Multiscale deformable registration for dual-energy x-ray imaging.

Dual-energy (DE) imaging of the chest improves the conspicuity of subtle lung nodules through the removal of overlying anatomical noise. Recent work has shown double-shot DE imaging (i.e., successive acquisition of low- and high-energy projections) to provide detective quantum efficiency, spectral separation (and therefore contrast), and radiation dose superior to single-shot DE imaging configurations (e.g., with a CR cassette). However, the temporal separation between high-energy (HE) and low-energy (LE) image acquisition can result in motion artifacts in the DE images, reducing image quality and diminishing diagnostic performance. This has motivated the development of a deformable registration technique that aligns the HE image onto the LE image before DE decomposition. The algorithm reported here operates in multiple passes at progressively smaller scales and increasing resolution. The first pass addresses large-scale motion by means of mutual information optimization, while successive passes (2-4) correct misregistration at finer scales by means of normalized cross correlation. Evaluation of registration performance in 129 patients imaged using an experimental DE imaging prototype demonstrated a statistically significant improvement in image alignment. Specific to the cardiac region, the registration algorithm was found to outperform a simple cardiac-gating system designed to trigger both HE and LE exposures during diastole. Modulation transfer function (MTF) analysis reveals additional advantages in DE image quality in terms of noise reduction and edge enhancement. This algorithm could offer an important tool in enhancing DE image quality and potentially improving diagnostic performance.

Dual-energy imaging of the chest: optimization of image acquisition techniques for the 'bone-only' image.

Experiments were conducted to determine optimal acquisition techniques for bone image decompositions for a prototype dual-energy (DE) imaging system. Technique parameters included kVp pair (denoted [kVp(L)/kVp(H)]) and dose allocation (the proportion of dose in low- and high-energy projections), each optimized to provide maximum signal difference-to-noise ratio in DE images. Experiments involved a chest phantom representing an average patient size and containing simulated ribs and lung nodules. Low- and high-energy kVp were varied from 60-90 and 120-150 kVp, respectively. The optimal kVp pair was determined to be [60/130] kVp, with image quality showing a strong dependence on low-kVp selection. Optimal dose allocation was approximately 0.5-i.e., an equal dose imparted by the low- and high-energy projections. The results complement earlier studies of optimal DE soft-tissue image acquisition, with differences attributed to the specific imaging task. Together, the results help to guide the development and implementation of high-performance DE imaging systems, with applications including lung nodule detection and diagnosis, pneumothorax identification, and musculoskeletal imaging (e.g., discrimination of rib fractures from metastasis).

Cardiac gating with a pulse oximeter for dual-energy imaging.

The development and evaluation of a prototype cardiac gating system for double-shot dual-energy (DE) imaging is described. By acquiring both low- and high-kVp images during the resting phase of the cardiac cycle (diastole), heart misalignment between images can be reduced, thereby decreasing the magnitude of cardiac motion artifacts. For this initial implementation, a fingertip pulse oximeter was employed to measure the peripheral pulse waveform ('plethysmogram'), offering potential logistic, cost and workflow advantages compared to an electrocardiogram. A gating method was developed that accommodates temporal delays due to physiological pulse propagation, oximeter waveform processing and the imaging system (software, filter-wheel, anti-scatter Bucky-grid and flat-panel detector). Modeling the diastolic period allowed the calculation of an implemented delay, t(imp), required to trigger correctly during diastole at any patient heart rate (HR). The model suggests a triggering scheme characterized by two HR regimes, separated by a threshold, HR(thresh). For rates at or below HR(thresh), sufficient time exists to expose on the same heartbeat as the plethysmogram pulse [t(imp)(HR) = 0]. Above HR(thresh), a characteristic t(imp)(HR) delays exposure to the subsequent heartbeat, accounting for all fixed and variable system delays. Performance was evaluated in terms of accuracy and precision of diastole-trigger coincidence and quantitative evaluation of artifact severity in gated and ungated DE images. Initial implementation indicated 85% accuracy in diastole-trigger coincidence. Through the identification of an improved HR estimation method (modified temporal smoothing of the oximeter waveform), trigger accuracy of 100% could be achieved with improved precision. To quantify the effect of the gating system on DE image quality, human observer tests were conducted to measure the magnitude of cardiac artifact under conditions of successful and unsuccessful diastolic gating. Six observers independently measured the artifact in 111 patient DE images. The data indicate that successful diastolic gating results in a statistically significant reduction (p < 0.001) in the magnitude of cardiac motion artifact, with residual artifact attributed primarily to gross patient motion.

Optimization of image acquisition techniques for dual-energy imaging of the chest.

Experimental and theoretical studies were conducted to determine optimal acquisition techniques for a prototype dual-energy (DE) chest imaging system. Technique factors investigated included the selection of added x-ray filtration, kVp pair, and the allocation of dose between low- and high-energy projections, with total dose equal to or less than that of a conventional chest radiograph. Optima were computed to maximize lung nodule detectability as characterized by the signal-difference-to-noise ratio (SDNR) in DE chest images. Optimal beam filtration was determined by cascaded systems analysis of DE image SDNR for filter selections across the periodic table (Z(filter) = 1-92), demonstrating the importance of differential filtration between low- and high-kVp projections and suggesting optimal high-kVp filters in the range Z(filter) = 25-50. For example, added filtration of approximately 2.1 mm Cu, approximately 1.2 mm Zr, approximately 0.7 mm Mo, and approximately 0.6 mm Ag to the high-kVp beam provided optimal (and nearly equivalent) soft-tissue SDNR. Optimal kVp pair and dose allocation were investigated using a chest phantom presenting simulated lung nodules and ribs for thin, average, and thick body habitus. Low- and high-energy techniques ranged from 60-90 kVp and 120-150 kVp, respectively, with peak soft-tissue SDNR achieved at [60/120] kVp for all patient thicknesses and all levels of imaging dose. A strong dependence on the kVp of the low-energy projection was observed. Optimal allocation of dose between low- and high-energy projections was such that approximately 30% of the total dose was delivered by the low-kVp projection, exhibiting a fairly weak dependence on kVp pair and dose. The results have guided the implementation of a prototype DE imaging system for imaging trials in early-stage lung nodule detection and diagnosis.

Optimal kvp selection for dual-energy imaging of the chest: evaluation by task-specific observer preference tests.

Human observer performance tests were conducted to identify optimal imaging techniques in dual-energy (DE) imaging of the chest with respect to a variety of visualization tasks for soft and bony tissue. Specifically, the effect of kVp selection in low- and high-energy projection pairs was investigated. DE images of an anthropomorphic chest phantom formed the basis for observer studies, decomposed from low-energy and high-energy projections in the range 60-90 kVp and 120-150 kVp, respectively, with total dose for the DE image equivalent to that of a single chest radiograph. Five expert radiologists participated in observer preference tests to evaluate differences in image quality among the DE images. For visualization of soft-tissue structures in the lung, the [60/130] kVp pair provided optimal image quality, whereas [60/140] kVp proved optimal for delineation of the descending aorta in the retrocardiac region. Such soft-tissue detectability tasks exhibited a strong dependence on the low-kVp selection (with 60 kVp providing maximum soft-tissue conspicuity) and a weaker dependence on the high-kVp selection (typically highest at 130-140 kVp). Qualitative examination of DE bone-only images suggests optimal bony visualization at a similar technique, viz., [60/140] kVp. Observer preference was largely consistent with quantitative analysis of contrast, noise, and contrast-to-noise ratio, with subtle differences likely related to the imaging task and spatial-frequency characteristics of the noise. Observer preference tests offered practical, semiquantitative identification of optimal, task-specific imaging techniques and will provide useful guidance toward clinical implementation of high-performance DE imaging systems.

Excitation energy transfer in the light-harvesting chlorophyll a/b.protein.

The "light-harvesting chlorophyll a/b.protein" described by Thornber has been prepared electrophoretically from spinach chloroplasts. The optical properties relevant to energy transfer have been measured in the red region (i.e. 600-700 nm). Measurements of the absorption spectrum, fluorescence excitation spectrum and excitation dependence of the fluorescence emission spectrum of this protein confirm that energy transfer from chlorophyll b to chlorophyll a is highly efficient, as is the case in concentrated chlorophyll solutions and in vivo. The excitiation dependence of the fluorescence polarization shows a minimum polarization of 1.9% at 650 nm which is the absorption maximum of chlorophyll b in the protein and rises steadily to a maximum value of 13.8% at 695 nm, the red edge of the chlorophyll a absorption band. Analysis of these measurements shows that at least two unresolved components must be responsible for the chlorophyll a absorption maximum. Comparison of polarization measurements with those observed in vivo shows that most of the depolarization observed in vivo can take place within a single protein. Circular dichroism measurements show a double structure in the chlorophyll b absorption band which suggest an exciton splitting not resolved in absorption. Analysis of these data yields information about the relative orientation of the So leads to S1 transition moments of the chlorophyll molecules within the protein.

Optimization of chest films of equalization radiography (advanced multiple beam equalization radiography). Comparison by means of a receiver operating characteristic study of simulated nodular interstitial disease.

RATIONALE AND OBJECTIVE: To optimize screen-film combinations for equalization radiography (advanced multiple beam equalization radiography [AMBER]), five different film-screen-technique combinations were compared by receiver operating characteristics study of simulated interstitial disease. MATERIALS AND METHODS: The Ortho C-Lanex Regular and the Insight Thoracic Imaging HC system were compared in conventional nonequalized technique; T-Mat G-Lanex Regular and T-Max L-Lanex Regular were compared in conventional, nonequalized, and AMBER technique; and an experimental high-contrast, low-noise, near-zero crossover film-screen combination was compared in AMBER technique. Interstitial disease was simulated by superimposing birdseed on the back of a humanoid phantom. Twenty-five posterior-anterior radiographs were made with each technique. Seven observers scored the presence of interstitial disease in each of the quadrants on a 5-point scale following receiver operating characteristic methodology. RESULTS: The highest performance was found with the experimental film-screen-AMBER combination (Az = 0.92) and the lowest with the T-Mat L-Lanex Regular-AMBER combination (Az = 0.83) and the Insight Thoracic Imaging HC system-conventional combination (Az = 0.85). T-Mat L-Lanex Regular-conventional ranked second (Az = 0.90) while T-Mat G-Lanex Regular-conventional (Az = 0.89), T-Mat L-Lanex Regular-AMBER (Az = 0.88) and Ortho-C-Lanex Regular-conventional (Az = 0.87) scored lower. CONCLUSION: Higher contrast films in AMBER improve diagnostic performance, whereas a loss of information is found if the AMBER system is combined with lower contrast films.

Describing the signal-transfer characteristics of asymmetrical radiographic screen-film systems.

The measurement of modulation transfer function for radiographic screen-film systems depends critically upon a proper linearization of the measured line spread function. This is normally done by photographic photometry (i.e., using the measured density versus log exposure relationship to transform the density line spread function into an exposure line spread function). It has been long appreciated that this procedure may fail for asymmetrical dual screen systems that use film with emulsion coated on both sides of the support. The advent of asymmetrical and near-zero crossover films that can be used with highly asymmetric screen pairs has prompted a reinvestigation of these concerns about the definition and measurement of modulation transfer function. For such cases, it is useful to define the contrast transfer function, which is a function of exposure and spatial frequency. When normalized by its zero frequency value the contrast transfer function can serve as the "effective MTF" for low-contrast input signals in such systems. In the limit of symmetrical systems this quantity approaches the conventionally measured MTF. The utility of this approach is demonstrated by applying it to a commercially available asymmetrical screen-film combination.

Objective performance characteristics of a new asymmetric screen-film system.

A study of the objective imaging characteristics of a new asymmetric screen-film system is presented herein. The system is characterized by high x-ray absorption asymmetric screens, and a low-noise, high-contrast asymmetric film having near-zero crossover. Comparisons are made with the imaging characteristics of two widely used conventional screen-film systems. Sensitometry, modulation transfer function, and noise power spectra were measured using standard methods. Granularity, noise equivalent quanta, and detective quantum efficiency were computed from these. The new screen-film system has an average gradient at lung-field densities between the two conventional systems studied, while the mediastinum contrast exceeds both conventional systems. The lung-field modulation transfer function half bandwidth of the new asymmetric system exceeds that of both conventional systems by 60%. At mid exposures the detective quantum efficiency of the new asymmetric system is comparable to those of the conventional systems studied. However, the exposure range over which detective quantum efficiency remains high is substantially wider.

An application of multivariate moment-generating functions to the analysis of signal and noise propagation in radiographic screen-film systems.

Previous studies [J. Opt. Soc. Am. 4A 895-901 (1987)] have shown the utility of multivariate moment-generating functions for analyzing the influence of stochastic amplifying and scattering mechanisms on the transfer of signal and noise through multistage imaging systems. Recently, these studies were extended to include cases in which the amplification or scattering parameters are themselves stochastic variables [J. Opt. Soc. Am. 6A, 1156-1164 (1989)]. In this paper we consider a special case in which amplification is followed by scattering such that the same random variable which characterizes the parameters of each amplification process also characterizes the parameters of the subsequent scattering of the amplified output events. In radiographic imaging, this can be used to describe the physics of the depth dependence of emission efficiency and light scatter in x-ray intensifying screens, which was originally treated by Lubberts [J. Opt. Soc. Am. 58, 1475-1483 (1968]). In this work Lubberts' original results are rederived in a more general form. They are then illustrated in terms of a diffusion model [Appl. Opt. 12, 1865-1870 (1973)] for light scatter within the intensifying screen.

Understanding the relative sensitivity of radiographic screens to scattered radiation.

This study compared the relative response of various screen-film and computed radiography (CR) systems to diagnostic radiation exposure. An analytic model was developed to calculate the total energy deposition within the depth of screen and the readout signal generated from this energy for the x-ray detection system. The model was used to predict the relative sensitivity of several screen-film and CR systems to scattered radiation as a function of various parameters, such as x-ray spectra, phantom thickness, phosphor composition, screen thickness, screen configuration (single front screen, single back screen, screen pair), and readout conditions. In addition, measurements of the scatter degradation factor (SDF) for different screen systems by using the beam stop technique with water phantoms were made to verify the model results. Theoretically calculated values of SDF were in good agreement with experimental data. These results are consistent with the common observation that rare-earth screens generally produce better image quality than calcium tungstate screens and the CR screen.

Detection of artificial air space opacities with digital radiography: ex vivo study on enhanced latitude post-processing.

PURPOSE: To evaluate in a.-p. digital chest radiograms of an ex vivo system if increased latitude and enhanced image detail contrast (EVP) improve the accuracy of detecting artificial air space opacities in parts of the lung that are superimposed by the diaphragm. MATERIALS AND METHODS: 19 porcine lungs were inflated inside a chest phantom, prepared with 20-50 ml gelatin-stabilized liquid to generate alveolar air space opacities, and examined with direct radiography (3.0 × 2.5 k detector/ 125 kVp/ 4 mAs). 276 a.-p. images with and without EVP of 1.0-3.0 were presented to 6 observers. 8 regions were read for opacities, the reference was defined by CT. Statistics included sensitivity/specificity, interobserver variability, and calculation of Az (area under ROC curve). RESULTS: Behind the diaphragm (opacities in 32/92 regions), the median sensitivity increased from 0.35 without EVP to 0.53-0.56 at EVP 1.5-3.0 (significant in 5/6 observers). The specificity decreased from 0.96 to 0.90 (significant in 6/6), and the Az value and interobserver correlation increased from 0.66 to 0.74 and 0.39 to 0.48, respectively. Above the diaphragm, the median sensitivity for artificial opacities (136/276 regions) increased from 0.71 to 0.77-0.82 with EVP (significant in 4/6 observers). The specificity and Az value decreased from 0.76 to 0.62 and 0.74 to 0.70, respectively, (significant in 3/6). CONCLUSION: In this ex vivo experiment, EVP improved the diagnostic accuracy for artificial air space opacities in the superimposed parts of the lung (area under the ROC curve). Above the diaphragm, the accuracy was not affected due to a tradeoff in sensitivity/specificity.

Noise power spectrum analysis of a scanning microdensitometer.

We have analyzed the principal sources of noise in a commercially available 2-D scanning microdensitometer which we use to estimate the noise power spectra of radiographic films. Two kinds of noise have been observed. One source, associated with the glass platen of the instrument, is correlated from scan to scan. This source of noise limits our ability to measure the NPS of film samples at low sample optical densities. The other major noise source is uncorrelated from scan to scan and increases exponentially with sample optical density. We have measured both of these component noise sources as well as the total instrument noise as a function of instrument density and spatial frequency. A method for minimizing the effects of instrument noise on estimates of the noise power of film samples is described and demonstrated.

Fluorescence of light-harvesting chlorophyll a/b-protein complexes: implications for the photosynthetic unit.

To the extent that extracted light-harvesting chlorophyll proteins (LHCPs) retain the chlorophyll configuration which they had in vivo, information on the optical properties of LHCPs is useful for an assessment of the transfer process of the primary excitation energy in photosynthesis. Within this context we report and discuss the implication of three kinds of data on spinach chloroplast LHCP. First, an analysis of the spectroscopic dependence of absorption, polarization and circular dichroism (reported recently by R.L.V.) suggests a model affording the possibility of easy chlorophyll a intercomplex transfer with chlorophyll b groups acting as local antitraps. Second, the ratio of LHCP emission and absorption probabilities obeys the Stepanov relation over a relatively wide range, an observation which suggests rapid Chl b-Chl a excitation equilibration. Finally, an LHCP absolute fluorescence yield as great as 10% has been measured, which provides a possible upper limit for the yield of the antenna fluorescence.

Detective quantum efficiency of imaging systems with amplifying and scattering mechanisms.

We have analyzed the influence of stochastic amplifying and scattering mechanisms on the transfer of signal and noise through multistage imaging systems in terms of multivariate moment-generating functions. Stochastic amplification of photon noise by one stage of an imaging system is shown to constitute an effective signal to the next, while the underlying photon-noise component is unaffected by a subsequent scattering process. In the case of stationary, photon-limited inputs, these considerations then lead to useful expressions for the noise power spectrum and detective quantum efficiency for multistage image systems. The application of these results to the analysis of radiographic screen-film imaging systems is discussed.

Analysis of the detective quantum efficiency of a radiographic screen-film combination.

Detective quantum efficiency provides a useful measure of the imaging efficiency of imaging systems. Methods for measuring the exposure and the spatial-frequency dependence of the contrast transfer function, the noise power spectrum, and the detective quantum efficiency are developed for x-ray imaging systems. These are applied to a high-resolution screen-film combination exposed to a 30-kV-peak x-ray spectrum. The major component sources of screen-film noise in this system are identified and quantified. These are interpreted in terms of a simple model to predict the screen-film noise power spectrum and detective quantum efficiency. Reasonable agreement is found between model predictions and experimental measurements.

Small object contrast in AMBER and conventional chest radiography.

The ability of a commercially available scanning equalization system for chest radiography to render small object contrast in the lung-, mediastinum-, and subdiaphragm-equivalent regions of an acrylic chest phantom was quantitatively evaluated. Images from nine chest phantoms that represented a wide range of patient sizes and dynamic ranges of x-ray transmittance were analyzed. Subject contrast was measured with a photostimulable phosphor detector, and images were acquired in both equalized and nonequalized (conventional) imaging modes. Available subject contrast in the lung-equivalent region was 8%-15% lower in the equalized images compared with the nonequalized images in all phantoms (patient types); contrast in the mediastinum-, retro-cardiac-, and subdiaphragm-equivalent regions was 11%-63% higher in the equalized images, with the degree of improvement increasing as patient size and dynamic range increased. Images of each phantom were also acquired with the screen-film systems currently in use at the authors' institution, permitting an assessment of the relative performance (in terms of radiographic contrast) of these imagers with and without use of equalization.

Effect of exposure variation on the clinical utility of chest radiographs.

PURPOSE: To study the effects of exposure error on the clinical utility of chest radiographs. MATERIALS AND METHODS: Under- and overexposed screen-film images were simulated by adding exposure offsets to the normalized CR log(10) exposure data from a computed radiography (CR) system and printed by using the sensitometric response of a medium-latitude system. The clinical utility of the overall image, lung, and soft tissue in 48 images were independently graded by eight radiologists. RESULTS: Most variability in image scores was due to differences in exposure. About 95% of the lung regions and 75% of the soft-tissue regions were rated as having good or ideal clinical utility at the nominal exposure. About 80% of the overall images were rated as good or better for exposures within 40% [0.15 log(10) exposure] of the nominal. The overall image scores were heavily influenced by the lung region, and reader tolerance for exposure error was greater for soft tissue than for lung. The optimal exposure for soft tissue was about 60% [0.25 log(10) exposure] greater than for lung; therefore, no single exposure was optimal for the entire image. CONCLUSION: Conventional medium-latitude screen-film systems have intrinsic limitations for capturing and displaying the wide transmittance differences in the thorax. The clinical utility of chest radiographs may be improved by developing better image capture and display techniques.