Fluoroscopy
9.0 INTRODUCTION
Fluoroscopy is a procedure that provides real-time transient x-ray imaging with permanent recording of appropriate images. Fluoroscopy is not generally intended to generally replace conventional projection radiography. The basic modes of operation of a fluoroscopic system are as follows:
Fluoroscopy using relatively low radiation dose rates provides real-time imaging for positioning with selective recording only when sequences contain sufficient diagnostic information for later use.
Fluorography (acquisition) using substantially higher radiation dose rates provides images for later use with image quality higher than that achievable using stored fluoroscopy.
Fluoroscopic systems are flexible general-purpose devices that are configurable to meet the requirements of a wide range of clinical procedures (Fig. 9-1).
 ▪ FIGURE 9-1 Block diagram of an interventional fluoroscope. Key components and control loops are shown. The operator is a key element with the ability to select and modify radiation factors, imaging geometry, image acquisition parameters, and image processing; These controls include (1) x-ray generator settings (kV, mA, continuous versus pulsed fluoro operation), (2) automatic exposure rate control (AERC) dose rates, (3) acquisition modes and frame rates, (4) image-receptor field of view and collimation, and (5) image processing methods and image display adjustments. |
9.1 FLUOROSCOPIC IMAGING CHAIN OVERVIEW
1. Principal feature of fluoroscopy over radiography is the ability to acquire real-time images.
a. Real-time frame rates of 30 frames/s and reduced rates (15, 7.5, 3.75, 2, 1 frames/s).
b. 20-minute fluoro “on-time” at 15 frames/s produces 18,000 individual images.
2. Images (frames) are produced with much less dose than comparable radiograph.
a. Fluoroscopy 10 to 100 nGy/image
b. Fluorography 100 to 1,000 nGy/image—digital acquisition (DA), digital subtraction angiography (DSA)
c. Radiography 3,000 to 10,000 nGy/image
3. Fluoroscopic imaging chain components of a modern system (Fig. 9-2).
a. High gain solid-state flat-panel image detector converts x-ray signal into a series of digital images.
b. Antiscatter grid reduces scatter to achieve good subject contrast.
c. Collimation limits the x-ray beam to the active field of view or to a clinically relevant area.
d. X-ray tube; beam attenuation filters used to reduce dose to patient.
 ▪ FIGURE 9-2 Diagram of a fluoroscopic system. The fluoroscopic imaging chain is illustrated, with the patient in the supine position. This system includes a flat-panel detector. The x-ray system includes a collimator with motorized blades that automatically adjust to conform to the current field of view (FOV) and source to image distance (SID). |
9.2 IMAGING CHAIN COMPONENTS
1. X-ray image intensifier (II)
a. Key component of a fluoroscopy system from approximately 1960 to 2010; many remain in use in 2020
b. Diagram of an II and its components (Fig. 9-3)
▪ FIGURE 9-3 Diagram of an image intensifier and evacuated insert tube: x-rays pass through the input window of the vacuum element and strike the input screen, producing light that stimulates the photocathode to emit electrons, which are ejected into the electronic lens system. The electronic lens system adds energy to the image-carrying electrons, while minifying the image onto a small output screen. The electron image is finally converted back to light at the output screen. These conversion steps are sketched on the bottom of the figure. The combination of energy input and minification in the electron optics yields an image 5 to 10 thousand times as bright as the image emerging from the input screen. The light image is detected by a CCD or CMOS camera to produce a digital video signal that is re-created on a display monitor in the examination room.
c. Input screen
(i) Includes a cesium-iodide (CsI) scintillator that converts the x-ray image into a visible light image and a photocathode that converts the light image into an electron image
(ii) Curved to meet the needs of the II’s electron optics, resulting in pincushion distortion of the image
d. Electron optics
(i) Add energy to the electrons emitted by the photocathode under a potential difference of 20 to 40 kV between the photocathode and the output screen/anode.
(ii) Focus the electrons generated in the active area of the input onto the fixed size output screen.
(iii) Multi-mode IIs have a separate set of electrodes for each input field size.
(iv) Concentrate all of the signal intensity produced at the input screen onto the smaller output screen.
(v) Intensity gain is proportional to the ratios of the areas of the input and output screens (minification).
e. Output screen
(i) Consists of a layer of ZnCdS coated onto either a glass or fiber optic output window.
(ii) Produces an image of fixed diameter irrespective of the size of the active input layer.
(iii) Limiting resolution of an II is inversely proportional to the input screen’s diameter.
f. Optical distributor
(i) Couples the output phosphor image to the video camera with optical lenses or direct fiberoptics.
(ii) Optical aperture adjusts the output light levels for fluoroscopy and fluorography.
(iii) Images are transferred from the output screen via a digital video camera.
(iv) The video image of the output screen is the same size for any FOV setting of the II.
g. Brightness gain = Minification gain × Electronic gain (acceleration of electrons) [9-1]
(i) The input dose level is inversely proportional to the area of the input screen.
(ii) Maintaining the constant light output requires a higher x-ray flux for magnified modes.
h. Conversion factor is the light intensity exiting the II divided by the exposure rate at the input.
(i) A measure of overall performance including brightness gain
(ii) Decreases over time due to degradation of the photocathode
2. Flat-Panel Detectors
a. Replacement for the II tube, optical system, and cameras.
(i) Directly converts x-rays into a digital form.
(ii) Carbon fiber covers reduce x-ray attenuation and increase DQE compared to II.
b. Composed of arrays of detector elements (DEXELS) packaged in a square or rectangular area (Fig. 9-4).
▪ FIGURE 9-4 Flat-panel detectors (FPD). A. Components in an indirect FPD. Note that the CsI layer is continuous and that there is space between individual dexels. The dexels in a direct x-ray system are also separated (not shown). B. Information flow through indirect and direct FPDs. C. The photodiode and some of the electronics in a single indirect dexel.
F. 9-5
c. Indirect systems have a phosphor layer that absorbs the x-rays and converts a fraction of their energy into light and a photodiode that converts the x-ray-induced light from the phosphor into a corresponding charge.
d. Direct systems have a semiconductor (e.g., selenium) that produces x-ray-induced charge directly, which is captured by the electronics within the same dexel as the x-ray absorption event.
e. FPD readout
(i) For fluoroscopy, after the capacitor is drained, the charge continues to be stored for the next frame.
(ii) Real-time x-ray data are collected as the charge continues to be stored in a steady-state manner.
f. Dexel arrays
(i) Reducing the FOV on spatial resolution differs for an image-intensifier and a FPD (Fig. 9-5).
(ii) In an II, the FOV is reduced by electro-optically focusing a smaller portion of its input screen onto the fixed size output screen, to provide a fixed image size (image-matrix size) for the camera and image processor.
 ▪ FIGURE 9-5 Pixel size and magnification mode—image intensifier and flat-panel detector. A-D. Image intensifier—analog and digital camera outputs. The full pixel array of the digital camera captures the output of the II at any magnification mode. E-G. Flatpanel detector with a fixed dexel array (e.g., Table 9-1 small detector). The same dexel covers the same portion of the patient irrespective of magnification mode (E and F). Resolution does not change when the image is digitally magnified. H. Flat-panel detector with unbundled dexels (e.g., Table 9-1 large detector, small FOV). Resolution is improved if the FPD unbundles dexels when using a small FOV.  F. 9-6 |
TABLE 9-1 DEXELS, PIXELS, AND FIELD OF VIEW | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
9.3 FLUOROSCOPIC X-RAY SOURCE ASSEMBLY
1. X-ray tube
a. Have higher heat-storage capacity and faster cooling rates than radiographic tubes.
b. Small focal spot is used for fluoroscopy and large for fluorography (CINE/DA/DSA).
(i) Spatial resolution is likely to decrease when going from fluoroscopy to fluorography.
(ii) Some tubes also have an additional micro-focus focal spot for geometrically magnified procedures.
c. For all modes of operation, newer systems select the smallest focal spot that can accommodate the immediate x-ray power level and may result in spatial resolution differences for thin and thick body parts.
2. Collimator
a. Dynamically adjusts the overall x-ray field size in response to a combination of inputs from both the operator and the system to confine the x-ray beam to the smallest area within the active area of the image receptor consistent with immediate clinical requirements.
b. Wedge filters are used to equalize image receptor inputs when anatomical structures providing very different attenuation (e.g., lung and mediastinum) are simultaneously in the beam.
c. Collimation automatically limits the x-ray beam to the active FOV whenever the FOV is changed and to changes in the source to image-receptor distance (SID).
3. X-ray spectral shaping filters
a. Increases the fraction of x-ray photons in the beam with energies slightly above the K absorption edge material of interest, to improve conspicuity of higher Z elements (e.g., iodine at 33.2 keV).
b. Copper filter (0.1 to 1.0 mm) and restricted x-ray tube voltage (lower kV, e.g., 60) reduce the number of photons in the beam below the iodine edge and limits the number of higher energy photons (Fig. 9-6).
▪ FIGURE 9-6 Shaping the spectrum of the x-ray beam. A. The x-ray spectrum emerging from the tube is modified by a copper layer. The overall intensity of the beam is reduced, and the resultant spectrum has shifted to higher average energy. This figure shows the original spectrum, the attenuation coefficient of iodine as a function of energy, and the resultant modified spectrum with a higher fraction of photons above the iodine K edge. B. Increasing kV shifts the spectrum further, but most of the higher-energy photons are far from the iodine K edge and do not contribute very much to iodine visualization. Tungsten K characteristic photons are seen here as well. C. Reducing kV moves more of the spectrum toward the iodine K edge. There are fewer low-energy photons that would contribute to skin dose but not to the image. The photons, with energies above 60 kV, shown in (B) are not produced. X-ray tubes need to have high power capabilities so that they can provide adequate x-ray flux without increasing kV.
F. 9-7
c. Most modern systems have variable thickness copper filters in the collimator assembly.
d. Acquisition selection (e.g., low-dose rate fluoro, standard-dose rate fluoro) and automatic exposure rate control (AERC) can determine filter thickness and beam HVL (see textbook Figure 9-8, page 320 for AERC control of added filtration).
e. Limited x-ray tube power output may require the filter thickness to decrease and the kV to increase.
f. HVL may decrease with an increase in kV if there is a simultaneous decrease in copper thickness.
9.4 CONTROLS
1. Automatic exposure rate control
a. Delivers constant x-ray intensity to the image receptor irrespective of tissue thickness in the beam path.
b. Goal is reasonably constant detector SNR over its working range by maintaining a constant detector entrance dose rate.
c. AERC collects detected x-ray intensity information observed by a predetermined subset of dexels.
(i) Low average signal causes AERC to command the generator to increase x-ray output.
(ii) High average signal causes AERC to decrease x-ray output.
d. AERC output adjustments
(i) Changes in FOV or frame rate adjust x-ray tube output to stabilize perceived image noise.
(ii) X-ray factors (kV, mA, pulse width), focal spot, and beam filtration provide different balances between patient dose and image quality for various system options (Fig. 9-7).
(iii) Image presentation to the operator is further influenced by these changes.
2. Field of view and magnification modes
a. The active FOV of an image intensifier is established by focusing a portion of its input screen onto the full area of the output screen, with an increase in image magnification as the FOV decreases (Fig. 9-5).
b. The active FOV of a flat-panel detector is reduced by accepting information from a smaller area of the image receptor, and the subsequent image is digitally resized to fill the fluoroscope’s display (Fig. 9-5).
c. The brightness gain of an II decreases as the magnification increases—the AERC compensates by boosting the x-ray exposure rate, typically by the square of the input phosphor diameter ratios.
d. Flat-panel systems adjust dose rate when pixels are unbinned, and additional exposure rate increases are applied for all FOVs to meet perceptual requirements, usually linear with changes in active FOV area.
 ▪ FIGURE 9-7 Output dose rates as a function of patient size are a configurable property of the fluoroscopic system. A. System with three maximum fluoroscopic dose rate modes. In this system, the dose rate is identical for all modes until the individual mode limit is exceeded. B. The lower two fluoroscopic modes converge near the regulatory limit. The high-dose-rate mode is higher for all patient thicknesses and has double the usual regulatory limit. C. Hypothetical trajectories. Keeping the kV low for as long as possible improves iodine contrast at the expense of higher skin dose for heavy patients (highest relative dose rate for any tissue thickness). Raising the kV quickly minimizes patient dose at the expense of iodine contrast for many patients (lowest relative dose rate). Balanced curves (intermediate relative dose rate) represent a different trade-off between dose and iodine visibility.  F. 9-9 |
9.5 MODES OF OPERATION
1. Continuous fluoroscopy
a. An uninterrupted x-ray beam with an image stream partitioned into discrete frames by the video system.
b. The typical frame rate is 30 frames per second (FPS).
c. Quantum noise and anatomical motion occurring in a single frame (1/30 s) is averaged and blurred.
d. A second stage of averaging occurs in the observer’s eye due to temporal averaging.
e. A single frame stored and viewed on a display monitor will appear nosier and sharper than the live image because observer temporal averaging no longer influences perception.
2. Pulsed fluoroscopy and fluorography
a. X-ray generator produces a series of short pulses with an on-time smaller than the frame time (Fig. 9-8A).
b. Short pulses (e.g., 3 to 10 ms) reduce blurring of moving objects in individual images.
c. Appearance of a moving object when viewed dynamically is dependent on the interactions of object motion, frame rate, processing of gap-filling display frames, and temporal filtering.
d. An object will often appear sharper when a single image is viewed because there is no temporal averaging in either the image processor or the human visual system.
e. The digital imaging system provides an output frame rate high enough to avoid image flicker by displaying each image multiple times when low acquisition frame rates are used (Fig. 9-8B).
▪ FIGURE 9-8 Pulse widths, frame rates, and dose rates. A. The same dose per pulse can be delivered using a high dose rate for a short time or a lower dose rate for a longer time. Short pulses minimize individual frame motion blur. B. The imaging system eliminates flicker at any frame rate by providing copies to fill the gaps at an appropriate rate. C. Some fluoroscopes adjust the dose rate to produce equal perceived noise at different frame rates.
F. 9-10
f. Pulsed imaging can reduce radiation because the acquisition frame rate is decoupled from the display rate.
g. Fluorographic images are intended to be viewed individually and therefore requires the same dose per pulse at any frame rate to maintain image SNR.
h. Fluoroscopic images will appear noisier if the same dose per pulse is used for all frame rates.
i. Increasing the fluoroscopy dose per pulse by about 1.4 when the FPS is halved maintains constant noise perception (Fig. 9-8C).
3. Fluorography (unsubtracted)
a. Fluorographic images, also known as digital angiography, DA images, or cine, should only be obtained if the intent is to store them for later review.
b. Enough dose per image is needed to reduce quantum noise to an acceptable value without relying on temporal averaging in the observer’s eye.
c. A single fluorographic frame (not intended for DSA) will require about 10 times the radiation needed for a comparable fluoroscopic frame.
d. Frames intended for DSA will require 50 to 100 times as much radiation as a fluoroscopic frame (Fig. 9-9).
▪ FIGURE 9-9 Fluoroscopic and fluorographic images. These two images were taken a few seconds apart. The fluorographic DA image (B) from a DSA series required approximately 50 times the dose needed for the single fluoroscopic frame (A). Note the increased quantum noise in the fluoroscopic image.
F 9-11
4. Fluorography (subtracted)
a. DSA is performed by acquiring a series of DA images (e.g., 1 to 6 images per second) while a contrast agent (e.g., iodine or CO2) is injected (Fig. 9-10).
▪ FIGURE 9-10 Image subtraction. A. Mask image: This image will be subtracted from a later image in the series (not shown) in which iodine is present in the patient’s arterial tree. B. Result of the subtraction: Structures common to both images have been removed. X-ray quantum noise from both images is present. The contrast of the image has been increased to improve the conspicuity of the vessels. It is necessary to use high radiation doses when acquiring the images to reduce the visibility of quantum noise. C. CO2 subtraction image of the abdominal aorta. The CO2 bubble does not fill the entire aorta at any one time. This image is reconstructed by subtracting the mask from a sequence of CO2 containing images and then combining them to reconstruct the vascular tree.
F. 9-12
b. The first few images contain no contrast agent, and one or an average of images is selected as the mask.
c. The contrast agent moves through the vessels, images contain both anatomy and contrast-enhanced vessels.
d. The mask image is subtracted (pixel by pixel) from the images with contrast images, resulting in the subtraction of the anatomical background.
e. Software tools permit automatic or manual reregistration of images to minimize motion artifacts.
f. DSA removes fixed information common to both the mask and live image.
g. Random noise adds and produces more quantum noise in the difference image than in either initial image.
h. Display contrast is increased to improve visibility of blood vessels or other moving structures. This increases the visibility of quantum noise in the image.
i. Suppressing quantum noise requires the use of more per-image radiation.
j. Nominally, up to 10 times the dose is needed for a single DSA frame than for a comparable unsubtracted fluorographic (DA) image.
k. Using low frame rates, where practicable, reduces radiation.
5. Cine fluorography (analog and digital)
a. The same digital image chain supports fluoroscopy, DSA, and cine. Imaging modes are changed by configuring fluoroscopic hardware and software.
b. Frame rates can be set to conform to procedure specific requirements. Reduction in cine frame rate produces a major reduction in patient dose.
c. Retrospectively stored fluoroscopy (if of adequate quality) can be substituted for acquisitions in the clinical image stream, saving both radiation and the volume of contrast media used in a procedure.
9.6 IMAGE PROCESSING
1. Single image processing
a. This section discusses blurring and unsharp masking applied uniformly across the entire image.
b. Newer algorithms use processing to differentially modify the visibility of objects in different parts of the same image.
c. Image blurring to reduce the visibility of noise (Fig. 9-11).
d. Unsharp masking parameters (Fig. 9-12).
▪ FIGURE 9-11 Image blurring to reduce the visibility of noise. The original image (512 × 512 pixels, 8-bit gray scale) is in the upper left corner. Columns: Gaussian noise (N) added to the original image at four levels (0%, 5%, 25%, 50%). Rows: The resultant images were then blurred using Gaussian blurring (B) using radii of 0 (no blur), 5, and 10 pixels.
F. 9-13
▪ FIGURE 9-12 Unsharp masking parameters. The original (unblurred) image is on the left. Each row has constant Gaussian blur radius M of 5, 20, 50 pixels. Each column has the same weighted combination of blurred and original image weighting factors W of 0% (blur only), 50%, 200%, 500%.
F. 9-14
2. Temporal image processing (multi-frame)
a. Individual fluoroscopic images are relatively noisy, with noise patterns statistically changing from frame to frame.
b. Appropriately combining several fluoroscopic frames will improve the perceived SNR while maintaining the system’s inherent spatial resolution at a potential cost of blurring moving objects.
c. Similar processing is seldom needed during fluorography (e.g., cine) because the higher dose per frame is intended to reduce noise.
d. Image processors facilitate task-specific temporal blurring such as recursive filtering (Fig. 9-13).Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
