Breast cancer in vivo: contrast enhancement with thermoacoustic CT at 434 MHz-feasibility study

The authors performed thermoacoustic computed tomography (CT) with 434-MHz radio waves in five patients with documented breast cancer. Three of the patients underwent imaging before chemotherapy was initiated and two at the conclusion of their primary chemotherapy. In the former three patients, thermoacoustic CT demonstrated contrast enhancement in the region of the tumor. In the latter two patients, no contrast enhancement was seen, and pathologic examination after surgical resection of the area of original tumor confirmed complete remission of disease.

Thermoacoustic computed tomography--technical considerations

We have constructed a thermoacoustic computed tomography scanner for imaging soft tissue in the human body. Thermoacoustic signals are induced in soft tissue by irradiation with 434 MHz rf energy. The thermoacoustic signals are detected by an array of transducers mounted on a hemispherical bowl. A three-dimensional, filtered backprojection algorithm is used to reconstruct rf absorption patterns within soft tissue. We have demonstrated soft tissue differentiation sufficient to delineate the normal internal structures of an excised lamb kidney using safe levels of rf radiation.

Thermoacoustic CT with radio waves: a medical imaging paradigm

The authors evaluated images obtained with a prototypic thermoacoustic computed tomographic (CT) scanner constructed for use at 434 MHz, a promising radio frequency for detecting breast cancer. In one excised porcine kidney, acoustic energy emanating from the kidney was detected with transducers. The resultant electric signals were used to create a three-dimensional data set. Two-dimensional images reconstructed in multiple planes were compared with state-of-the-art T1- and T2-weighted magnetic resonance images. The renal outline, parenchyma, and collecting system were clearly delineated on the thermoacoustic CT images.

World Wide Web interface to digital imaging and communication in medicine-capable image servers

As a trial project, the Indiana University Department of Radiology has develop[ed a low-cost manner of distributing radiological images throughout a medical environment using the World Wide Web (WWW). The interface requires the user to have a WWW-browser client, such as Netscape, running on UNIX, PC, or Macintosh platforms. A forms-based interface allows the user to query several DICOM-capable machines at the machine, patient, study, series, and image levels. Once an image transfer is initiated, images are prewindowed from 16- to 8-bits, compressed using public domain Joint Photographic Expert Group (JPEG) compression routines, transferred to the WWW client program, and decompressed and displayed using a locally selected image viewing program. At the currently implemented level of compression (75% quality), the entire fetch-transform-JPEG-display process takes 2 to 5 seconds over Ethernet, depending on the platform used.

Image intensifier-based computed tomography volume scanner for angiography

RATIONALE AND OBJECTIVES

A prototype volume computed tomography (CT) system for use in angiography was designed, constructed, and tested. The system consisted of a fixed X-ray tube, a conventional image intensifier (II) coupled to a charge-coupled device camera, and a computer-controlled turntable on which phantoms were placed. We wanted to predict, through phantom studies, the imaging performance of an II-based volume CT for direct three-dimensional (3D) reconstruction of vascular structures.

METHODS

To explore the imaging performance of the system for reconstructing a vascular structure, two sets of projection images of a vascular phantom, acquired over 250 projection angles with two different-sized IIs, were digitized and used for a direct 3D conebeam reconstruction. The signal-to-noise ratio (SNR) of each reconstructed image was measured. From these measurements, image quality was accessed as a function of the number of reconstructions averaged and the different orientations. The spatial resolution limits of the system were measured from the 3D reconstructed images of a specially designed resolution phantom for different orientations and locations.

RESULTS

The measured SNRs of all direct 3D reconstruction images were reasonably good, and back-ground noise levels measured from 3D reconstruction images were almost 30 Hounsfield units. The measured spatial resolution of the system was 0.5 line pairs per millimeter. However, spatial resolution was reduced around the edge of the II to nearly half that measured in the central area of the field of view.

CONCLUSION

An II-based volume CT scanner can produce direct 3D reconstructions of vascular structures with good image quality for intraarterial angiography.

Dual-slice spiral versus single-slice spiral scanning: comparison of the physical performance of two computed tomography scanners

In this paper we deal with two types of spiral scanners; one is a single-slice spiral scanner, while the other employs dual-slice technology into spiral scanning. Physical performance parameters, including image noise, contrast resolution, spatial resolution (transversal and longitudinal), and radiation exposure are measured. Computer simulations based on two interpolation methods (180 degrees and 360 degrees linear interpolation) are also used in evaluating the slice-sensitivity profile (SSP) and noise. The results show that the noise behaves in the same way for both types of scanners. The noise change, relative to that of the standard scan with the same scanning parameters, depends solely on the interpolation algorithm. Table speed and scanner geometry (either single slice or dual slice) have no effect on the noise value. For the given table speed, as well as individual detector collimation (slice width) the dual-slice scan results in better longitudinal resolution (SSP) compared to a single-slice scan if the scan is obtained with nonoverlapping slices (pitch greater than 2). This is because the dual-slice scan obtains twice the number of nonoverlapped projections for the same length, which reduces the degradation of the slice profile by using more densely arranged projections (in the longitudinal direction) for the interpolation. In the dual-slice scanner the workable scan rate is extended up to pitch 4 compared to a pitch of 2 for the single-slice scanner. Therefore, the dual-slice spiral scanner is preferred in applications requiring an increased scan rate with comparative image quality.

Photoacoustic ultrasound (PAUS)--reconstruction tomography

The theoretical underpinnings of photoacoustic ultrasound (PAUS) reconstruction tomography are presented. A formal relationship between PAUS signals and the heterogeneous distribution of optical absorption within the object being investigated is developed. Based on this theory, a reconstruction approach, analogous to that used in x-ray computed tomography, is suggested. Initial experimental results suggest that this approach produces "reasonable" reconstructions for absorbers distributed within a narrow plane embedded within a highly scattering medium.

Volume detection threshold: quantitative comparison of computed radiography and screen-film radiography in detection of pneumothoraces in an animal model that simulates the neonate

PURPOSE

To quantitatively compare computed radiography (CR) and screen-film radiography (SFR) in the detection of pneumothorax in an animal model that simulates the neonate.

MATERIALS AND METHODS

Three rabbits underwent placement of 5-F catheters in the right pleural space. Eight CR and eight SFR images were obtained. Volume-controlled pneumothorax was induced by incrementally injecting 4 cm3 of air (24 cm3 total) and four CR and four SFR images were obtained after each injection. Four radiologists blindly viewed the images twice (eight readings per image, 1,600 total observations). A dichotomous yes-no score for the presence of pneumothorax was used to calculate the volume detection threshold (VDT), defined as the air volume at which 50% of the pneumothoraces can be detected. The authors plotted the likelihood of a pneumothorax against the air volume by using the confidence indicator.

RESULTS

Intraobserver VDT values for CR and SFR were not significantly different.

CONCLUSION

CR is as accurate as SFR in the detection of pneumothorax in this quantitative animal model.

Photoacoustic ultrasound: pulse production and detection of 0.5% Liposyn

Theoretical predictions and experimental measurements of photoacoustic pulse production within a 0.5% solution of Liposyn, a highly scattering, optical propagation medium, are reported. A simple model for photoacoustic energetics is developed that predicts photoacoustic signal pressure as a function of depth within a turbid medium following surface irradiation from an infrared source. The model is valid for very short irradiation duration. The model predicts that the acoustic pressure produced at a distance r from the center of a small, highly absorbing sphere of radius R consists of two, opposite polarity pulses, one originating from the near and one from the far side of the sphere. The magnitude of these biphasic pulses is expected to be proportional to the energy fluence (E) incident on the surface of the sphere and to the ratio, R/r. Furthermore, the energy fluence (E) that reaches the sphere is roughly proportional to e-mu effZ, where mu eff is the effective attenuation coefficient of the turbid medium and Z is the depth of the embedded sphere below the irradiated surface. The variation of E with depth within the absorber and biphasic acoustic pulse production have been verified experimentally. Further experiments demonstrate that a small (3-mm diameter), highly absorbing sphere can be detected and localized at a depth of 37.5 mm within a 0.5% solution of Liposyn with a spatial resolution of 1 x 6 mm2, using a biologically safe level of infrared irradiation (lambda = 1064 nm) and a conventional ultrasound transducer (frequency = 2.25 MHz). These results suggest that photoacoustic ultrasound imaging may have application to biologic systems such as the human breast.

Rabbit pneumothorax. Feasibility of an animal model to simulate neonatal pneumothorax

RATIONALE AND OBJECTIVES

The feasibility of developing an animal model to simulate radiographic features of neonatal pneumothorax was investigated, and supine computed radiography (CR) and film-screen radiography (FS) images were compared.

METHODS

5-Fr pigtail catheters were inserted into the right pleural space of three New Zealand white rabbits. Two CR and two FS supine images were obtained using identical technique. After serial 4-mL injections of air, one CR and one FS supine image were obtained. The images were masked and randomized, and four radiologists reviewed the images for typical signs of pneumothorax. Chest computed tomography (CT) was performed before the injection of any air and after completion of the study.

RESULTS

At zero-mL air volume, radiologists did not perceive a pneumothorax. As the air volume increased, the number of signs of a pneumothorax increased; there was no significant difference between CR and FS in the number of signs of pneumothorax detected at any volume. Chest CT confirmed that there was no unexplained air leak. The volume of air injected correlated with the calculated volume from the CT.

CONCLUSIONS

This induced, volume-controlled, animal model simulated the supine radiographic features of neonatal pneumothorax.

Photoacoustic ultrasound

Differential absorption has been detected and localized in three-dimensions by recording the photoacoustic pulses that were produced when short-duration (approximately 1 microsecond) pulses of electromagnetic energy were absorbed regionally within a turbid medium. These absorption sites were localized with a spatial resolution of approximately 6 mm within a 20 x 20 x 7.5-cm3 volume of 0.3% Liposyn solution, a highly scattering medium. A Xenon flashlamp, delivering a nominal 1 microsecond pulse of broadband light, was used to irradiate the Liposyn solution. Photoacoustic echoes were recorded with a focused, ultrasonic transducer, tuned to a nominal frequency of 0.5 MHz. The spatial resolution that was demonstrated is consistent with the expected ultrasonic properties of the transducer. Improved spatial resolution can be expected with shorter-duration radiation exposure and higher-frequency transducer designs. The technique is generalizable to any electromagnetic energy range (including long-wavelength microwaves) that penetrates the medium and produces differential, regional absorption of energy.

The problem of recurrent stenosis following carotid endarterectomy

From 1977 to 1985, 306 consecutive patients underwent carotid endarterectomy by a single surgeon. The post-operative stroke and death rates were 1.3% and 2.6%. Of the 184 patients with follow-up duplex scanning, 24 (13%) had a recurrent stenosis of 50% or greater. Of 15 possible risk factors studied to assess a possible relationship with recurrent stenosis, four were definitely associated and a fifth probably implicated. The four definitely associated risk factors for recurrent stenosis were an age of 70 years or older (p = 0.0025), female gender (p = 0.0001), ulcerated lesions (p = 0.013), and asymptomatic lesions (p = 0.041). The fifth risk factor that may play a role (though not reaching statistical significance) was combined carotid endarterectomy and myocardial revascularization (p = 0.066). In these patients, we recommend using vein patch angioplasty to reduce the recurrence rate.

Time resolved imaging through a highly scattering medium

Transmission images through a highly scattering medium have been obtained using picosecond pulses of visible light. The imaging method involves recording and discriminating between the times-of-flight of photons that penetrate th medium and using a fraction of the light with the shortest travel times to construct an image. The technique is being developed as a possible alternative method of screening for breast cancer without using potentially harmful x-rays. One- and two-dimensional images are presented of objects whose optical thicknesses are comparable with those of the human breast at visible wavelengths.

Light equalization radiography

An electro-optical, photographic dodging technique, called light equalization radiography (LER), has been developed. The use of LER extends the dynamic range of radiographic film by enhancing the film contrast in the "toe" region of a radiographic film. Contrast recorded above some predetermined optical density is unaffected. The use of LER permits high-contrast films to be used in radiographic exams where low-contrast, "latitude" film types have been used traditionally, e.g., in chest radiography.

Emergency eye injuries

This study analyses all patients presenting with eye complaints to the casualty section of a Brisbane Hospital during a one month period. Eye complaints constituted 3.6 per cent of all patients. A foreign body was involved in 57 per cent of all eye injuries. The patients were subject to a trial assessing the effectiveness of antibiotic treatment following removal of the foreign body. There was no significant difference between antibiotic and placebo (sterile saline).

Transillumination imaging performance: a time-of-flight imaging system

A series of Monte Carlo simulations have been performed in order to investigate the feasibility of constructing a time-of-flight transillumination imaging system that could be used as an effective tool for screening for breast disease. The conceptual design of such a system is described, and simulated images are presented that demonstrate its likely performance. It is found that, whereas the spatial resolution achievable with such a system is only dependent upon its temporal resolution, the scattering characteristics of the tissue being imaged will strongly affect the ultimate imaging performance of such a system.

Transillumination imaging performance: spatial resolution simulation studies

Monte Carlo calculations were used to simulate the propagation of visible and near-infrared light through homogenous tissue in order to quantitate the potential spatial resolution performance for transillumination imaging. Specifically, the relative effectiveness of coaxial collimation and time of flight (TOF) detection for improving spatial resolution was investigated. The results demonstrate that significant improvements in spatial resolution can be achieved through these techniques, with TOF methods offering superior performance for a given level of detected signal intensity.

Comparison of dual-energy and conventional chest radiography for nodule detection

This paper presents the results of a nodule detection study, using a Humanoid chest phantom, which was designed to evaluate the performance of two types of dual-energy and conventional (single-energy) chest radiography. The film-screen apparatis were used as image detectors for all imaging modalities. The area under the ROC curve and the cumulative true-positive fraction both were used as performance indexes in the evaluation. Because of the small number of false-positive responses in the observer studies, the cumulative true-positive fraction was eventually regarded as a more conclusive index of accuracy than the area under the ROC curve to make a reasonable conclusion. Both dual-energy techniques, dual- and single-exposure, were found to be superior to conventional chest radiography, P less than 0.0005 and P less than 0.006 for dual- and single-exposure techniques, respectively. The difference between the two dual-energy techniques was statistically insignificant, P less than 0.47. We concluded that the dual-energy, single-exposure technique is worthy of further clinical study based on these encouraging results and because of the ease with which the technique can be implemented.

Comparison of dual and single exposure techniques in dual-energy chest radiography

Conventional chest radiography is the most effective tool for lung cancer detection and diagnosis; nevertheless, a high percentage of lung cancer tumors are missed because of the overlap of lung nodule image contrast with bone image contrast in a chest radiograph. Two different energy subtraction strategies, dual exposure and single exposure techniques, were studied for decomposing a radiograph into bone-free and soft tissue-free images to address this problem. For comparing the efficiency of these two techniques in lung nodule detection, the performances of the techniques were evaluated on the basis of residual tissue contrast, energy separation, and signal-to-noise ratio. The evaluation was based on both computer simulation and experimental verification. The dual exposure technique was found to be better than the single exposure technique because of its higher signal-to-noise ratio and greater residual tissue contrast. However, x-ray tube loading and patient motion are problems.

An electrodeless measuring technique for determining conductivity of biological tissues at radio frequencies

An electrodeless measuring technique for determining the conductivity of biological tissues at radio frequencies is described. The technique is based on measuring magnetic power dissipation in a conducting sample using a tuned circuit when the sample is immersed in a time-varying magnetic field. Theory regarding the measurements is presented. Consideration for coil design is given. Coil construction and measurement techniques are described in detail. The technique has three advantages: errors resulting from electrode lead inductance are completely removed; contacts between tissue surfaces and electrode surfaces are not needed; and maintaining water content and concentration of pH solution inside the tissues during measurements becomes easily facilitated. With the technique, conductivity values of dog kidney at room temperature were obtained in good agreement with previous work.

Medical imaging strategies

Computed tomography (CT), intravenous digital subtraction angiography (IV-DSA), digital radiographic image processing and dual energy subtraction are four examples of medical imaging strategies that have met with various degrees of success as judged by diagnostic performance. The success of CT has been spectacular; IV-DSA has provided modest benefits; digital image processing of chest radiographs has been singularly disappointing; and the verdict on dual energy subtraction is undecided. The degree of success of each of these techniques can be understood by considering the degree to which each simplifies image interpretation or isolates a fingerprint of disease.

Computer simulation of image intensifier-based computed tomography detector: vascular application

This present study reports the results of a computer simulation whose aim was to predict the low-contrast imaging performance of which a conventional x-ray image intensifier with charge coupled device (CCD) camera would be capable if incorporated into a computed tomography (CT) volume imager. A vascular imaging task was modeled in our simulation. The effects of detector noise, x-ray exposure levels, analog-to-digital conversion (ADC) precision and residual levels of detected x-ray scatter were considered. The results of this simulation indicate that the low-contrast imaging performance of an image intensifier-based CT system was most limited by the CCD detector readout noise. Given this limitation the detection of greater than about 100,000 detected photons/pixel/projection gave marginal improvement in low-contrast resolution. At these exposures 12 bit ADC precision resulted in little additional image noise. The effects of detecting scattered x rays are twofold; decreasing the signal-to-noise ratio associated with our modeled artery and introducing a cupping artifact. Based on the results from the simulation, it appears that an image intensifier-based CT system is a feasible concept from a noise viewpoint, if the anticipated imaging task is intravenous angiography.

Quantitative digital subtraction coronary angiography using videodensitometry. An in vivo analysis

A videodensitometric method for measuring absolute cross-sectional area and diameter has been tested in living dogs with coronary artery stenoses created surgically by placement of small Silastic cuffs. Coronary arteriograms were performed using a circular tomographic unit to provide multiple views of each lesion, and measurements were made from logarithmically subtracted digital images. Dimensions of 13 stenoses of cross section 1 to 5 mm2 and adjacent reference segments (2 to 9 mm2) were determined by histologic sectioning of the segments after injection with a rapidly hardening plastic fixative under physiologic pressure. Two different methods were tested for calculating cross-sectional area. On 238 measurements, 102 of normal vessel segments and 135 of stenoses, both methods showed good correlation with histologic measurements, with slopes of 0.929 + (SD) 0.037 (r = 0.8563) and 0.948 + (SD) 0.037 (r = 0.8554). Multiple measurements of each segment produced values within 30% of the true absolute cross sectional area in most cases. The method shows promise as a means for quantitating absolute dimensions of vessels in clinical arteriography.

Convolutional reconstruction algorithm for fan beam concave and convex circular detectors

A modified convolution-backprojection reconstruction algorithm for a circular detector with an arbitrary radius of curvature has been derived for fan-beam geometries. The algorithm reduces to conventional fan-beam algorithms for flat detectors and for curved detectors having whose radii of curvature coincident with the X-ray tube focal spot. The algorithm substitutes a power series of convolution integrals in place of a single convolution integral. Computer simulations have verified the validity of the algorithm for a detector that curves away from the X-ray source. In this case, it has been demonstrated that good reconstructions are obtained when only a few of the terms of the power series are retained. The development of this algorithm represents a first step toward a cone-beam reconstruction algorithm that eventually will be used for an image-intensifier-based computerized-tomography volume imager being developed in the authors' laboratory.

Reconstruction of blood vessels from x-ray subtraction projections: limited angle geometry

Several algorithms have been investigated for reconstructing blood vessels from a limited number of x-ray subtraction projections, distributed over a limited range of angles. Both computer simulations and an in vivo animal study were carried out. The best reconstruction performance was achieved using an algorithm that folded in two pieces of a priori knowledge of the vascular density distributions: (1) the object is dilute, consisting mainly of a void; and (2) the density distribution in the reconstructions is most likely to be non-negative. Both the signal-to-noise ratio (SNR) and the signal to out-of-focus blur were quantitated. Compared to tomosynthetic reconstruction (backprojection), the amount of residual blur from out-of-focus planes was significantly reduced with only a small penalty in diminished SNR. The combined effect resulted in significant qualitative image improvement for real arterial distributions as demonstrated in a canine arterial imaging example.

Dual-energy electronic scanning-slit fluorography for the determination of vertebral bone mineral content

A technique for performing accurate bone mineral content (BMC) determination, using dual-energy scanning-slit fluorography, has been developed. X-ray scatter and veiling glare are suppressed using a multiple scanning-slit device interfaced to a conventional digital subtraction angiography system. The technique employs a pair of calibration step wedges, one tissue and one bone equivalent. The proper use of the step wedges obviates the need for accurate kVp calibration or absolute calibration of the system's response to BMC. Phantom studies intended to simulate vertebral BMC measurement indicate that the technique is capable of measuring BMC with less than 30 mg/cm2 inaccuracy over a wide range of BMC.

Scatter estimation for a digital radiographic system using convolution filtering

The use of a convolution-filtering method to estimate the scatter distribution in images acquired with a digital subtraction angiography (DSA) imaging system has been studied. Investigation of more than 175 convolution kernels applied to images of anthropomorphic head, chest, and pelvic phantoms using 15-, 25-, and 36-cm fields of view (digitized onto a 512 X 512 pixel image matrix) showed that two-dimensional exponential kernels with a full width at half maximum (FWHM) of 50-150 pixels best reproduced the scatter fields within these images with a root-mean-square percentage error from 4% to 8%. A two-dimensional exponential kernal with a FWHM of 75 pixels in each dimension applied to ten different anatomic presentations and fields of view, resulted in an average root-mean-square percentage error of 6.6% for the ten cases studied. The method should be implementable using an array of small lead beam stops placed in the field of only a single mask image and the above described convolution kernel applied to both mask and postopacification images. The mask beam-stop data are used to scale both mask and postopacification convolution-filtered images. This scaled, convolution-filtered image is then subtracted from the original image to produce a largely scatter-corrected image.

Fast circular tomography device for cardiac imaging: image deflection mechanism and evaluation

A circular tomography system for cardiac imaging is described and measurements of the detector system are presented. Fast tomographic motion is obtained with a custom-made rotating focal spot X-ray tube combined with servo-operated scanning mirrors that deflect the optical image between a large field of view image intensifier and television camera. The motion of the deflection mirrors is locked in frequency and phase to the position of the focal spot with precision analog electronics. Initial measurements on the system components indicate that circular tomograms of the beating heart can be obtained in 1/60 s at rates as high as 30/s. Image degradation due to focal spot tracking error is minimal; however, the optical temporal lag of a P-20 type image intensifier output phosphor results in a factor of three contrast loss for millimeter-sized objects. A linear system analysis in conjunction with the measured temporal step response of the phosphor is in good agreement with the magnitude of this contrast loss.

Electronic scanning-slit fluorography: design and performance of a prototype unit

Electronic scanning-slit fluorography involves replacing paired fore and aft slits for scatter rejection with only one beam-defining tantalum fore aperture. Since the video signal within the projection of the aperture on the image intensifier is much more intense than behind the tantalum, one can discriminate electronically between these two signals and thus eliminate the unwanted x-ray scatter and veiling glare. The general features of a prototype unit are described along with the rationale for the choice of design factors employed. Imaging time of 1-2 s has been achieved using multiple scanning slits. Small focal-spot size and large number of pixels are favored for higher dose utilization, shorter imaging time, and lower x-ray tube loading, as well as for better spatial resolution. Images of a chest phantom show better visibility of low-contrast details, especially in poorly penetrated areas, when compared with the image obtained and displayed under the same conditions, but using a conventional grid to reject scattered radiation.

A signal processing model of diagnostic x-ray scatter

A model of scatter is developed from a signal processing approach. The scattering process is viewed as a nonlinear filter (NLF), which transforms a two-dimensional signal representing phantom thickness into a two-dimensional signal of scattered radiation. The NLF point spread function (PSF) is derived from a single scattering model, using the principles of Compton scattering and x-ray attenuation. The PSF is characterized by three approximations: a constant geometric shape, a volume that depends on the phantom thickness, and a width that depends on the phantom-to-detector distance. This leads to a closed form expression for the scatter-to-primary ratio as a function of phantom thickness, field size, photon energy, source-to-phantom distance, and phantom-to-detector distance. The NLF model is compared with previously reported measurements using constant thickness phantoms, and discrepancies are discussed. The good agreement found between the NLF model and measured data shows that the functional dependence of scatter on the above parameters, previously only explained in terms of empirical models or Monte Carlo simulations, can be incorporated into a signal processing model.

Cross-sectional area measurements by digital subtraction videodensitometry

A method for measurement of absolute cross-section areas in digital images of iodine-opacified vessels was tested in phantoms containing "normal" segments of circular cross-section and either circular or noncircular "stenoses" over a wide range of iodine concentrations. Accuracy varied with iodine concentration, with a slight tendency to overestimate the area at high concentration (185 mg I/ml) and to underestimate at low concentration (23 mg I/ml). Reproducibility was improved at the higher concentrations, with standard deviation ranging from 2% to 7% of the measured value, compared with 7% to 27% at the lowest concentration. Accuracy and reproducibility of area measurements on either normal or stenotic segments were unchanged by rotation of the vessel by up to 25 degrees out of the plane of the image. This method is easily implemented using computerized image processing techniques and has advantages over either manual or automated edge detection because it eliminates the need for an edge criterion and is independent of the viewing projection.

Scatter rejection by electronic collimation

An electronic scanning-slit technology is under development which involves replacing paired fore and aft slits for scatter rejection with only one fore slit. As the slit scans across the portion of the patient to be imaged, pulsed x-ray exposures produce images of the slit in successive positions, which are digitized and stored in computer memory. Software techniques are used for tracking the slit image, and discriminating against scatter and veiling glare lying outside the slit image. Such "electronic collimation" does not require synchronization between the slit-scanning and detector readout, which makes it much simpler than alternative methods and potentially adaptable to any digital fluorography system. The performance characteristics of a prototype unit are described and images of a low-contrast phantom are presented.

Tomographic digital subtraction angiography: initial clinical studies using tomosynthesis. Work in progress

We have developed a method for acquiring multiple tomographic subtraction images using a rapid, repetitive, circular tomographic motion. The method combines the principles of digital subtraction angiography (DSA) and electronic tomosynthesis. Fifteen patients were examined with the technique using single intravenous bolus injections of contrast material. The image sequence obtained during each injection was first processed with a nontomographic mask subtraction, and the result was then compared with the tomographic DSA scans synthesized from the same sequence. The effective section thickness was approximately 0.5 cm, with each section being 0.5-1.0 cm apart. Twelve of the intravenous DSA scans provided the necessary diagnostic or clinically useful information. Two of the three nondiagnostic scans were caused by avoidable technical reasons. In eight cases, the tomographic DSA scans were superior in quality to the nontomographic scans, exhibited significantly less artifact from patient motion and overlying bowel gas, and were effective in separating overlapping vessels. Tomosynthesis permits multiple electronic imaging of the area of interest without reinjection of contrast material and appears to be more informative than nontomographic intravenous DSA imaging.

Computerized tomographic determination of vertebral density after total hip arthroplasty

Computerized tomography (CT) provides the capability to determine vertebral bone density with a high degree of accuracy and precision, detecting density changes as small as 5% in serial testing. In this pilot study, CT was used to determine the effects of total hip arthroplasty and its consequent alterations in physical activity levels on the vertebral bone density of 13 postoperative patients. Vertebral bone density was also evaluated in six preoperative patients with degenerative joint disease of the hip and five control patients. The preoperative group and controls were then compared with the postoperative patients, who were divided into two groups--those who had achieved good activity levels and those whose levels of activity remained poor. CT measurement of bone density in the central bodies of T12, L1, and L2 showed that the preoperative group was similar to controls. This suggests that patients accept total hip arthroplasty rather than endure significant activity limitations for extended periods of time. When activity level was restored by arthroplasty, vertebral bone density was preserved. If the patient continued at a poor activity level postoperation, significant loss of bone density occurred secondary to disuse.

Vessel diameter measurement using digital subtraction radiography

A method for obtaining absolute diameter and cross-sectional area measurements on subtraction digital images is described and tested in phantom vessels from 1.5 to 5.5 mm in diameter filled with iodine contrast at concentrations from 23 to 185 mg I/ml. A highly linear correlation of true vs. calculated diameter is demonstrated, with accuracy and reproducibility of the method varying from +/- 1% to 2% at the highest iodine concentration to +/- 30% in the smallest tube at the lowest concentration. A method is described for correction of the observed video density values to allow for nonlinearity of response of the imaging system to iodine density, and its effect on the measured diameters is demonstrated.

Selective plane removal in limited angle tomographic imaging

A method of improving a planar image in limited angle tomography by removing blurred image information from selected out-of-focus planes is discussed. Focused tomosynthesized images rather than individual projections are used. The necessary equations for removing the information from either two or four adjacent planes, produced with a symmetrical, but otherwise arbitrary blurring function, are developed and specialized to the geometry of circular tomography. Results of a phantom experiment illustrating the technique are presented.

Digital subtraction coronary angiography using high-pass temporal filtration: a comparison with cineangiography

Selective coronary angiograms were obtained using a real-time high-pass temporal filtration digital subtraction technique with videotape storage and display and compared to simultaneously recorded 35-mm cineangiograms for 32 stenotic lesions in 15 patients. Both methods were evaluated by three independent observers using caliper measurement of percent diameter reduction for each lesion. There was a good correlation between the two imaging methods for individual observers, though considerable variability was seen, r = .73, standard error of estimate (SEE) = 9.1%. The average severity of stenosis and the interobserver variability were similar between methods. This digital subtraction technique for selective coronary angiography compares favorably with a conventional film-based technique for evaluation of coronary stenoses and offers advantages of real-time image processing, limited tolerance to patient motion, and relatively small digital memory requirements. In addition to further improvements in image quality, more objective computer-aided scoring methods are needed to reduce the variability in lesion analysis.

Digital angiography: matched filtration versus mask-mode subtraction

Matched filtration was compared to mask-mode subtraction in patients undergoing routine intravenous digital subtraction angiography. Images were analyzed for noise and edge sharpness, and the two techniques were compared subjectively by three radiologists. The quantitative results indicated that matched filtration increased the signal-to-noise ratio by an average of 2.23 but did not significantly increase vessel blurring. Subjectively, the observers favored matched filtration over mask-mode subtraction for overall image quality in 26 out of 30 cases (p less than 0.005) in terms of both sharpness and freedom from motion artifacts.

Tomosynthesis applied to digital subtraction angiography

This extension of our previous work on tomographic digital subtraction angiography (DSA) describes the theory of tomosynthetic DSA image reconstruction techniques. In addition to developing the resolution limits resulting from x-ray exposure length and image intensifier field curvature, we describe one method of image formation and show tomosynthetic DSA images of animal and human anatomy. Methods for improving the present technique are discussed.

Real-time digital angiocardiography using a temporal high-pass filter

A temporal high-pass filtration technique for digital subtraction angiocardiography was studied, using real-time digital studies performed simultaneously with routine cineangiocardiography (cine) for qualitative image comparison. The digital studies showed increased contrast and suppression of background anatomy and also enhanced detection of wall motion abnormalities when compared with cine. This digital technique is compatible with panning the image intensifier during an injection. The digital images are comparable with, and in some cases better than, cine images. Clinical efficacy of this digital technique is currently being evaluated. Real-time display, as well as potential reductions in radiation and contrast material doses, may make digital angiocardiography an attractive technique.

Digital angiography

The many available methods of digital subtraction angiography (DSA) are briefly reviewed. At present the most commonly used are temporal filtration techniques, which include conventional subtraction, integrated remasking , and various types of filtering. Their present use in intravenous, as well as intra-arterial, DSA is shown. The "moving mask" subtraction technique for cardiac and coronary studies is of particular interest. The current status of second-order subtraction techniques such as tomographic DSA and parametric digital imaging is presented. The latter method is particularly useful for demonstration of shunts. Finally, several examples of non-angiographic and future applications of digital radiography are presented.

Blood flow measurement using digital angiography and parametric imaging

A method of blood flow measurement is described that determines flow through a particular artery in absolute units or as a fraction of the cardiac output. For measuring flow as a percentage of cardiac output, no assumption about the vessel's cross-sectional shape is necessary. If one assumes that the vessel's cross section is circular, then absolute flow determination is possible. All methods would require only a single intravenous injection of contrast material. Details of the theory of these methods are presented and alternative data analysis options are developed and discussed.

Tomographic DSA using temporal filtration: initial neurovascular application

A preliminary study showed that encouraging laboratory results reported previously using tomographic digital subtraction angiography (DSA) can be transferred to clinical application for neurovascular imaging. Tomography may show cervical carotid disease more clearly than standard DSA images, and it eliminates the interference caused by overlapping vessels. Production of multiple tomographic image planes from a single set of projection data, tomosynthesis, must be incorporated into this imaging system before tomographic DSA becomes clinically useful. This is a practical reality with the present equipment; clinical evaluation of this new capability is underway.

Image data acquisition, processing, storage and display

It is important to optimize each stage of image formation to minimize and take the greatest advantage of X-ray dose used. For each stage (acquisition, processing, and display) an optimum exists within the limits of physical laws and current technology. These laws must be understood and the technologic limitations imposed by them appreciated to acquire the maximum information from digital subtraction angiography (DSA) examinations.

Blood flow determination using recursive processing: a digital radiographic method

Temporal filtration of fluoroscopic video sequences is being used as an alternative to pulsed digital subtraction angiography. Using the same image processing architecture and a slight modification in processing logic a parametric image can be synthesized from such a temporally filtered image sequence in virtual real time, i.e., an image sequence that spans T seconds takes exactly T seconds to process. Off-line computer processing is not required. Initial phantom studies imply that the time to maximum opacification (tmax) can be used to determine absolute and relative blood flow with a high confidence level (r greater than .989). Phantom and animal examples are presented.

Clinical intra-arterial digital subtraction imaging. Use of small volumes of iodinated contrast material or carbon dioxide

Forty patients were examined with intra-arterial digital subtraction angiography (DSA) using meglumine iothalamate (202 mg iodine/ml) or CO2 as a contrast medium. Diagnostic studies were obtained in all patients. The use of CO2 resulted in diagnostic examinations of the common femoral artery to the level of the trifurcation. The use of CO2 and low-iodine-content contrast material reduced the iodine dosage, and temporal bandpass filtration DSA permitted a lower entrance exposure than pulsed systems. It is concluded that intra-arterial DSA, using the techniques described in this paper, will become a more important and widely used procedure than it currently is.