Improved PAA algorithm for breast mass detection in mammograms.

Mammography screening is instrumental in the early detection and diagnosis of breast cancer by identifying masses in mammograms. With the rapid development of deep learning, numerous deep learning-based object detection algorithms have been explored for mass detection studies. However, these methods often yield a high false positive rate per image (FPPI) while achieving a high true positive rate (TPR). To maintain a higher TPR while also ensuring lower FPPI, we improved the Probability Anchor Assignment (PAA) algorithm to enhance the detection capability for mammographic characteristics with our previous work. We considered three dimensions: the backbone network, feature fusion module, and dense detection heads. The final experiment showed the effectiveness of the proposed method, and the TPR/FPPI values of the final improved PAA algorithm were 0.96/0.56 on the INbreast datasets. Compared to other methods, our method stands distinguished with its effectiveness in addressing the imbalance between positive and negative classes in cases of single lesion detection.

Deep learning for differentiation of osteolytic osteosarcoma and giant cell tumor around the knee joint on radiographs: a multicenter study.

OBJECTIVES: To develop a deep learning (DL) model for differentiating between osteolytic osteosarcoma (OS) and giant cell tumor (GCT) on radiographs. METHODS: Patients with osteolytic OS and GCT proven by postoperative pathology were retrospectively recruited from four centers (center A, training and internal testing; centers B, C, and D, external testing). Sixteen radiologists with different experiences in musculoskeletal imaging diagnosis were divided into three groups and participated with or without the DL model's assistance. DL model was generated using EfficientNet-B6 architecture, and the clinical model was trained using clinical variables. The performance of various models was compared using McNemar's test. RESULTS: Three hundred thirty-three patients were included (mean age, 27 years ± 12 [SD]; 186 men). Compared to the clinical model, the DL model achieved a higher area under the curve (AUC) in both the internal (0.97 vs. 0.77, p = 0.008) and external test set (0.97 vs. 0.64, p < 0.001). In the total test set (including the internal and external test sets), the DL model achieved higher accuracy than the junior expert committee (93.1% vs. 72.4%; p < 0.001) and was comparable to the intermediate and senior expert committee (93.1% vs. 88.8%, p = 0.25; 87.1%, p = 0.35). With DL model assistance, the accuracy of the junior expert committee was improved from 72.4% to 91.4% (p = 0.051). CONCLUSION: The DL model accurately distinguished osteolytic OS and GCT with better performance than the junior radiologists, whose own diagnostic performances were significantly improved with the aid of the model, indicating the potential for the differential diagnosis of the two bone tumors on radiographs. CRITICAL RELEVANCE STATEMENT: The deep learning model can accurately distinguish osteolytic osteosarcoma and giant cell tumor on radiographs, which may help radiologists improve the diagnostic accuracy of two types of tumors. KEY POINTS: • The DL model shows robust performance in distinguishing osteolytic osteosarcoma and giant cell tumor. • The diagnosis performance of the DL model is better than junior radiologists'. • The DL model shows potential for differentiating osteolytic osteosarcoma and giant cell tumor.

Comparison of six machine learning methods for differentiating benign and malignant thyroid nodules using ultrasonographic characteristics.

BACKGROUND: Several machine learning (ML) classifiers for thyroid nodule diagnosis have been compared in terms of their accuracy, sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and area under the receiver operating curve (AUC). A total of 525 patients with thyroid nodules (malignant, n = 228; benign, n = 297) underwent conventional ultrasonography, strain elastography, and contrast-enhanced ultrasound. Six algorithms were compared: support vector machine (SVM), linear discriminant analysis (LDA), random forest (RF), logistic regression (LG), GlmNet, and K-nearest neighbors (K-NN). The diagnostic performances of the 13 suspicious sonographic features for discriminating benign and malignant thyroid nodules were assessed using different ML algorithms. To compare these algorithms, a 10-fold cross-validation paired t-test was applied to the algorithm performance differences. RESULTS: The logistic regression algorithm had better diagnostic performance than the other ML algorithms. However, it was only slightly higher than those of GlmNet, LDA, and RF. The accuracy, sensitivity, specificity, NPV, PPV, and AUC obtained by running logistic regression were 86.48%, 83.33%, 88.89%, 87.42%, 85.20%, and 92.84%, respectively. CONCLUSIONS: The experimental results indicate that GlmNet, SVM, LDA, LG, K-NN, and RF exhibit slight differences in classification performance.

Viscoelasticity measured by shear wave elastography in a rat model of nonalcoholic fatty liver disease: comparison with dynamic mechanical analysis.

BACKGROUND: Nonalcoholic fatty liver disease (NAFLD) is rapidly becoming one of the most common liver diseases. Ultrasound elastography has been used for the diagnosis of NAFLD. However, clinical research on steatosis by elastography technology has mainly focused on steatosis with fibrosis or non-alcoholic steatohepatitis (NASH), while steatosis without fibrosis has been poorly studied. Moreover, the relationship between liver viscoelasticity and steatosis grade is not clear. In this study, we evaluated the degree of liver steatosis in a simple steatosis rat model using shear wave elastography (SWE). RESULTS: The viscoelasticity values of 69 rats with hepatic steatosis were measured quantitatively by SWE in vivo and validated by a dynamic mechanical analysis (DMA) test. Pathological sections were used to determine the steatosis grade for each rat. The results showed that the elasticity values µ obtained by the two methods followed the same trend, and µ is significantly correlated with liver steatosis. The Pearson's correlation coefficients indicate that [Formula: see text] obtained by SWE is positively linear correlated with DMA (r = 0.628, p = 7.85 × 10-9). However, the viscosity values [Formula: see text] obtained by SWE were relatively independent of those obtained by DMA with a correlation coefficient of - 0.01. The combined Voigt elasticity measurements have high validity in the prediction of steatosis (S0 vs. S1-S4), with an AUROC of 0.755 (95% CI 0.6175-0.8925, p < 0.01) and the optimal cutoff value was 2.08 kPa with a sensitivity of 78% and specificity of 63%. CONCLUSION: SWE might have the feasibility to be introduced as an auxiliary technique for NAFLD patients in clinical settings. However, the viscosity results measured by SWE and DMA are significantly different, because the two methods work in different frequency bands.

Accurate and Feasible Deep Learning Based Semi-Automatic Segmentation in CT for Radiomics Analysis in Pancreatic Neuroendocrine Neoplasms.

Current clinical practice or radiomics studies of pancreatic neuroendocrine neoplasms (pNENs) require manual delineation of the lesions in computed tomography (CT) images, which is time-consuming and subjective. We used a semi-automatic deep learning (DL) method for segmentation of pNENs and verified its feasibility in radiomics analysis. This retrospective study included two datasets: Dataset 1, contrast-enhanced CT images (CECT) of 80 and 18 patients respectively collected from two centers; and Dataset 2, CECT of 56 and 16 patients respectively from two centers. A DL-based semi-automatic segmentation model was developed and validated with Dataset 1 and Dataset 2, and the segmentation results were used for radiomics analysis from which the performance was compared against that based on manual segmentation. The mean Dice similarity coefficient of the trained segmentation model was 81.8% and 74.8% for external validation with Dataset 1 and Dataset 2 respectively. Four classifiers frequently used in radiomics studies were trained and tested with leave-one-out cross-validation strategy. For pathological grading prediction with Dataset 1, the area under the receiver operating characteristic curve (AUC) with semi-automatic segmentation was up to 0.76 and 0.87 respectively for internal and external validation. For recurrence study with Dataset 2, the AUC with semi-automatic segmentation was up to 0.78. All these AUCs were not statistically significant from the corresponding results based on manual segmentation. Our study showed that DL-based semi-automatic segmentation is accurate and feasible for the radiomics analysis in pNENs.

An in vitro study on the antitumor effect of sonodynamic therapy using sinoporphyrin sodium on human glioblastoma cells.

Sonodynamic therapy (SDT) is a promising modality for cancer treatment. Sinoporphyrin sodium (DVDMS), purified from Photofrin II, shows great potential in SDT evidenced by growing studies. The purpose of the current study was to investigate the antitumor effect of SDT combined with DVDMS on human glioblastoma (U87 MG) cell line in vitro. The cellular uptake of DVDMS was investigated by confocal microscopy and IVIS spectrum imaging system. In addition, DVDMS toxicity and anti-tumor effect of SDT were assessed by flow cytometry. The generation of intracellular reactive oxygen species (ROS) was determined using DCFH-DA staining. Simultaneously, fluorescence microscopy was performed to access the destabilization of mitochondrial membrane potential (MMP). The results showed that DVDMS could easily enter the cells and accumulated in the cytoplasm, especially the mitochondria. And the intracellular DVDMS increased with incubation time or concentrations. The results also showed remarkable cytotoxicity of DVDMS-mediated SDT (center frequency: 0.970 MHz; peak-rarefactional pressure: 0.52-MPa; acoustic power: 0.32 W; pulse repetition frequency: 1 Hz; duty cycle: 1-30%; duration: 3 min) on U87 MG cells, while DVDMS alone was non-toxic to the cells. In comparison with the control group, the SDT-treated group showed significant generation of intracellular ROS and loss of MMP at 1 h post-treatment. These results indicated that DVDMS-mediated SDT could induce great cytotoxicity in U87 MG cells via the production of ROS and showed potentials in the treatment for glioblastoma.

Gear Shifting of Quadriceps during Isometric Knee Extension Disclosed Using Ultrasonography.

Ultrasonography has been widely employed to estimate the morphological changes of muscle during contraction. To further investigate the motion pattern of quadriceps during isometric knee extensions, we studied the relative motion pattern between femur and quadriceps under ultrasonography. An interesting observation is that although the force of isometric knee extension can be controlled to change almost linearly, femur in the simultaneously captured ultrasound video sequences has several different piecewise moving patterns. This phenomenon is like quadriceps having several forward gear ratios like a car starting from rest towards maximal voluntary contraction (MVC) and then returning to rest. Therefore, to verify this assumption, we captured several ultrasound video sequences of isometric knee extension and collected the torque/force signal simultaneously. Then we extract the shapes of femur from these ultrasound video sequences using video processing techniques and study the motion pattern both qualitatively and quantitatively. The phenomenon can be seen easier via a comparison between the torque signal and relative spatial distance between femur and quadriceps. Furthermore, we use cluster analysis techniques to study the process and the clustering results also provided preliminary support to the conclusion that, during both ramp increasing and decreasing phases, quadriceps contraction may have several forward gear ratios relative to femur.

Enhanced delivery of paclitaxel liposomes using focused ultrasound with microbubbles for treating nude mice bearing intracranial glioblastoma xenografts.

Paclitaxel liposomes (PTX-LIPO) are a clinically promising antineoplastic drug formulation for the treatment of various extracranial cancers, excluding glioblastoma. A main reason for this is the presence of the blood-brain barrier (BBB) or blood-tumor barrier (BTB), preventing liposomal drugs from crossing at a therapeutically meaningful level. Focused ultrasound (FUS) in conjunction with microbubbles (MBs) has been suggested in many studies to be an effective approach to increase the BBB or BTB permeability. In this study, we investigated the feasibility of enhancing the delivery of PTX-LIPO in intracranial glioblastoma-bearing nude mice using pulsed low-intensity FUS exposure in the presence of MBs. Our results showed that the delivery efficiency of PTX-LIPO could be effectively improved in terms of the penetration of both the BBB in vitro and BTB in vivo by pulsed FUS sonication with a 10 ms pulse length and 1 Hz pulse repetition frequency at 0.64 MPa peak-rarefactional pressure in the presence of MBs. Quantitative analysis showed that a 2-fold higher drug concentration had accumulated in the glioblastoma 3 h after FUS treatment, with 7.20±1.18 µg PTX per g glioma tissue. Longitudinal magnetic resonance imaging analysis illustrated that the intracranial glioblastoma progression in nude mice treated with PTX-LIPO delivered via FUS with MBs was suppressed consistently for 4 weeks compared to the untreated group. The medium survival time of these tumor-bearing nude mice was significantly prolonged by 20.8%, compared to the untreated nude mice. Immunohistochemical analysis further confirmed the antiproliferation effect and cell apoptosis induction. Our study demonstrated that noninvasive low-intensity FUS with MBs can be used as an effective approach to deliver PTX-LIPO in order to improve their chemotherapy efficacy toward glioblastoma.

An ultrasound transient elastography system with coded excitation.

BACKGROUND: Ultrasound transient elastography technology has found its place in elastography because it is safe and easy to operate. However, it's application in deep tissue is limited. The aim of this study is to design an ultrasound transient elastography system with coded excitation to obtain greater detection depth. METHODS: The ultrasound transient elastography system requires tissue vibration to be strictly synchronous with ultrasound detection. Therefore, an ultrasound transient elastography system with coded excitation was designed. A central component of this transient elastography system was an arbitrary waveform generator with multi-channel signals output function. This arbitrary waveform generator was used to produce the tissue vibration signal, the ultrasound detection signal and the synchronous triggering signal of the radio frequency data acquisition system. The arbitrary waveform generator can produce different forms of vibration waveform to induce different shear wave propagation in the tissue. Moreover, it can achieve either traditional pulse-echo detection or a phase-modulated or a frequency-modulated coded excitation. A 7-chip Barker code and traditional pulse-echo detection were programmed on the designed ultrasound transient elastography system to detect the shear wave in the phantom excited by the mechanical vibrator. Then an elasticity QA phantom and sixteen in vitro rat livers were used for performance evaluation of the two detection pulses. RESULTS: The elasticity QA phantom's results show that our system is effective, and the rat liver results show the detection depth can be increased more than 1 cm. In addition, the SNR (signal-to-noise ratio) is increased by 15 dB using the 7-chip Barker coded excitation. CONCLUSIONS: Applying 7-chip Barker coded excitation technique to the ultrasound transient elastography can increase the detection depth and SNR. Using coded excitation technology to assess the human liver, especially in obese patients, may be a good choice.

Improved shear wave motion detection using coded excitation for transient elastography.

Transient elastography (TE) is well adapted for use in studying liver elasticity. However, because the shear wave motion signal is extracted from the ultrasound signal, the weak ultrasound signal can significantly deteriorate the shear wave motion tracking process and make it challenging to detect the shear wave motion in a severe noise environment, such as within deep tissues and within obese patients. This paper, therefore, investigated the feasibility of implementing coded excitation in TE for shear wave detection, with the hypothesis that coded ultrasound signals can provide robustness to weak ultrasound signals compared with traditional short pulse. The Barker 7, Barker 13, and short pulse were used for detecting the shear wave in the TE application. Two phantom experiments and one in vitro liver experiment were done to explore the performances of the coded excitation in TE measurement. The results show that both coded pulses outperform the short pulse by providing superior shear wave signal-to-noise ratios (SNR), robust shear wave speed measurement, and higher penetration intensity. In conclusion, this study proved the feasibility of applying coded excitation in shear wave detection for TE application. The proposed method has the potential to facilitate robust shear elasticity measurements of tissue.

Dynamic mechanical analysis to assess viscoelasticity of liver tissue in a rat model of nonalcoholic fatty liver disease.

Nonalcoholic fatty liver disease (NAFLD) is the most common liver disorder in both developed and developing countries. A noninvasive method of detecting early stage NAFLD and distinguishing non-alcoholic steatohepatitis (NASH) from simple steatosis (SS) would be useful. The over-accumulation of fat in hepatocytes alters the physical microstructure and chemical contents of the liver tissue. This study included dynamic mechanical analysis (DMA) testing on liver samples from a rat model of NAFLD to determine whether the tissue shows any significant changes in viscoelasticity due to the histological changes. Liver steatosis was induced in 57 rats by gavage feeding of a high fat emulsion; 12 rats received a standard diet only and served as controls. Each rat provided 2 or 3 samples for DMA tests. The shear modulus and loss modulus were measured at 9 frequency points evenly-spaced in the range from 1Hz to 41Hz. The phase velocity of shear wave was calculated from the measured modulus. Multivariate T2 test was used to assess the significance of intra-group difference. The results showed significant changes (p < 0.05) in storage modulus in livers with moderate to severe (S2 to S4) steatosis in comparison with livers without steatosis (S0), while the loss modulus demonstrated significant changes earlier in stage S1, indicating that fat accumulation affects the mechanical properties of liver, particularly viscosity. However, no significant differences were observed between the steatosis grades. These results also suggest that mild inflammation may affect the mechanical properties, which requires further verification. These findings provide new information about the mechanical properties of livers with NAFLD in low frequency range and suggest that it is possible to distinguish normal livers from livers with NAFLD.

Model-dependent and model-independent approaches for evaluating hepatic fibrosis in rat liver using shearwave dispersion ultrasound vibrometry.

This study assesses gradations of hepatic fibrosis in rat livers using both model-dependent and model-independent approaches. Liver fibrosis was induced in 37 rats using carbon tetrachloride (CCl4); 6 rats served as the controls. Shear wave velocity as a function of frequency, referred to as velocity dispersion, was measured in vitro by an ultrasound elastography method called shearwave dispersion ultrasound vibrometry (SDUV). For the model-dependent approach, the velocity dispersion data were fit to the Voigt model to solve the viscoelastic modulus. For the model-independent approach, the pattern of the velocity dispersion data was analyzed by linear regression to extract the slope and intercept features. The parameters obtained by both approaches were evaluated separately using a receiver operating characteristic (ROC) curve analysis. The results show that, of all the parameters for differentiating between grade F0-F1 and grade F2-F4 fibrosis, the intercept had the greatest value for the area under the ROC curve. This finding suggests that the model-independent approach may provide an alternative method to the model-dependent approach for staging liver fibrosis.

Measurement of quantitative viscoelasticity of bovine corneas based on lamb wave dispersion properties.

The viscoelastic properties of the human cornea can provide valuable information for clinical applications such as the early detection of corneal diseases, better management of corneal surgery and treatment and more accurate measurement of intra-ocular pressure. However, few techniques are capable of quantitatively and non-destructively assessing corneal biomechanics in vivo. The cornea can be regarded as a thin plate in which the vibration induced by an external vibrator propagates as a Lamb wave, the properties of which depend on the thickness and biomechanics of the tissue. In this study, pulses of ultrasound radiation force with a repetition frequency of 100 or 200 Hz were applied to the apex of corneas, and the linear-array transducer of a SonixRP system was used to track the tissue motion in the radial direction. Shear elasticity and viscosity were estimated from the phase velocities of the A0 Lamb waves. To assess the effectiveness of the method, some of the corneas were subjected to collagen cross-linking treatment, and the changes in mechanical properties were validated with a tensile test. The results indicated that the shear modulus was 137 ± 37 kPa and the shear viscosity was 3.01 ± 2.45 mPa · s for the group of untreated corneas and 1145 ± 267 kPa and was 0.16 ± 0.11 mPa · s for the treated group, respectively, implying a significant increase in elasticity and a significant decrease in viscosity after collagen cross-linking treatment. This result is in agreement with the results of the mechanical tensile test and with reports in the literature. This initial investigation illustrated the ability of this ultrasound-based method, which uses the velocity dispersion of low-frequency A0 Lamb waves, to quantitatively assess both the elasticity and viscosity of corneas. Future studies could discover ways to optimize this system and to determine the feasibility of using this method in clinical situations.

Preliminary study on the effect of stiffness on lamb wave propagation in bovine corneas.

The viscoelastic properties of human cornea could provide valuable information for various clinical applications. Particularly, it will be helpful to achieve a patient-specific biomechanical optimization in LASIK refractive surgery, early detection of corneal ecstatic disease or improved accuracy of intraocular pressure (IOP) measurement. However, there are few techniques that are capable of accurately assessing the corneal elasticity in situ in a nondestructive fashion. In order to develop a quantitative method for assessing both elasticity and viscosity of the cornea, we use ultrasound radiation force to excite Lamb waves in cornea, and a pulse echo transducer to track the tissue vibration. The fresh postmortem bovine eyes were treated via collagen cross-linking to make the cornea stiff. The effect of stiffness was studied by comparing the propagation of Lamb waves in normal and treated corneas. It was found that the waveform of generated Lamb waves changed significantly due to the increase in higher modes in treated corneas. This result indicated that the generated waveform was a complex of multiple harmonics and the varied stiffness will affect the energy distribution over different components. Therefore, it is important for assessing the viscoelastic properties of the cornea to know the components of Lamb wave and calculate the phase velocity appropriately.

[Finite element analysis of the dynamic behaviour of a microbubble within a rigid micro-tube].

The dynamic behaviour of a microbubble confined within a rigid micro-tube was studied using finite element method. The results indicated that the microbubble oscillation was limited when constrained within the micro-tube. Both the expansion ratio of its effective radius and natural frequency decreased with the decrease of the tube radius. Meanwhile, the deformation of the microbubble was non-spherical and became more significant when the ultrasound pressure amplitude increased. The dynamic behaviour in micro-tube was different from that in infinite liquid.

Highly sensitive computer aided diagnosis system for breast tumor based on color Doppler flow images.

A computer-aided diagnosis (CAD) system for breast tumor based on color Doppler flow images is proposed. Our system consists of automatic segmentation, feature extraction, and classification of breast tumors. First, the B-mode grayscale image containing anatomical information was separated from a color Doppler flow image (CDFI). Second, the boundary of the breast tumor was automatically defined in the B-mode image and then morphologic and gray features were extracted. Third, an optimal feature vector was created using K-means cluster algorithm. Then a back-propagation (BP) artificial neural network (ANN) was used to classify breast tumors as benign, malignant or uncertain. Finally, the blood flow feature was extracted selectively from the CDFI, and was used to classify the uncertain tumor as benign or malignant. Experiments on 500 cases show that the proposed system yields an accuracy of 100% for the malignant and 80.8% for the benign classification. Comparing with other systems, the advantage of our system is that it has a much lower percentage of malignant tumor misdiagnosis.

[Narrow-band level set and visualization technology for cochlea segmentation].

Segmentation is a precondition step for modeling and analysis of the cochlea. In this paper, an interactive segmentation approach with a combination of 3D (three-dimensional) narrow-band level set algorithm and visualization technology was adopted to separate the cochlea. The 3D narrow-band level set algorithm was used to separate the objective of interest from volume data, and the visualization technology was used to display the segmented result. Users could modify intermediate parameters based on the direct 3D visual feedback until getting satisfying result. The basic principle and characteristics of level set and narrow-band level set algorithms were described in detail. The 3D narrow-band level set algorithm was successfully used to separate the cochlea from spiral CT images of the temporal bone. The experimental results show that the interactive method combining the narrow-band level set algorithm with visualization technique is capable of segmenting the cochlea from the medical volume data. Compared with the method of segmenting object from image volume slice by slice, our method can save much time.