An automated bladder volume measurement algorithm by pixel classification using random forests.

Residual bladder volume measurement is a very important marker for patients with urinary retention problems. To be able to monitor patients with these conditions at the bedside by nurses or in an out patient setting by general physicians, hand held ultrasound devices will be extremely useful. However to increase the usage of these devices by non traditional users, automated tools that can aid them in the scanning and measurement process will be of great help. In our paper, we have developed a robust segmentation algorithm to automatically measure bladder volume by segmenting bladder contours from sagittal and transverse ultrasound views using a combination of machine learning and active contour algorithms. The algorithm is tested on 50 unseen images and 23 transverse and longitudinal image pairs and the performance is reported.

Automated interventricular septum thickness measurement from B-mode echocardiograms.

In this work, we address the problem of automated measurement of the interventricular septum thickness, one of the key parameters in cardiology, from B-mode echocardiograms. The problem is challenging due to high levels of noise, multi modal intensity, weak contrast due to near field haze, and non rigid motion of the septum across frames. We introduce a complete system for automated measurement of septum thickness from B-mode echocardiograms incorporating three main components: a 1D curve evolution algorithm using region statistics for segmenting the septum, a motion clustering method to locate the mitral valve, and a robust method to calculate the septum width from these inputs in accordance with medical standards. Our method effectively handles the challenges of such measurements and runs in near real time. Results on 57 patient recordings showed excellent agreement of the automated measurements with expert manual measurements.

Real-time simultaneous triplane contrast echocardiography gives rapid, accurate, and reproducible assessment of left ventricular volumes and ejection fraction: a comparison with magnetic resonance imaging.

OBJECTIVE: We sought to compare the feasibility, accuracy, and reproducibility of simultaneous triplane echocardiography for measurements of left ventricular (LV) volumes and ejection fraction (EF) with reference to magnetic resonance imaging (MRI). METHODS: Digital echocardiography recordings of apical LV views with and without intravenous contrast were collected from 53 consecutive patients with conventional 2-dimensional (2D) imaging and with simultaneous triplane imaging. MRI of multiple LV short-axis sections was performed with a 1.5-T scanner. Endocardial borders were manually traced, and LV volumes and EF from 2D biplane echocardiography and MRI were calculated by method of disks. On triplane data, a triangular mesh was constructed by 3-dimensional interpolation and volumes calculated by the divergence theorem. RESULTS: Triplane image acquisition was less time-consuming than 2D biplane. Precontrast feasibility was 72% for triplane and 82% for 2D biplane images, increasing to 98% and 100% with contrast, respectively. Bland-Altman analysis demonstrated LV volume underestimation by echocardiography versus MRI, which was significantly reduced by contrast and triplane imaging. The 95% limits of agreement for EF between echocardiography and MRI narrowed using triplane compared with 2D biplane (precontrast -12.5 to 6.7% vs -17.2 to 9.9%, and with contrast -7.1 to 5.8% vs -9.4 to 6.4%, respectively). At intraobserver and interobserver analysis of 20 patients, limits of agreement for EF narrowed with contrast triplane compared with 2D biplane. CONCLUSION: Simultaneous LV triplane imaging is feasible with simple and rapid image acquisition and volume analysis, and with contrast enhancement it gives accurate and reproducible LV EF measurements compared with MRI.

Quantitative adenosine real-time myocardial contrast echocardiography for detection of angiographically significant coronary artery disease.

BACKGROUND: Real-time (RT) myocardial contrast echocardiography (MCE) is a novel method for assessment of regional myocardial perfusion. We sought to evaluate the feasibility and diagnostic accuracy of quantitative adenosine RT MCE in predicting significant coronary stenoses, with reference to quantitative coronary angiography. METHODS: Low-power RT MCE was performed in 43 patients scheduled for quantitative coronary angiography. Peak signal intensity (A), rate of signal intensity increase (beta), A x beta (myocardial blood flow), and their hyperemic reserves were estimated and compared with angiographic data. RESULTS: The feasibility of quantitative stress RT MCE covering all coronary territories was 77% of patients with adequate baseline image quality. At rest we found no significant difference for any of the perfusion parameters between the normal and stenosed coronary territories. During hyperemia, beta and A x beta, but not A, increased significantly in normal coronary territories. In the regions subtended by significantly stenosed arteries, there were no significant increases in beta and A x beta. Receiver operating characteristic curves indicated that beta- and A x beta-reserves, but not A-reserve, could be sensitive parameters for detecting flow-limiting coronary stenosis in selected patients, particularly if significant left anterior descending coronary artery disease was involved. CONCLUSION: Quantitative assessment of myocardial blood flow and its velocity reserve by RT MCE has the potential to detect significant coronary artery disease, but because of imaging and technical problems it is not yet robust enough for clinical use in unselected patients.

Effects of ultrasound contrast during tissue velocity imaging on regional left ventricular velocity, strain, and strain rate measurements.

BACKGROUND: Strain (epsilon) rate (SR) imaging and left ventricular (LV) opacification with intravenous (IV) contrast both potentially decrease operator dependency in interpretation of stress echocardiography. The aim of this study was to evaluate whether contrast present during tissue velocity imaging (TVI) significantly affected measurements of velocity, epsilon, and SR. Secondly, we sought to evaluate whether increased scan line density improved feasibility of simultaneous TVI and contrast echocardiography. METHODS: The 4-chamber LV view in 15 healthy volunteers and 25 patients was acquired at rest before and after IV injections of contrast using: (1) conventional TVI; (2) LV opacification with standard TVI added; and (3) modified LV opacification with doubled TVI line density. Velocity, SR, and epsilon curves, along with peak systolic velocity, peak systolic SR, and end-systolic epsilon, were assessed from midwall segments. RESULTS: IV contrast significantly reduced feasibility of TVI with standard settings, giving noisy data for SR and epsilon, particularly in the septum. Absolute values of peak systolic SR and end-systolic epsilon from adequately shaped curves were significantly higher with contrast compared with baseline. However, increased TVI line density significantly improved feasibility of velocity traces with contrast and decreased the level of noise in SR and epsilon. Furthermore, higher line density improved agreement between peak systolic velocity, peak systolic SR, and end-systolic epsilon measured with contrast, and corresponding precontrast values from the conventional TVI setting. CONCLUSIONS: SR imaging was not feasible performed with IV contrast during conventional TVI settings, and we do not recommend the clinical use of this combination. Increased TVI line density made velocity curves with contrast feasible and resulted in less noisy SR and epsilon curves, but variability in SR and epsilon measurements with contrast is still too high for clinical use.

Automatic timing of aortic valve closure in apical tissue Doppler images.

Ultrasound color tissue Doppler imaging (TDI) can be used to estimate velocities of moving left ventricular cardiac tissue. Aortic valve closure (AVC) can be observed as a notch in apical TDI velocity/time curves occurring after ejection, but before early relaxation. This work sought to evaluate automatic and automated algorithms using TDI for timing AVC. Mitral valve position and the time point of early relaxation were extracted and used to accomplish the task. To test the algorithms, phonocardiogram of the second heart sound was recorded simultaneously with TDI and used as a reference method. The algorithms were tested on apical views of 16 healthy subjects. In 98% of the cardiac cycles, the automatic algorithm estimated the time point of AVC within 25 ms of the reference. Automatic detection of AVC might save manual effort and provide a marker separating ejection and diastole for further automated analysis.

Quantification of resting myocardial blood flow velocity in normal humans using real-time contrast echocardiography. A feasibility study.

BACKGROUND: Real-time myocardial contrast echocardiography (MCE) is a novel method for assessing myocardial perfusion. The aim of this study was to evaluate the feasibility of a very low-power real-time MCE for quantification of regional resting myocardial blood flow (MBF) velocity in normal human myocardium. METHODS: Twenty study subjects with normal left ventricular (LV) wall motion and normal coronary arteries, underwent low-power real-time MCE based on color-coded pulse inversion Doppler. Standard apical LV views were acquired during constant IV. infusion of SonoVue. Following transient microbubble destruction, the contrast replenishment rate (beta), reflecting MBF velocity, was derived by plotting signal intensity vs. time and fitting data to the exponential function; y (t) =A (1-e(-beta(t-t0))) + C. RESULTS: Quantification was feasible in 82%, 49% and 63% of four-chamber, two-chamber and apical long-axis view segments, respectively. The LAD (left anterior descending artery) and RCA (right coronary artery) territories could potentially be evaluated in most, but contrast detection in the LCx (left circumflex artery) bed was poor. Depending on localisation and which frames to be analysed, mean values of beta were 0.21-0.69 s(-1), with higher values in medial than lateral, and in basal compared to apical regions of scan plane (p = 0.03 and p < 0.01). Higher beta-values were obtained from end-diastole than end-systole (p < 0.001), values from all-frames analysis lying between. CONCLUSION: Low-power real-time MCE did have the potential to give contrast enhancement for quantification of resting regional MBF velocity. However, the technique is difficult and subjected to several limitations. Significant variability in beta suggests that this parameter is best suited for with-in patient changes, comparing values of stress studies to baseline.

Accurate and reproducible measurement of left ventricular volume and ejection fraction by contrast echocardiography: a comparison with magnetic resonance imaging.

OBJECTIVES: We evaluated the accuracy and reproducibility of contrast echocardiography versus tissue harmonic imaging for measurements of left ventricular (LV) volumes and ejection fraction (EF) compared to magnetic resonance imaging (MRI). METHODS: Digital echo recordings of apical LV views before and after intravenous contrast were collected from 110 consecutive patients. Magnetic resonance imaging of multiple short-axis LV sections was performed with a 1.5-T scanner. Left ventricular volumes and EF were calculated offline by method of discs. Thirty randomly selected patients were reanalyzed for intraobserver and interobserver variability. RESULTS: Compared with baseline, contrast echo increased feasibility for single-plane and biplane volume analysis from 87% to 100% and from 79% to 95%, respectively. The Bland-Altman analysis demonstrated volume underestimation by echo, but much less pronounced with contrast. Limits of agreement between echo and MRI narrowed significantly with contrast: from -18.1% to 8.3% to -7.7% to 4.1% (EF), from -98.2 to -11.7 ml to -59.0 to 10.7 ml (end-diastolic volume), and from -58.8 to 21.8 ml to -38.6 to 23.9 ml (end-systolic volume). Ejection fraction from precontrast echo and MRI differed by > or =10% (EF units) in 23 patients versus 0 after contrast (p < 0.001). At intraobserver and interobserver analysis, limits of agreement for EF narrowed significantly with contrast. CONCLUSIONS: The two-dimensional echocardiographic evaluation of LV volumes and EF in non-selected cardiac patients was found to be more accurate and reproducible when adding an intravenous contrast agent.

Improved contrast to tissue ratio at higher harmonics.

The challenge in ultrasound contrast imaging is a better discrimination between the perfused tissue and the contrast bubbles, which is usually expressed by contrast to tissue ratio (CTR). Imaging based on the second harmonic frequency showed a higher CTR than imaging at the fundamental frequency. However, because of nonlinear propagation of ultrasound waves, harmonic frequencies are generated. These harmonic frequencies will be linearly reflected by the tissue and therefore limit the CTR at the second harmonic frequency. In order to reduce the scattering of tissue at harmonic frequencies and by that increase the CTR, nonlinear distortion has to be reduced. We demonstrate in this study that the CTR increases with the harmonic number. The increase is substantial when transmitting at lower frequencies. To take advantage of the higher harmonics (third, fourth, fifth and the ultraharmonics and termed here super harmonics), we have developed a new phased array transducer with a wide frequency band. In-vitro measurements using the new probe show an increase of 40 dB of the CTR for super harmonic components over the conventional second harmonic system. The increase in CTR is in agreement with the calculations using existing models for the response of encapsulated bubbles and known theory of nonlinear propagation.

Determinants of myocardial hypoperfusion analyzed for the interventricular septum using power Doppler harmonic imaging with contrast echocardiography in humans: a methodologic approach for clinical practice.

BACKGROUND: To evaluate determinants of myocardial hypoperfusion using power Doppler harmonic imaging (PDHI) with myocardial contrast echocardiography (MCE) in clinical practice, a retrospective clinical study was performed comparing echocardiographic and angiographic data. Angiographic data of patients with a normal coronary angiogram (non-CAD) and symptomatic patients with low flow conditions caused by a stenosis of the left anterior descending coronary artery (LAD) or occlusion, or TIMI-II-flow in the LAD were compared with the PDHI data. METHODS AND RESULTS: In 32 patients, MCE was performed with a System Five Performance ultrasound system (GE Vingmed Ultrasound, Horten, Norway). Myocardial perfusion was semiquantitatively analyzed with the EchoPac 6.2b.134 software, bolus injection with Optison (0.35 mL with 5 mL saline flush), and continuous infusion with Levovist (400 mg/mL(-1); 3.5-5 mL/min(-1)) were performed (8 non-CAD patients, 8 CAD patients, respectively). After bolus injection, Doppler intensity (DI) kinetics showed a significant decrease of maximum DI wash-in rate (eg, apical septum [AS]: 4.9 +/- 3.3 vs 2.4 +/- 1.9 dB/s(-1)), of peak maximum DI (eg, AS: 25.3 +/- 6.3 vs 16.4 +/- 5.7 dB), and of DI determined 10 and 20 seconds after peak maximum DI (eg, AS: 22.1 +/- 4.9 vs 10.8 +/- 4.6 dB; AS: 20.4 +/- 5.3 vs 8.0 +/- 3.8 dB, respectively) using a trigger interval once every 3 cardiac cycles when normal perfused areas were compared with hypoperfused areas. During infusion coronary transit time (3.3 +/- 0.9 vs 7.0 +/- 3.6 seconds), maximum DI wash-in rate (eg, AS: 3.2 +/- 1.3 vs 1.3 +/- 0.8 dB/s(-1)) and DI-maximum plateau (eg, AS: 28.6 +/- 4.7 vs 18.3 +/- 6.4 dB) significantly decreased, respectively. CONCLUSION: Regional myocardial hypoperfusion at rest can be detected by using PDHI with MCE in clinical practice, according to a standardized methodologic protocol.

Super harmonic imaging: a new imaging technique for improved contrast detection.

For ultrasound contrast agents (UCA), nonlinear imaging now has become fundamental. All of the current contrast-imaging methods are dominantly based on the nonlinear response of UCA bubbles. The discrimination between the perfused tissue and the UCA is the challenge in the field of UCA-imaging. This differentiation is usually associated or expressed by the ratio of the scattered power from the contrast agent to the scattered power from the tissue and is termed "contrast-to-tissue ratio" (CTR). Second harmonic imaging showed a better discrimination between tissue and UCA than fundamental imaging because of a higher CTR. We demonstrate, in this study, that the CTR increases as a function of the order of the harmonic frequency. Currently, due to the limited bandwidth of the transducers, only the second harmonic is selectively imaged, resulting in images with a superior quality to fundamental images, but still degraded and not optimal because of the harmonic generation in the underlying tissue (due to nonlinear propagation) and hence giving a limited CTR. To increase the CTR and to take advantage of the higher harmonics (third, fourth, fifth and the ultraharmonics and termed here super harmonics), we have developed a new phased array transducer. The array transducer contains two different types of elements arranged in an interleaved pattern (odd and even elements). The total number of elements is 96. The elements can operate separately and at a distinct frequency, enabling separate transmission and reception modes. The odd elements (48) operate at typically 2.8 MHz center frequency and 80% bandwidth. The even elements (48) have a center frequency of 900 kHz with a bandwidth of 50%. In vitro measurements using the dual frequency probe show an increase of 40 dB in the CTR for super harmonic components over the conventional second harmonic system. The increase in CTR is in agreement with the calculations using existing models for the response of encapsulated bubbles and known theory of nonlinear propagation. Animal experiments have demonstrated the feasibility of this approach using commercially available UCA and showed a similar increase of the CTR.