SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Photonics West 2019 | Register Today

SPIE Defense + Commercial Sensing 2019 | Call for Papers

2019 SPIE Optics + Photonics | Call for Papers



Print PageEmail PageView PDF

Electronic Imaging & Signal Processing

Designing and testing a microangiographic imaging system

An assessment method that incorporates focal-spot blurring and scattered radiation optimizes the performance of a fluoroscopic detector.
27 March 2008, SPIE Newsroom. DOI: 10.1117/2.1200803.1041

Minimally invasive interventions are replacing major surgery for many diseases, including stroke and neurovascular disease. This trend has driven demand for high-resolution, real-time imaging systems to perform fluoroscopic angiography, a procedure in which fluorescent dye is injected into blood vessels and viewed in rapid-sequence x-rays. Yet existing state-of-the-art x-ray image intensifiers and flat-panel detectors suffer from limitations that slow progress in patient treatment and diagnosis. In addition, the current evaluation method, called the objective assessment, only accounts for the detector. It fails to capture overall performance in clinical conditions, where final image quality is also affected by many other features of the setup.

To address these difficulties, we designed and tested a new high-sensitivity microangiographic fluoroscopic (HSMAF) detector. We have demonstrated its utility in image-guided interventions, and have also developed a new technique, called the generalized objective assessment, to measure and optimize its performance. The assessment is tailored to the total clinical configuration and thus incorporates properties not only of the detector but also the x-ray tube, patient-specific scattered radiation, and imaging geometry.1

Figure 1. Schematic diagram of the HSMAF detector. A cesium iodide (CsI) scintillator connects to a microchannel plate light image intensifier (MCP LII). The direct coupling between the fiber-optic taper and the charge-coupled device (CCD) sensor provides wide dynamic range.

Figure 2. The generalized modulation transfer function (GMTF) vs. spatial frequency (f) in simulated neurovascular angiography. The detector MTF, which does not account for scatter or focal-spot blurring, overestimates the resolution. FOV: Field of view.

The HMSAF (see Figure 1) couples a 300μm-thick, thallium-doped cesium iodide (CsI) scintillator to a 40mm-diameter, variable-gain, dual-stage microchannel plate (MCP) light image intensifier (LII). A 2.88:1 minifying fiber-optic taper connects the amplified output to a custom-bonded fiber-optic window on the chip of a charge-coupled device (CCD) camera.

Because of the direct fiber-optic coupling between the amplifying stage and the sensor, the LII provides wide dynamic range, and achieves high sensitivity without excessive noise. The HSMAF has an effective pixel size of 35μm (referenced to the detector input) in its finest-resolution mode, and can employ pixel binning for greater sensitivity. A high-speed frame-grabber board converts the camera signals to digital form for post-processing and display. A custom graphical user interface facilitates the acquisition and processing required by interventional imaging protocol.

We have also developed a new set of generalized parameters to assess image quality. They consider the focal-spot blurring and the scatter-induced image degradation as parallel processes. The detector degradation follows these in the chain. For example, the modulation transfer functions (MTFs) for the detector (MTFD), the focal spot (MTFF) and the scatter (MTFS) are combined with the scatter fraction (ρ) to define the generalized MTF (GMTF) in the object plane,2


where m is a magnification factor and f represents the object spatial frequency in mm−1. We have also formulated similar metrics for the normalized noise power spectrum, the noise-equivalent quanta, and the detective quantum efficiency.2

We define scatter fraction as ρ = S/(S+P), where S and P are the scatter and primary components detected by the image receptor, respectively. The scatter fraction and magnification can be expressed in terms of geometry-dependent variables such as air gap and irradiation field area for a given source-to-image distance, and location of the object of interest inside the patient.

Figure 2 shows the GMTF as a function of spatial frequency at two different air gaps (2.5 and 7.5cm) for a 0.3mm focal spot in simulated neurovascular angiography. Compared to the total system resolution, the detector MTF, which neglects scatter and geometric blurring, overestimates the actual resolution. Increasing the air gap simultaneously contributes two factors to the GMTF: scattered radiation, dominant at low spatial frequencies, and focal-spot blurring, dominant at higher spatial frequencies. At smaller air gaps, increased scatter degrades the GMTF at low spatial frequencies. Yet at higher frequencies, the small air gap results in reduced geometric magnification and improved resolution. For the cases studied, the HSMAF GMTF, which includes focal-spot blurring and scatter, remains high even at spatial frequencies above 3lp/mm. Conventional detectors are often not useful above this range.

The HSMAF's performance, even at high spatial frequencies, demonstrates its potential for neurovascular interventions. In addition, the generalized objective assessment method can provide a more accurate evaluation of high-resolution systems, leading to improved designs tailored to the imaging task.

Currently, the HMSAF is being used in animal models for the placement of stents and other devices within the vasculature. Our next step, a human study, may help improve cerebral stroke treatment by providing real-time image guidance. Future work may also employ the detector for high-resolution, 3D volume image acquisition.

Girijesh Yadava, Stephen Rudin, Andrew Kuhls-Gilcrist, Daniel Bednarek 
Toshiba Stroke Research Center
University at Buffalo (SUNY)
Buffalo, NY