Search

Purpose: The purpose of this study is to develop a method to evaluate the clinical utility of a high resolution 75µ pixel MAF-CMOS image receptor compared to a 200µ flat-panel detector (FPD). Methods: A simulated procedure evaluation method was developed using an aneurysm phantom since every clinical device-deployment procedure is different and to faithfully compare the performance of two detectors in a clinical setting would require the device deployment to be stopped at different stages so that images could be acquired with both detectors. Not only is this extremely challenging to implement for a large case cohort, it also requires additional exposure and risk to the patient. To simulate a treatment procedure, a 3D printed aneurysm phantom based on patient morphology was used. An anthropomorphic head phantom was placed underneath the aneurysm phantom to simulate patient attenuation and anatomical features. An endovascular coil was deployed into the aneurysm under the guidance of both the MAF-CMOS and the FPD. To qualitatively evaluate the performance of the detectors based on images, a questionnaire containing visualization criteria was developed and a grading scale from 1 (lowest) to 10 (highest) was assigned to each question. The images from these procedures along with the questionnaire will be presented to observers to qualitatively evaluate and compare the performance of the two detectors. Results: Realistic clinical interventions were simulated using a patient-specific aneurysm phantom. Using a common set of criteria in a questionnaire to individually evaluate these images and then comparing the results for a large cohort can be an effective method to determine the significance of using the MAF-CMOS detector during clinical procedures. Conclusion: The above method can be used to effectively evaluate the clinical utility of the MAF-CMOS detector without additional risk to the patient. partial support from NIH Grant R01-EB002873 and equipment grant from Toshiba Medical Systems Corp.

Purpose: Signal and noise aliasing are major issues with high resolution direct and indirect detectors for which the Apodized Aperture Pixel (AAP) design provides an alternative solution. Methods: High resolution detectors for neurovascular interventions have MTFs that remain significant even at the Nyquist frequency. Under-sampling in such detectors, leads to aliasing of object frequencies and noise. Frequencies above the Nyquist wrap into the lower frequency range leading to aliasing. The conventional approach of convolving images with low-pass filters such as pixel binning helps in noise reduction but still allows aliasing. The AAP design proposed by Cunningham et.al, uses super sampled images with subpixels from a high resolution detector, convolved with a sinc function to get an image of the desired pixel size while apodizing the MTF beyond the Nyquist. The performance of the AAP filter in eliminating signal aliasing with respect to a standard filter such as simple 2x pixel binning are compared. Results: Three detectors were considered for this study including two indirect detectors with varying thickness of CsI (350µm, HR) and 500µm, HL) and a direct detector (aSe - 1000µm thickness – MTF after Rivetti et.al,). The presampled MTFs of respective detectors were convolved with a 2x pixel binning kernel and the AAP filter. The Fourier transform of an AAP filter, being the box function, the MTF remains high, up to Nyquist and then drops off drastically to zero. Detailed comparisons of resulting MTFs along with the original presampled MTF are shown. The AAP filter resulted in the least degradation of spatial resolution while eliminating aliasing thus outperforming 2x pixel binning Conclusion: In neurovascular imaging using high resolution detectors, the AAP design provides a viable option to effectively remove signal aliasing without compromising high spatial resolution. Partial support from NIH Grant R01-EB002873 and Toshiba Medical Systems Corp.

Purpose: To evaluate the task specific imaging performance of a new 25µm pixel pitch, 1000µm thick amorphous selenium direct detection system with CMOS readout for typical angiographic exposure parameters using the relative object detectability (ROD) metric. Methods: The ROD metric uses a simulated object function weighted at each spatial frequency by the detectors’ detective quantum efficiency (DQE), which is an intrinsic performance metric. For this study, the simulated objects were aluminum spheres of varying diameter (0.05–0.6mm). The weighted object function is then integrated over the full range of detectable frequencies inherent to each detector, and a ratio is taken of the resulting value for two detectors. The DQE for the 25µm detector was obtained from a simulation of a proposed a-Se detector using an exposure of 200µR for a 50keV x-ray beam. This a-Se detector was compared to two microangiographic fluoroscope (MAF) detectors [the MAF-CCD with pixel size of 35µm and Nyquist frequency of 14.2 cycles/mm and the MAF-CMOS with pixel size of 75µm and Nyquist frequency of 6.6 cycles/mm] and a standard flat-panel detector (FPD with pixel size of 194µm and Nyquist frequency of 2.5cycles/mm). Results: ROD calculations indicated vastly superior performance by the a-Se detector in imaging small aluminum spheres. For the 50µm diameter sphere, the ROD values for the a-Se detector compared to the MAF-CCD, the MAF-CMOS, and the FPD were 7.3, 9.3 and 58, respectively. Detector performance in the low frequency regime was dictated by each detector’s DQE(0) value. Conclusion: The a-Se with CMOS readout is unique and appears to have distinctive advantages of incomparable high resolution, low noise, no readout lag, and expandable design. The a-Se direct detection system will be a powerful imaging tool in angiography, with potential break-through applications in diagnosis and treatment of neuro-vascular disease. Supported by NIH Grant: 2R01EB002873 and an equipment grant from Toshiba Medical Systems Corporation

Purpose: Demonstrate and quantify the augmented resolution due to focalspot size decrease in images acquired on the anode side of the field, for both small and medium (0.3 and 0.6mm) focal-spot sizes using the experimental task-based GM-ROD metric. Theoretical calculations have shown that a medium focal-spot can achieve the resolution of a small focal-spot if acquired with a tilted anode, effectively providing a higher-output small focal-spot. Methods: The MAF-CMOS (micro-angiographic fluoroscopic complementary-metal-oxide semiconductor) detector (75µm pixel pitch) imaged two copper wire segments of different diameter and a pipeline stent at the central axis and on the anode side of the beam, achieved by tilting the x-ray C-arm (Toshiba Infinix) to 6° and realigning the detector with the perpendicular ray to correct for x-ray obliquity. The relative gain in resolution was determined using the GM-ROD metric, which compares images on the basis of the Fourier transform of the image and the measured NNPS. To emphasize the geometric unsharpness, images were acquired at a magnification of two. Results: Images acquired on the anode side were compared to those acquired on the central axis with the same target-area focal-spot to consider the effect of an angled tube, and for all three objects the advantage of the smaller effective focal-spot was clear, showing a maximum improvement of 36% in GM-ROD. The images obtained with the small focal-spot at the central axis were compared to those of the medium focal-spot at the anode side and, for all objects, the relative performance was comparable. Conclusion: For three objects, the GM-ROD demonstrated the advantage of the anode side focal-spot. The comparable performance of the medium focal-spot on the anode side will allow for a high-output small focal-spot; a necessity in endovascular image-guided interventions. Partial support from an NIH grant R01EB002873 and an equipment grant from Toshiba Medical Systems Corp.

Purpose: The Generalized Relative Object Detectability (G-ROD) family of ideal observer metrics are a well-characterized set of task-based metrics used for quantitatively comparing imaging systems that include the detector, focal spot geometry and scatter characteristics on the basis of system abilities in imaging a given simulated object. The G-ROD metric takes the integral over spatial frequencies of the Fourier transform of a simulated object function weighted with the detector generalized DQE (GDQE), divided by the comparable integral for another detector. When a measured image of an actual object is used, the generalized measured-ROD (GM-ROD) metric can compare detector systems without explicitly determining detector MTF. In this work, the results of simulated and measured ROD calculations for the same object are compared for the first time. Methods: G-ROD calculations were performed to compare a high resolution microangiographic fluoroscopic (MAF) detector to a flat panel detector (FPD) using 5mm segments of simulated copper wires of varying diameter (d=0.05–0.6mm) as the object function for three different focal spot sizes at two magnifications. A GM-ROD experiment with similar parameters was performed to compare the same detector system using three actual 5mm copper wires (d=0.113mm,0.238mm,0.558mm) imaged under identical but simulated conditions used in G-ROD calculations to determine the agreement between simulated and experimentally obtained results. Results: The G-ROD and GM-ROD calculations examined relative detector system imaging performance. The GM-ROD results for all three different sized wires imaged with low magnification and a small focal spot size were 2.1, 1.38, and 1.1, respectively, and the G-ROD results for the same calculation was 1.81, 1.3, and 1.07, corresponding to percent deviations of 13%, 5.8%, and 2.7%. Conclusion: These results indicate experimental validation of the simulated metric results. Partial support from NIH grant R01EB002873 and an equipment grant from Toshiba Medical Systems Corp.

Purpose: In endovascular image-guided neuro-interventions, visualization of fine detail is paramount. For example, the ability of the interventionist to visualize the stent struts depends heavily on the x-ray imaging detector performance. Methods: A study to examine the relative performance of the high resolution MAF-CMOS (pixel size 75µm, Nyquist frequency 6.6 cycles/mm) and a standard Flat Panel Detector (pixel size 194µm, Nyquist frequency 2.5 cycles/mm) detectors in imaging a neuro stent was done using the Generalized Measured Relative Object Detectability (GM-ROD) metric. Low quantum noise images of a deployed stent were obtained by averaging 95 frames obtained by both detectors without changing other exposure or geometric parameters. The square of the Fourier transform of each image is taken and divided by the generalized normalized noise power spectrum to give an effective measured task-specific signal-to-noise ratio. This expression is then integrated from 0 to each of the detector’s Nyquist frequencies, and the GM-ROD value is determined by taking a ratio of the integrals for the MAF-CMOS to that of the FPD. The lower bound of integration can be varied to emphasize high frequencies in the detector comparisons. Results: The MAF-CMOS detector exhibits vastly superior performance over the FPD when integrating over allmore » frequencies, yielding a GM-ROD value of 63.1. The lower bound of integration was stepped up in increments of 0.5 cycles/mm for higher frequency comparisons. As the lower bound increased, the GM-ROD value was augmented, reflecting the superior performance of the MAF-CMOS in the high frequency regime. Conclusion: GM-ROD is a versatile metric that can provide quantitative detector and task dependent comparisons that can be used as a basis for detector selection. Supported by NIH Grant: 2R01EB002873 and an equipment grant from Toshiba Medical Systems Corporation.« less

Purpose: CMOS-based aSe detectors compared to CsI-TFT-based flat panels have the advantages of higher spatial sampling due to smaller pixel size and decreased blurring characteristic of direct rather than indirect detection. For systems with such detectors, the limiting factor degrading image resolution then becomes the focal-spot geometric unsharpness. This effect can seriously limit the use of such detectors in areas such as cone beam computed tomography, clinical fluoroscopy and angiography. In this work a technique to remove the effect of focal-spot blur is presented for a simulated aSe detector. Method: To simulate images from an aSe detector affected with focal-spot blur, first a set of high-resolution images of a stent (FRED from Microvention, Inc.) were acquired using a 75µm pixel size Dexela-Perkin-Elmer detector and averaged to reduce quantum noise. Then the averaged image was blurred with a known Gaussian blur at two different magnifications to simulate an idealized focal spot. The blurred images were then deconvolved with a set of different Gaussian blurs to remove the effect of focal-spot blurring using a threshold-based, inverse-filtering method. Results: The blur was removed by deconvolving the images using a set of Gaussian functions for both magnifications. Selecting the correct function resulted in an image close to the original; however, selection of too wide a function would cause severe artifacts. Conclusion: Experimentally, focal-spot blur at different magnifications can be measured using a pin hole with a high resolution detector. This spread function can be used to deblur the input images that are acquired at corresponding magnifications to correct for the focal spot blur. For CBCT applications, the magnification of specific objects can be obtained using initial reconstructions then corrected for focal-spot blurring to improve resolution. Similarly, if object magnification can be determined such correction may be applied in fluoroscopy and angiography.

Purpose: The development of patient specific models for endovascular procedures using additive manufacturing. Methods: The work to be presented is composed of two parts: manufacturing and testing of patientspecific phantoms. Phantom manufacturing began with acquisition of 3D patient data and segmentation for the generation of a printable STL file. The models were printed using a Polyjet printer, Objet Eden260V. Once printed, the phantoms were cleaned of support material, and tested for geometric accuracy as compared to the original patient specific data. Imaging validation using x-ray angiography and qualitative evaluation of mechanical properties were also performed during an endovascular image-guided intervention. Several iterations of phantoms have been printed and tested, and optimal design parameters have been determined for a favorable manufactured output. Results: The most challenging aspect of phantom manufacturing, as experienced in both simple and complex phantoms, is the removal of support material. Cleaning support material from inside of vessels can be mechanically difficult due to tortuous designs, and must be performed delicately to prevent wall rupture. Phantoms required NaOH baths and high-pressure washing to eliminate all support materials. Phantom accuracy testing showed size variations of 120μm, a very good agreement with the original design. Interventionists experienced similar back pressure from phantoms when testing mechanical behavior in simulated clinical procedures. Conclusion: This process can serve as a viable tool for procedures and devices, and present unique learning opportunities for the endovascular field as a whole. Supported by NIH Grant: 2R01EB002873 and an equipment grant from Toshiba Medical Systems Corporation

Purpose: To investigate the variations in bone volume calculations in mice involved in aging research when changing cone beam micro-CT x-ray and reconstruction parameters. Methods: Mouse spines were placed on an indexed turn table that rotated 0.5° per projection and imaged by a self-built micro CT machine containing a CCD-based high-resolution x-ray detector. After the full 360° rotation data set of object images was obtained, a standard filtered back-projection cone beam reconstruction was performed. Four different kVp's between 40–70 kVp in 10kVp increments were selected. For each kVp two mAs settings were used. Each acquisition was reconstructed using two voxel sizes (12 and 25μm) and two step angles, 0.5° and 1°, respectively. A LabView program was written to determine the total bone volume contained in the mouse's total spine volume (bone plus gaps) as a measure of spine health. First, the user selected the desired 512×512 reconstruction to view the whole spine volume which was then used to select a gray-level threshold that allowed for viewing of the bone structure, then another threshold to include gaps. The program returned bone volume, bone × gap volume, and their ratio, BVF. Results: The calculated bone volume fractions were compared as a functionmore » of tube potential. Cases with 25μm slice thickness showed trials with lower kVp's had greater image contrast, which resulted in higher calculated bone volume fractions. Cases with 12μm reconstructed slice thickness were significantly noisier, and showed no clear maximum BVF. Conclusion: Using the projection images and reconstructions acquired from the micro CT, it can be shown that the micro-CT x-ray and reconstruction parameters significantly affect the total bone volume calculations. When comparing mice cohorts treated with different therapies researchers need to be aware of such details and use volumes which were acquired and processed in identical conditions.« less