Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “deconvolution”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,009 records · Page 56Linked to original sources
PubMedDate not supplied

1975 Memorial Award Paper. Image generation and display techniques for CT scan data. Thin transverse and reconstructed coronal and sagittal planes.

The various limitations to computerized axial tomographic (CT) interpretation are due in part to the 8-13 mm standard tissue plane thickness and in part to the absence of alternative planes of view, such as coronal or sagittal images. This paper describes a method for gathering multiple overlapped 8 mm transverse sections, subjecting these data to a deconvolution process, and then displaying thin (1 mm) transverse as well as reconstructed coronal and sagittal CT images. Verification of the deconvolution technique with phantom experiments is described. Application of the phantom results to human post mortem CT scan data illustrates this method's faithful reconstruction of coronal and sagittal tissue densities when correlated with actual specimen photographs of a sectioned brain. A special CT procedure, limited basal overlap scanning, is proposed for use on current first generation CT scanners without hardware modification.

Adolescent↗
PubMed1999

Kinetic analysis of [11C]McN5652: a serotonin transporter radioligand.

The impulse response function of a radioligand is the most fundamental way to describe its pharmacokinetics and to assess its tissue uptake and retention pattern. This study investigates the impulse response function of [11C](+)McN5652, a radioligand used for positron emission tomography (PET) imaging of the serotonin transporter (SERT) in the brain. Dynamic PET studies were performed in eight healthy volunteers injected with [11C](+)McN5652 and subsequently with its pharmacologically inactive enantiomer [11C](-)McN5652. The impulse response function was calculated by deconvolution analysis of regional time-activity curves, and its peak value (f(max)), its retention value at 75 minutes (fT), and its normalized retention (f(rel) = fT/f(max)) were obtained. Alternatively, compartmental models were applied to calculate the apparent total distribution volume (DV(T)) and its specific binding component (DV(S)). Both the noncompartmental (fT,f(rel)) and the compartmental parameters (DV) were investigated with and without correction for nonspecific binding by simple subtraction of the corresponding value obtained with [11C](-)McN5652. The impulse response function obtained by deconvolution analysis demonstrated high tracer extraction followed by a slow decline in the form of a monoexponential function. Statistical analysis revealed that the best compartmental model in terms of analysis of variance F and condition number of the parameter variance-covariance matrix was the one that was based on a single tissue compartment with parameters k1 and k2 and that also included the parameter of regional cerebral blood volume (BV). The parameter f(rel) demonstrated low between-subject variance (coefficient of variation [CV] = 19%), a midbrain to cerebellum ratio of 1.85, and high correlation with the known density of SERT (r = 0.787 where r is the coefficient of linear correlation between the parameter and the known density of SERT). After correction for nonspecific binding, f(rel) demonstrated further improvement in correlation (r = 0.814) and midbrain to cerebellum ratio (3.09). The variance of the distribution volumes was acceptable when the logarithmic transform lnDV was used instead of DV (17% for the three-parameter model), but correlation of this compartmental parameter was slightly less (r = 0.652 for the three-parameter model) than the correlation of the noncompartmental f(rel) with the known density of SERT, and the midbrain to cerebellum ratio was only 1.5 (uncorrected) and 1.8 (corrected). At the expense of increasing variance, the correlation was increased after correction for nonspecific binding using the inactive enantiomer (r = 0.694; CV = 22%). These results indicate that the kinetics of [11C](+)McN5652 can best be described by a one-tissue compartment model with three parameters (k1, k2, and BV), and that both the noncompartmental parameter f(rel) and the compartmental distribution volumes have the potential for quantitative estimation of the density of SERT. Further validation of the radioligand in experimental and clinical situations is warranted.

Adult↗
PubMedDate not supplied

A knowledge-based approach to deconvolve the water component in in vivo proton MR spectroscopy.

OBJECTIVE: The water component dominates in in vivo proton (1H) spectroscopy. This consists of a prominent central peak with large "wings," and consequently metabolites that lie on the "wings" become difficult to quantitate. A method has been developed to deconvolve the water component in in vivo 1H spectroscopy, hence highlighting the metabolite information. MATERIALS AND METHODS: Spectra were acquired from volunteers and patients using 4D chemical shift imaging. The water deconvolution procedure employed knowledge-based data processing in the frequency domain and was fully automated. This involved describing the water component as a coarse function consisting of a "bulk" region and "wing" areas. Points were identified in the spectrum that fit this description and then linked together to produce the water component. The latter was smoothed and then subtracted from the original spectra to produce good water deconvolution. RESULTS: Over 2,000 in vivo 1H spectra have been subjected to this algorithm. The method took approximately 5 s to execute per spectrum consisting of 2,048 data points. CONCLUSION: Knowledge-based data processing has provided a fast, efficient, and robust procedure to deconvolve the water component.

Adult↗
PubMedDate not supplied

Quantitative assessment of cerebral hemodynamics using CT: stability, accuracy, and precision studies in dogs.

PURPOSE: The limited clinical availability of currently used methods to measure regional cerebral blood volume (CBV) and cerebral blood flow (CBF) represents an important restriction. We undertook this study to evaluate a new dynamic CT method to measure CBV and CBF in normal and ischemic tissue. METHOD: A total of 21 dynamic CT studies were performed in seven male beagles. The contrast enhancement curves of the carotid arteries and of various brain regions were deconvolved to obtain CBV and CBF. The stability of the deconvolution method employed was assessed by comparing three data sets obtained by analyses of one, two, and four regions of interest (ROIs), all covering the entire brain area. The accuracy of CT-derived CBF was analyzed for normal (n = 5 studies) and ischemic (n = 7 studies) brain tissue using fluorescent microspheres. Repetitive CT studies were performed to evaluate the precision of the CT measurements. RESULTS: The stability of the deconvolution method was high with variabilities of 2.3% (CBV), 5.9% (CBF), and 8.9% (mean transit time), respectively. The correlation between the CT and the microsphere measurements was good for both normal and ischemia studies (r > 0.78, slope > 0.9). The variability of the CT CBF (30.6%) was higher than that of the CT CBV (12.3%) measurements. CONCLUSION: Our novel dynamic CT method is stable with respect to the sizes of ROIs used, allowing for accurate measurements of CBV and CBF in both normal and ischemic tissue. Further studies are necessary to evaluate the variability of this method under controlled physiologic conditions.

Animals↗
PubMed1988

Theoretical and practical considerations of measuring extravascular lung water.

The volume of extravascular lung water is currently measured in vivo from the difference in mean transit times of the extrapolated first-pass dilution curves of two indicators, one diffusible and the other confined to the intravascular space. To overcome the limitations of this method, one can prolong the measurement interval, introduce a highly diffusible indicator, or both. In the first case, recirculating indicators are measured and included in the computation by deconvolution of the mean transit time through the lung. In the second case, heat is used as the water indicator. In the third case, not yet explored, recirculating heat would be measured and long thermal transit times uncovered. In view of the complexity of the deconvolution method and the pitfalls of the thermal dilution method, a radiographic score of pulmonary edema may be more useful clinically to assess the volume of extravascular lung water in patients with heart disease or adult respiratory distress syndrome.

Animals↗
PubMed2002

Circulating white blood cells affect red cell pulmonary transit times in endurance athletes during intense exercise.

PURPOSE: The aim of this study was to determine the relationship between the right-to-left ventricular red cell pulmonary transit times (PTT) during intense exercise and circulating white blood cell (WBC) counts in highly trained endurance athletes. We postulated that high levels of WBCs preexercise would slow PTT. Eleven endurance-trained athletes (VO2max = 69.6 +/- 7.7 mL.kg-1.min-1; weight = 75.0 +/- 6.2 kg; height = 181.0 +/- 7.1 cm) performed 6.5 min constant-load, near-maximal cycling exercise (approximately 92% VO2max) on two different days. Preexercise WBC counts were measured in arterial blood drawn from the radial artery 30 min before exercise. PTT was measured during the 3rd min of exercise by first-pass radionuclide cardiography using centroid and deconvolution analysis, whereas cardiac output (Q) was measured during the last 2.5 min of exercise via a count-based ratio method from the MUGA technique. RESULTS: Combined mean PTT from both deconvolution and centroid analysis at minute three of exercise was 2.45 +/- 0.21 s, whereas the preexercise WBC count was 5.3 +/- 1.6 x 109.L-1. Cardiopulmonary blood volume at minute three of exercise was 1.22 +/- 0.13 L, VO2 was 4.58 +/- 0.44 L.min-1, and Q was 30.2 +/- 4.2 L.min-1. We found that PTT was negatively correlated with circulating WBC (r = -0.61; adjusted r2 = 0.30; P = 0.04; N = 11) but not with the dispersion (spread) of transit times around the mean (r = 0.19; P = 0.57). CONCLUSION: This suggests that athletes with higher circulating numbers of WBCs preexercise have faster (shorter) red cell transit times through the lung during intense exercise.

Cardiac Output↗
PubMed1996

Comparison of a corticopelvic transfer index with renal transit times.

The aim of this study was to compare a simple measure obtained visually from serial images, the corticopelvic transfer index (CPI), with renal transit times obtained by deconvolution. The CPI was defined as the time interval between the appearance of the tracer in the renal cortex to its appearance in the renal pelvis. Four groups of subjects as determined by clinical data and renal transit times were investigated retrospectively: 86 normal kidneys, 20 kidneys with renovascular disorder, 17 kidneys with obstructive nephropathy and 26 kidneys with dilated pelves without obstructive nephropathy. There were good correlations only between the CPI and mean parenchymal transit time (MPTT; r = 0.6) and the whole kidney transit time index (WKTTI; r = 0.8). It was concluded that in centres where gamma-camera studies with deconvolution analysis are not possible, the CPI could be a useful parameter in renal radionuclide scintigraphy for suspected renovascular disorders.

Humans↗
PubMed1991

Prolonged maximal stimulation of insulin secretion in healthy subjects does not provoke preferential release of proinsulin.

Release of immature secretory granules rich in incompletely processed proinsulin has been proposed to explain the relative hyperproinsulinemia in type 2 diabetic and insulinoma patients because of a constant secretory drive resulting from hyperglycemia and autonomous secretion, respectively. To test this hypothesis, insulin secretion was stimulated by a combination of hyperglycemia (11 mmol/L clamp), intravenous (i.v.) tolbutamide (1 g), and i.v. glucagon (initial bolus 10 micrograms/kg body weight, maintenance infusion 2 micrograms/kg body weight per hour) for 3 h. Circulating IR-insulin and IR-C-peptide concentrations increased 89-fold and 14-fold over basal values, respectively, but IR-proinsulin concentrations increased only ninefold over basal values. Estimation of the amount of insulin secreted (based on deconvolution analysis of plasma C-peptide values) showed that approximately 76 +/- 21 U were secreted during the stimulation period. This amount is a significant proportion of pancreatic insulin content in normal humans. In molar terms, IR-proinsulin (integrated incremental response multiplied by metabolic clearance rate of proinsulin) relative to IR-C-peptide (= insulin) secretion (deconvolution analysis) was estimated to be equal or even lower than the known proportion in islets (0.22 +/- 0.05%). Thus, using a near-maximal stimulation of insulin secretion maintained long enough to cause release of amounts of insulin approaching the estimated pancreatic content, no preferential release of proinsulin was observed in normal humans. Therefore, the hyperproinsulinemia of type 2 diabetes and in insulinoma patients may be caused by additional defects in the proinsulin to insulin conversion process.

Adult↗
PubMed2002

Noninvasive measurement of cerebral blood flow in adults using near-infrared spectroscopy and indocyanine green: a pilot study.

This pilot study was designed to determine the feasibility of measuring cerebral blood flow noninvasively after an intravenous bolus of indocyanine green using near-infrared spectroscopy and pulse dye-densitometry. Feasibility aside, this study did not attempt to validate the measured values of cerebral blood flow against an established method of measurement. Twelve healthy volunteers were investigated after peripheral intravenous injection of indocyanine green. Arterial and cerebral changes in indocyanine green concentration were measured using pulse dye-densitometry and near-infrared spectroscopy, respectively. Two methods of calculating cerebral blood flow were used, and a blood flow index was also estimated. Absolute cerebral blood flow was calculated using a modification of the Fick principle and a deconvolution algorithm to derive the impulse residue function. Mean (range) estimated cerebral blood flow for the Fick method was 8.2 mL/100 g/min (4.2-16.2 mL/100 g/min) and 8.3 mL/100 g/min (4.7-15.3 mL/100 g/min) for the impulse residue function method. The impulse residue function method provided a more precise intrasubject estimation of cerebral blood flow compared with the modified Fick principle, with a coefficient of variation of 10.1% versus 25.5%. The blood flow index was 8.6 mg/sec (range: 5.6-17.3 mg/sec) with an intrasubject coefficient of variation of 12.0%. Estimation of cerebral blood flow using near-infrared spectroscopy and pulse dye-densitometry can be made at the bedside after intravenous injection of indocyanine green, and the precision can be improved using a deconvolution algorithm. Notwithstanding the low values obtained for absolute cerebral blood flow, further investigation and validation of this bedside technique is warranted.

Adult↗
PubMed2004

Enteral glutamine but not alanine maintains small bowel barrier function after ischemia/reperfusion injury in rats.

We previously demonstrated that glucose and glutamine, solutes metabolized by the gut, replenish ATP and enhance gut function compared with alanine, a solute not metabolized by the gut, following mesenteric ischemia/reperfusion (I/R). The purpose of the present study was to determine if the nonmetabolizable solute alanine differentially modulates cytoskeletal organization and paracellular small intestinal permeability compared with the metabolizable solutes glucose and glutamine following mesenteric I/R. At laparotomy, rats had jejunal sacs filled with 10 mM glucose, glutamine, alanine, or magnesium sulfate (5 mm, osmotic control) followed by superior mesenteric artery clamping for 60 min and 30 min of reperfusion or sham laparotomy. Jejunum was harvested for evaluation by deconvolution microscopy, fluorescent measurement of F:G actin ratio, or mounted in an Ussing chamber for determination of intestinal permeability. Deconvolution microscopy revealed that the actin cytoskeleton was preserved by enteral glutamine, comparable to shams, but disrupted by enteral alanine. Glucose and controls resulted in comparable disruption, which was less than that with alanine. The F:G actin ratio was highest for glutamine and lowest for alanine; glucose was comparable to controls. Intestinal permeability was highest for alanine and lowest for glutamine, which was comparable to shams. Permeability following glucose and controls was higher than that following glutamine but lower than that following alanine. The nonmetabolizable solute alanine resulted in disruption of the actin cytoskeleton and enhanced intestinal permeability under conditions of mesenteric I/R. The metabolizable solute glutamine was protective under these conditions, whereas glucose exerted minimal effect on the integrity of the cytoskeleton and intestinal permeability. The individual components of enteral diets may differentially modulate intestinal barrier function, which could have important implications when administered to critically injured patients.

Alanine↗
PubMed1995

Comparative Pharmacokinetics and Bioavailability of Dilacor XR and Cardizem CD in Healthy Volunteers.

The objective of this investigation was to compare the single-dose and steady-state pharmacokinetic profiles of Dilacor XR to Cardizem CD. The study enrolled 24 healthy males and was divided into three parts: a single intravenous 25-mg bolus dose of diltiazem HCl (Cardizem Injectable) followed by a two-way crossover comparison of single and multiple once-daily 240-mg oral doses of Dilacor XR and Cardizem CD. Plasma samples were analyzed for diltiazem using a specific and sensitive HPLC assay. Statistical analysis and deconvolution were performed on the data. A 1- and 3-hour lag time in diltiazem absorption was noted following the administration of Dilacor XR and Cardizem CD, respectively. Statistically significant differences were noted in mean single- and multiple-dose t(max) values with Cardizem CD taking approximately twice as long as Dilacor XR to reach C(max). Dilacor XR was equivalent to Cardizem CD with respect to AUC((0--infty infinity)) and C(max). Equivalent minimum and average steady-state plasma diltiazem concentrations were noted after multiple-dose administration. Deconvolution of the single-dose data also showed similar mean bioavailabilities for the respective formulations but revealed dissimilarities in each product's absorption profile that may reflect observed differences in absorption lag time and t(max).

Journal Article↗
PubMed2004

The immune-enhancing enteral agents arginine and glutamine differentially modulate gut barrier function following mesenteric ischemia/reperfusion.

BACKGROUND: Immune-enhancing enteral diets have been shown to improve patient outcome. One contributing mechanism may be via maintenance of gut barrier function. While recent data has shown that glutamine is beneficial, arginine may be harmful. We therefore hypothesized that the immune-enhancing agents, glutamine and arginine, differentially modulate gut barrier function. METHODS: At laparotomy, rats had jejunal sacs filled with 10 mmol/L glutamine, arginine, fructose, or magnesium sulfate (osmotic control) followed by 60 minutes of superior mesenteric artery occlusion and 2 hours of reperfusion. Jejunum was harvested for histology, deconvolution microscopy, F:G actin, ATP, and permeability measurements. RESULTS: Glutamine and fructose minimized mucosal injury compared with controls and arginine. Deconvolution microscopy confirmed that glutamine and fructose preserved the actin cytoskeleton but there was disruption by arginine which correlated with F:G actin ratios and tissue ATP levels. Permeability was enhanced by arginine compared with the other groups. CONCLUSION: Arginine resulted in worsened mucosal injury, disruption of the actin cytoskeleton, decreased tissue ATP and enhanced permeability compared with glutamine which appeared protective. The immune-enhancing agent arginine results in breakdown of gut barrier function which may have important implications for critically injured patients.

Actins↗
PubMed2004

Dynamic imaging of perfusion and oxygenation by functional magnetic resonance imaging.

Cerebral blood flow can be measured with magnetic resonance imaging (MRI) by arterial spin labeling techniques, where magnetic labeling of flowing spins in arterial blood water functions as the endogenous tracer upon mixing with the unlabeled stationary spins of tissue water. The consequence is that the apparent longitudinal relaxation time (T1) of tissue water is attenuated. A modified functional MRI scheme for dynamic CBF measurement is proposed that depends on extraction of T1 weighting from the blood oxygenation level-dependent (BOLD) image contrast, because the functional MRI signal also has an intrinsic T1 weighting that can be altered by variations of the excitation flip angle. In the alpha-chloralose-anesthetized rat model at 7T, the authors show that the stimulation-induced BOLD signal change measured with two different flip angles can be combined to obtain a T1-weighted MRI signal, reflecting the magnitude of the CBF change, which can be deconvolved to obtain dynamic changes in CBF. The deconvolution of the T1-weighted MRI signal, which is a necessary step for accurate reflection of the dynamic changes in CBF, was made possible by a transfer function obtained from parallel laser-Doppler flowmetry experiments. For all stimulus durations (ranging from 4 to 32 seconds), the peak CBF response measured by MRI after the deconvolution was reached at 4.5 +/- 1.0 seconds, which is in good agreement with (present and prior) laser-Doppler measurements. Because the low flip angle data can also provide dynamic changes of the conventional BOLD image contrast, this method can be used for simultaneous imaging of CBF and BOLD dynamics.

Animals↗
PubMed2005

Bilinear dynamical systems.

In this paper, we propose the use of bilinear dynamical systems (BDS)s for model-based deconvolution of fMRI time-series. The importance of this work lies in being able to deconvolve haemodynamic time-series, in an informed way, to disclose the underlying neuronal activity. Being able to estimate neuronal responses in a particular brain region is fundamental for many models of functional integration and connectivity in the brain. BDSs comprise a stochastic bilinear neurodynamical model specified in discrete time, and a set of linear convolution kernels for the haemodynamics. We derive an expectation-maximization (EM) algorithm for parameter estimation, in which fMRI time-series are deconvolved in an E-step and model parameters are updated in an M-Step. We report preliminary results that focus on the assumed stochastic nature of the neurodynamic model and compare the method to Wiener deconvolution.

Algorithms↗
PubMed2003

Experimental and calculated Stark widths within the Kr I spectrum.

On the basis of the precisely recorded 20 neutral krypton (Kr I) line shapes (in the 5s-5p and 5s-6p transitions), we have obtained the basic plasma parameters, i.e., electron temperature (T) and electron density (N) using our line deconvolution procedure in a plasma created in a linear, low-pressure, pulsed arc discharge operated in krypton. The mentioned plasma parameters have also been measured using independent experimental diagnostics techniques. Agreement has been found among the two sets of the obtained parameters. This recommends our deconvolution procedure for plasma diagnostical purposes, especially in astrophysics where direct measurements of the main plasma parameters (T and N) are not possible. On the basis of the observed asymmetry of the Stark broadened line profile, we have obtained not only its ion broadening parameter (A) which is caused by influence of the ion-microfield over the line broadening mechanism but also the influence of the ion-dynamic effect (D) over the line shape. The separate electron (W(e)) and ion (W(i)) contributions to the total Stark width, which have not been measured so far, have also been obtained. Stark widths are calculated using the semiclassical perturbation formalism for electrons, protons, and helium ions as perturbers.

Journal Article↗
PubMed2000

The high-mosaicity illusion: revealing the true physical characteristics of macromolecular crystals.

Typical measurements of macromolecular crystal mosaicity are dominated by the characteristics of the X-ray beam and as a result the mosaicity value given during data processing can be an artifact of the instrumentation rather than the sample. For physical characterization of crystals, an experimental system and software have been developed to simultaneously measure the diffraction resolution and mosaic spread of macromolecular crystals. The contributions of the X-ray beam to the reflection angular widths were minimized by using a highly parallel, highly monochromatic synchrotron source. Hundreds of reflection profiles over a wide resolution range were rapidly measured using a charge-coupled device (CCD) area detector in combination with superfine phi-slicing data collection. The Lorentz effect and beam contributions were evaluated and deconvoluted from the recorded data. Data collection and processing is described. From 1 degrees of superfine phi-slice data collected on a crystal of manganese superoxide dismutase, the mosaicities of 260 reflections were measured. The average mosaicity was 0.0101 degrees (s.d. 0.0035 degrees ) measured as the full-width at half-maximum (FWHM) and ranged from 0.0011 to 0. 0188 degrees. Each reflection profile was individually fitted with two Gaussian profiles, with the first Gaussian contributing 55% (s.d. 9%) and the second contributing 35% (s.d. 9%) of the reflection. On average, the deconvoluted width of the first Gaussian was 0.0054 degrees (s.d. 0.0015 degrees ) and the second was 0.0061 degrees (s. d. 0.0023 degrees ). The mosaicity of the crystal was anisotropic, with FWHM values of 0.0068, 0.0140 and 0.0046 degrees along the a, b and c axes, respectively. The anisotropic mosaicity analysis indicates that the crystal is most perfect in the direction that corresponds to the favored growth direction of the crystal.

Crystallography, X-Ray↗
PubMed2001

Model-based estimation of ultrasonic echoes. Part II: Nondestructive evaluation applications.

Accurate estimation of the ultrasonic echo pattern leading to the physical property of the object is desirable for ultrasonic NDE (nondestructive evaluation) applications. In Part I of this study, we have presented a generalized parametric ultrasonic echo model, composed of a number of Gaussian echoes corrupted by noise, and algorithms for accurately estimating the parameters. In Part II of this study, we explore the merits of this model-based estimation method in ultrasonic applications. This method produces high resolution and accurate estimates for ultrasonic echo parameters, i.e., time of flight (TOF) amplitude, center frequency, bandwidth, and phase. Furthermore, it offers a solution to the deconvolution problem for restoration of the target response, i.e., ultrasonic reflection and transmission properties of materials, from the backscattered echoes. The model-based estimation method makes deconvolution possible in the presence of significant noise. It can also restore closely spaced overlapping echoes beyond the resolution of the measuring system. These properties of the estimation method are investigated in various ultrasonic applications such as transducer pulse-echo wavelet estimation, subsample time delay estimation, and thickness sizing of thin layers.

Journal Article↗
PubMed2004

Time delay for arrival of MR contrast agent in collateral-dependent myocardium.

An analysis of the kinetics of myocardial contrast enhancement is an important component of myocardial perfusion studies. The contrast enhancement can be modeled by a linear time-invariant system, and the myocardial impulse response, calculated by deconvolution of the measured tissue response with an arterial input, gives a direct estimate of myocardial blood flow. In this paper, we analyze the effects of delays in the contrast enhancement, that occur in collateral-dependent myocardium, where the tracer reaches the tissue region only through branches from other coronary arteries that form natural bypass vessels. We investigate how the delayed arrival of tracer alters the myocardial impulse response. Model-independent deconvolution is applied to determine the lag between arterial input and tissue enhancement. Experimental data in a porcine model of collateral development indicate that the delayed arrival of an injected tracer, measured at rest, is a useful marker to identify collateral-dependent myocardium, and predict its flow capacitance.

Algorithms↗