Left ventricular ejection fraction estimation using mutual information on technetium-99m multiple-gated SPECT scans.

BACKGROUND: A new non-linear approach was applied to calculate the left ventricular ejection fraction (LVEF) using multigated acquisition (MUGA) images. METHODS: In this study, 50 patients originally for the estimation of the percentage of LVEF to monitor the effects of various cardiotoxic drugs in chemotherapy were retrospectively selected. All patients had both MUGA and echocardiography examinations (ECHO LVEF) at the same time. Mutual information (MI) theory was utilized to calculate the LVEF using MUGA imaging (MUGA MI). RESULTS: MUGA MI estimation was significantly different from MUGA LVEF and ECHO LVEF, respectively (p < 0.005). The higher repeatability for MUGA MI can be observed in the figure by the higher correlation coefficient for MUGA MI (r = 0.95) compared with that of MUGA LVEF (r = 0.80). Again, the reproducibility was better for MUGA MI (r = 0.90, 0.92) than MUGA LVEF (r = 0.77, 0.83). The higher correlation coefficients were obtained between proposed MUGA MI and ECHO LVEF compared to that between the conventional MUGA LVEF and ECHO LVEF. CONCLUSIONS: MUGA image with the aid of MI is promising to be more interchangeable LVEF to ECHO LVEF measurement as compared with the conventional approach on MUGA image.

Left ventricle segmentation in dynamic cardiac CT using random walks method.

PURPOSE: Segmentation of the left ventricle (LV) in cardiac CT (CCT) images is difficult due to the intensity heterogeneity arising from accumulation of contrast agent in papillary muscle and trabeculae carneae. In this study, we demonstrated the random walks method for LV segmentation in CCT through cardiac phases. METHODS: 63 CCT data sets from 7 patients with 9 cardiac phases were included in this study. All cardiac CT examinations were performed with GE 64-detector CT scanner with ECG gating. In each patient, 60-80 ml iohexol was injected at a flow rate of 5 ml/sec followed by 60 ml normal saline solution. Random walks (RW) based on probability of labels was used for LV segmentation. The LV delineations generated by the experienced physician (MD), conventional image-based method (IB), and RW were compared. RESULTS: In general the contours segment the LV closely by RW and MD, but the discrepancies in papillary muscle and trabeculae carneae were observed while using the IB method. CONCLUSION: We showed the RW method potentially improved LV segmentation as compared to the volume by conventional IB method. In this study, we demonstrated the clinical feasibility of LV volume segmentation using random walks algorithm.

Ventricular hemodynamics using cardiac computed tomography and optical flow method.

Ventricular hemodynamics plays an important role in assessing cardiac function in clinical practice. The aim of this study was to determine the ventricular hemodynamics based on contrast movement in the left ventricle (LV) between the phases in a cardiac cycle recorded using an electrocardiography (ECG) with cardiac computed tomography (CT) and optical flow method. Cardiac CT data were acquired at 120 kV and 280 mA with a 350 ms gantry rotation, which covered one cardiac cycle, on the 640-slice CT scanner with ECG for a selected patient without heart disease. Ventricular hemodynamics (mm/phase) were calculated using the optical flow method based on contrast changes with ECG phases in anterior-posterior, lateral and superior-inferior directions. Local hemodynamic information of the LV with color coating was presented. The visualization of the functional information made the hemodynamic observation easy.

Lapatinib-mediated cyclooxygenase-2 expression via epidermal growth factor receptor/HuR interaction enhances the aggressiveness of triple-negative breast cancer cells.

Lapatinib, a dual epidermal growth factor receptor (EGFR)/human epidermal growth factor receptor 2 (HER2) kinase inhibitor, showed clinical benefits in advanced HER2-positive breast cancer patients. Because some triple-negative breast cancers (TNBCs) frequently overexpress EGFR, the antitumor activity of lapatinib in such diseases was also tested. However, the results showed a worse event-free survival rate. It remains unknown whether and how lapatinib elicits the aggressiveness of such cancer cells. In this study, our results demonstrated that lapatinib facilitated axillary and lung metastases of triple-negative MDA-MB-231 breast cancer cells without affecting their viability, leading to worse survival in orthotopic xenograft mice. The lapatinib-increased motility was attributed by the elevation of EGFR through the downregulation of microRNA-7 and by the subsequent overexpression of cyclooxygenase-2 (COX-2). Strikingly, independent of its kinase activity, the elevated EGFR at least partly stabilized COX-2 expression by enhancing the binding of HuR to COX-2 mRNA. Our results suggest that lapatinib may increase the migration and invasion of MDA-MB-231 cells by upregulating EGFR and COX-2 through the downregulation of microRNA-7, providing a potential explanation for the worse clinical outcome of TNBC patients who receive lapatinib-based treatment. These findings also shed new light on the molecular mechanism of COX-2 mRNA stabilization by EGFR in a kinase-independent manner.

Quantitative flow measurement by digital subtraction angiography in cerebral carotid stenosis using optical flow method.

OBJECTIVE: We analyzed intracranial regional blood flows using an optical flow method (OFM) and digital subtraction angiography in patients with internal carotid artery (ICA) stenosis. We also retrospectively explored the correlation between the patients' diagnoses and the severity of the ICA stenoses. MATERIALS AND METHODS: OFM, an image-processing algorithm to estimate motion, was applied to determine the mean velocity V(mean) in the vessels. A group of 40 patients without vascular anomalies acted as the control group. The patients were classified as having either moderate stenosis (< 80%, n=14) or severe stenosis (> 80%, n=23). RESULTS: The V(mean) of the ICAs was significantly lower in the stenotic group compared with the control group (p< 0.01). The V(mean) of the ICAs was inversely correlated with the severity of the stenosis (p< 0.05). The receiver operating characteristic curve of the V(mean) in an AP view showed substantial discriminatory power, with an optimal cutoff value of 3.48 pixels/frame for the detection of patients with carotid stenosis. The sensitivity and specificity were 84% and 50%, respectively. On a lateral view, the best cutoff for the V(mean) was 4.01 pixels/frame, and the sensitivity and specificity were 92% and 43%, respectively. CONCLUSIONS: Digital subtraction angiography combined with the OFM is a feasible parametric method for intracranial blood flow measurements in patients with moderate to severe carotid stenosis.

Peritherapeutic quantitative flow analysis of arteriovenous malformation on digital subtraction angiography.

BACKGROUND: Digital subtraction angiography (DSA) provides detailed hemodynamic information. However, the imaging interpretation is mainly based on the physician's experience and observation. We aimed to quantitatively study the peritherapeutic blood flow changes of a cerebral arteriovenous malformation (AVM) treated by embolization using optical flow estimation on DSA. METHODS: A 37-year-old woman with an AVM in the right frontal lobe of her brain was enrolled. The optical flow method with a pixel-by-pixel measurement was applied to determine the blood flow in brain vessels on anterior-posterior and lateral DSA views before and after embolization. RESULTS: A return toward normalization of blood flow as a result of embolization was determined semiquantitatively on the posttherapeutic DSA. CONCLUSIONS: Optical flow analysis on DSA illustrated the potential of quantifying intracranial blood flows in patients with cerebral vascular disorders and the therapeutic effects.

Quantitative portal vein velocity of liver cancer patients with transcatheter arterial chemoembolization on angiography.

OBJECTIVE: We applied optical flow method (OFM) to quantify relative velocities of blood flow using digital subtraction angiography (DSA) in the vascular analysis of hepatocellular carcinoma (HCC) patients who underwent transarterial chemoembolization (TACE) treatment. METHODS: A total of 40 HCC patients treated by TACE were analyzed in this study. DSA imaging with a 12-inch field of view, 1024 × 1024 pixels and 4 frames/second was acquired. OFM developed for motion estimation is applied for blood flow estimation. Two acrylic phantoms were built to validate the method. RESULTS: The relationship between the OFM and Doppler measurements was found linear with R(2) = 0.99 for both straight and curved tube phantoms. Quantitative blood flow distribution images of the portal vein region were presented. After TACE, the minimum, maximum and mean velocities in the portal vein all decreased (P < 0.05). Additionally, the velocity in the portal vein is significantly lower with a higher Child-Pugh score (P < 0.01). CONCLUSIONS: The present technique provides add-on quantitative information of flows to DSA and the hemodynamic analysis in relative quantifications of blood flow in portal vein of hepatocellular carcinoma patients using DSA.

Hemodynamic analysis of vascular stenting treatment outcome: computational fluid dynamics method vs optical flow method.

BACKGROUND AND PURPOSE: Computational fluid dynamics method (CFDM) and optical flow method (OFM) effectively provide the hemodynamic information based on the digital subtraction angiogram (DSA). However, the quantitative analysis in comparison of CFDM and OFM is still absent. The goal of this study is to apply CFDM and OFM in quantitative analysis of stenting treatment. MATERIAL AND METHOD: A left carotid stenosis patient underwent stenting of percutaneous transluminal angioplasty was analyzed as an example. CFDM and OFM for hemodynamic analysis on digital subtraction angiography before and after stenting treatment were presented. RESULTS: Improvement gains of blood flow velocities on left internal carotid artery after stenting treatment for different initial conditions on the common carotid artery were 1.91 ∼ 2.13, 1.62 ∼ 2.09, and 0.69 by CFDM with Newtonian and non-Newtonian fluids and OFM, respectively. With the CFDM analysis, the flow mapping by OFM using time resolved DSA data on the fly to estimate hemodynamic significance of a cervical carotid stenosis was explained. CONCLUSION: Quantificative blood flow estimations by CFDM and OFM to evaluate the treatment outcomes to patient with carotid stenosis are practical. Both methods are able to provide quantitative information of blood flow for stenting treatment. It is advantagious to use both methods in treatment evaluation.

Hemodynamic analysis of capillary in finger nail-fold using computational fluid dynamics and image estimation.

Red blood cell (RBC) dynamics in capillaries is a useful diagnostic tool for many diseases. Previous study showed that optical flow estimation (OFE) is capable of accurately calculating RBC velocities using image registration technique. The computational fluid dynamics (CFD) method is explored in this study to calculate the RBC velocity in capillaries of finger nail-fold for six cases. The two-dimensional capillary images were reconstructed to three-dimensional, assuming circular cross sections. The no-slip boundary conditions were applied on the vessel walls. The initial velocity of the RBC going into each capillary was calculated by OFE. The velocities of multiple points along each capillary calculated by CFD, V(CFD), were compared with OFE calculations, V(OFE). The calculated RBC velocity was in the range of 56-685µm/s. The average difference (V(CFD) - V(OFE)) with one standard deviation is -2.66±18.61µm/s for all the 48 calculation points, and 0.03±0.12µm/s for all except one points (47 points), indicating that CFD can provide a reasonable accuracy in RBC velocity calculation in finger nail-fold capillaries.

Accuracy evaluation of RBC velocity measurement in nail-fold capillaries.

Cutaneous red blood cell velocity in vivo can be measured by using capillaroscopy with image processing techniques. However, unlike simulated blood flow images, there is no standard to determine the accuracy of the techniques for computing blood flow velocities. In this paper, we quantitatively evaluated the accuracy of previously proposed optical flow method for measuring red blood cell velocity in nail-fold capillaries. Blood flow images of subjects under normal and occlusion-release conditions were examined by a capillaroscope. To obtain velocity values, the images were further analyzed by using optical flow, cross-correlation and visual inspection methods, respectively. Visual inspection method was taken as the golden standard to determine the accuracy of blood flow velocity measurement using optical flow and cross-correlation techniques. Results showed that optical flow estimation provided superior accuracy to cross-correlation when assessing real blood flow velocity in nail-fold capillaries. Optical flow estimation is able to measure red blood cell velocity with a high accuracy of 91% and 86% when the observed velocity is less than 0.5mm/s under normal and occlusion-release conditions, respectively. In addition, optical flow method showed good agreement with visual inspection in determining blood flow velocity in both normal and occlusion-release conditions when the high-velocity zone is excluded.

Usefulness of Machine Learning-Based Detection and Classification of Cardiac Arrhythmias With 12-Lead Electrocardiograms.

BACKGROUND: Deep-learning algorithms to annotate electrocardiograms (ECGs) and classify different types of cardiac arrhythmias with the use of a single-lead ECG input data set have been developed. It remains to be determined whether these algorithms can be generalized to 12-lead ECG-based rhythm classification. METHODS: We used a long short-term memory (LSTM) model to detect 12 heart rhythm classes with the use of 65,932 digital 12-lead ECG signals from 38,899 patients, using annotations obtained by consensus of 3 board-certified electrophysiologists as the criterion standard. RESULTS: The accuracy of the LSTM model for the classification of each of the 12 heart rhythms was ≥ 0.982 (range 0.982-1.0), with an area under the receiver operating characteristic curve of ≥ 0.987 (range 0.987-1.0). The precision and recall ranged from 0.692 to 1 and from 0.625 to 1, respectively, with an F1 score of ≥ 0.777 (range 0.777-1.0). The accuracy of the model (0.90) was superior to the mean accuracies of internists (0.55), emergency physicians (0.73), and cardiologists (0.83). CONCLUSIONS: We demonstrated the feasibility and effectiveness of the deep-learning LSTM model for interpreting 12 common heart rhythms according to 12-lead ECG signals. The findings may have clinical relevance for the early diagnosis of cardiac rhythm disorders.

Experimental estimation of blood flow velocity through simulation of intravital microscopic imaging in micro-vessels by different image processing methods.

Quantization of red blood cell (RBC) velocity in micro-vessel is one of the techniques for dynamic observation of microvascular mechanisms. The flow measurement of RBC in micro-vessels is still a challenge nowadays. Image processing for velocity measurement using a frame by frame analysis is a common approach. The accuracy of the calculations, which is algorithm dependant, has rarely been examined. In this paper, we evaluated the accuracy of the existing methods, which includes cross correlation method, Hough transform method, and optical flow method, by applying these methods to simulated micro-vessel image sequences. Simulated experiments in various micro-vessels with random RBC motion were applied in the evaluation. The blood flow variation in the same micro-vessels with different RBC densities and velocities was considered in the simulations. The calculation accuracy of different flow patterns and vessel shapes were also examined, respectively. Based on the comparison, the use of an optical flow method, which is superior to a cross-correlation method or a Hough transform method, is proposed for measuring RBC velocity. The study indicated that the optical flow method is suitable for accurately measuring the velocity of the RBCs in small or large micro-vessels.

Red blood cell velocity measurements of complete capillary in finger nail-fold using optical flow estimation.

A new approach for the measurement of the red blood cell (RBC) velocity from capillary video by using optical flow estimation has been developed. An image registration function based on mutual information was used for stabilizing images in order to cope with slight finger movement during video acquisition. After image alignment, a skeleton extraction algorithm implemented by thinning was followed which enabled tracking blood flow entirely in arteriolar and venular limbs, and the curved segment as well. Optical flow and cross-correlation approaches were applied individually for velocity estimation of twelve microcirculation videos acquired independently from three healthy volunteers. The RBC velocity of 12 vessels at three given measurement sites (arteriolar, curve and venular sites) in a 45-second period of occlusion-release condition of vessel were examined. There were four stages of flow conditions: resting (T(1)), pre-occlusion (T(2)), post-occlusion (T(3)) and release (T(4)). The results from both approaches revealed that the velocity difference among the three sites were not significant. The pattern of distribution of RBC velocity was also reported. The correlation coefficient (r) of the velocity calculated using optical flow and cross-correlation in four stages of blood flow conditions and the overall correlation were: 1-window: r(T1)=0.68, r(T2)=0.67, r(T3)=0.92, r(T4)=0.88 and r(All)=0.79; 2-window: r(T1)=0.84, r(T2)=0.88, r(T3)=0.87, r(T4)=0.93 and r(All)=0.88. The averaged velocity results showed no significant differences between optical flow and 2-window cross-correlation in all flow conditions. Optical flow estimation is not only independent to the direction of flow, but also able to calculate the intensity displacement of all pixels. The proposed velocity measurement system has been shown to provide complete velocity information for the whole vessel limb which demonstrates the advantage of measuring blood flow at the level of microcirculation more accurately.

Detection and classification the breast tumors using mask R-CNN on sonograms.

Breast cancer is one of the most harmful diseases for women with the highest morbidity. An efficient way to decrease its mortality is to diagnose cancer earlier by screening. Clinically, the best approach of screening for Asian women is ultrasound images combined with biopsies. However, biopsy is invasive and it gets incomprehensive information of the lesion. The aim of this study is to build a model for automatic detection, segmentation, and classification of breast lesions with ultrasound images. Based on deep learning, a technique using Mask regions with convolutional neural network was developed for lesion detection and differentiation between benign and malignant. The mean average precision was 0.75 for the detection and segmentation. The overall accuracy of benign/malignant classification was 85%. The proposed method provides a comprehensive and noninvasive way to detect and classify breast lesions.

Radiomics features analysis of PET images in oropharyngeal and hypopharyngeal cancer.

This study used radiomics image analysis to examine the differences of texture feature values extracted from oropharyngeal and hypopharyngeal cancer positron emission tomography (PET) images on various tumor segmentations, and finds the proper and stable feature groups. A total of 80 oropharyngeal and hypopharyngeal cancer cases were retrospectively recruited. Radiomics method was applied to the PET image for the 80 oropharyngeal and hypopharyngeal cancer cases to extract texture features from various defined metabolic volumes. Kruskal-Wallis one-way analysis of variance method was used to test whether feature value difference exists between groups, which were grouped by stage, response to treatment, and recurrence. If there was a significant difference, the corresponding feature cutoff value was applied to the Kaplan-Meier estimator to estimate the survival functions. For the various defined metabolic volumes, there were 16 features that had significant differences between early (T1, T2) and late tumor stages (T3, T4). Five images and 2 textural features were found to be able to predict the tumor response and recurrence, respectively, with the areas under the receiver operating characteristic curves reaching 0.7. The histogram entropy was found to be a good predictor of overall survival (OS) and primary relapse-free survival (PRFS) of oropharyngeal and hypopharyngeal cancer patients. Textural features from PET images provide predictive and prognostic information in tumor staging, tumor response, recurrence, and have the potential to be a prognosticator for OS and PRFS in oropharyngeal and hypopharyngeal cancer.

Use of three-dimensional (3D) optical flow method in mapping 3D anatomic structure and tumor contours across four-dimensional computed tomography data.

A 3-dimensional (3D) optical flow program that includes a multi-resolution feature has been developed and applied to 3D anatomical structure and gross tumor volume (GTV) contour mapping for 4-dimensional (4D) CT data. The study includes contour mapping for 3 real patient CT data sets, and also for a thoracic phantom in which the displacement for each voxel is known. Of the real patient CT data sets, one set has been used to map contours of lung and GTV over all the respiration phases, while the others were studied using only the end inspiration and end expiration phases, in which the displacement between the phases were the largest. Including the residual motion in the 4D CT data and motion table shaking, the optical flow calculation agrees to within 1 mm with the known displacement. Excluding those errors that are not introduced by optical flow algorithm, the agreement can be within 0.1 mm with a displacement magnitude of 24 mm. The mapped contours of lungs, liver, esophagus, GTV, etc. in real patient 4D CT images were acceptable to clinicians. The 3D optical flow program is a good tool for anatomical structure and tumor volume contour mapping across 4D CT scans.

Dose mapping: validation in 4D dosimetry with measurements and application in radiotherapy follow-up evaluation.

The purpose of this paper is to validate a dose mapping program using optical flow method (OFM), and to demonstrate application of the program in radiotherapy follow-up evaluation. For the purpose of validation, the deformation matrices between four-dimensional (4D) CT data of different simulated respiration phases of a phantom were calculated using OFM. The matrices were then used to map doses of all phases to a single-phase image, and summed in equal time weighting. The calculated dose should closely represent the dose delivered to the moving phantom if the deformation matrices are accurately calculated. The measured point doses agreed with the OFM calculations better than 2% at isocenters, and dose distributions better than 1mm for the 50% isodose line. To demonstrate proof-of-concept for the use of deformable image registration in dose mapping for treatment evaluation, the treatment-planning CT was registered with the post-treatment CT image from the positron emission tomography (PET)/CT resulting in a deformation matrix. The dose distribution from the treatment plan was then mapped onto the restaging PET/CT using the deformation matrix. Two cases in which patients had thoracic malignancies are presented. Each patient had CT-based treatment planning for radiotherapy and restaging fluorodeoxy glucose (FDG)-PET/CT imaging 4-6 weeks after completion of treatments. Areas of pneumonitis and recurrence were identified radiographically on both PET and CT restaging images. Local dose and standard uptake values for pneumonitis and recurrence were studied as a demonstration of this method. By comparing the deformable mapped dose to measurement, the treatment evaluation method which is introduced in this manuscript proved to be accurate. It thus provides a more accurate analysis than other rigid or linear dose-image registration when used in studying treatment outcome versus dose.

Quantitative Motion Analysis of Tai Chi Chuan: The Upper Extremity Movement.

The quantitative and reproducible analysis of the standard body movement in Tai Chi Chuan (TCC) was performed in this study. We aimed to provide a reference of the upper extremities for standardizing TCC practice. Microsoft Kinect was used to record the motion during the practice of TCC. The preparation form and eight essential forms of TCC performed by an instructor and 101 practitioners were analyzed in this study. The instructor completed an entire TCC practice cycle and performed the cycle 12 times. An entire cycle of TCC was performed by practitioners and images were recorded for statistics analysis. The performance of the instructor showed high similarity (Pearson correlation coefficient (r) = 0.71 ~ 0.84) to the first practice cycle. Among the 9 forms, lay form had the highest similarity (rmean = 0.90) and push form had the lowest similarity (rmean = 0.52). For the practitioners, ward off form (rmean = 0.51) and roll back form (rmean = 0.45) had the highest similarity with moderate correlation. We used Microsoft Kinect to record the spatial coordinates of the upper extremity joints during the practice of TCC and the data to perform quantitative and qualitative analysis of the joint positions and elbow joint angle.

Radiomic features analysis in computed tomography images of lung nodule classification.

PURPOSE: Radiomics, which extract large amount of quantification image features from diagnostic medical images had been widely used for prognostication, treatment response prediction and cancer detection. The treatment options for lung nodules depend on their diagnosis, benign or malignant. Conventionally, lung nodule diagnosis is based on invasive biopsy. Recently, radiomics features, a non-invasive method based on clinical images, have shown high potential in lesion classification, treatment outcome prediction. METHODS: Lung nodule classification using radiomics based on Computed Tomography (CT) image data was investigated and a 4-feature signature was introduced for lung nodule classification. Retrospectively, 72 patients with 75 pulmonary nodules were collected. Radiomics feature extraction was performed on non-enhanced CT images with contours which were delineated by an experienced radiation oncologist. RESULT: Among the 750 image features in each case, 76 features were found to have significant differences between benign and malignant lesions. A radiomics signature was composed of the best 4 features which included Laws_LSL_min, Laws_SLL_energy, Laws_SSL_skewness and Laws_EEL_uniformity. The accuracy using the signature in benign or malignant classification was 84% with the sensitivity of 92.85% and the specificity of 72.73%. CONCLUSION: The classification signature based on radiomics features demonstrated very good accuracy and high potential in clinical application.

The clinical outcome correlations between radiation dose and pretreatment metabolic tumor volume for radiotherapy in head and neck cancer: A retrospective analysis.

This study was to investigate the clinical outcomes between radiation dose and pretreatment metabolic tumor volume (MTV) in patients with head and neck cancer treated with definitive chemoradiotherapy.Thirty-four patients received pretreatment F-fluorodeoxyglucose (F-FDG) positron emission tomography-computed tomography (PET/CT) were recruited for this study. The CT-based volume (gross tumor volume of the primary [GTVp]) and 4 types of MTVs were measured on the basis of either a maximal standardized uptake value (SUVmax) of 2.5 (MTV2.5), 3.0 (MTV3.0), or a fixed threshold of 40% (MTV40%), 50% (MTV50%). F-FDG PET-CT images before treatment, and data including response to treatment, local recurrence, death due to the cancer, disease-free survival (DFS) and primary relapse-free survival (PRFS), were collected for analysis.The Wilcoxon rank test showed that all values determined by the different delineation techniques were significantly different from the GTVp (P < .05). Tumor volume and the homogeneity of target dose of MTV2.5, MTV3.0, MTV40%, and MTV50% were significantly different between the 2 groups of patients through treatment outcomes (P < .05).The survival curves for DFS and PRFS demonstrated that the homogeneity of the target dose in MTVs was a good indicator. The homogeneity of target dose in the tumor is a potential indicator of DSF and PRFS in patients with head and neck cancer who underwent radiotherapy.