mass spectrometry-hybrid linear trap quadrupole orbitrap (LTQ-Orbitrap) mass spectrometry and triple-quadrupole tandem (QQQ) mass spectrometry-were employed to acquire structural information of TM-2 metabolites. A total of 17 components were identified as the metabolites of TM-2 in bile, feces, and urine samples. Accurate mass measurement using LC-LTQ-Orbitrap-MS was used to determine the accurate mass data and elemental composition of metabolites thereby confirming the proposed structures of the metabolites. The metabolites proposed were mainly oxidates of TM-2, including methoxy, hydroxyl, dihydroxy, and trihydroxyl analogues. The major metabolic pathway of TM-2 was the hydroxylation of the taxane ring or the lateral chain. These important metabolic data serve as a useful resource to support further research of TM-2.

iGAIT: an interactive accelerometer based gait analysis system.

This paper presents a software program (iGAIT) developed in MATLAB, for the analysis of gait patterns extracted from accelerometer recordings. iGAIT provides a user-friendly graphical interface to display and analyse gait acceleration data recorded by an accelerometer attached to the lower back of subjects. The core function of iGAIT is gait feature extraction, which can be used to derive 31 features from acceleration data, including 6 spatio-temporal features, 7 regularity and symmetry features, and 18 spectral features. Features extracted are summarised and displayed on screen, as well as an option to be stored in text files for further review or analysis if required. Another unique feature of iGAIT is that it provides interactive functionality allowing users to manually adjust the analysis process according to their requirements. The system has been tested under Window XP, Vista and Window 7 using three different types of accelerometer data. It is designed for analysis of accelerometer data recorded with sample frequencies ranging from 5 Hz to 200 Hz.

A machine learning approach to assessing gait patterns for Complex Regional Pain Syndrome.

Complex Regional Pain Syndrome (CRPS) is a condition that causes a long-term burning pain in a limb or part of a limb and it can cause various degrees of the physical functional performance deterioration. Objective assessment of physical functional performance of patients is one critical component to evaluate the therapy outcome for CRPS. This paper aims to investigate the feasibility of assessing the physical performance of patients with Complex Regional Pain Syndrome based on the analysis of gait data recorded by an accelerometer in short walking distances. Ten subjects with CRPS and ten control subjects were recruited. Thirty three features were extracted from each recording. A machine learning method, Multilayer perceptron neural-networks (MLP), was applied to classify the normal and abnormal gait patterns from data obtained on a 2.4m performance evaluation test. The best classification accuracy (99.38%) was achieved using 3 features selected by a step-wise-forward method. To further validate its performance, an independent test set including 14 cases extracted from data obtained on a 20 m performance evaluation test was adopted. A prediction accuracy of 85.7% was obtained.

Mutual information-based approach to the analysis of dynamic electrocardiograms.

Dynamic electrocardiogram (ECG) plays an important role in the analysis of heart movement and cardio-diseases. In an attempt to obtain a better understanding of the mechanisms of patterns and differences found in dynamic ECGs, techniques based on different theories such as chaos and fractal theory have been used to extract nonlinear information encoding in dynamics ECG signals. In this paper, we propose an information theory approach to supporting the analysis of dynamic ECG recorded during different time of a day. Mutual information of R-R intervals extracted from four subject groups were calculated and analysed. Results indicate that heart movement is similar to chaotic movement in many ways. Moreover, the mutual information of R-R intervals exhibits different patterns over different periods of a day and different subject groups, suggesting that it would be a useful tool to support classification analysis of heart movement and cardio related diseases.